Bump to 0.1.11

Closes #288 and #287

.bumpversion.cfg

@@ -0,0 +1,12 @@
+[bumpversion]
+current_version = 0.1.13
+commit = True
+message = Bump version: {current_version} → {new_version}
+tag = True
+tag_name = v{new_version}
+
+[bumpversion:file:node/package.json]
+
+[bumpversion:file:rust/ffi/node/Cargo.toml]
+
+[bumpversion:file:rust/vectordb/Cargo.toml]

.github/workflows/cargo-publish.yml

@@ -0,0 +1,29 @@
+name: Cargo Publish
+
+on:
+  release:
+    types: [ published ]
+
+env:
+  # This env var is used by Swatinem/rust-cache@v2 for the cache
+  # key, so we set it to make sure it is always consistent.
+  CARGO_TERM_COLOR: always
+
+jobs:
+  build:
+    runs-on: ubuntu-22.04
+    timeout-minutes: 30
+    # Only runs on tags that matches the make-release action
+    if: startsWith(github.ref, 'refs/tags/v')
+    steps:
+      - uses: actions/checkout@v3
+      - uses: Swatinem/rust-cache@v2
+        with:
+          workspaces: rust
+      - name: Install dependencies
+        run: |
+          sudo apt update
+          sudo apt install -y protobuf-compiler libssl-dev
+      - name: Publish the package
+        run: |
+          cargo publish -p vectordb --all-features --token ${{ secrets.CARGO_REGISTRY_TOKEN }}

.github/workflows/docs.yml

@@ -39,6 +39,28 @@ jobs:
         run: |
           python -m pip install -e .
           python -m pip install -r ../docs/requirements.txt
+      - name: Set up node
+        uses: actions/setup-node@v3
+        with:
+          node-version: ${{ matrix.node-version }}
+          cache: 'npm'
+          cache-dependency-path: node/package-lock.json
+      - uses: Swatinem/rust-cache@v2
+      - name: Install node dependencies
+        working-directory: node
+        run: |
+          sudo apt update
+          sudo apt install -y protobuf-compiler libssl-dev
+      - name: Build node
+        working-directory: node
+        run: |
+          npm ci
+          npm run build
+          npm run tsc
+      - name: Create markdown files
+        working-directory: node
+        run: |
+          npx typedoc --plugin typedoc-plugin-markdown --out ../docs/src/javascript src/index.ts
       - name: Build docs
         run: |
           PYTHONPATH=. mkdocs build -f docs/mkdocs.yml

.github/workflows/docs_test.yml

@@ -0,0 +1,93 @@
+name: Documentation Code Testing
+
+on:
+  push:
+    branches:
+      - main
+    paths:
+      - docs/**
+      - .github/workflows/docs_test.yml
+  pull_request:
+    paths:
+      - docs/**
+      - .github/workflows/docs_test.yml
+
+  # Allows you to run this workflow manually from the Actions tab
+  workflow_dispatch:
+
+env:
+  # Disable full debug symbol generation to speed up CI build and keep memory down
+  # "1" means line tables only, which is useful for panic tracebacks.
+  RUSTFLAGS: "-C debuginfo=1"
+  RUST_BACKTRACE: "1"
+
+jobs:
+  test-python:
+    name: Test doc python code
+    runs-on: ${{ matrix.os }}
+    strategy:
+      matrix:
+        python-minor-version: [ "11" ]
+        os: ["ubuntu-22.04"]
+    steps:
+    - name: Checkout
+      uses: actions/checkout@v3
+    - name: Set up Python
+      uses: actions/setup-python@v4
+      with:
+        python-version: 3.${{ matrix.python-minor-version }}
+        cache: "pip"
+        cache-dependency-path: "docs/test/requirements.txt"
+    - name: Build Python
+      working-directory: docs/test
+      run:
+        python -m pip install -r requirements.txt
+    - name: Create test files
+      run: |
+        cd docs/test
+        python md_testing.py
+    - name: Test
+      run: |
+        cd docs/test/python
+        for d in *; do cd "$d"; echo "$d".py; python "$d".py; cd ..; done
+  test-node:
+    name: Test doc nodejs code
+    runs-on: ${{ matrix.os }}
+    strategy:
+      matrix:
+        node-version: [ "18" ]
+        os: ["ubuntu-22.04"]
+    steps:
+    - name: Checkout
+      uses: actions/checkout@v3
+      with:
+        fetch-depth: 0
+        lfs: true
+    - name: Set up Node
+      uses: actions/setup-node@v3
+      with:
+        node-version: ${{ matrix.node-version }}
+    - name: Install dependecies needed for ubuntu
+      if: ${{ matrix.os == 'ubuntu-22.04' }}
+      run: |
+        sudo apt install -y protobuf-compiler libssl-dev
+    - name: Install node dependencies
+      run: |
+        cd docs/test
+        npm install 
+    - name: Rust cache
+      uses: swatinem/rust-cache@v2
+    - name: Install LanceDB
+      run: |
+        cd docs/test/node_modules/vectordb 
+        npm ci
+        npm run build-release
+        npm run tsc
+    - name: Create test files
+      run: |
+        cd docs/test
+        node md_testing.js
+    - name: Test
+      run: |
+        cd docs/test/node
+        for d in *; do cd "$d"; echo "$d".js; node "$d".js; cd ..; done

.github/workflows/make-release-commit.yml

@@ -0,0 +1,55 @@
+name: Create release commit
+
+on:
+  workflow_dispatch:
+    inputs:
+      dry_run:
+        description: 'Dry run (create the local commit/tags but do not push it)'
+        required: true
+        default: "false"
+        type: choice
+        options:
+          - "true"
+          - "false"
+      part:
+        description: 'What kind of release is this?'
+        required: true
+        default: 'patch'
+        type: choice
+        options:
+          - patch
+          - minor
+          - major
+
+jobs:
+  bump-version:
+    runs-on: ubuntu-latest
+    steps:
+    - name: Check out main
+      uses: actions/checkout@v3
+      with:
+        ref: main
+        persist-credentials: false
+        fetch-depth: 0
+        lfs: true
+    - name: Set git configs for bumpversion
+      shell: bash
+      run: |
+        git config user.name 'Lance Release'
+        git config user.email 'lance-dev@lancedb.com'
+    - name: Set up Python 3.10
+      uses: actions/setup-python@v4
+      with:
+        python-version: "3.10"
+    - name: Bump version, create tag and commit
+      run: |
+        pip install bump2version
+        bumpversion --verbose ${{ inputs.part }}
+    - name: Push new version and tag
+      if: ${{ inputs.dry_run }} == "false"
+      uses: ad-m/github-push-action@master
+      with:
+        github_token: ${{ secrets.LANCEDB_RELEASE_TOKEN }}
+        branch: main
+        tags: true
+

.github/workflows/node.yml

@@ -67,8 +67,12 @@ jobs:
     - name: Build
       run: |
         npm ci
-        npm run build
         npm run tsc
+        npm run build
+        npm run pack-build
+        npm install --no-save ./dist/vectordb-*.tgz
+        # Remove index.node to test with dependency installed
+        rm index.node
     - name: Test
       run: npm run test
   macos:
@@ -94,8 +98,12 @@ jobs:
     - name: Build
       run: |
         npm ci
-        npm run build
         npm run tsc
+        npm run build
+        npm run pack-build
+        npm install --no-save ./dist/vectordb-*.tgz
+        # Remove index.node to test with dependency installed
+        rm index.node
     - name: Test
       run: |
         npm run test

.github/workflows/npm-publish.yml

@@ -0,0 +1,171 @@
+name: NPM Publish
+
+on:
+  release:
+    types: [ published ]
+
+jobs:
+  node:
+    runs-on: ubuntu-latest
+    # Only runs on tags that matches the make-release action
+    if: startsWith(github.ref, 'refs/tags/v')
+    defaults:
+      run:
+        shell: bash
+        working-directory: node
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v3
+      - uses: actions/setup-node@v3
+        with:
+          node-version: 20
+          cache: 'npm'
+          cache-dependency-path: node/package-lock.json
+      - name: Install dependencies
+        run: |
+          sudo apt update
+          sudo apt install -y protobuf-compiler libssl-dev
+      - name: Build
+        run: |
+          npm ci
+          npm run tsc
+          npm pack
+      - name: Upload Linux Artifacts
+        uses: actions/upload-artifact@v3
+        with:
+          name: node-package
+          path: |
+            node/vectordb-*.tgz
+
+  node-macos:
+    runs-on: macos-12
+    # Only runs on tags that matches the make-release action
+    if: startsWith(github.ref, 'refs/tags/v')
+    strategy:
+      fail-fast: false
+      matrix:
+        target: [x86_64-apple-darwin, aarch64-apple-darwin]
+    steps:
+    - name: Checkout
+      uses: actions/checkout@v3
+    - name: Install system dependencies
+      run: brew install protobuf
+    - name: Install npm dependencies
+      run: |
+        cd node
+        npm ci
+    - name: Install rustup target
+      if: ${{ matrix.target == 'aarch64-apple-darwin' }}
+      run: rustup target add aarch64-apple-darwin
+    - name: Build MacOS native node modules
+      run: bash ci/build_macos_artifacts.sh ${{ matrix.target }}
+    - name: Upload Darwin Artifacts
+      uses: actions/upload-artifact@v3
+      with:
+        name: darwin-native
+        path: |
+          node/dist/vectordb-darwin*.tgz
+
+  node-linux:
+    name: node-linux (${{ matrix.arch}}-unknown-linux-${{ matrix.libc }})
+    runs-on: ubuntu-latest
+    # Only runs on tags that matches the make-release action
+    if: startsWith(github.ref, 'refs/tags/v')
+    strategy:
+      fail-fast: false
+      matrix:
+        libc:
+          - gnu
+          # TODO: re-enable musl once we have refactored to pre-built containers
+          # Right now we have to build node from source which is too expensive.
+          # - musl
+        arch:
+          - x86_64
+          # Building on aarch64 is too slow for now
+          # - aarch64
+    steps:
+    - name: Checkout
+      uses: actions/checkout@v3
+    - name: Change owner to root (for npm)
+      # The docker container is run as root, so we need the files to be owned by root
+      # Otherwise npm is a nightmare: https://github.com/npm/cli/issues/3773
+      run: sudo chown -R root:root .
+    - name: Set up QEMU
+      if: ${{ matrix.arch == 'aarch64' }}
+      uses: docker/setup-qemu-action@v2
+      with:
+        platforms: arm64
+    - name: Build Linux GNU native node modules
+      if: ${{ matrix.libc == 'gnu' }}
+      run: |
+        docker run \
+          -v $(pwd):/io -w /io \
+          rust:1.70-bookworm \
+          bash ci/build_linux_artifacts.sh ${{ matrix.arch }}-unknown-linux-gnu
+    - name: Build musl Linux native node modules
+      if: ${{ matrix.libc == 'musl' }}
+      run: |
+        docker run --platform linux/arm64/v8 \
+          -v $(pwd):/io -w /io \
+          quay.io/pypa/musllinux_1_1_${{ matrix.arch }} \
+          bash ci/build_linux_artifacts.sh ${{ matrix.arch }}-unknown-linux-musl
+    - name: Upload Linux Artifacts
+      uses: actions/upload-artifact@v3
+      with:
+        name: linux-native
+        path: |
+          node/dist/vectordb-linux*.tgz
+
+  node-windows:
+    runs-on: windows-2022
+    # Only runs on tags that matches the make-release action
+    if: startsWith(github.ref, 'refs/tags/v')
+    strategy:
+      fail-fast: false
+      matrix:
+        target: [x86_64-pc-windows-msvc]
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v3
+      - name: Install Protoc v21.12
+        working-directory: C:\
+        run: |
+          New-Item -Path 'C:\protoc' -ItemType Directory
+          Set-Location C:\protoc
+          Invoke-WebRequest https://github.com/protocolbuffers/protobuf/releases/download/v21.12/protoc-21.12-win64.zip -OutFile C:\protoc\protoc.zip
+          7z x protoc.zip
+          Add-Content $env:GITHUB_PATH "C:\protoc\bin"
+        shell: powershell
+      - name: Install npm dependencies
+        run: |
+          cd node
+          npm ci
+      - name: Build Windows native node modules
+        run: .\ci\build_windows_artifacts.ps1 ${{ matrix.target }}
+      - name: Upload Windows Artifacts
+        uses: actions/upload-artifact@v3
+        with:
+          name: windows-native
+          path: |
+            node/dist/vectordb-win32*.tgz
+
+  release:
+    needs: [node, node-macos, node-linux]
+    runs-on: ubuntu-latest
+    # Only runs on tags that matches the make-release action
+    if: startsWith(github.ref, 'refs/tags/v')
+    steps:
+      - uses: actions/download-artifact@v3
+      - name: Display structure of downloaded files
+        run: ls -R
+      - uses: actions/setup-node@v3
+        with:
+          node-version: 20
+      - name: Publish to NPM
+        env:
+          NODE_AUTH_TOKEN: ${{ secrets.LANCEDB_NPM_REGISTRY_TOKEN }}
+        run: |
+          mv */*.tgz .
+          for filename in *.tgz; do
+            npm publish $filename
+          done

.github/workflows/pypi-publish.yml

@@ -0,0 +1,31 @@
+name: PyPI Publish
+
+on:
+  release:
+    types: [ published ]
+
+jobs:
+  publish:
+    runs-on: ubuntu-latest
+    # Only runs on tags that matches the python-make-release action
+    if: startsWith(github.ref, 'refs/tags/python-v')
+    defaults:
+      run:
+        shell: bash
+        working-directory: python
+    steps:
+      - uses: actions/checkout@v3
+      - name: Set up Python
+        uses: actions/setup-python@v4
+        with:
+          python-version: "3.8"
+      - name: Build distribution
+        run: |
+          ls -la
+          pip install wheel setuptools --upgrade
+          python setup.py sdist bdist_wheel
+      - name: Publish
+        uses: pypa/gh-action-pypi-publish@v1.8.5
+        with:
+          password: ${{ secrets.LANCEDB_PYPI_API_TOKEN }}
+          packages-dir: python/dist

.github/workflows/python.yml

@@ -32,9 +32,11 @@ jobs:
       run: |
         pip install -e .
         pip install tantivy@git+https://github.com/quickwit-oss/tantivy-py#164adc87e1a033117001cf70e38c82a53014d985
-        pip install pytest pytest-mock black
+        pip install pytest pytest-mock black isort
     - name: Black
       run: black --check --diff --no-color --quiet .
+    - name: isort
+      run: isort --check --diff --quiet .
     - name: Run tests
       run: pytest -x -v --durations=30 tests
     - name: doctest
@@ -59,6 +61,8 @@ jobs:
       run: |
         pip install -e .
         pip install tantivy@git+https://github.com/quickwit-oss/tantivy-py#164adc87e1a033117001cf70e38c82a53014d985
-        pip install pytest pytest-mock
+        pip install pytest pytest-mock black
+    - name: Black
+      run: black --check --diff --no-color --quiet .
     - name: Run tests
       run: pytest -x -v --durations=30 tests

.github/workflows/rust.yml

@@ -0,0 +1,89 @@
+name: Rust
+
+on:
+  push:
+    branches:
+      - main
+  pull_request:
+    paths:
+      - Cargo.toml
+      - rust/**
+      - .github/workflows/rust.yml
+
+env:
+  # This env var is used by Swatinem/rust-cache@v2 for the cache
+  # key, so we set it to make sure it is always consistent.
+  CARGO_TERM_COLOR: always
+  # Disable full debug symbol generation to speed up CI build and keep memory down
+  # "1" means line tables only, which is useful for panic tracebacks.
+  RUSTFLAGS: "-C debuginfo=1"
+  RUST_BACKTRACE: "1"
+
+jobs:
+  linux:
+    timeout-minutes: 30
+    runs-on: ubuntu-22.04
+    defaults:
+      run:
+        shell: bash
+        working-directory: rust
+    steps:
+    - uses: actions/checkout@v3
+      with:
+          fetch-depth: 0
+          lfs: true
+    - uses: Swatinem/rust-cache@v2
+      with:
+          workspaces: rust
+    - name: Install dependencies
+      run: |
+          sudo apt update
+          sudo apt install -y protobuf-compiler libssl-dev
+    - name: Build
+      run: cargo build --all-features
+    - name: Run tests
+      run: cargo test --all-features
+  macos:
+    runs-on: macos-12
+    timeout-minutes: 30
+    defaults:
+      run:
+        shell: bash
+        working-directory: rust
+    steps:
+      - uses: actions/checkout@v3
+        with:
+            fetch-depth: 0
+            lfs: true
+      - name: CPU features
+        run: sysctl -a | grep cpu
+      - uses: Swatinem/rust-cache@v2
+        with:
+          workspaces: rust
+      - name: Install dependencies
+        run: brew install protobuf
+      - name: Build
+        run: cargo build --all-features
+      - name: Run tests
+        run: cargo test --all-features
+  windows:
+    runs-on: windows-2022
+    steps:
+      - uses: actions/checkout@v3
+      - uses: Swatinem/rust-cache@v2
+        with:
+          workspaces: rust
+      - name: Install Protoc v21.12
+        working-directory: C:\
+        run: |
+          New-Item -Path 'C:\protoc' -ItemType Directory
+          Set-Location C:\protoc
+          Invoke-WebRequest https://github.com/protocolbuffers/protobuf/releases/download/v21.12/protoc-21.12-win64.zip -OutFile C:\protoc\protoc.zip
+          7z x protoc.zip
+          Add-Content $env:GITHUB_PATH "C:\protoc\bin"
+        shell: powershell
+      - name: Run tests
+        run: |
+          $env:VCPKG_ROOT = $env:VCPKG_INSTALLATION_ROOT
+          cargo build
+          cargo test

.gitignore

@@ -3,6 +3,9 @@
 *.egg-info
 **/__pycache__
 .DS_Store
+venv
+
+.vscode
 
 rust/target
 rust/Cargo.lock
@@ -30,3 +33,4 @@ node/examples/**/dist
 ## Rust
 target
 
+Cargo.lock

.pre-commit-config.yaml

@@ -9,3 +9,13 @@ repos:
     rev: 22.12.0
     hooks:
     -   id: black
+-   repo: https://github.com/astral-sh/ruff-pre-commit
+    # Ruff version.
+    rev: v0.0.277
+    hooks:
+    - id: ruff
+-   repo: https://github.com/pycqa/isort
+    rev: 5.12.0
+    hooks:
+    - id: isort
+      name: isort (python)

Cargo.toml

@@ -4,3 +4,13 @@ members = [
     "rust/ffi/node"
 ]
 resolver = "2"
+
+[workspace.dependencies]
+lance = "=0.5.8"
+arrow-array = "42.0"
+arrow-data = "42.0"
+arrow-schema = "42.0"
+arrow-ipc = "42.0"
+half = { "version" = "2.2.1", default-features = false }
+object_store = "0.6.1"
+

README.md

@@ -65,7 +65,7 @@ pip install lancedb
 ```python
 import lancedb
 
-uri = "/tmp/lancedb"
+uri = "data/sample-lancedb"
 db = lancedb.connect(uri)
 table = db.create_table("my_table",
                          data=[{"vector": [3.1, 4.1], "item": "foo", "price": 10.0},

ci/build_linux_artifacts.sh

@@ -0,0 +1,72 @@
+#!/bin/bash
+# Builds the Linux artifacts (node binaries).
+# Usage: ./build_linux_artifacts.sh [target]
+# Targets supported: 
+# - x86_64-unknown-linux-gnu:centos
+# - aarch64-unknown-linux-gnu:centos
+# - aarch64-unknown-linux-musl
+# - x86_64-unknown-linux-musl
+
+# TODO: refactor this into a Docker container we can pull
+
+set -e
+
+setup_dependencies() {
+    echo "Installing system dependencies..."
+    if [[ $1 == *musl ]]; then
+        # musllinux
+        apk add openssl-dev
+    else
+        # rust / debian
+        apt update
+        apt install -y libssl-dev protobuf-compiler
+    fi
+}
+
+install_node() {
+    echo "Installing node..."
+    curl -o- https://raw.githubusercontent.com/nvm-sh/nvm/v0.34.0/install.sh | bash
+    source "$HOME"/.bashrc
+
+    if [[ $1 == *musl ]]; then
+        # This node version is 15, we need 16 or higher:
+        # apk add nodejs-current npm 
+        # So instead we install from source (nvm doesn't provide binaries for musl):
+        nvm install -s --no-progress 17
+    else
+        nvm install --no-progress 17 # latest that supports glibc 2.17
+    fi
+}
+
+build_node_binary() {
+    echo "Building node library for $1..."
+    pushd node
+
+    npm ci
+    
+    if [[ $1 == *musl ]]; then
+        # This is needed for cargo to allow build cdylibs with musl
+        export RUSTFLAGS="-C target-feature=-crt-static"
+    fi
+
+    # Cargo can run out of memory while pulling dependencies, especially when running
+    # in QEMU. This is a workaround for that.
+    export CARGO_NET_GIT_FETCH_WITH_CLI=true
+
+    # We don't pass in target, since the native target here already matches
+    # We need to pass OPENSSL_LIB_DIR and OPENSSL_INCLUDE_DIR for static build to work https://github.com/sfackler/rust-openssl/issues/877
+    OPENSSL_STATIC=1 OPENSSL_LIB_DIR=/usr/lib/x86_64-linux-gnu OPENSSL_INCLUDE_DIR=/usr/include/openssl/ npm run build-release
+    npm run pack-build
+
+    popd
+}
+
+TARGET=${1:-x86_64-unknown-linux-gnu}
+# Others:
+# aarch64-unknown-linux-gnu
+# x86_64-unknown-linux-musl
+# aarch64-unknown-linux-musl
+
+setup_dependencies $TARGET
+install_node $TARGET
+build_node_binary $TARGET

ci/build_macos_artifacts.sh

@@ -0,0 +1,33 @@
+# Builds the macOS artifacts (node binaries).
+# Usage: ./ci/build_macos_artifacts.sh [target]
+# Targets supported: x86_64-apple-darwin aarch64-apple-darwin
+
+prebuild_rust() {
+    # Building here for the sake of easier debugging.
+    pushd rust/ffi/node
+    echo "Building rust library for $1"
+    export RUST_BACKTRACE=1
+    cargo build --release --target $1
+    popd
+}
+
+build_node_binaries() {
+    pushd node
+    echo "Building node library for $1"
+    npm run build-release -- --target $1
+    npm run pack-build -- --target $1
+    popd
+}
+
+if [ -n "$1" ]; then
+    targets=$1
+else
+    targets="x86_64-apple-darwin aarch64-apple-darwin"
+fi
+
+echo "Building artifacts for targets: $targets"
+for target in $targets
+    do
+    prebuild_rust $target
+    build_node_binaries $target
+done

ci/build_windows_artifacts.ps1

@@ -0,0 +1,41 @@
+# Builds the Windows artifacts (node binaries).
+# Usage:  .\ci\build_windows_artifacts.ps1 [target]
+# Targets supported:
+# - x86_64-pc-windows-msvc
+# - i686-pc-windows-msvc
+
+function Prebuild-Rust {
+    param (
+        [string]$target
+    )
+
+    # Building here for the sake of easier debugging.
+    Push-Location -Path "rust/ffi/node"
+    Write-Host "Building rust library for $target"
+    $env:RUST_BACKTRACE=1
+    cargo build --release --target $target
+    Pop-Location
+}
+
+function Build-NodeBinaries {
+    param (
+        [string]$target
+    )
+
+    Push-Location -Path "node"
+    Write-Host "Building node library for $target"
+    npm run build-release -- --target $target
+    npm run pack-build -- --target $target
+    Pop-Location
+}
+
+$targets = $args[0]
+if (-not $targets) {
+    $targets = "x86_64-pc-windows-msvc"
+}
+
+Write-Host "Building artifacts for targets: $targets"
+foreach ($target in $targets) {
+    Prebuild-Rust $target
+    Build-NodeBinaries $target
+}

docs/mkdocs.yml

@@ -6,11 +6,13 @@ docs_dir: src
 theme:
   name: "material"
   logo: assets/logo.png
+  favicon: assets/logo.png
   features:
     - content.code.copy
     - content.tabs.link
   icon:
     repo: fontawesome/brands/github
+  custom_dir: overrides
 
 plugins:
 - search
@@ -36,6 +38,7 @@ plugins:
 
 markdown_extensions:
 - admonition
+- footnotes
 - pymdownx.superfences
 - pymdownx.details
 - pymdownx.highlight:
@@ -47,13 +50,19 @@ markdown_extensions:
 - pymdownx.superfences
 - pymdownx.tabbed:
     alternate_style: true
+- md_in_html
 
 nav:
 - Home: index.md
 - Basics: basic.md
 - Embeddings: embedding.md
 - Python full-text search: fts.md
-- Python integrations: integrations.md
+- Python integrations:
+  - Pandas and PyArrow: python/arrow.md
+  - DuckDB: python/duckdb.md
+  - LangChain 🦜️🔗: https://python.langchain.com/en/latest/modules/indexes/vectorstores/examples/lancedb.html
+  - LlamaIndex 🦙: https://gpt-index.readthedocs.io/en/latest/examples/vector_stores/LanceDBIndexDemo.html
+  - Pydantic: python/pydantic.md
 - Python examples:
   - YouTube Transcript Search: notebooks/youtube_transcript_search.ipynb
   - Documentation QA Bot using LangChain: notebooks/code_qa_bot.ipynb
@@ -62,8 +71,10 @@ nav:
   - Serverless QA Bot with Modal: examples/serverless_qa_bot_with_modal_and_langchain.md
 - Javascript examples:
   - YouTube Transcript Search: examples/youtube_transcript_bot_with_nodejs.md
+  - TransformersJS Embedding Search: examples/transformerjs_embedding_search_nodejs.md
 - References:
   - Vector Search: search.md
+  - SQL filters: sql.md
   - Indexing: ann_indexes.md
 - API references:
   - Python API: python/python.md

docs/overrides/partials/header.html

@@ -0,0 +1,176 @@
+<!--
+  Copyright (c) 2016-2023 Martin Donath <martin.donath@squidfunk.com>
+
+  Permission is hereby granted, free of charge, to any person obtaining a copy
+  of this software and associated documentation files (the "Software"), to
+  deal in the Software without restriction, including without limitation the
+  rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+  sell copies of the Software, and to permit persons to whom the Software is
+  furnished to do so, subject to the following conditions:
+
+  The above copyright notice and this permission notice shall be included in
+  all copies or substantial portions of the Software.
+
+  THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+  IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+  FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL THE
+  AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+  LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+  FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+  IN THE SOFTWARE.
+-->
+
+{% set class = "md-header" %}
+{% if "navigation.tabs.sticky" in features %}
+  {% set class = class ~ " md-header--shadow md-header--lifted" %}
+{% elif "navigation.tabs" not in features %}
+  {% set class = class ~ " md-header--shadow" %}
+{% endif %}
+
+<!-- Header -->
+<header class="{{ class }}" data-md-component="header">
+  <nav
+    class="md-header__inner md-grid"
+    aria-label="{{ lang.t('header') }}"
+  >
+
+    <!-- Link to home -->
+    <a
+      href="{{ config.extra.homepage | d(nav.homepage.url, true) | url }}"
+      title="{{ config.site_name | e }}"
+      class="md-header__button md-logo"
+      aria-label="{{ config.site_name }}"
+      data-md-component="logo"
+    >
+      {% include "partials/logo.html" %}
+    </a>
+
+    <!-- Button to open drawer -->
+    <label class="md-header__button md-icon" for="__drawer">
+      {% include ".icons/material/menu" ~ ".svg" %}
+    </label>
+
+    <!-- Header title -->
+    <div class="md-header__title" style="width: auto !important;" data-md-component="header-title">
+      <div class="md-header__ellipsis">
+        <div class="md-header__topic">
+          <span class="md-ellipsis">
+            {{ config.site_name }}
+          </span>
+        </div>
+        <div class="md-header__topic" data-md-component="header-topic">
+          <span class="md-ellipsis">
+            {% if page.meta and page.meta.title %}
+              {{ page.meta.title }}
+            {% else %}
+              {{ page.title }}
+            {% endif %}
+          </span>
+        </div>
+      </div>
+    </div>
+
+    <!-- Color palette -->
+    {% if config.theme.palette %}
+      {% if not config.theme.palette is mapping %}
+        <form class="md-header__option" data-md-component="palette">
+          {% for option in config.theme.palette %}
+            {% set scheme  = option.scheme  | d("default", true) %}
+            {% set primary = option.primary | d("indigo", true) %}
+            {% set accent  = option.accent  | d("indigo", true) %}
+            <input
+              class="md-option"
+              data-md-color-media="{{ option.media }}"
+              data-md-color-scheme="{{ scheme | replace(' ', '-') }}"
+              data-md-color-primary="{{ primary | replace(' ', '-') }}"
+              data-md-color-accent="{{ accent | replace(' ', '-') }}"
+              {% if option.toggle %}
+                aria-label="{{ option.toggle.name }}"
+              {% else  %}
+                aria-hidden="true"
+              {% endif %}
+              type="radio"
+              name="__palette"
+              id="__palette_{{ loop.index }}"
+            />
+            {% if option.toggle %}
+              <label
+                class="md-header__button md-icon"
+                title="{{ option.toggle.name }}"
+                for="__palette_{{ loop.index0 or loop.length }}"
+                hidden
+              >
+                {% include ".icons/" ~ option.toggle.icon ~ ".svg" %}
+              </label>
+            {% endif %}
+          {% endfor %}
+        </form>
+      {% endif %}
+    {% endif %}
+
+    <!-- Site language selector -->
+    {% if config.extra.alternate %}
+      <div class="md-header__option">
+        <div class="md-select">
+          {% set icon = config.theme.icon.alternate or "material/translate" %}
+          <button
+            class="md-header__button md-icon"
+            aria-label="{{ lang.t('select.language') }}"
+          >
+            {% include ".icons/" ~ icon ~ ".svg" %}
+          </button>
+          <div class="md-select__inner">
+            <ul class="md-select__list">
+              {% for alt in config.extra.alternate %}
+                <li class="md-select__item">
+                  <a
+                    href="{{ alt.link | url }}"
+                    hreflang="{{ alt.lang }}"
+                    class="md-select__link"
+                  >
+                    {{ alt.name }}
+                  </a>
+                </li>
+              {% endfor %}
+            </ul>
+          </div>
+        </div>
+      </div>
+    {% endif %}
+
+    <!-- Button to open search modal -->
+    {% if "material/search" in config.plugins %}
+      <label class="md-header__button md-icon" for="__search">
+        {% include ".icons/material/magnify.svg" %}
+      </label>
+
+      <!-- Search interface -->
+      {% include "partials/search.html" %}
+    {% endif %}
+
+    <div style="margin-left: 10px; margin-right: 5px;">
+        <a href="https://discord.com/invite/zMM32dvNtd" target="_blank" rel="noopener noreferrer">
+            <svg fill="#FFFFFF" xmlns="http://www.w3.org/2000/svg"  viewBox="0 0 50 50" width="25px" height="25px"><path d="M 41.625 10.769531 C 37.644531 7.566406 31.347656 7.023438 31.078125 7.003906 C 30.660156 6.96875 30.261719 7.203125 30.089844 7.589844 C 30.074219 7.613281 29.9375 7.929688 29.785156 8.421875 C 32.417969 8.867188 35.652344 9.761719 38.578125 11.578125 C 39.046875 11.867188 39.191406 12.484375 38.902344 12.953125 C 38.710938 13.261719 38.386719 13.429688 38.050781 13.429688 C 37.871094 13.429688 37.6875 13.378906 37.523438 13.277344 C 32.492188 10.15625 26.210938 10 25 10 C 23.789063 10 17.503906 10.15625 12.476563 13.277344 C 12.007813 13.570313 11.390625 13.425781 11.101563 12.957031 C 10.808594 12.484375 10.953125 11.871094 11.421875 11.578125 C 14.347656 9.765625 17.582031 8.867188 20.214844 8.425781 C 20.0625 7.929688 19.925781 7.617188 19.914063 7.589844 C 19.738281 7.203125 19.34375 6.960938 18.921875 7.003906 C 18.652344 7.023438 12.355469 7.566406 8.320313 10.8125 C 6.214844 12.761719 2 24.152344 2 34 C 2 34.175781 2.046875 34.34375 2.132813 34.496094 C 5.039063 39.605469 12.972656 40.941406 14.78125 41 C 14.789063 41 14.800781 41 14.8125 41 C 15.132813 41 15.433594 40.847656 15.621094 40.589844 L 17.449219 38.074219 C 12.515625 36.800781 9.996094 34.636719 9.851563 34.507813 C 9.4375 34.144531 9.398438 33.511719 9.765625 33.097656 C 10.128906 32.683594 10.761719 32.644531 11.175781 33.007813 C 11.234375 33.0625 15.875 37 25 37 C 34.140625 37 38.78125 33.046875 38.828125 33.007813 C 39.242188 32.648438 39.871094 32.683594 40.238281 33.101563 C 40.601563 33.515625 40.5625 34.144531 40.148438 34.507813 C 40.003906 34.636719 37.484375 36.800781 32.550781 38.074219 L 34.378906 40.589844 C 34.566406 40.847656 34.867188 41 35.1875 41 C 35.199219 41 35.210938 41 35.21875 41 C 37.027344 40.941406 44.960938 39.605469 47.867188 34.496094 C 47.953125 34.34375 48 34.175781 48 34 C 48 24.152344 43.785156 12.761719 41.625 10.769531 Z M 18.5 30 C 16.566406 30 15 28.210938 15 26 C 15 23.789063 16.566406 22 18.5 22 C 20.433594 22 22 23.789063 22 26 C 22 28.210938 20.433594 30 18.5 30 Z M 31.5 30 C 29.566406 30 28 28.210938 28 26 C 28 23.789063 29.566406 22 31.5 22 C 33.433594 22 35 23.789063 35 26 C 35 28.210938 33.433594 30 31.5 30 Z"/></svg>
+        </a>
+    </div>
+    <div style="margin-left: 5px; margin-right: 5px;">
+        <a href="https://twitter.com/lancedb" target="_blank" rel="noopener noreferrer">
+            <svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" viewBox="0,0,256,256" width="25px" height="25px" fill-rule="nonzero"><g fill-opacity="0" fill="#ffffff" fill-rule="nonzero" stroke="none" stroke-width="1" stroke-linecap="butt" stroke-linejoin="miter" stroke-miterlimit="10" stroke-dasharray="" stroke-dashoffset="0" font-family="none" font-weight="none" font-size="none" text-anchor="none" style="mix-blend-mode: normal"><path d="M0,256v-256h256v256z" id="bgRectangle"></path></g><g fill="#ffffff" fill-rule="nonzero" stroke="none" stroke-width="1" stroke-linecap="butt" stroke-linejoin="miter" stroke-miterlimit="10" stroke-dasharray="" stroke-dashoffset="0" font-family="none" font-weight="none" font-size="none" text-anchor="none" style="mix-blend-mode: normal"><g transform="scale(4,4)"><path d="M57,17.114c-1.32,1.973 -2.991,3.707 -4.916,5.097c0.018,0.423 0.028,0.847 0.028,1.274c0,13.013 -9.902,28.018 -28.016,28.018c-5.562,0 -12.81,-1.948 -15.095,-4.423c0.772,0.092 1.556,0.138 2.35,0.138c4.615,0 8.861,-1.575 12.23,-4.216c-4.309,-0.079 -7.946,-2.928 -9.199,-6.84c1.96,0.308 4.447,-0.17 4.447,-0.17c0,0 -7.7,-1.322 -7.899,-9.779c2.226,1.291 4.46,1.231 4.46,1.231c0,0 -4.441,-2.734 -4.379,-8.195c0.037,-3.221 1.331,-4.953 1.331,-4.953c8.414,10.361 20.298,10.29 20.298,10.29c0,0 -0.255,-1.471 -0.255,-2.243c0,-5.437 4.408,-9.847 9.847,-9.847c2.832,0 5.391,1.196 7.187,3.111c2.245,-0.443 4.353,-1.263 6.255,-2.391c-0.859,3.44 -4.329,5.448 -4.329,5.448c0,0 2.969,-0.329 5.655,-1.55z"></path></g></g></svg>
+        </a>
+    </div>
+
+    <!-- Repository information -->
+    {% if config.repo_url %}
+      <div class="md-header__source" style="margin-left: -5px !important;">
+        {% include "partials/source.html" %}
+      </div>
+    {% endif %}
+  </nav>
+
+  <!-- Navigation tabs (sticky) -->
+  {% if "navigation.tabs.sticky" in features %}
+    {% if "navigation.tabs" in features %}
+      {% include "partials/tabs.html" %}
+    {% endif %}
+  {% endif %}
+</header>

docs/src/ann_indexes.md

@@ -1,7 +1,7 @@
 # ANN (Approximate Nearest Neighbor) Indexes
 
 You can create an index over your vector data to make search faster.
-Vector indexes are faster but less accurate than exhaustive search.
+Vector indexes are faster but less accurate than exhaustive search (KNN or Flat Search).
 LanceDB provides many parameters to fine-tune the index's size, the speed of queries, and the accuracy of results.
 
 Currently, LanceDB does *not* automatically create the ANN index.
@@ -10,7 +10,18 @@ If you can live with <100ms latency, skipping index creation is a simpler workfl
 
 In the future we will look to automatically create and configure the ANN index.
 
-## Creating an ANN Index
+## Types of Index
+
+Lance can support multiple index types, the most widely used one is `IVF_PQ`.
+
+* `IVF_PQ`: use **Inverted File Index (IVF)** to first divide the dataset into `N` partitions,
+   and then use **Product Quantization** to compress vectors in each partition.
+* `DISKANN` (**Experimental**): organize the vector as a on-disk graph, where the vertices approximately
+   represent the nearest neighbors of each vector.
+
+## Creating an IVF_PQ Index
+
+Lance supports `IVF_PQ` index type by default.
 
 === "Python"
      Creating indexes is done via the [create_index](https://lancedb.github.io/lancedb/python/#lancedb.table.LanceTable.create_index) method.
@@ -23,7 +34,7 @@ In the future we will look to automatically create and configure the ANN index.
 
      # Create 10,000 sample vectors
      data = [{"vector": row, "item": f"item {i}"}
-        for i, row in enumerate(np.random.random((10_000, 768)).astype('float32'))]
+        for i, row in enumerate(np.random.random((10_000, 1536)).astype('float32'))]
 
      # Add the vectors to a table
      tbl = db.create_table("my_vectors", data=data)
@@ -41,19 +52,22 @@ In the future we will look to automatically create and configure the ANN index.
      for (let i = 0; i < 10_000; i++) {
          data.push({vector: Array(1536).fill(i), id: `${i}`, content: "", longId: `${i}`},)
      }
-     const table = await db.createTable('vectors', data)
+     const table = await db.createTable('my_vectors', data)
-     await table.create_index({ type: 'ivf_pq', column: 'vector', num_partitions: 256, num_sub_vectors: 96 })
+     await table.createIndex({ type: 'ivf_pq', column: 'vector', num_partitions: 256, num_sub_vectors: 96 })
      ```
 
-Since `create_index` has a training step, it can take a few minutes to finish for large tables. You can control the index
+- **metric** (default: "L2"): The distance metric to use. By default it uses euclidean distance "`L2`".
-creation by providing the following parameters:
+We also support "cosine" and "dot" distance as well.
+- **num_partitions** (default: 256): The number of partitions of the index.
+- **num_sub_vectors** (default: 96): The number of sub-vectors (M) that will be created during Product Quantization (PQ).
+For D dimensional vector, it will be divided into `M` of `D/M` sub-vectors, each of which is presented by
+a single PQ code.
+
+<figure markdown>
+  ![IVF PQ](./assets/ivf_pq.png)
+  <figcaption>IVF_PQ index with <code>num_partitions=2, num_sub_vectors=4</code></figcaption>
+</figure>
 
-- **metric** (default: "L2"): The distance metric to use. By default we use euclidean distance. We also support "cosine" distance.
-- **num_partitions** (default: 256): The number of partitions of the index. The number of partitions should be configured so each partition has 3-5K vectors. For example, a table
-with ~1M vectors should use 256 partitions. You can specify arbitrary number of partitions but powers of 2 is most conventional.
-A higher number leads to faster queries, but it makes index generation slower.
-- **num_sub_vectors** (default: 96): The number of subvectors (M) that will be created during Product Quantization (PQ). A larger number makes
-search more accurate, but also makes the index larger and slower to build.
 
 ## Querying an ANN Index
 
@@ -67,18 +81,19 @@ There are a couple of parameters that can be used to fine-tune the search:
   e.g., for 1M vectors divided up into 256 partitions, nprobes should be set to ~20-40.<br/>
   Note: nprobes is only applicable if an ANN index is present. If specified on a table without an ANN index, it is ignored.
 - **refine_factor** (default: None): Refine the results by reading extra elements and re-ranking them in memory.<br/>
-  A higher number makes search more accurate but also slower. If you find the recall is less than idea, try refine_factor=10 to start.<br/>
+  A higher number makes search more accurate but also slower. If you find the recall is less than ideal, try refine_factor=10 to start.<br/>
   e.g., for 1M vectors divided into 256 partitions, if you're looking for top 20, then refine_factor=200 reranks the whole partition.<br/>
   Note: refine_factor is only applicable if an ANN index is present. If specified on a table without an ANN index, it is ignored.
 
 === "Python"
      ```python
-     tbl.search(np.random.random((768))) \
+     tbl.search(np.random.random((1536))) \
          .limit(2) \
          .nprobes(20) \
          .refine_factor(10) \
          .to_df()
-
+     ```
+     ```
                                               vector       item       score
      0  [0.44949695, 0.8444449, 0.06281311, 0.23338133...  item 1141  103.575333
      1  [0.48587373, 0.269207, 0.15095535, 0.65531915,...  item 3953  108.393867
@@ -86,8 +101,8 @@ There are a couple of parameters that can be used to fine-tune the search:
 
 === "Javascript"
      ```javascript
-     const results = await table
+     const results_1 = await table
-         .search(Array(768).fill(1.2))
+         .search(Array(1536).fill(1.2))
          .limit(2)
          .nprobes(20)
          .refineFactor(10)
@@ -104,14 +119,14 @@ You can further filter the elements returned by a search using a where clause.
 
 === "Python"
      ```python
-     tbl.search(np.random.random((768))).where("item != 'item 1141'").to_df()
+     tbl.search(np.random.random((1536))).where("item != 'item 1141'").to_df()
      ```
 
 === "Javascript"
      ```javascript
-     const results = await table
+     const results_2 = await table
          .search(Array(1536).fill(1.2))
-         .where("item != 'item 1141'")
+         .where("id != '1141'")
          .execute()
      ```
 
@@ -121,7 +136,9 @@ You can select the columns returned by the query using a select clause.
 
 === "Python"
      ```python
-     tbl.search(np.random.random((768))).select(["vector"]).to_df()
+     tbl.search(np.random.random((1536))).select(["vector"]).to_df()
+     ```
+     ```
      vector      score
      0  [0.30928212, 0.022668175, 0.1756372, 0.4911822...  93.971092
      1  [0.2525465, 0.01723831, 0.261568, 0.002007689,...  95.173485
@@ -130,8 +147,36 @@ You can select the columns returned by the query using a select clause.
 
 === "Javascript"
      ```javascript
-     const results = await table
+     const results_3 = await table
          .search(Array(1536).fill(1.2))
          .select(["id"])
          .execute()
      ```
+
+## FAQ
+
+### When is it necessary to create an ANN vector index.
+
+`LanceDB` has manually tuned SIMD code for computing vector distances.
+In our benchmarks, computing 100K pairs of 1K dimension vectors only take less than 20ms.
+For small dataset (<100K rows) or the applications which can accept 100ms latency, vector indices are usually not necessary.
+
+For large-scale or higher dimension vectors, it is beneficial to create vector index.
+
+### How big is my index, and how many memory will it take.
+
+In LanceDB, all vector indices are disk-based, meaning that when responding to a vector query, only the relevant pages from the index file are loaded from disk and cached in memory. Additionally, each sub-vector is usually encoded into 1 byte PQ code.
+
+For example, with a 1024-dimension dataset, if we choose `num_sub_vectors=64`, each sub-vector has `1024 / 64 = 16` float32 numbers.
+Product quantization can lead to approximately `16 * sizeof(float32) / 1 = 64` times of space reduction.
+
+### How to choose `num_partitions` and `num_sub_vectors` for `IVF_PQ` index.
+
+`num_partitions` is used to decide how many partitions the first level `IVF` index uses.
+Higher number of partitions could lead to more efficient I/O during queries and better accuracy, but it takes much more time to train.
+On `SIFT-1M` dataset, our benchmark shows that keeping each partition 1K-4K rows lead to a good latency / recall.
+
+`num_sub_vectors` decides how many Product Quantization code to generate on each vector. Because
+Product Quantization is a lossy compression of the original vector, the more `num_sub_vectors` usually results to
+less space distortion, and thus yield better accuracy. However, similarly, more `num_sub_vectors` causes heavier I/O and
+more PQ computation, thus, higher latency. `dimension / num_sub_vectors` should be aligned with 8 for better SIMD efficiency.

docs/src/basic.md

@@ -23,7 +23,7 @@ We'll cover the basics of using LanceDB on your local machine in this section.
 === "Python"
       ```python
       import lancedb
-      uri = "~/.lancedb"
+      uri = "data/sample-lancedb"
       db = lancedb.connect(uri)
       ```
 
@@ -35,7 +35,7 @@ We'll cover the basics of using LanceDB on your local machine in this section.
       ```javascript
       const lancedb = require("vectordb");
 
-      const uri = "~./lancedb";
+      const uri = "data/sample-lancedb";
       const db = await lancedb.connect(uri);
       ```
       
@@ -102,7 +102,7 @@ Once created, you can open a table using the following code:
       If you forget the name of your table, you can always get a listing of all table names:
 
       ```javascript
-      console.log(db.tableNames());
+      console.log(await db.tableNames());
       ```
 
 ## How to add data to a table
@@ -118,7 +118,7 @@ After a table has been created, you can always add more data to it using
 
 === "Javascript"
       ```javascript
-      await tbl.add([vector: [1.3, 1.4], item: "fizz", price: 100.0},
+      await tbl.add([{vector: [1.3, 1.4], item: "fizz", price: 100.0},
               {vector: [9.5, 56.2], item: "buzz", price: 200.0}])
       ```
 

docs/src/embedding.md

@@ -98,7 +98,7 @@ You can also use an external API like OpenAI to generate embeddings
      embededings for your data.
 
       ```javascript
-      const db = await lancedb.connect("/tmp/lancedb");
+      const db = await lancedb.connect("data/sample-lancedb");
       const data = [
         { text: 'pepperoni'  },
         { text: 'pineapple' }
@@ -126,7 +126,7 @@ belong in the same latent space and your results will be nonsensical.
 === "Javascript"
      ```javascript
       const results = await table
-        .search('What's the best pizza topping?')
+        .search("What's the best pizza topping?")
         .limit(10)
         .execute()
      ```

docs/src/examples/modal_langchain.py

@@ -1,18 +1,19 @@
-import sys
-from modal import Secret, Stub, Image, web_endpoint
-import lancedb
-import re
 import pickle
-import requests
+import re
+import sys
 import zipfile
 from pathlib import Path
 
+import requests
+from langchain.chains import RetrievalQA
 from langchain.document_loaders import UnstructuredHTMLLoader
 from langchain.embeddings import OpenAIEmbeddings
+from langchain.llms import OpenAI
 from langchain.text_splitter import RecursiveCharacterTextSplitter
 from langchain.vectorstores import LanceDB
-from langchain.llms import OpenAI
+from modal import Image, Secret, Stub, web_endpoint
-from langchain.chains import RetrievalQA
+
+import lancedb
 
 lancedb_image = Image.debian_slim().pip_install(
     "lancedb", "langchain", "openai", "pandas", "tiktoken", "unstructured", "tabulate"
@@ -78,10 +79,7 @@ def qanda_langchain(query):
     download_docs()
     docs = store_docs()
 
-    text_splitter = RecursiveCharacterTextSplitter(
+    text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=200,)
-        chunk_size=1000,
-        chunk_overlap=200,
-    )
     documents = text_splitter.split_documents(docs)
     embeddings = OpenAIEmbeddings()
 

docs/src/examples/transformerjs_embedding_search_nodejs.md

@@ -0,0 +1,121 @@
+# Vector embedding search using TransformersJS
+
+## Embed and query data from LacneDB using TransformersJS
+
+<img id="splash" width="400" alt="transformersjs" src="https://github.com/lancedb/lancedb/assets/43097991/88a31e30-3d6f-4eef-9216-4b7c688f1b4f">
+
+This example shows how to use the [transformers.js](https://github.com/xenova/transformers.js) library to perform vector embedding search using LanceDB's Javascript API.
+
+
+### Setting up
+First, install the dependencies:
+```bash
+npm install vectordb
+npm i @xenova/transformers
+```
+
+We will also be using the [all-MiniLM-L6-v2](https://huggingface.co/Xenova/all-MiniLM-L6-v2) model to make it compatible with Transformers.js
+
+Within our `index.js` file we will import the necessary libraries and define our model and database:
+
+```javascript
+const lancedb = require('vectordb')
+const { pipeline } = await import('@xenova/transformers')
+const pipe = await pipeline('feature-extraction', 'Xenova/all-MiniLM-L6-v2');
+```
+
+### Creating the embedding function
+
+Next, we will create a function that will take in a string and return the vector embedding of that string. We will use the `pipe` function we defined earlier to get the vector embedding of the string.
+
+```javascript
+// Define the function. `sourceColumn` is required for LanceDB to know
+// which column to use as input.
+const embed_fun = {}
+embed_fun.sourceColumn = 'text'
+embed_fun.embed = async function (batch) {
+    let result = []
+    // Given a batch of strings, we will use the `pipe` function to get
+    // the vector embedding of each string.
+    for (let text of batch) {
+        // 'mean' pooling and normalizing allows the embeddings to share the
+        // same length.
+        const res = await pipe(text, { pooling: 'mean', normalize: true })
+        result.push(Array.from(res['data']))
+    }
+    return (result)
+}
+```
+
+### Creating the database
+
+Now, we will create the LanceDB database and add the embedding function we defined earlier.
+
+```javascript
+// Link a folder and create a table with data
+const db = await lancedb.connect('data/sample-lancedb')
+
+// You can also import any other data, but make sure that you have a column
+// for the embedding function to use.
+const data = [
+    { id: 1, text: 'Cherry', type: 'fruit' },
+    { id: 2, text: 'Carrot', type: 'vegetable' },
+    { id: 3, text: 'Potato', type: 'vegetable' },
+    { id: 4, text: 'Apple', type: 'fruit' },
+    { id: 5, text: 'Banana', type: 'fruit' }
+]
+
+// Create the table with the embedding function
+const table = await db.createTable('food_table', data, "create", embed_fun)
+```
+
+### Performing the search
+
+Now, we can perform the search using the `search` function. LanceDB automatically uses the embedding function we defined earlier to get the vector embedding of the query string.
+
+```javascript
+// Query the table
+const results = await table
+    .search("a sweet fruit to eat")
+    .metricType("cosine")
+    .limit(2)
+    .execute()
+console.log(results.map(r => r.text))
+```
+```bash
+[ 'Banana', 'Cherry' ]
+```
+
+Output of `results`:
+```bash
+[
+  {
+    vector: Float32Array(384) [
+      -0.057455405592918396,
+      0.03617725893855095,
+      -0.0367760956287384,
+      ... 381 more items
+    ],
+    id: 5,
+    text: 'Banana',
+    type: 'fruit',
+    score: 0.4919965863227844
+  },
+  {
+    vector: Float32Array(384) [
+      0.0009714411571621895,
+      0.008223623037338257,
+      0.009571489877998829,
+      ... 381 more items
+    ],
+    id: 1,
+    text: 'Cherry',
+    type: 'fruit',
+    score: 0.5540297031402588
+  }
+]
+```
+
+### Wrapping it up
+
+In this example, we showed how to use the `transformers.js` library to perform vector embedding search using LanceDB's Javascript API. You can find the full code for this example on [Github](https://github.com/lancedb/lancedb/blob/main/node/examples/js-transformers/index.js)!

docs/src/fts.md

@@ -18,6 +18,20 @@ Assume:
 1. `table` is a LanceDB Table
 2. `text` is the name of the Table column that we want to index
 
+For example,
+
+```python
+import lancedb
+
+uri = "data/sample-lancedb"
+db = lancedb.connect(uri)
+
+table = db.create_table("my_table",
+            data=[{"vector": [3.1, 4.1], "text": "Frodo was a happy puppy"},
+                  {"vector": [5.9, 26.5], "text": "There are several kittens playing"}])
+
+```
+
 To create the index:
 
 ```python

docs/src/index.md

@@ -14,7 +14,7 @@ The key features of LanceDB include:
 
 * Zero-copy, automatic versioning, manage versions of your data without needing extra infrastructure.
 
-* Ecosystem integrations with [LangChain 🦜️🔗](https://python.langchain.com/en/latest/modules/indexes/vectorstores/examples/lanecdb.html), [LlamaIndex 🦙](https://gpt-index.readthedocs.io/en/latest/examples/vector_stores/LanceDBIndexDemo.html), Apache-Arrow, Pandas, Polars, DuckDB and more on the way.
+* Ecosystem integrations with [LangChain 🦜️🔗](https://python.langchain.com/en/latest/modules/indexes/vectorstores/examples/lancedb.html), [LlamaIndex 🦙](https://gpt-index.readthedocs.io/en/latest/examples/vector_stores/LanceDBIndexDemo.html), Apache-Arrow, Pandas, Polars, DuckDB and more on the way.
 
 LanceDB's core is written in Rust 🦀 and is built using <a href="https://github.com/lancedb/lance">Lance</a>, an open-source columnar format designed for performant ML workloads.
 
@@ -28,7 +28,7 @@ LanceDB's core is written in Rust 🦀 and is built using <a href="https://githu
       ```python
       import lancedb
 
-      uri = "/tmp/lancedb"
+      uri = "data/sample-lancedb"
       db = lancedb.connect(uri)
       table = db.create_table("my_table",
                               data=[{"vector": [3.1, 4.1], "item": "foo", "price": 10.0},
@@ -44,7 +44,7 @@ LanceDB's core is written in Rust 🦀 and is built using <a href="https://githu
       ```javascript
       const lancedb = require("vectordb");
 
-      const uri = "/tmp/lancedb";
+      const uri = "data/sample-lancedb";
       const db = await lancedb.connect(uri);
       const table = await db.createTable("my_table",
             [{ id: 1, vector: [3.1, 4.1], item: "foo", price: 10.0 },
@@ -67,6 +67,6 @@ LanceDB's core is written in Rust 🦀 and is built using <a href="https://githu
 * [`Embedding Functions`](embedding.md) - functions for working with embeddings.
 * [`Indexing`](ann_indexes.md) - create vector indexes to speed up queries.
 * [`Full text search`](fts.md) - [EXPERIMENTAL] full-text search API
-* [`Ecosystem Integrations`](integrations.md) - integrating LanceDB with python data tooling ecosystem.
+* [`Ecosystem Integrations`](python/integration.md) - integrating LanceDB with python data tooling ecosystem.
 * [`Python API Reference`](python/python.md) - detailed documentation for the LanceDB Python SDK.
 * [`Node API Reference`](javascript/modules.md) - detailed documentation for the LanceDB Python SDK.

docs/src/integrations.md

@@ -1,108 +0,0 @@
-# Integrations
-
-Built on top of Apache Arrow, `LanceDB` is easy to integrate with the Python ecosystem, including Pandas, PyArrow and DuckDB.
-
-## Pandas and PyArrow
-
-First, we need to connect to a `LanceDB` database.
-
-``` py
-
-import lancedb
-
-db = lancedb.connect("/tmp/lancedb")
-```
-
-And write a `Pandas DataFrame` to LanceDB directly.
-
-```py
-import pandas as pd
-
-data = pd.DataFrame({
-    "vector": [[3.1, 4.1], [5.9, 26.5]],
-    "item": ["foo", "bar"],
-    "price": [10.0, 20.0]
-})
-table = db.create_table("pd_table", data=data)
-```
-
-You will find detailed instructions of creating dataset and index in [Basic Operations](basic.md) and [Indexing](indexing.md)
-sections.
-
-
-We can now perform similarity searches via `LanceDB`.
-
-```py
-# Open the table previously created.
-table = db.open_table("pd_table")
-
-query_vector = [100, 100]
-# Pandas DataFrame
-df = table.search(query_vector).limit(1).to_df()
-print(df)
-```
-
-```
-    vector     item  price        score
-0  [5.9, 26.5]  bar   20.0  14257.05957
-```
-
-If you have a simple filter, it's faster to provide a where clause to `LanceDB`'s search query.
-If you have more complex criteria, you can always apply the filter to the resulting pandas `DataFrame` from the search query.
-
-```python
-
-# Apply the filter via LanceDB
-results = table.search([100, 100]).where("price < 15").to_df()
-assert len(results) == 1
-assert results["item"].iloc[0] == "foo"
-
-# Apply the filter via Pandas
-df = results = table.search([100, 100]).to_df()
-results = df[df.price < 15]
-assert len(results) == 1
-assert results["item"].iloc[0] == "foo"
-```
-
-## DuckDB
-
-`LanceDB` works with `DuckDB` via [PyArrow integration](https://duckdb.org/docs/guides/python/sql_on_arrow).
-
-Let us start with installing `duckdb` and `lancedb`.
-
-```shell
-pip install duckdb lancedb
-```
-
-We will re-use the dataset created previously
-
-```python
-import lancedb
-
-db = lancedb.connect("/tmp/lancedb")
-table = db.open_table("pd_table")
-arrow_table = table.to_arrow()
-```
-
-`DuckDB` can directly query the `arrow_table`:
-
-```python
-In [15]: duckdb.query("SELECT * FROM t")
-Out[15]:
-┌─────────────┬─────────┬────────┐
-│   vector    │  item   │ price  │
-│   float[]   │ varchar │ double │
-├─────────────┼─────────┼────────┤
-│ [3.1, 4.1]  │ foo     │   10.0 │
-│ [5.9, 26.5] │ bar     │   20.0 │
-└─────────────┴─────────┴────────┘
-
-In [16]: duckdb.query("SELECT mean(price) FROM t")
-Out[16]:
-┌─────────────┐
-│ mean(price) │
-│   double    │
-├─────────────┤
-│        15.0 │
-└─────────────┘
-```

docs/src/javascript/README.md

@@ -16,9 +16,11 @@ npm install vectordb
 
 ```javascript
 const lancedb = require('vectordb');
-const db = lancedb.connect('<PATH_TO_LANCEDB_DATASET>');
+const db = await lancedb.connect('data/sample-lancedb');
-const table = await db.openTable('my_table');
+const table = await db.createTable("my_table",
-const query = await table.search([0.1, 0.3]).setLimit(20).execute();
+      [{ id: 1, vector: [0.1, 1.0], item: "foo", price: 10.0 },
+      { id: 2, vector: [3.9, 0.5], item: "bar", price: 20.0 }])
+const results = await table.search([0.1, 0.3]).limit(20).execute();
 console.log(results);
 ```
 
@@ -26,12 +28,6 @@ The [examples](./examples) folder contains complete examples.
 
 ## Development
 
-The LanceDB javascript is built with npm:
-
-```bash
-npm run tsc
-```
-
 Run the tests with
 
 ```bash

docs/src/javascript/classes/Connection.md

@@ -1,211 +0,0 @@
-[vectordb](../README.md) / [Exports](../modules.md) / Connection
-
-# Class: Connection
-
-A connection to a LanceDB database.
-
-## Table of contents
-
-### Constructors
-
-- [constructor](Connection.md#constructor)
-
-### Properties
-
-- [\_db](Connection.md#_db)
-- [\_uri](Connection.md#_uri)
-
-### Accessors
-
-- [uri](Connection.md#uri)
-
-### Methods
-
-- [createTable](Connection.md#createtable)
-- [createTableArrow](Connection.md#createtablearrow)
-- [openTable](Connection.md#opentable)
-- [tableNames](Connection.md#tablenames)
-
-## Constructors
-
-### constructor
-
-• **new Connection**(`db`, `uri`)
-
-#### Parameters
-
-| Name | Type |
-| :------ | :------ |
-| `db` | `any` |
-| `uri` | `string` |
-
-#### Defined in
-
-[index.ts:46](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L46)
-
-## Properties
-
-### \_db
-
-• `Private` `Readonly` **\_db**: `any`
-
-#### Defined in
-
-[index.ts:44](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L44)
-
-___
-
-### \_uri
-
-• `Private` `Readonly` **\_uri**: `string`
-
-#### Defined in
-
-[index.ts:43](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L43)
-
-## Accessors
-
-### uri
-
-• `get` **uri**(): `string`
-
-#### Returns
-
-`string`
-
-#### Defined in
-
-[index.ts:51](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L51)
-
-## Methods
-
-### createTable
-
-▸ **createTable**(`name`, `data`): `Promise`<[`Table`](Table.md)<`number`[]\>\>
-
-Creates a new Table and initialize it with new data.
-
-#### Parameters
-
-| Name | Type | Description |
-| :------ | :------ | :------ |
-| `name` | `string` | The name of the table. |
-| `data` | `Record`<`string`, `unknown`\>[] | Non-empty Array of Records to be inserted into the Table |
-
-#### Returns
-
-`Promise`<[`Table`](Table.md)<`number`[]\>\>
-
-#### Defined in
-
-[index.ts:91](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L91)
-
-▸ **createTable**<`T`\>(`name`, `data`, `embeddings`): `Promise`<[`Table`](Table.md)<`T`\>\>
-
-Creates a new Table and initialize it with new data.
-
-#### Type parameters
-
-| Name |
-| :------ |
-| `T` |
-
-#### Parameters
-
-| Name | Type | Description |
-| :------ | :------ | :------ |
-| `name` | `string` | The name of the table. |
-| `data` | `Record`<`string`, `unknown`\>[] | Non-empty Array of Records to be inserted into the Table |
-| `embeddings` | [`EmbeddingFunction`](../interfaces/EmbeddingFunction.md)<`T`\> | An embedding function to use on this Table |
-
-#### Returns
-
-`Promise`<[`Table`](Table.md)<`T`\>\>
-
-#### Defined in
-
-[index.ts:99](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L99)
-
-___
-
-### createTableArrow
-
-▸ **createTableArrow**(`name`, `table`): `Promise`<[`Table`](Table.md)<`number`[]\>\>
-
-#### Parameters
-
-| Name | Type |
-| :------ | :------ |
-| `name` | `string` |
-| `table` | `Table`<`any`\> |
-
-#### Returns
-
-`Promise`<[`Table`](Table.md)<`number`[]\>\>
-
-#### Defined in
-
-[index.ts:109](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L109)
-
-___
-
-### openTable
-
-▸ **openTable**(`name`): `Promise`<[`Table`](Table.md)<`number`[]\>\>
-
-Open a table in the database.
-
-#### Parameters
-
-| Name | Type | Description |
-| :------ | :------ | :------ |
-| `name` | `string` | The name of the table. |
-
-#### Returns
-
-`Promise`<[`Table`](Table.md)<`number`[]\>\>
-
-#### Defined in
-
-[index.ts:67](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L67)
-
-▸ **openTable**<`T`\>(`name`, `embeddings`): `Promise`<[`Table`](Table.md)<`T`\>\>
-
-Open a table in the database.
-
-#### Type parameters
-
-| Name |
-| :------ |
-| `T` |
-
-#### Parameters
-
-| Name | Type | Description |
-| :------ | :------ | :------ |
-| `name` | `string` | The name of the table. |
-| `embeddings` | [`EmbeddingFunction`](../interfaces/EmbeddingFunction.md)<`T`\> | An embedding function to use on this Table |
-
-#### Returns
-
-`Promise`<[`Table`](Table.md)<`T`\>\>
-
-#### Defined in
-
-[index.ts:74](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L74)
-
-___
-
-### tableNames
-
-▸ **tableNames**(): `Promise`<`string`[]\>
-
-Get the names of all tables in the database.
-
-#### Returns
-
-`Promise`<`string`[]\>
-
-#### Defined in
-
-[index.ts:58](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L58)

docs/src/javascript/classes/LocalConnection.md

@@ -0,0 +1,294 @@
+[vectordb](../README.md) / [Exports](../modules.md) / LocalConnection
+
+# Class: LocalConnection
+
+A connection to a LanceDB database.
+
+## Implements
+
+- [`Connection`](../interfaces/Connection.md)
+
+## Table of contents
+
+### Constructors
+
+- [constructor](LocalConnection.md#constructor)
+
+### Properties
+
+- [\_db](LocalConnection.md#_db)
+- [\_uri](LocalConnection.md#_uri)
+
+### Accessors
+
+- [uri](LocalConnection.md#uri)
+
+### Methods
+
+- [createTable](LocalConnection.md#createtable)
+- [createTableArrow](LocalConnection.md#createtablearrow)
+- [dropTable](LocalConnection.md#droptable)
+- [openTable](LocalConnection.md#opentable)
+- [tableNames](LocalConnection.md#tablenames)
+
+## Constructors
+
+### constructor
+
+• **new LocalConnection**(`db`, `uri`)
+
+#### Parameters
+
+| Name | Type |
+| :------ | :------ |
+| `db` | `any` |
+| `uri` | `string` |
+
+#### Defined in
+
+[index.ts:132](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L132)
+
+## Properties
+
+### \_db
+
+• `Private` `Readonly` **\_db**: `any`
+
+#### Defined in
+
+[index.ts:130](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L130)
+
+___
+
+### \_uri
+
+• `Private` `Readonly` **\_uri**: `string`
+
+#### Defined in
+
+[index.ts:129](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L129)
+
+## Accessors
+
+### uri
+
+• `get` **uri**(): `string`
+
+#### Returns
+
+`string`
+
+#### Implementation of
+
+[Connection](../interfaces/Connection.md).[uri](../interfaces/Connection.md#uri)
+
+#### Defined in
+
+[index.ts:137](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L137)
+
+## Methods
+
+### createTable
+
+▸ **createTable**(`name`, `data`, `mode?`): `Promise`<[`Table`](../interfaces/Table.md)<`number`[]\>\>
+
+Creates a new Table and initialize it with new data.
+
+#### Parameters
+
+| Name | Type | Description |
+| :------ | :------ | :------ |
+| `name` | `string` | The name of the table. |
+| `data` | `Record`<`string`, `unknown`\>[] | Non-empty Array of Records to be inserted into the Table |
+| `mode?` | [`WriteMode`](../enums/WriteMode.md) | The write mode to use when creating the table. |
+
+#### Returns
+
+`Promise`<[`Table`](../interfaces/Table.md)<`number`[]\>\>
+
+#### Implementation of
+
+[Connection](../interfaces/Connection.md).[createTable](../interfaces/Connection.md#createtable)
+
+#### Defined in
+
+[index.ts:177](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L177)
+
+▸ **createTable**(`name`, `data`, `mode`): `Promise`<[`Table`](../interfaces/Table.md)<`number`[]\>\>
+
+#### Parameters
+
+| Name | Type |
+| :------ | :------ |
+| `name` | `string` |
+| `data` | `Record`<`string`, `unknown`\>[] |
+| `mode` | [`WriteMode`](../enums/WriteMode.md) |
+
+#### Returns
+
+`Promise`<[`Table`](../interfaces/Table.md)<`number`[]\>\>
+
+#### Implementation of
+
+Connection.createTable
+
+#### Defined in
+
+[index.ts:178](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L178)
+
+▸ **createTable**<`T`\>(`name`, `data`, `mode`, `embeddings`): `Promise`<[`Table`](../interfaces/Table.md)<`T`\>\>
+
+Creates a new Table and initialize it with new data.
+
+#### Type parameters
+
+| Name |
+| :------ |
+| `T` |
+
+#### Parameters
+
+| Name | Type | Description |
+| :------ | :------ | :------ |
+| `name` | `string` | The name of the table. |
+| `data` | `Record`<`string`, `unknown`\>[] | Non-empty Array of Records to be inserted into the Table |
+| `mode` | [`WriteMode`](../enums/WriteMode.md) | The write mode to use when creating the table. |
+| `embeddings` | [`EmbeddingFunction`](../interfaces/EmbeddingFunction.md)<`T`\> | An embedding function to use on this Table |
+
+#### Returns
+
+`Promise`<[`Table`](../interfaces/Table.md)<`T`\>\>
+
+#### Implementation of
+
+Connection.createTable
+
+#### Defined in
+
+[index.ts:188](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L188)
+
+___
+
+### createTableArrow
+
+▸ **createTableArrow**(`name`, `table`): `Promise`<[`Table`](../interfaces/Table.md)<`number`[]\>\>
+
+#### Parameters
+
+| Name | Type |
+| :------ | :------ |
+| `name` | `string` |
+| `table` | `Table`<`any`\> |
+
+#### Returns
+
+`Promise`<[`Table`](../interfaces/Table.md)<`number`[]\>\>
+
+#### Implementation of
+
+[Connection](../interfaces/Connection.md).[createTableArrow](../interfaces/Connection.md#createtablearrow)
+
+#### Defined in
+
+[index.ts:201](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L201)
+
+___
+
+### dropTable
+
+▸ **dropTable**(`name`): `Promise`<`void`\>
+
+Drop an existing table.
+
+#### Parameters
+
+| Name | Type | Description |
+| :------ | :------ | :------ |
+| `name` | `string` | The name of the table to drop. |
+
+#### Returns
+
+`Promise`<`void`\>
+
+#### Implementation of
+
+[Connection](../interfaces/Connection.md).[dropTable](../interfaces/Connection.md#droptable)
+
+#### Defined in
+
+[index.ts:211](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L211)
+
+___
+
+### openTable
+
+▸ **openTable**(`name`): `Promise`<[`Table`](../interfaces/Table.md)<`number`[]\>\>
+
+Open a table in the database.
+
+#### Parameters
+
+| Name | Type | Description |
+| :------ | :------ | :------ |
+| `name` | `string` | The name of the table. |
+
+#### Returns
+
+`Promise`<[`Table`](../interfaces/Table.md)<`number`[]\>\>
+
+#### Implementation of
+
+[Connection](../interfaces/Connection.md).[openTable](../interfaces/Connection.md#opentable)
+
+#### Defined in
+
+[index.ts:153](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L153)
+
+▸ **openTable**<`T`\>(`name`, `embeddings`): `Promise`<[`Table`](../interfaces/Table.md)<`T`\>\>
+
+Open a table in the database.
+
+#### Type parameters
+
+| Name |
+| :------ |
+| `T` |
+
+#### Parameters
+
+| Name | Type | Description |
+| :------ | :------ | :------ |
+| `name` | `string` | The name of the table. |
+| `embeddings` | [`EmbeddingFunction`](../interfaces/EmbeddingFunction.md)<`T`\> | An embedding function to use on this Table |
+
+#### Returns
+
+`Promise`<[`Table`](../interfaces/Table.md)<`T`\>\>
+
+#### Implementation of
+
+Connection.openTable
+
+#### Defined in
+
+[index.ts:160](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L160)
+
+___
+
+### tableNames
+
+▸ **tableNames**(): `Promise`<`string`[]\>
+
+Get the names of all tables in the database.
+
+#### Returns
+
+`Promise`<`string`[]\>
+
+#### Implementation of
+
+[Connection](../interfaces/Connection.md).[tableNames](../interfaces/Connection.md#tablenames)
+
+#### Defined in
+
+[index.ts:144](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L144)

docs/src/javascript/classes/LocalTable.md

@@ -0,0 +1,289 @@
+[vectordb](../README.md) / [Exports](../modules.md) / LocalTable
+
+# Class: LocalTable<T\>
+
+A LanceDB Table is the collection of Records. Each Record has one or more vector fields.
+
+## Type parameters
+
+| Name | Type |
+| :------ | :------ |
+| `T` | `number`[] |
+
+## Implements
+
+- [`Table`](../interfaces/Table.md)<`T`\>
+
+## Table of contents
+
+### Constructors
+
+- [constructor](LocalTable.md#constructor)
+
+### Properties
+
+- [\_embeddings](LocalTable.md#_embeddings)
+- [\_name](LocalTable.md#_name)
+- [\_tbl](LocalTable.md#_tbl)
+
+### Accessors
+
+- [name](LocalTable.md#name)
+
+### Methods
+
+- [add](LocalTable.md#add)
+- [countRows](LocalTable.md#countrows)
+- [createIndex](LocalTable.md#createindex)
+- [delete](LocalTable.md#delete)
+- [overwrite](LocalTable.md#overwrite)
+- [search](LocalTable.md#search)
+
+## Constructors
+
+### constructor
+
+• **new LocalTable**<`T`\>(`tbl`, `name`)
+
+#### Type parameters
+
+| Name | Type |
+| :------ | :------ |
+| `T` | `number`[] |
+
+#### Parameters
+
+| Name | Type |
+| :------ | :------ |
+| `tbl` | `any` |
+| `name` | `string` |
+
+#### Defined in
+
+[index.ts:221](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L221)
+
+• **new LocalTable**<`T`\>(`tbl`, `name`, `embeddings`)
+
+#### Type parameters
+
+| Name | Type |
+| :------ | :------ |
+| `T` | `number`[] |
+
+#### Parameters
+
+| Name | Type | Description |
+| :------ | :------ | :------ |
+| `tbl` | `any` |  |
+| `name` | `string` |  |
+| `embeddings` | [`EmbeddingFunction`](../interfaces/EmbeddingFunction.md)<`T`\> | An embedding function to use when interacting with this table |
+
+#### Defined in
+
+[index.ts:227](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L227)
+
+## Properties
+
+### \_embeddings
+
+• `Private` `Optional` `Readonly` **\_embeddings**: [`EmbeddingFunction`](../interfaces/EmbeddingFunction.md)<`T`\>
+
+#### Defined in
+
+[index.ts:219](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L219)
+
+___
+
+### \_name
+
+• `Private` `Readonly` **\_name**: `string`
+
+#### Defined in
+
+[index.ts:218](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L218)
+
+___
+
+### \_tbl
+
+• `Private` `Readonly` **\_tbl**: `any`
+
+#### Defined in
+
+[index.ts:217](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L217)
+
+## Accessors
+
+### name
+
+• `get` **name**(): `string`
+
+#### Returns
+
+`string`
+
+#### Implementation of
+
+[Table](../interfaces/Table.md).[name](../interfaces/Table.md#name)
+
+#### Defined in
+
+[index.ts:234](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L234)
+
+## Methods
+
+### add
+
+▸ **add**(`data`): `Promise`<`number`\>
+
+Insert records into this Table.
+
+#### Parameters
+
+| Name | Type | Description |
+| :------ | :------ | :------ |
+| `data` | `Record`<`string`, `unknown`\>[] | Records to be inserted into the Table |
+
+#### Returns
+
+`Promise`<`number`\>
+
+The number of rows added to the table
+
+#### Implementation of
+
+[Table](../interfaces/Table.md).[add](../interfaces/Table.md#add)
+
+#### Defined in
+
+[index.ts:252](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L252)
+
+___
+
+### countRows
+
+▸ **countRows**(): `Promise`<`number`\>
+
+Returns the number of rows in this table.
+
+#### Returns
+
+`Promise`<`number`\>
+
+#### Implementation of
+
+[Table](../interfaces/Table.md).[countRows](../interfaces/Table.md#countrows)
+
+#### Defined in
+
+[index.ts:278](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L278)
+
+___
+
+### createIndex
+
+▸ **createIndex**(`indexParams`): `Promise`<`any`\>
+
+Create an ANN index on this Table vector index.
+
+**`See`**
+
+VectorIndexParams.
+
+#### Parameters
+
+| Name | Type | Description |
+| :------ | :------ | :------ |
+| `indexParams` | `IvfPQIndexConfig` | The parameters of this Index, |
+
+#### Returns
+
+`Promise`<`any`\>
+
+#### Implementation of
+
+[Table](../interfaces/Table.md).[createIndex](../interfaces/Table.md#createindex)
+
+#### Defined in
+
+[index.ts:271](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L271)
+
+___
+
+### delete
+
+▸ **delete**(`filter`): `Promise`<`void`\>
+
+Delete rows from this table.
+
+#### Parameters
+
+| Name | Type | Description |
+| :------ | :------ | :------ |
+| `filter` | `string` | A filter in the same format used by a sql WHERE clause. |
+
+#### Returns
+
+`Promise`<`void`\>
+
+#### Implementation of
+
+[Table](../interfaces/Table.md).[delete](../interfaces/Table.md#delete)
+
+#### Defined in
+
+[index.ts:287](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L287)
+
+___
+
+### overwrite
+
+▸ **overwrite**(`data`): `Promise`<`number`\>
+
+Insert records into this Table, replacing its contents.
+
+#### Parameters
+
+| Name | Type | Description |
+| :------ | :------ | :------ |
+| `data` | `Record`<`string`, `unknown`\>[] | Records to be inserted into the Table |
+
+#### Returns
+
+`Promise`<`number`\>
+
+The number of rows added to the table
+
+#### Implementation of
+
+[Table](../interfaces/Table.md).[overwrite](../interfaces/Table.md#overwrite)
+
+#### Defined in
+
+[index.ts:262](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L262)
+
+___
+
+### search
+
+▸ **search**(`query`): [`Query`](Query.md)<`T`\>
+
+Creates a search query to find the nearest neighbors of the given search term
+
+#### Parameters
+
+| Name | Type | Description |
+| :------ | :------ | :------ |
+| `query` | `T` | The query search term |
+
+#### Returns
+
+[`Query`](Query.md)<`T`\>
+
+#### Implementation of
+
+[Table](../interfaces/Table.md).[search](../interfaces/Table.md#search)
+
+#### Defined in
+
+[index.ts:242](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L242)

docs/src/javascript/classes/OpenAIEmbeddingFunction.md

@@ -40,7 +40,7 @@ An embedding function that automatically creates vector representation for a giv
 
 #### Defined in
 
-[embedding/openai.ts:21](https://github.com/lancedb/lancedb/blob/31dab97/node/src/embedding/openai.ts#L21)
+[embedding/openai.ts:21](https://github.com/lancedb/lancedb/blob/7247834/node/src/embedding/openai.ts#L21)
 
 ## Properties
 
@@ -50,7 +50,7 @@ An embedding function that automatically creates vector representation for a giv
 
 #### Defined in
 
-[embedding/openai.ts:19](https://github.com/lancedb/lancedb/blob/31dab97/node/src/embedding/openai.ts#L19)
+[embedding/openai.ts:19](https://github.com/lancedb/lancedb/blob/7247834/node/src/embedding/openai.ts#L19)
 
 ___
 
@@ -60,7 +60,7 @@ ___
 
 #### Defined in
 
-[embedding/openai.ts:18](https://github.com/lancedb/lancedb/blob/31dab97/node/src/embedding/openai.ts#L18)
+[embedding/openai.ts:18](https://github.com/lancedb/lancedb/blob/7247834/node/src/embedding/openai.ts#L18)
 
 ___
 
@@ -76,7 +76,7 @@ The name of the column that will be used as input for the Embedding Function.
 
 #### Defined in
 
-[embedding/openai.ts:50](https://github.com/lancedb/lancedb/blob/31dab97/node/src/embedding/openai.ts#L50)
+[embedding/openai.ts:50](https://github.com/lancedb/lancedb/blob/7247834/node/src/embedding/openai.ts#L50)
 
 ## Methods
 
@@ -102,4 +102,4 @@ Creates a vector representation for the given values.
 
 #### Defined in
 
-[embedding/openai.ts:38](https://github.com/lancedb/lancedb/blob/31dab97/node/src/embedding/openai.ts#L38)
+[embedding/openai.ts:38](https://github.com/lancedb/lancedb/blob/7247834/node/src/embedding/openai.ts#L38)

docs/src/javascript/classes/Query.md

@@ -18,7 +18,6 @@ A builder for nearest neighbor queries for LanceDB.
 
 ### Properties
 
-- [\_columns](Query.md#_columns)
 - [\_embeddings](Query.md#_embeddings)
 - [\_filter](Query.md#_filter)
 - [\_limit](Query.md#_limit)
@@ -27,7 +26,9 @@ A builder for nearest neighbor queries for LanceDB.
 - [\_query](Query.md#_query)
 - [\_queryVector](Query.md#_queryvector)
 - [\_refineFactor](Query.md#_refinefactor)
+- [\_select](Query.md#_select)
 - [\_tbl](Query.md#_tbl)
+- [where](Query.md#where)
 
 ### Methods
 
@@ -37,6 +38,7 @@ A builder for nearest neighbor queries for LanceDB.
 - [metricType](Query.md#metrictype)
 - [nprobes](Query.md#nprobes)
 - [refineFactor](Query.md#refinefactor)
+- [select](Query.md#select)
 
 ## Constructors
 
@@ -60,27 +62,17 @@ A builder for nearest neighbor queries for LanceDB.
 
 #### Defined in
 
-[index.ts:241](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L241)
+[index.ts:362](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L362)
 
 ## Properties
 
-### \_columns
-
-• `Private` `Optional` `Readonly` **\_columns**: `string`[]
-
-#### Defined in
-
-[index.ts:236](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L236)
-
-___
-
 ### \_embeddings
 
 • `Private` `Optional` `Readonly` **\_embeddings**: [`EmbeddingFunction`](../interfaces/EmbeddingFunction.md)<`T`\>
 
 #### Defined in
 
-[index.ts:239](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L239)
+[index.ts:360](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L360)
 
 ___
 
@@ -90,7 +82,7 @@ ___
 
 #### Defined in
 
-[index.ts:237](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L237)
+[index.ts:358](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L358)
 
 ___
 
@@ -100,7 +92,7 @@ ___
 
 #### Defined in
 
-[index.ts:233](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L233)
+[index.ts:354](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L354)
 
 ___
 
@@ -110,7 +102,7 @@ ___
 
 #### Defined in
 
-[index.ts:238](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L238)
+[index.ts:359](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L359)
 
 ___
 
@@ -120,7 +112,7 @@ ___
 
 #### Defined in
 
-[index.ts:235](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L235)
+[index.ts:356](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L356)
 
 ___
 
@@ -130,7 +122,7 @@ ___
 
 #### Defined in
 
-[index.ts:231](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L231)
+[index.ts:352](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L352)
 
 ___
 
@@ -140,7 +132,7 @@ ___
 
 #### Defined in
 
-[index.ts:232](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L232)
+[index.ts:353](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L353)
 
 ___
 
@@ -150,7 +142,17 @@ ___
 
 #### Defined in
 
-[index.ts:234](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L234)
+[index.ts:355](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L355)
+
+___
+
+### \_select
+
+• `Private` `Optional` **\_select**: `string`[]
+
+#### Defined in
+
+[index.ts:357](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L357)
 
 ___
 
@@ -160,7 +162,33 @@ ___
 
 #### Defined in
 
-[index.ts:230](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L230)
+[index.ts:351](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L351)
+
+___
+
+### where
+
+• **where**: (`value`: `string`) => [`Query`](Query.md)<`T`\>
+
+#### Type declaration
+
+▸ (`value`): [`Query`](Query.md)<`T`\>
+
+A filter statement to be applied to this query.
+
+##### Parameters
+
+| Name | Type | Description |
+| :------ | :------ | :------ |
+| `value` | `string` | A filter in the same format used by a sql WHERE clause. |
+
+##### Returns
+
+[`Query`](Query.md)<`T`\>
+
+#### Defined in
+
+[index.ts:410](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L410)
 
 ## Methods
 
@@ -182,7 +210,7 @@ Execute the query and return the results as an Array of Objects
 
 #### Defined in
 
-[index.ts:301](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L301)
+[index.ts:433](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L433)
 
 ___
 
@@ -204,7 +232,7 @@ A filter statement to be applied to this query.
 
 #### Defined in
 
-[index.ts:284](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L284)
+[index.ts:405](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L405)
 
 ___
 
@@ -226,7 +254,7 @@ Sets the number of results that will be returned
 
 #### Defined in
 
-[index.ts:257](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L257)
+[index.ts:378](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L378)
 
 ___
 
@@ -252,7 +280,7 @@ MetricType for the different options
 
 #### Defined in
 
-[index.ts:293](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L293)
+[index.ts:425](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L425)
 
 ___
 
@@ -274,7 +302,7 @@ The number of probes used. A higher number makes search more accurate but also s
 
 #### Defined in
 
-[index.ts:275](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L275)
+[index.ts:396](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L396)
 
 ___
 
@@ -296,4 +324,26 @@ Refine the results by reading extra elements and re-ranking them in memory.
 
 #### Defined in
 
-[index.ts:266](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L266)
+[index.ts:387](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L387)
+
+___
+
+### select
+
+▸ **select**(`value`): [`Query`](Query.md)<`T`\>
+
+Return only the specified columns.
+
+#### Parameters
+
+| Name | Type | Description |
+| :------ | :------ | :------ |
+| `value` | `string`[] | Only select the specified columns. If not specified, all columns will be returned. |
+
+#### Returns
+
+[`Query`](Query.md)<`T`\>
+
+#### Defined in
+
+[index.ts:416](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L416)

docs/src/javascript/classes/Table.md

@@ -1,215 +0,0 @@
-[vectordb](../README.md) / [Exports](../modules.md) / Table
-
-# Class: Table<T\>
-
-## Type parameters
-
-| Name | Type |
-| :------ | :------ |
-| `T` | `number`[] |
-
-## Table of contents
-
-### Constructors
-
-- [constructor](Table.md#constructor)
-
-### Properties
-
-- [\_embeddings](Table.md#_embeddings)
-- [\_name](Table.md#_name)
-- [\_tbl](Table.md#_tbl)
-
-### Accessors
-
-- [name](Table.md#name)
-
-### Methods
-
-- [add](Table.md#add)
-- [create\_index](Table.md#create_index)
-- [overwrite](Table.md#overwrite)
-- [search](Table.md#search)
-
-## Constructors
-
-### constructor
-
-• **new Table**<`T`\>(`tbl`, `name`)
-
-#### Type parameters
-
-| Name | Type |
-| :------ | :------ |
-| `T` | `number`[] |
-
-#### Parameters
-
-| Name | Type |
-| :------ | :------ |
-| `tbl` | `any` |
-| `name` | `string` |
-
-#### Defined in
-
-[index.ts:121](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L121)
-
-• **new Table**<`T`\>(`tbl`, `name`, `embeddings`)
-
-#### Type parameters
-
-| Name | Type |
-| :------ | :------ |
-| `T` | `number`[] |
-
-#### Parameters
-
-| Name | Type | Description |
-| :------ | :------ | :------ |
-| `tbl` | `any` |  |
-| `name` | `string` |  |
-| `embeddings` | [`EmbeddingFunction`](../interfaces/EmbeddingFunction.md)<`T`\> | An embedding function to use when interacting with this table |
-
-#### Defined in
-
-[index.ts:127](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L127)
-
-## Properties
-
-### \_embeddings
-
-• `Private` `Optional` `Readonly` **\_embeddings**: [`EmbeddingFunction`](../interfaces/EmbeddingFunction.md)<`T`\>
-
-#### Defined in
-
-[index.ts:119](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L119)
-
-___
-
-### \_name
-
-• `Private` `Readonly` **\_name**: `string`
-
-#### Defined in
-
-[index.ts:118](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L118)
-
-___
-
-### \_tbl
-
-• `Private` `Readonly` **\_tbl**: `any`
-
-#### Defined in
-
-[index.ts:117](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L117)
-
-## Accessors
-
-### name
-
-• `get` **name**(): `string`
-
-#### Returns
-
-`string`
-
-#### Defined in
-
-[index.ts:134](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L134)
-
-## Methods
-
-### add
-
-▸ **add**(`data`): `Promise`<`number`\>
-
-Insert records into this Table.
-
-#### Parameters
-
-| Name | Type | Description |
-| :------ | :------ | :------ |
-| `data` | `Record`<`string`, `unknown`\>[] | Records to be inserted into the Table |
-
-#### Returns
-
-`Promise`<`number`\>
-
-The number of rows added to the table
-
-#### Defined in
-
-[index.ts:152](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L152)
-
-___
-
-### create\_index
-
-▸ **create_index**(`indexParams`): `Promise`<`any`\>
-
-Create an ANN index on this Table vector index.
-
-**`See`**
-
-VectorIndexParams.
-
-#### Parameters
-
-| Name | Type | Description |
-| :------ | :------ | :------ |
-| `indexParams` | `IvfPQIndexConfig` | The parameters of this Index, |
-
-#### Returns
-
-`Promise`<`any`\>
-
-#### Defined in
-
-[index.ts:171](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L171)
-
-___
-
-### overwrite
-
-▸ **overwrite**(`data`): `Promise`<`number`\>
-
-Insert records into this Table, replacing its contents.
-
-#### Parameters
-
-| Name | Type | Description |
-| :------ | :------ | :------ |
-| `data` | `Record`<`string`, `unknown`\>[] | Records to be inserted into the Table |
-
-#### Returns
-
-`Promise`<`number`\>
-
-The number of rows added to the table
-
-#### Defined in
-
-[index.ts:162](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L162)
-
-___
-
-### search
-
-▸ **search**(`query`): [`Query`](Query.md)<`T`\>
-
-Creates a search query to find the nearest neighbors of the given search term
-
-#### Parameters
-
-| Name | Type | Description |
-| :------ | :------ | :------ |
-| `query` | `T` | The query search term |
-
-#### Returns
-
-[`Query`](Query.md)<`T`\>
-
-#### Defined in
-
-[index.ts:142](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L142)

docs/src/javascript/enums/MetricType.md

@@ -9,6 +9,7 @@ Distance metrics type.
 ### Enumeration Members
 
 - [Cosine](MetricType.md#cosine)
+- [Dot](MetricType.md#dot)
 - [L2](MetricType.md#l2)
 
 ## Enumeration Members
@@ -21,7 +22,19 @@ Cosine distance
 
 #### Defined in
 
-[index.ts:341](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L341)
+[index.ts:481](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L481)
+
+___
+
+### Dot
+
+• **Dot** = ``"dot"``
+
+Dot product
+
+#### Defined in
+
+[index.ts:486](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L486)
 
 ___
 
@@ -33,4 +46,4 @@ Euclidean distance
 
 #### Defined in
 
-[index.ts:336](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L336)
+[index.ts:476](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L476)

docs/src/javascript/enums/WriteMode.md

@@ -2,11 +2,14 @@
 
 # Enumeration: WriteMode
 
+Write mode for writing a table.
+
 ## Table of contents
 
 ### Enumeration Members
 
 - [Append](WriteMode.md#append)
+- [Create](WriteMode.md#create)
 - [Overwrite](WriteMode.md#overwrite)
 
 ## Enumeration Members
@@ -15,9 +18,23 @@
 
 • **Append** = ``"append"``
 
+Append new data to the table.
+
 #### Defined in
 
-[index.ts:326](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L326)
+[index.ts:466](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L466)
+
+___
+
+### Create
+
+• **Create** = ``"create"``
+
+Create a new [Table](../interfaces/Table.md).
+
+#### Defined in
+
+[index.ts:462](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L462)
 
 ___
 
@@ -25,6 +42,8 @@ ___
 
 • **Overwrite** = ``"overwrite"``
 
+Overwrite the existing [Table](../interfaces/Table.md) if presented.
+
 #### Defined in
 
-[index.ts:325](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L325)
+[index.ts:464](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L464)

docs/src/javascript/interfaces/Connection.md

@@ -0,0 +1,152 @@
+[vectordb](../README.md) / [Exports](../modules.md) / Connection
+
+# Interface: Connection
+
+A LanceDB Connection that allows you to open tables and create new ones.
+
+Connection could be local against filesystem or remote against a server.
+
+## Implemented by
+
+- [`LocalConnection`](../classes/LocalConnection.md)
+
+## Table of contents
+
+### Properties
+
+- [uri](Connection.md#uri)
+
+### Methods
+
+- [createTable](Connection.md#createtable)
+- [createTableArrow](Connection.md#createtablearrow)
+- [dropTable](Connection.md#droptable)
+- [openTable](Connection.md#opentable)
+- [tableNames](Connection.md#tablenames)
+
+## Properties
+
+### uri
+
+• **uri**: `string`
+
+#### Defined in
+
+[index.ts:45](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L45)
+
+## Methods
+
+### createTable
+
+▸ **createTable**<`T`\>(`name`, `data`, `mode?`, `embeddings?`): `Promise`<[`Table`](Table.md)<`T`\>\>
+
+Creates a new Table and initialize it with new data.
+
+#### Type parameters
+
+| Name |
+| :------ |
+| `T` |
+
+#### Parameters
+
+| Name | Type | Description |
+| :------ | :------ | :------ |
+| `name` | `string` | The name of the table. |
+| `data` | `Record`<`string`, `unknown`\>[] | Non-empty Array of Records to be inserted into the table |
+| `mode?` | [`WriteMode`](../enums/WriteMode.md) | The write mode to use when creating the table. |
+| `embeddings?` | [`EmbeddingFunction`](EmbeddingFunction.md)<`T`\> | An embedding function to use on this table |
+
+#### Returns
+
+`Promise`<[`Table`](Table.md)<`T`\>\>
+
+#### Defined in
+
+[index.ts:65](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L65)
+
+___
+
+### createTableArrow
+
+▸ **createTableArrow**(`name`, `table`): `Promise`<[`Table`](Table.md)<`number`[]\>\>
+
+#### Parameters
+
+| Name | Type |
+| :------ | :------ |
+| `name` | `string` |
+| `table` | `Table`<`any`\> |
+
+#### Returns
+
+`Promise`<[`Table`](Table.md)<`number`[]\>\>
+
+#### Defined in
+
+[index.ts:67](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L67)
+
+___
+
+### dropTable
+
+▸ **dropTable**(`name`): `Promise`<`void`\>
+
+Drop an existing table.
+
+#### Parameters
+
+| Name | Type | Description |
+| :------ | :------ | :------ |
+| `name` | `string` | The name of the table to drop. |
+
+#### Returns
+
+`Promise`<`void`\>
+
+#### Defined in
+
+[index.ts:73](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L73)
+
+___
+
+### openTable
+
+▸ **openTable**<`T`\>(`name`, `embeddings?`): `Promise`<[`Table`](Table.md)<`T`\>\>
+
+Open a table in the database.
+
+#### Type parameters
+
+| Name |
+| :------ |
+| `T` |
+
+#### Parameters
+
+| Name | Type | Description |
+| :------ | :------ | :------ |
+| `name` | `string` | The name of the table. |
+| `embeddings?` | [`EmbeddingFunction`](EmbeddingFunction.md)<`T`\> | An embedding function to use on this table |
+
+#### Returns
+
+`Promise`<[`Table`](Table.md)<`T`\>\>
+
+#### Defined in
+
+[index.ts:55](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L55)
+
+___
+
+### tableNames
+
+▸ **tableNames**(): `Promise`<`string`[]\>
+
+#### Returns
+
+`Promise`<`string`[]\>
+
+#### Defined in
+
+[index.ts:47](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L47)

docs/src/javascript/interfaces/EmbeddingFunction.md

@@ -45,7 +45,7 @@ Creates a vector representation for the given values.
 
 #### Defined in
 
-[embedding/embedding_function.ts:27](https://github.com/lancedb/lancedb/blob/31dab97/node/src/embedding/embedding_function.ts#L27)
+[embedding/embedding_function.ts:27](https://github.com/lancedb/lancedb/blob/7247834/node/src/embedding/embedding_function.ts#L27)
 
 ___
 
@@ -57,4 +57,4 @@ The name of the column that will be used as input for the Embedding Function.
 
 #### Defined in
 
-[embedding/embedding_function.ts:22](https://github.com/lancedb/lancedb/blob/31dab97/node/src/embedding/embedding_function.ts#L22)
+[embedding/embedding_function.ts:22](https://github.com/lancedb/lancedb/blob/7247834/node/src/embedding/embedding_function.ts#L22)

docs/src/javascript/interfaces/Table.md

@@ -0,0 +1,195 @@
+[vectordb](../README.md) / [Exports](../modules.md) / Table
+
+# Interface: Table<T\>
+
+A LanceDB Table is the collection of Records. Each Record has one or more vector fields.
+
+## Type parameters
+
+| Name | Type |
+| :------ | :------ |
+| `T` | `number`[] |
+
+## Implemented by
+
+- [`LocalTable`](../classes/LocalTable.md)
+
+## Table of contents
+
+### Properties
+
+- [add](Table.md#add)
+- [countRows](Table.md#countrows)
+- [createIndex](Table.md#createindex)
+- [delete](Table.md#delete)
+- [name](Table.md#name)
+- [overwrite](Table.md#overwrite)
+- [search](Table.md#search)
+
+## Properties
+
+### add
+
+• **add**: (`data`: `Record`<`string`, `unknown`\>[]) => `Promise`<`number`\>
+
+#### Type declaration
+
+▸ (`data`): `Promise`<`number`\>
+
+Insert records into this Table.
+
+##### Parameters
+
+| Name | Type | Description |
+| :------ | :------ | :------ |
+| `data` | `Record`<`string`, `unknown`\>[] | Records to be inserted into the Table |
+
+##### Returns
+
+`Promise`<`number`\>
+
+The number of rows added to the table
+
+#### Defined in
+
+[index.ts:95](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L95)
+
+___
+
+### countRows
+
+• **countRows**: () => `Promise`<`number`\>
+
+#### Type declaration
+
+▸ (): `Promise`<`number`\>
+
+Returns the number of rows in this table.
+
+##### Returns
+
+`Promise`<`number`\>
+
+#### Defined in
+
+[index.ts:115](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L115)
+
+___
+
+### createIndex
+
+• **createIndex**: (`indexParams`: `IvfPQIndexConfig`) => `Promise`<`any`\>
+
+#### Type declaration
+
+▸ (`indexParams`): `Promise`<`any`\>
+
+Create an ANN index on this Table vector index.
+
+**`See`**
+
+VectorIndexParams.
+
+##### Parameters
+
+| Name | Type | Description |
+| :------ | :------ | :------ |
+| `indexParams` | `IvfPQIndexConfig` | The parameters of this Index, |
+
+##### Returns
+
+`Promise`<`any`\>
+
+#### Defined in
+
+[index.ts:110](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L110)
+
+___
+
+### delete
+
+• **delete**: (`filter`: `string`) => `Promise`<`void`\>
+
+#### Type declaration
+
+▸ (`filter`): `Promise`<`void`\>
+
+Delete rows from this table.
+
+##### Parameters
+
+| Name | Type | Description |
+| :------ | :------ | :------ |
+| `filter` | `string` | A filter in the same format used by a sql WHERE clause. |
+
+##### Returns
+
+`Promise`<`void`\>
+
+#### Defined in
+
+[index.ts:122](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L122)
+
+___
+
+### name
+
+• **name**: `string`
+
+#### Defined in
+
+[index.ts:81](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L81)
+
+___
+
+### overwrite
+
+• **overwrite**: (`data`: `Record`<`string`, `unknown`\>[]) => `Promise`<`number`\>
+
+#### Type declaration
+
+▸ (`data`): `Promise`<`number`\>
+
+Insert records into this Table, replacing its contents.
+
+##### Parameters
+
+| Name | Type | Description |
+| :------ | :------ | :------ |
+| `data` | `Record`<`string`, `unknown`\>[] | Records to be inserted into the Table |
+
+##### Returns
+
+`Promise`<`number`\>
+
+The number of rows added to the table
+
+#### Defined in
+
+[index.ts:103](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L103)
+
+___
+
+### search
+
+• **search**: (`query`: `T`) => [`Query`](../classes/Query.md)<`T`\>
+
+#### Type declaration
+
+▸ (`query`): [`Query`](../classes/Query.md)<`T`\>
+
+Creates a search query to find the nearest neighbors of the given search term
+
+##### Parameters
+
+| Name | Type | Description |
+| :------ | :------ | :------ |
+| `query` | `T` | The query search term |
+
+##### Returns
+
+[`Query`](../classes/Query.md)<`T`\>
+
+#### Defined in
+
+[index.ts:87](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L87)

docs/src/javascript/modules.md

@@ -11,14 +11,16 @@
 
 ### Classes
 
-- [Connection](classes/Connection.md)
+- [LocalConnection](classes/LocalConnection.md)
+- [LocalTable](classes/LocalTable.md)
 - [OpenAIEmbeddingFunction](classes/OpenAIEmbeddingFunction.md)
 - [Query](classes/Query.md)
-- [Table](classes/Table.md)
 
 ### Interfaces
 
+- [Connection](interfaces/Connection.md)
 - [EmbeddingFunction](interfaces/EmbeddingFunction.md)
+- [Table](interfaces/Table.md)
 
 ### Type Aliases
 
@@ -36,13 +38,13 @@
 
 #### Defined in
 
-[index.ts:224](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L224)
+[index.ts:345](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L345)
 
 ## Functions
 
 ### connect
 
-▸ **connect**(`uri`): `Promise`<[`Connection`](classes/Connection.md)\>
+▸ **connect**(`uri`): `Promise`<[`Connection`](interfaces/Connection.md)\>
 
 Connect to a LanceDB instance at the given URI
 
@@ -54,8 +56,8 @@ Connect to a LanceDB instance at the given URI
 
 #### Returns
 
-`Promise`<[`Connection`](classes/Connection.md)\>
+`Promise`<[`Connection`](interfaces/Connection.md)\>
 
 #### Defined in
 
-[index.ts:34](https://github.com/lancedb/lancedb/blob/31dab97/node/src/index.ts#L34)
+[index.ts:34](https://github.com/lancedb/lancedb/blob/7247834/node/src/index.ts#L34)

docs/src/notebooks/diffusiondb/datagen.py

@@ -21,12 +21,13 @@ from argparse import ArgumentParser
 from multiprocessing import Pool
 
 import lance
-import lancedb
 import pyarrow as pa
 from datasets import load_dataset
 from PIL import Image
 from transformers import CLIPModel, CLIPProcessor, CLIPTokenizerFast
 
+import lancedb
+
 MODEL_ID = "openai/clip-vit-base-patch32"
 
 device = "cuda"

docs/src/python/arrow.md

@@ -0,0 +1,101 @@
+# Pandas and PyArrow
+
+
+Built on top of [Apache Arrow](https://arrow.apache.org/),
+`LanceDB` is easy to integrate with the Python ecosystem, including [Pandas](https://pandas.pydata.org/)
+and PyArrow.
+
+## Create dataset
+
+First, we need to connect to a `LanceDB` database.
+
+```py
+
+import lancedb
+
+db = lancedb.connect("data/sample-lancedb")
+```
+
+Afterwards, we write a `Pandas DataFrame` to LanceDB directly.
+
+```py
+import pandas as pd
+
+data = pd.DataFrame({
+    "vector": [[3.1, 4.1], [5.9, 26.5]],
+    "item": ["foo", "bar"],
+    "price": [10.0, 20.0]
+})
+table = db.create_table("pd_table", data=data)
+```
+
+Similar to [`pyarrow.write_dataset()`](https://arrow.apache.org/docs/python/generated/pyarrow.dataset.write_dataset.html),
+[db.create_table()](../python/#lancedb.db.DBConnection.create_table) accepts a wide-range of forms of data.
+
+For example, if you have a dataset that is larger than memory size, you can create table with `Iterator[pyarrow.RecordBatch]`,
+to lazily generate data:
+
+```py
+
+from typing import Iterable
+import pyarrow as pa
+import lancedb
+
+def make_batches() -> Iterable[pa.RecordBatch]:
+    for i in range(5):
+        yield pa.RecordBatch.from_arrays(
+            [
+                pa.array([[3.1, 4.1], [5.9, 26.5]]),
+                pa.array(["foo", "bar"]),
+                pa.array([10.0, 20.0]),
+            ],
+            ["vector", "item", "price"])
+
+schema=pa.schema([
+    pa.field("vector", pa.list_(pa.float32())),
+    pa.field("item", pa.utf8()),
+    pa.field("price", pa.float32()),
+])
+
+table = db.create_table("iterable_table", data=make_batches(), schema=schema)
+```
+
+You will find detailed instructions of creating dataset in
+[Basic Operations](../basic.md) and [API](../python/#lancedb.db.DBConnection.create_table)
+sections.
+
+## Vector Search
+
+We can now perform similarity search via `LanceDB` Python API.
+
+```py
+# Open the table previously created.
+table = db.open_table("pd_table")
+
+query_vector = [100, 100]
+# Pandas DataFrame
+df = table.search(query_vector).limit(1).to_df()
+print(df)
+```
+
+```
+    vector     item  price        score
+0  [5.9, 26.5]  bar   20.0  14257.05957
+```
+
+If you have a simple filter, it's faster to provide a `where clause` to `LanceDB`'s search query.
+If you have more complex criteria, you can always apply the filter to the resulting Pandas `DataFrame`.
+
+```python
+
+# Apply the filter via LanceDB
+results = table.search([100, 100]).where("price < 15").to_df()
+assert len(results) == 1
+assert results["item"].iloc[0] == "foo"
+
+# Apply the filter via Pandas
+df = results = table.search([100, 100]).to_df()
+results = df[df.price < 15]
+assert len(results) == 1
+assert results["item"].iloc[0] == "foo"
+```

docs/src/python/duckdb.md

@@ -0,0 +1,56 @@
+# DuckDB
+
+`LanceDB` works with `DuckDB` via [PyArrow integration](https://duckdb.org/docs/guides/python/sql_on_arrow).
+
+Let us start with installing `duckdb` and `lancedb`.
+
+```shell
+pip install duckdb lancedb
+```
+
+We will re-use [the dataset created previously](./arrow.md):
+
+```python
+import pandas as pd
+import lancedb
+
+db = lancedb.connect("data/sample-lancedb")
+data = pd.DataFrame({
+    "vector": [[3.1, 4.1], [5.9, 26.5]],
+    "item": ["foo", "bar"],
+    "price": [10.0, 20.0]
+})
+table = db.create_table("pd_table", data=data)
+arrow_table = table.to_arrow()
+```
+
+`DuckDB` can directly query the `arrow_table`:
+
+```python
+import duckdb
+
+duckdb.query("SELECT * FROM arrow_table")
+```
+
+```
+┌─────────────┬─────────┬────────┐
+│   vector    │  item   │ price  │
+│   float[]   │ varchar │ double │
+├─────────────┼─────────┼────────┤
+│ [3.1, 4.1]  │ foo     │   10.0 │
+│ [5.9, 26.5] │ bar     │   20.0 │
+└─────────────┴─────────┴────────┘
+```
+
+```py
+duckdb.query("SELECT mean(price) FROM arrow_table")
+```
+
+```
+┌─────────────┐
+│ mean(price) │
+│   double    │
+├─────────────┤
+│        15.0 │
+└─────────────┘
+```

docs/src/python/integration.md

@@ -0,0 +1,7 @@
+# Integration
+
+Built on top of [Apache Arrow](https://arrow.apache.org/),
+`LanceDB` is very easy to be integrate with Python ecosystems.
+
+* [Pandas and Arrow Integration](./arrow.md)
+* [DuckDB Integration](./duckdb.md)

docs/src/python/pydantic.md

@@ -0,0 +1,35 @@
+# Pydantic
+
+[Pydantic](https://docs.pydantic.dev/latest/) is a data validation library in Python.
+
+## Schema
+
+LanceDB supports to create Apache Arrow Schema from a
+[Pydantic BaseModel](https://docs.pydantic.dev/latest/api/main/#pydantic.main.BaseModel)
+via [pydantic_to_schema()](python.md##lancedb.pydantic.pydantic_to_schema) method.
+
+::: lancedb.pydantic.pydantic_to_schema
+
+## Vector Field
+
+LanceDB provides a [`vector(dim)`](python.md#lancedb.pydantic.vector) method to define a
+vector Field in a Pydantic Model.
+
+::: lancedb.pydantic.vector
+
+## Type Conversion
+
+LanceDB automatically convert Pydantic fields to
+[Apache Arrow DataType](https://arrow.apache.org/docs/python/generated/pyarrow.DataType.html#pyarrow.DataType).
+
+Current supported type conversions:
+
+| Pydantic Field Type | PyArrow Data Type |
+| ------------------- | ----------------- |
+| `int`               | `pyarrow.int64`   |
+| `float`              | `pyarrow.float64`  |
+| `bool`              | `pyarrow.bool`    |
+| `str`               | `pyarrow.utf8()`    |
+| `list`              | `pyarrow.List`    |
+| `BaseModel`         | `pyarrow.Struct`    |
+| `vector(n)`         | `pyarrow.FixedSizeList(float32, n)` |

docs/src/python/python.md

@@ -10,14 +10,16 @@ pip install lancedb
 
 ::: lancedb.connect
 
-::: lancedb.LanceDBConnection
+::: lancedb.db.DBConnection
 
 ## Table
 
-::: lancedb.table.LanceTable
+::: lancedb.table.Table
 
 ## Querying
 
+::: lancedb.query.Query
+
 ::: lancedb.query.LanceQueryBuilder
 
 ::: lancedb.query.LanceFtsQueryBuilder
@@ -41,3 +43,17 @@ pip install lancedb
 ::: lancedb.fts.populate_index
 
 ::: lancedb.fts.search_index
+
+## Utilities
+
+::: lancedb.vector
+
+## Integrations
+
+### Pydantic
+
+::: lancedb.pydantic.pydantic_to_schema
+
+::: lancedb.pydantic.vector
+
+

docs/src/search.md

@@ -18,26 +18,55 @@ Currently, we support the following metrics:
 | ----------- | ------------------------------------ |
 | `L2`        | [Euclidean / L2 distance](https://en.wikipedia.org/wiki/Euclidean_distance) |
 | `Cosine`    | [Cosine Similarity](https://en.wikipedia.org/wiki/Cosine_similarity)|
+| `Dot`       | [Dot Production](https://en.wikipedia.org/wiki/Dot_product) |
 
 
 ## Search
 
 ### Flat Search
 
+If LanceDB does not create a vector index, LanceDB would need to scan (`Flat Search`) the entire vector column
+and compute the distance for each vector in order to find the closest matches.
 
-If there is no [vector index is created](ann_indexes.md), LanceDB will just brute-force scan
-the vector column and compute the distance.
 
+<!-- Setup Code
+```python
+import lancedb
+import numpy as np
+uri = "data/sample-lancedb"
+db = lancedb.connect(uri)
+
+data = [{"vector": row, "item": f"item {i}"}
+     for i, row in enumerate(np.random.random((10_000, 1536)).astype('float32'))]
+
+db.create_table("my_vectors", data=data)
+```
+-->
+<!-- Setup Code
+```javascript
+const vectordb_setup = require('vectordb')
+const db_setup = await vectordb_setup.connect('data/sample-lancedb')
+
+let data = []
+for (let i = 0; i < 10_000; i++) {
+     data.push({vector: Array(1536).fill(i), id: `${i}`, content: "", longId: `${i}`},)
+}
+await db_setup.createTable('my_vectors', data)
+```
+-->
 === "Python"
 
+
     ```python
     import lancedb
+    import numpy as np
+
     db = lancedb.connect("data/sample-lancedb")
 
     tbl = db.open_table("my_vectors")
 
-    df = tbl.search(np.random.random((768)))
+    df = tbl.search(np.random.random((1536))) \
-        .limit(10)
+        .limit(10) \
         .to_df()
     ```
 
@@ -47,10 +76,10 @@ the vector column and compute the distance.
     const vectordb = require('vectordb')
     const db = await vectordb.connect('data/sample-lancedb')
 
-    tbl = db.open_table("my_vectors")
+    const tbl = await db.openTable("my_vectors")
 
-    const results = await tbl.search(Array(768))
+    const results_1 = await tbl.search(Array(1536).fill(1.2))
-        .limit(20)
+        .limit(10)
         .execute()
     ```
 
@@ -60,26 +89,33 @@ as well.
 === "Python"
 
     ```python
-    df = tbl.search(np.random.random((768)))
+    df = tbl.search(np.random.random((1536))) \
-        .metric("cosine")
+        .metric("cosine") \
-        .limit(10)
+        .limit(10) \
         .to_df()
     ```
 
+
 === "JavaScript"
 
     ```javascript
-    const vectordb = require('vectordb')
+    const results_2 = await tbl.search(Array(1536).fill(1.2))
-    const db = await vectordb.connect('data/sample-lancedb')
+        .metricType("cosine")
-
+        .limit(10)
-    tbl = db.open_table("my_vectors")
-
-    const results = await tbl.search(Array(768))
-        .metric("cosine")
-        .limit(20)
         .execute()
     ```
 
-### Search with Vector Index.
+
+### Approximate Nearest Neighbor (ANN) Search with Vector Index.
+
+To accelerate vector retrievals, it is common to build vector indices.
+A vector index is a data structure specifically designed to efficiently organize and
+search vector data based on their similarity or distance metrics.
+By constructing a vector index, you can reduce the search space and avoid the need
+for brute-force scanning of the entire vector column.
+
+However, fast vector search using indices often entails making a trade-off with accuracy to some extent.
+This is why it is often called **Approximate Nearest Neighbors (ANN)** search, while the Flat Search (KNN)
+always returns 100% recall.
 
 See [ANN Index](ann_indexes.md) for more details.

docs/src/sql.md

@@ -0,0 +1,120 @@
+# SQL filters
+
+LanceDB embraces the utilization of standard SQL expressions as predicates for hybrid
+filters. It can be used during hybrid vector search and deletion operations.
+
+Currently, Lance supports a growing list of expressions.
+
+* ``>``, ``>=``, ``<``, ``<=``, ``=``
+* ``AND``, ``OR``, ``NOT``
+* ``IS NULL``, ``IS NOT NULL``
+* ``IS TRUE``, ``IS NOT TRUE``, ``IS FALSE``, ``IS NOT FALSE``
+* ``IN``
+* ``LIKE``, ``NOT LIKE``
+* ``CAST``
+* ``regexp_match(column, pattern)``
+
+For example, the following filter string is acceptable:
+<!-- Setup Code
+```python 
+import lancedb
+import numpy as np
+uri = "data/sample-lancedb"
+db = lancedb.connect(uri)
+
+data = [{"vector": row, "item": f"item {i}"}
+     for i, row in enumerate(np.random.random((10_000, 2)).astype('int'))]
+
+tbl = db.create_table("my_vectors", data=data)
+```
+-->
+<!-- Setup Code
+```javascript 
+const vectordb = require('vectordb')
+const db = await vectordb.connect('data/sample-lancedb')
+
+let data = []
+for (let i = 0; i < 10_000; i++) {
+     data.push({vector: Array(1536).fill(i), id: `${i}`, content: "", longId: `${i}`},)
+}
+const tbl = await db.createTable('my_vectors', data)
+```
+-->
+=== "Python"
+
+    ```python
+    tbl.search([100, 102]) \
+       .where("""(
+        (label IN [10, 20])
+        AND
+        (note.email IS NOT NULL)
+        ) OR NOT note.created
+        """)
+
+    ```
+=== "Javascript"
+
+    ```javascript
+    tbl.search([100, 102])
+       .where(`(
+            (label IN [10, 20])
+            AND
+            (note.email IS NOT NULL)
+        ) OR NOT note.created
+       `)
+    ```
+
+
+If your column name contains special characters or is a [SQL Keyword](https://docs.rs/sqlparser/latest/sqlparser/keywords/index.html),
+you can use backtick (`` ` ``) to escape it. For nested fields, each segment of the
+path must be wrapped in backticks.
+
+=== "SQL"
+    ```sql
+    `CUBE` = 10 AND `column name with space` IS NOT NULL
+      AND `nested with space`.`inner with space` < 2
+    ```
+
+!!! warning
+    Field names containing periods (``.``) are not supported.
+
+Literals for dates, timestamps, and decimals can be written by writing the string
+value after the type name. For example
+
+=== "SQL"
+    ```sql
+    date_col = date '2021-01-01'
+    and timestamp_col = timestamp '2021-01-01 00:00:00'
+    and decimal_col = decimal(8,3) '1.000'
+    ```
+
+For timestamp columns, the precision can be specified as a number in the type
+parameter. Microsecond precision (6) is the default.
+
+| SQL              | Time unit    |
+|------------------|--------------|
+| ``timestamp(0)`` | Seconds      |
+| ``timestamp(3)`` | Milliseconds |
+| ``timestamp(6)`` | Microseconds |
+| ``timestamp(9)`` | Nanoseconds  |
+
+LanceDB internally stores data in [Apache Arrow](https://arrow.apache.org/) format.
+The mapping from SQL types to Arrow types is:
+
+| SQL type | Arrow type |
+|----------|------------|
+| ``boolean`` | ``Boolean`` |
+| ``tinyint`` / ``tinyint unsigned`` | ``Int8`` / ``UInt8`` |
+| ``smallint`` / ``smallint unsigned`` | ``Int16`` / ``UInt16`` |
+| ``int`` or ``integer`` / ``int unsigned`` or ``integer unsigned`` | ``Int32`` / ``UInt32`` |
+| ``bigint`` / ``bigint unsigned`` | ``Int64`` / ``UInt64`` |
+| ``float`` | ``Float32`` |
+| ``double`` | ``Float64`` |
+| ``decimal(precision, scale)`` | ``Decimal128`` |
+| ``date`` | ``Date32`` |
+| ``timestamp`` | ``Timestamp`` [^1] |
+| ``string`` | ``Utf8`` |
+| ``binary`` | ``Binary`` |
+
+[^1]: See precision mapping in previous table.
+

docs/test/md_testing.js

@@ -0,0 +1,52 @@
+const glob = require("glob");
+const fs = require("fs");
+const path = require("path");
+
+const excludedFiles = [
+  "../src/fts.md",
+  "../src/embedding.md",
+  "../src/examples/serverless_lancedb_with_s3_and_lambda.md",
+  "../src/examples/serverless_qa_bot_with_modal_and_langchain.md",
+  "../src/examples/transformerjs_embedding_search_nodejs.md",
+  "../src/examples/youtube_transcript_bot_with_nodejs.md",
+];
+const nodePrefix = "javascript";
+const nodeFile = ".js";
+const nodeFolder = "node";
+const globString = "../src/**/*.md";
+const asyncPrefix = "(async () => {\n";
+const asyncSuffix = "})();";
+
+function* yieldLines(lines, prefix, suffix) {
+  let inCodeBlock = false;
+  for (const line of lines) {
+    if (line.trim().startsWith(prefix + nodePrefix)) {
+      inCodeBlock = true;
+    } else if (inCodeBlock && line.trim().startsWith(suffix)) {
+      inCodeBlock = false;
+      yield "\n";
+    } else if (inCodeBlock) {
+      yield line;
+    }
+  }
+}
+
+const files = glob.sync(globString, { recursive: true });
+
+for (const file of files.filter((file) => !excludedFiles.includes(file))) {
+  const lines = [];
+  const data = fs.readFileSync(file, "utf-8");
+  const fileLines = data.split("\n");
+
+  for (const line of yieldLines(fileLines, "```", "```")) {
+    lines.push(line);
+  }
+
+  if (lines.length > 0) {
+    const fileName = path.basename(file, ".md");
+    const outPath = path.join(nodeFolder, fileName, `${fileName}${nodeFile}`);
+    console.log(outPath)
+    fs.mkdirSync(path.dirname(outPath), { recursive: true });
+    fs.writeFileSync(outPath, asyncPrefix + "\n" + lines.join("\n") + asyncSuffix);
+  }
+}

docs/test/md_testing.py

@@ -0,0 +1,41 @@
+import glob
+from typing import Iterator
+from pathlib import Path
+
+excluded_files = [
+    "../src/fts.md",
+    "../src/embedding.md",
+    "../src/examples/serverless_lancedb_with_s3_and_lambda.md",
+    "../src/examples/serverless_qa_bot_with_modal_and_langchain.md",
+    "../src/examples/youtube_transcript_bot_with_nodejs.md"
+]
+
+python_prefix = "py"
+python_file = ".py"
+python_folder = "python"
+glob_string = "../src/**/*.md"
+
+def yield_lines(lines: Iterator[str], prefix: str, suffix: str):
+    in_code_block = False
+    # Python code has strict indentation
+    strip_length = 0
+    for line in lines:
+        if line.strip().startswith(prefix + python_prefix):
+            in_code_block = True
+            strip_length = len(line) - len(line.lstrip())
+        elif in_code_block and line.strip().startswith(suffix):
+            in_code_block = False
+            yield "\n"
+        elif in_code_block:
+            yield line[strip_length:]
+
+for file in filter(lambda file: file not in excluded_files, glob.glob(glob_string, recursive=True)):
+    with open(file, "r") as f:
+        lines = list(yield_lines(iter(f), "```", "```"))
+
+    if len(lines) > 0:
+        out_path = Path(python_folder) / Path(file).name.strip(".md") / (Path(file).name.strip(".md") + python_file)
+        print(out_path)
+        out_path.parent.mkdir(exist_ok=True, parents=True)
+        with open(out_path, "w") as out:
+            out.writelines(lines)

docs/test/package.json

@@ -0,0 +1,13 @@
+{
+  "name": "lancedb-docs-test",
+  "version": "1.0.0",
+  "description": "",
+  "author": "",
+  "license": "ISC",
+  "dependencies": {
+    "fs": "^0.0.1-security",
+    "glob": "^10.2.7",
+    "path": "^0.12.7",
+    "vectordb": "https://gitpkg.now.sh/lancedb/lancedb/node?main"
+  }
+}

docs/test/requirements.txt

@@ -0,0 +1,5 @@
+lancedb @ git+https://github.com/lancedb/lancedb.git#egg=subdir&subdirectory=python
+numpy
+pandas
+pylance
+duckdb

node/.eslintrc.js

@@ -12,5 +12,6 @@ module.exports = {
     sourceType: 'module'
   },
   rules: {
+    "@typescript-eslint/method-signature-style": "off",
   }
 }

node/.npmignore

@@ -0,0 +1,4 @@
+gen_test_data.py
+index.node
+dist/lancedb*.tgz
+vectordb*.tgz

node/README.md

@@ -8,15 +8,21 @@ A JavaScript / Node.js library for [LanceDB](https://github.com/lancedb/lancedb)
 npm install vectordb
 ```
 
+This will download the appropriate native library for your platform. We currently
+support x86_64 Linux, aarch64 Linux, Intel MacOS, and ARM (M1/M2) MacOS. We do not
+yet support Windows or musl-based Linux (such as Alpine Linux).
+
 ## Usage
 
 ### Basic Example
 
 ```javascript
 const lancedb = require('vectordb');
-const db = lancedb.connect('<PATH_TO_LANCEDB_DATASET>');
+const db = await lancedb.connect('data/sample-lancedb');
-const table = await db.openTable('my_table');
+const table = await db.createTable("my_table",
-const query = await table.search([0.1, 0.3]).setLimit(20).execute();
+      [{ id: 1, vector: [0.1, 1.0], item: "foo", price: 10.0 },
+      { id: 2, vector: [3.9, 0.5], item: "bar", price: 20.0 }])
+const results = await table.search([0.1, 0.3]).limit(20).execute();
 console.log(results);
 ```
 
@@ -24,17 +30,33 @@ The [examples](./examples) folder contains complete examples.
 
 ## Development
 
-The LanceDB javascript is built with npm:
+To build everything fresh:
+
+```bash
+npm install
+npm run tsc
+npm run build
+```
+
+Then you should be able to run the tests with:
+
+```bash
+npm test
+```
+
+### Rebuilding Rust library
+
+```bash
+npm run build
+```
+
+### Rebuilding Typescript
 
 ```bash
 npm run tsc
 ```
 
-Run the tests with
+### Fix lints
-
-```bash
-npm test
-```
 
 To run the linter and have it automatically fix all errors
 

node/examples/js-transformers/index.js

@@ -0,0 +1,66 @@
+// Copyright 2023 Lance Developers.
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+'use strict'
+
+
+async function example() {
+
+    const lancedb = require('vectordb')
+
+    // Import transformers and the all-MiniLM-L6-v2 model (https://huggingface.co/Xenova/all-MiniLM-L6-v2)
+    const { pipeline } = await import('@xenova/transformers')
+    const pipe = await pipeline('feature-extraction', 'Xenova/all-MiniLM-L6-v2');
+
+
+    // Create embedding function from pipeline which returns a list of vectors from batch
+    // sourceColumn is the name of the column in the data to be embedded
+    //
+    // Output of pipe is a Tensor { data: Float32Array(384) }, so filter for the vector
+    const embed_fun = {}
+    embed_fun.sourceColumn = 'text'
+    embed_fun.embed = async function (batch) {
+        let result = []
+        for (let text of batch) {
+            const res = await pipe(text, { pooling: 'mean', normalize: true })
+            result.push(Array.from(res['data']))
+        }
+        return (result)
+    }
+
+    // Link a folder and create a table with data
+    const db = await lancedb.connect('data/sample-lancedb')
+
+    const data = [
+        { id: 1, text: 'Cherry', type: 'fruit' },
+        { id: 2, text: 'Carrot', type: 'vegetable' },
+        { id: 3, text: 'Potato', type: 'vegetable' },
+        { id: 4, text: 'Apple', type: 'fruit' },
+        { id: 5, text: 'Banana', type: 'fruit' }
+    ]
+
+    const table = await db.createTable('food_table', data, "create", embed_fun)
+
+
+    // Query the table
+    const results = await table
+        .search("a sweet fruit to eat")
+        .metricType("cosine")
+        .limit(2)
+        .execute()
+    console.log(results.map(r => r.text))
+
+}
+
+example().then(_ => { console.log("Done!") })

node/examples/js-transformers/package.json

@@ -0,0 +1,16 @@
+{
+  "name": "vectordb-example-js-transformers",
+  "version": "1.0.0",
+  "description": "Example for using transformers.js with lancedb",
+  "main": "index.js",
+  "scripts": {
+    "test": "echo \"Error: no test specified\" && exit 1"
+  },
+  "author": "Lance Devs",
+  "license": "Apache-2.0",
+  "dependencies": {
+    "@xenova/transformers": "^2.4.1",
+    "vectordb": "^0.1.12"
+  }
+
+}

node/native.js

@@ -12,29 +12,26 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+const { currentTarget } = require('@neon-rs/load');
+
 let nativeLib;
 
-function getPlatformLibrary() {
-    if (process.platform === "darwin" && process.arch == "arm64") {
-        return require('./aarch64-apple-darwin.node');
-    } else if (process.platform === "darwin" && process.arch == "x64") {
-        return require('./x86_64-apple-darwin.node');
-    } else if (process.platform === "linux" && process.arch == "x64") {
-        return require('./x86_64-unknown-linux-gnu.node');
-    } else {
-        throw new Error(`vectordb: unsupported platform ${process.platform}_${process.arch}. Please file a bug report at https://github.com/lancedb/lancedb/issues`)
-    }
-}
-
 try {
-    nativeLib = require('./index.node')
+    nativeLib = require(`vectordb-${currentTarget()}`);
 } catch (e) {
-    if (e.code === "MODULE_NOT_FOUND") {
+    try {
-        nativeLib = getPlatformLibrary();
+        // Might be developing locally, so try that. But don't expose that error
-    } else {
+        // to the user.
-        throw new Error('vectordb: failed to load native library. Please file a bug report at https://github.com/lancedb/lancedb/issues');
+        nativeLib = require("./index.node");
+    } catch {
+        throw new Error(`vectordb: failed to load native library.
+  You may need to run \`npm install vectordb-${currentTarget()}\`.
+      
+  If that does not work, please file a bug report at https://github.com/lancedb/lancedb/issues
+      
+  Source error: ${e}`);
     }   
 }
 
-module.exports = nativeLib
+// Dynamic require for runtime.
-
+module.exports = nativeLib;

node/package-lock.json

@@ -1,19 +1,31 @@
 {
   "name": "vectordb",
-  "version": "0.1.5",
+  "version": "0.1.13",
   "lockfileVersion": 2,
   "requires": true,
   "packages": {
     "": {
       "name": "vectordb",
-      "version": "0.1.5",
+      "version": "0.1.13",
+      "cpu": [
+        "x64",
+        "arm64"
+      ],
       "license": "Apache-2.0",
+      "os": [
+        "darwin",
+        "linux",
+        "win32"
+      ],
       "dependencies": {
         "@apache-arrow/ts": "^12.0.0",
+        "@neon-rs/load": "^0.0.74",
         "apache-arrow": "^12.0.0"
       },
       "devDependencies": {
+        "@neon-rs/cli": "^0.0.74",
         "@types/chai": "^4.3.4",
+        "@types/chai-as-promised": "^7.1.5",
         "@types/mocha": "^10.0.1",
         "@types/node": "^18.16.2",
         "@types/sinon": "^10.0.15",
@@ -21,9 +33,10 @@
         "@typescript-eslint/eslint-plugin": "^5.59.1",
         "cargo-cp-artifact": "^0.1",
         "chai": "^4.3.7",
+        "chai-as-promised": "^7.1.1",
         "eslint": "^8.39.0",
         "eslint-config-standard-with-typescript": "^34.0.1",
-        "eslint-plugin-import": "^2.27.5",
+        "eslint-plugin-import": "^2.26.0",
         "eslint-plugin-n": "^15.7.0",
         "eslint-plugin-promise": "^6.1.1",
         "mocha": "^10.2.0",
@@ -35,6 +48,13 @@
         "typedoc": "^0.24.7",
         "typedoc-plugin-markdown": "^3.15.3",
         "typescript": "*"
+      },
+      "optionalDependencies": {
+        "vectordb-darwin-arm64": "0.1.13",
+        "vectordb-darwin-x64": "0.1.13",
+        "vectordb-linux-arm64-gnu": "0.1.13",
+        "vectordb-linux-x64-gnu": "0.1.13",
+        "vectordb-win32-x64-msvc": "0.1.13"
       }
     },
     "node_modules/@apache-arrow/ts": {
@@ -202,6 +222,20 @@
         "@jridgewell/sourcemap-codec": "^1.4.10"
       }
     },
+    "node_modules/@neon-rs/cli": {
+      "version": "0.0.74",
+      "resolved": "https://registry.npmjs.org/@neon-rs/cli/-/cli-0.0.74.tgz",
+      "integrity": "sha512-9lPmNmjej5iKKOTMPryOMubwkgMRyTWRuaq1yokASvI5mPhr2kzPN7UVjdCOjQvpunNPngR9yAHoirpjiWhUHw==",
+      "dev": true,
+      "bin": {
+        "neon": "index.js"
+      }
+    },
+    "node_modules/@neon-rs/load": {
+      "version": "0.0.74",
+      "resolved": "https://registry.npmjs.org/@neon-rs/load/-/load-0.0.74.tgz",
+      "integrity": "sha512-/cPZD907UNz55yrc/ud4wDgQKtU1TvkD9jeqZWG6J4IMmZkp6zgjkQcKA8UvpkZlcpPHvc8J17sGzLFbP/LUYg=="
+    },
     "node_modules/@nodelib/fs.scandir": {
       "version": "2.1.5",
       "resolved": "https://registry.npmjs.org/@nodelib/fs.scandir/-/fs.scandir-2.1.5.tgz",
@@ -311,6 +345,15 @@
       "integrity": "sha512-KnRanxnpfpjUTqTCXslZSEdLfXExwgNxYPdiO2WGUj8+HDjFi8R3k5RVKPeSCzLjCcshCAtVO2QBbVuAV4kTnw==",
       "dev": true
     },
+    "node_modules/@types/chai-as-promised": {
+      "version": "7.1.5",
+      "resolved": "https://registry.npmjs.org/@types/chai-as-promised/-/chai-as-promised-7.1.5.tgz",
+      "integrity": "sha512-jStwss93SITGBwt/niYrkf2C+/1KTeZCZl1LaeezTlqppAKeoQC7jxyqYuP72sxBGKCIbw7oHgbYssIRzT5FCQ==",
+      "dev": true,
+      "dependencies": {
+        "@types/chai": "*"
+      }
+    },
     "node_modules/@types/command-line-args": {
       "version": "5.2.0",
       "resolved": "https://registry.npmjs.org/@types/command-line-args/-/command-line-args-5.2.0.tgz",
@@ -787,24 +830,6 @@
         "url": "https://github.com/sponsors/ljharb"
       }
     },
-    "node_modules/array.prototype.flatmap": {
-      "version": "1.3.1",
-      "resolved": "https://registry.npmjs.org/array.prototype.flatmap/-/array.prototype.flatmap-1.3.1.tgz",
-      "integrity": "sha512-8UGn9O1FDVvMNB0UlLv4voxRMze7+FpHyF5mSMRjWHUMlpoDViniy05870VlxhfgTnLbpuwTzvD76MTtWxB/mQ==",
-      "dev": true,
-      "dependencies": {
-        "call-bind": "^1.0.2",
-        "define-properties": "^1.1.4",
-        "es-abstract": "^1.20.4",
-        "es-shim-unscopables": "^1.0.0"
-      },
-      "engines": {
-        "node": ">= 0.4"
-      },
-      "funding": {
-        "url": "https://github.com/sponsors/ljharb"
-      }
-    },
     "node_modules/assertion-error": {
       "version": "1.1.0",
       "resolved": "https://registry.npmjs.org/assertion-error/-/assertion-error-1.1.0.tgz",
@@ -960,6 +985,18 @@
         "node": ">=4"
       }
     },
+    "node_modules/chai-as-promised": {
+      "version": "7.1.1",
+      "resolved": "https://registry.npmjs.org/chai-as-promised/-/chai-as-promised-7.1.1.tgz",
+      "integrity": "sha512-azL6xMoi+uxu6z4rhWQ1jbdUhOMhis2PvscD/xjLqNMkv3BPPp2JyyuTHOrf9BOosGpNQ11v6BKv/g57RXbiaA==",
+      "dev": true,
+      "dependencies": {
+        "check-error": "^1.0.2"
+      },
+      "peerDependencies": {
+        "chai": ">= 2.1.2 < 5"
+      }
+    },
     "node_modules/chalk": {
       "version": "4.1.2",
       "resolved": "https://registry.npmjs.org/chalk/-/chalk-4.1.2.tgz",
@@ -1633,25 +1670,23 @@
       }
     },
     "node_modules/eslint-plugin-import": {
-      "version": "2.27.5",
+      "version": "2.26.0",
-      "resolved": "https://registry.npmjs.org/eslint-plugin-import/-/eslint-plugin-import-2.27.5.tgz",
+      "resolved": "https://registry.npmjs.org/eslint-plugin-import/-/eslint-plugin-import-2.26.0.tgz",
-      "integrity": "sha512-LmEt3GVofgiGuiE+ORpnvP+kAm3h6MLZJ4Q5HCyHADofsb4VzXFsRiWj3c0OFiV+3DWFh0qg3v9gcPlfc3zRow==",
+      "integrity": "sha512-hYfi3FXaM8WPLf4S1cikh/r4IxnO6zrhZbEGz2b660EJRbuxgpDS5gkCuYgGWg2xxh2rBuIr4Pvhve/7c31koA==",
       "dev": true,
       "dependencies": {
-        "array-includes": "^3.1.6",
+        "array-includes": "^3.1.4",
-        "array.prototype.flat": "^1.3.1",
+        "array.prototype.flat": "^1.2.5",
-        "array.prototype.flatmap": "^1.3.1",
+        "debug": "^2.6.9",
-        "debug": "^3.2.7",
         "doctrine": "^2.1.0",
-        "eslint-import-resolver-node": "^0.3.7",
+        "eslint-import-resolver-node": "^0.3.6",
-        "eslint-module-utils": "^2.7.4",
+        "eslint-module-utils": "^2.7.3",
         "has": "^1.0.3",
-        "is-core-module": "^2.11.0",
+        "is-core-module": "^2.8.1",
         "is-glob": "^4.0.3",
         "minimatch": "^3.1.2",
-        "object.values": "^1.1.6",
+        "object.values": "^1.1.5",
-        "resolve": "^1.22.1",
+        "resolve": "^1.22.0",
-        "semver": "^6.3.0",
         "tsconfig-paths": "^3.14.1"
       },
       "engines": {
@@ -1662,12 +1697,12 @@
       }
     },
     "node_modules/eslint-plugin-import/node_modules/debug": {
-      "version": "3.2.7",
+      "version": "2.6.9",
-      "resolved": "https://registry.npmjs.org/debug/-/debug-3.2.7.tgz",
+      "resolved": "https://registry.npmjs.org/debug/-/debug-2.6.9.tgz",
-      "integrity": "sha512-CFjzYYAi4ThfiQvizrFQevTTXHtnCqWfe7x1AhgEscTz6ZbLbfoLRLPugTQyBth6f8ZERVUSyWHFD/7Wu4t1XQ==",
+      "integrity": "sha512-bC7ElrdJaJnPbAP+1EotYvqZsb3ecl5wi6Bfi6BJTUcNowp6cvspg0jXznRTKDjm/E7AdgFBVeAPVMNcKGsHMA==",
       "dev": true,
       "dependencies": {
-        "ms": "^2.1.1"
+        "ms": "2.0.0"
       }
     },
     "node_modules/eslint-plugin-import/node_modules/doctrine": {
@@ -1682,14 +1717,11 @@
         "node": ">=0.10.0"
       }
     },
-    "node_modules/eslint-plugin-import/node_modules/semver": {
+    "node_modules/eslint-plugin-import/node_modules/ms": {
-      "version": "6.3.0",
+      "version": "2.0.0",
-      "resolved": "https://registry.npmjs.org/semver/-/semver-6.3.0.tgz",
+      "resolved": "https://registry.npmjs.org/ms/-/ms-2.0.0.tgz",
-      "integrity": "sha512-b39TBaTSfV6yBrapU89p5fKekE2m/NwnDocOVruQFS1/veMgdzuPcnOM34M6CwxW8jH/lxEa5rBoDeUwu5HHTw==",
+      "integrity": "sha512-Tpp60P6IUJDTuOq/5Z8cdskzJujfwqfOTkrwIwj7IRISpnkJnT6SyJ4PCPnGMoFjC9ddhal5KVIYtAt97ix05A==",
-      "dev": true,
+      "dev": true
-      "bin": {
-        "semver": "bin/semver.js"
-      }
     },
     "node_modules/eslint-plugin-n": {
       "version": "15.7.0",
@@ -3619,9 +3651,9 @@
       }
     },
     "node_modules/semver": {
-      "version": "7.5.0",
+      "version": "7.5.3",
-      "resolved": "https://registry.npmjs.org/semver/-/semver-7.5.0.tgz",
+      "resolved": "https://registry.npmjs.org/semver/-/semver-7.5.3.tgz",
-      "integrity": "sha512-+XC0AD/R7Q2mPSRuy2Id0+CGTZ98+8f+KvwirxOKIEyid+XSx6HbC63p+O4IndTHuX5Z+JxQ0TghCkO5Cg/2HA==",
+      "integrity": "sha512-QBlUtyVk/5EeHbi7X0fw6liDZc7BBmEaSYn01fMU1OUYbf6GPsbTtd8WmnqbI20SeycoHSeiybkE/q1Q+qlThQ==",
       "dev": true,
       "dependencies": {
         "lru-cache": "^6.0.0"
@@ -4256,6 +4288,42 @@
       "integrity": "sha512-wa7YjyUGfNZngI/vtK0UHAN+lgDCxBPCylVXGp0zu59Fz5aiGtNXaq3DhIov063MorB+VfufLh3JlF2KdTK3xg==",
       "dev": true
     },
+    "node_modules/vectordb-darwin-arm64": {
+      "version": "0.1.13",
+      "resolved": "https://registry.npmjs.org/vectordb-darwin-arm64/-/vectordb-darwin-arm64-0.1.13.tgz",
+      "integrity": "sha512-9lLuX5P8m75EfP85pfC4LxO9J7Tzu4LngX55BVAdFe6qPRHu+iHmLw0QYYSVDqNm3GtDr2qFJlL2ILlsApyYyg==",
+      "cpu": [
+        "arm64"
+      ],
+      "optional": true,
+      "os": [
+        "darwin"
+      ]
+    },
+    "node_modules/vectordb-darwin-x64": {
+      "version": "0.1.13",
+      "resolved": "https://registry.npmjs.org/vectordb-darwin-x64/-/vectordb-darwin-x64-0.1.13.tgz",
+      "integrity": "sha512-5mkhBJlcfAqcty7Ww2csgYogq+b0NhtllAbag9IIznvqfcrvITU0H0vm5LGWbRuE/BUUxC25MJhm93YWBzqEVA==",
+      "cpu": [
+        "x64"
+      ],
+      "optional": true,
+      "os": [
+        "darwin"
+      ]
+    },
+    "node_modules/vectordb-linux-x64-gnu": {
+      "version": "0.1.13",
+      "resolved": "https://registry.npmjs.org/vectordb-linux-x64-gnu/-/vectordb-linux-x64-gnu-0.1.13.tgz",
+      "integrity": "sha512-fU+sIHUkXyMdrWjggT93p0blKD+pbgr+x01tn9d2/pbA1ePo2AwuE86rYPA+BjyCUE1QifPgKadzGVVpqWYmnQ==",
+      "cpu": [
+        "x64"
+      ],
+      "optional": true,
+      "os": [
+        "linux"
+      ]
+    },
     "node_modules/vscode-oniguruma": {
       "version": "1.7.0",
       "resolved": "https://registry.npmjs.org/vscode-oniguruma/-/vscode-oniguruma-1.7.0.tgz",
@@ -4601,6 +4669,17 @@
         "@jridgewell/sourcemap-codec": "^1.4.10"
       }
     },
+    "@neon-rs/cli": {
+      "version": "0.0.74",
+      "resolved": "https://registry.npmjs.org/@neon-rs/cli/-/cli-0.0.74.tgz",
+      "integrity": "sha512-9lPmNmjej5iKKOTMPryOMubwkgMRyTWRuaq1yokASvI5mPhr2kzPN7UVjdCOjQvpunNPngR9yAHoirpjiWhUHw==",
+      "dev": true
+    },
+    "@neon-rs/load": {
+      "version": "0.0.74",
+      "resolved": "https://registry.npmjs.org/@neon-rs/load/-/load-0.0.74.tgz",
+      "integrity": "sha512-/cPZD907UNz55yrc/ud4wDgQKtU1TvkD9jeqZWG6J4IMmZkp6zgjkQcKA8UvpkZlcpPHvc8J17sGzLFbP/LUYg=="
+    },
     "@nodelib/fs.scandir": {
       "version": "2.1.5",
       "resolved": "https://registry.npmjs.org/@nodelib/fs.scandir/-/fs.scandir-2.1.5.tgz",
@@ -4703,6 +4782,15 @@
       "integrity": "sha512-KnRanxnpfpjUTqTCXslZSEdLfXExwgNxYPdiO2WGUj8+HDjFi8R3k5RVKPeSCzLjCcshCAtVO2QBbVuAV4kTnw==",
       "dev": true
     },
+    "@types/chai-as-promised": {
+      "version": "7.1.5",
+      "resolved": "https://registry.npmjs.org/@types/chai-as-promised/-/chai-as-promised-7.1.5.tgz",
+      "integrity": "sha512-jStwss93SITGBwt/niYrkf2C+/1KTeZCZl1LaeezTlqppAKeoQC7jxyqYuP72sxBGKCIbw7oHgbYssIRzT5FCQ==",
+      "dev": true,
+      "requires": {
+        "@types/chai": "*"
+      }
+    },
     "@types/command-line-args": {
       "version": "5.2.0",
       "resolved": "https://registry.npmjs.org/@types/command-line-args/-/command-line-args-5.2.0.tgz",
@@ -5038,18 +5126,6 @@
         "es-shim-unscopables": "^1.0.0"
       }
     },
-    "array.prototype.flatmap": {
-      "version": "1.3.1",
-      "resolved": "https://registry.npmjs.org/array.prototype.flatmap/-/array.prototype.flatmap-1.3.1.tgz",
-      "integrity": "sha512-8UGn9O1FDVvMNB0UlLv4voxRMze7+FpHyF5mSMRjWHUMlpoDViniy05870VlxhfgTnLbpuwTzvD76MTtWxB/mQ==",
-      "dev": true,
-      "requires": {
-        "call-bind": "^1.0.2",
-        "define-properties": "^1.1.4",
-        "es-abstract": "^1.20.4",
-        "es-shim-unscopables": "^1.0.0"
-      }
-    },
     "assertion-error": {
       "version": "1.1.0",
       "resolved": "https://registry.npmjs.org/assertion-error/-/assertion-error-1.1.0.tgz",
@@ -5172,6 +5248,15 @@
         "type-detect": "^4.0.5"
       }
     },
+    "chai-as-promised": {
+      "version": "7.1.1",
+      "resolved": "https://registry.npmjs.org/chai-as-promised/-/chai-as-promised-7.1.1.tgz",
+      "integrity": "sha512-azL6xMoi+uxu6z4rhWQ1jbdUhOMhis2PvscD/xjLqNMkv3BPPp2JyyuTHOrf9BOosGpNQ11v6BKv/g57RXbiaA==",
+      "dev": true,
+      "requires": {
+        "check-error": "^1.0.2"
+      }
+    },
     "chalk": {
       "version": "4.1.2",
       "resolved": "https://registry.npmjs.org/chalk/-/chalk-4.1.2.tgz",
@@ -5707,35 +5792,33 @@
       }
     },
     "eslint-plugin-import": {
-      "version": "2.27.5",
+      "version": "2.26.0",
-      "resolved": "https://registry.npmjs.org/eslint-plugin-import/-/eslint-plugin-import-2.27.5.tgz",
+      "resolved": "https://registry.npmjs.org/eslint-plugin-import/-/eslint-plugin-import-2.26.0.tgz",
-      "integrity": "sha512-LmEt3GVofgiGuiE+ORpnvP+kAm3h6MLZJ4Q5HCyHADofsb4VzXFsRiWj3c0OFiV+3DWFh0qg3v9gcPlfc3zRow==",
+      "integrity": "sha512-hYfi3FXaM8WPLf4S1cikh/r4IxnO6zrhZbEGz2b660EJRbuxgpDS5gkCuYgGWg2xxh2rBuIr4Pvhve/7c31koA==",
       "dev": true,
       "requires": {
-        "array-includes": "^3.1.6",
+        "array-includes": "^3.1.4",
-        "array.prototype.flat": "^1.3.1",
+        "array.prototype.flat": "^1.2.5",
-        "array.prototype.flatmap": "^1.3.1",
+        "debug": "^2.6.9",
-        "debug": "^3.2.7",
         "doctrine": "^2.1.0",
-        "eslint-import-resolver-node": "^0.3.7",
+        "eslint-import-resolver-node": "^0.3.6",
-        "eslint-module-utils": "^2.7.4",
+        "eslint-module-utils": "^2.7.3",
         "has": "^1.0.3",
-        "is-core-module": "^2.11.0",
+        "is-core-module": "^2.8.1",
         "is-glob": "^4.0.3",
         "minimatch": "^3.1.2",
-        "object.values": "^1.1.6",
+        "object.values": "^1.1.5",
-        "resolve": "^1.22.1",
+        "resolve": "^1.22.0",
-        "semver": "^6.3.0",
         "tsconfig-paths": "^3.14.1"
       },
       "dependencies": {
         "debug": {
-          "version": "3.2.7",
+          "version": "2.6.9",
-          "resolved": "https://registry.npmjs.org/debug/-/debug-3.2.7.tgz",
+          "resolved": "https://registry.npmjs.org/debug/-/debug-2.6.9.tgz",
-          "integrity": "sha512-CFjzYYAi4ThfiQvizrFQevTTXHtnCqWfe7x1AhgEscTz6ZbLbfoLRLPugTQyBth6f8ZERVUSyWHFD/7Wu4t1XQ==",
+          "integrity": "sha512-bC7ElrdJaJnPbAP+1EotYvqZsb3ecl5wi6Bfi6BJTUcNowp6cvspg0jXznRTKDjm/E7AdgFBVeAPVMNcKGsHMA==",
           "dev": true,
           "requires": {
-            "ms": "^2.1.1"
+            "ms": "2.0.0"
           }
         },
         "doctrine": {
@@ -5747,10 +5830,10 @@
             "esutils": "^2.0.2"
           }
         },
-        "semver": {
+        "ms": {
-          "version": "6.3.0",
+          "version": "2.0.0",
-          "resolved": "https://registry.npmjs.org/semver/-/semver-6.3.0.tgz",
+          "resolved": "https://registry.npmjs.org/ms/-/ms-2.0.0.tgz",
-          "integrity": "sha512-b39TBaTSfV6yBrapU89p5fKekE2m/NwnDocOVruQFS1/veMgdzuPcnOM34M6CwxW8jH/lxEa5rBoDeUwu5HHTw==",
+          "integrity": "sha512-Tpp60P6IUJDTuOq/5Z8cdskzJujfwqfOTkrwIwj7IRISpnkJnT6SyJ4PCPnGMoFjC9ddhal5KVIYtAt97ix05A==",
           "dev": true
         }
       }
@@ -7078,9 +7161,9 @@
       }
     },
     "semver": {
-      "version": "7.5.0",
+      "version": "7.5.3",
-      "resolved": "https://registry.npmjs.org/semver/-/semver-7.5.0.tgz",
+      "resolved": "https://registry.npmjs.org/semver/-/semver-7.5.3.tgz",
-      "integrity": "sha512-+XC0AD/R7Q2mPSRuy2Id0+CGTZ98+8f+KvwirxOKIEyid+XSx6HbC63p+O4IndTHuX5Z+JxQ0TghCkO5Cg/2HA==",
+      "integrity": "sha512-QBlUtyVk/5EeHbi7X0fw6liDZc7BBmEaSYn01fMU1OUYbf6GPsbTtd8WmnqbI20SeycoHSeiybkE/q1Q+qlThQ==",
       "dev": true,
       "requires": {
         "lru-cache": "^6.0.0"
@@ -7536,6 +7619,24 @@
       "integrity": "sha512-wa7YjyUGfNZngI/vtK0UHAN+lgDCxBPCylVXGp0zu59Fz5aiGtNXaq3DhIov063MorB+VfufLh3JlF2KdTK3xg==",
       "dev": true
     },
+    "vectordb-darwin-arm64": {
+      "version": "0.1.13",
+      "resolved": "https://registry.npmjs.org/vectordb-darwin-arm64/-/vectordb-darwin-arm64-0.1.13.tgz",
+      "integrity": "sha512-9lLuX5P8m75EfP85pfC4LxO9J7Tzu4LngX55BVAdFe6qPRHu+iHmLw0QYYSVDqNm3GtDr2qFJlL2ILlsApyYyg==",
+      "optional": true
+    },
+    "vectordb-darwin-x64": {
+      "version": "0.1.13",
+      "resolved": "https://registry.npmjs.org/vectordb-darwin-x64/-/vectordb-darwin-x64-0.1.13.tgz",
+      "integrity": "sha512-5mkhBJlcfAqcty7Ww2csgYogq+b0NhtllAbag9IIznvqfcrvITU0H0vm5LGWbRuE/BUUxC25MJhm93YWBzqEVA==",
+      "optional": true
+    },
+    "vectordb-linux-x64-gnu": {
+      "version": "0.1.13",
+      "resolved": "https://registry.npmjs.org/vectordb-linux-x64-gnu/-/vectordb-linux-x64-gnu-0.1.13.tgz",
+      "integrity": "sha512-fU+sIHUkXyMdrWjggT93p0blKD+pbgr+x01tn9d2/pbA1ePo2AwuE86rYPA+BjyCUE1QifPgKadzGVVpqWYmnQ==",
+      "optional": true
+    },
     "vscode-oniguruma": {
       "version": "1.7.0",
       "resolved": "https://registry.npmjs.org/vscode-oniguruma/-/vscode-oniguruma-1.7.0.tgz",

node/package.json

@@ -1,16 +1,18 @@
 {
   "name": "vectordb",
-  "version": "0.1.5",
+  "version": "0.1.13",
   "description": " Serverless, low-latency vector database for AI applications",
   "main": "dist/index.js",
   "types": "dist/index.d.ts",
   "scripts": {
     "tsc": "tsc -b",
-    "build": "cargo-cp-artifact --artifact cdylib vectordb-node index.node -- cargo build --message-format=json-render-diagnostics",
+    "build": "cargo-cp-artifact --artifact cdylib vectordb-node index.node -- cargo build --message-format=json",
     "build-release": "npm run build -- --release",
-    "test": "mocha -recursive dist/test",
+    "test": "npm run tsc && mocha -recursive dist/test",
     "lint": "eslint src --ext .js,.ts",
-    "clean": "rm -rf node_modules *.node dist/"
+    "clean": "rm -rf node_modules *.node dist/",
+    "pack-build": "neon pack-build",
+    "check-npm": "printenv && which node && which npm && npm --version"
   },
   "repository": {
     "type": "git",
@@ -25,7 +27,9 @@
   "author": "Lance Devs",
   "license": "Apache-2.0",
   "devDependencies": {
+    "@neon-rs/cli": "^0.0.74",
     "@types/chai": "^4.3.4",
+    "@types/chai-as-promised": "^7.1.5",
     "@types/mocha": "^10.0.1",
     "@types/node": "^18.16.2",
     "@types/sinon": "^10.0.15",
@@ -33,9 +37,10 @@
     "@typescript-eslint/eslint-plugin": "^5.59.1",
     "cargo-cp-artifact": "^0.1",
     "chai": "^4.3.7",
+    "chai-as-promised": "^7.1.1",
     "eslint": "^8.39.0",
     "eslint-config-standard-with-typescript": "^34.0.1",
-    "eslint-plugin-import": "^2.27.5",
+    "eslint-plugin-import": "^2.26.0",
     "eslint-plugin-n": "^15.7.0",
     "eslint-plugin-promise": "^6.1.1",
     "mocha": "^10.2.0",
@@ -50,6 +55,32 @@
   },
   "dependencies": {
     "@apache-arrow/ts": "^12.0.0",
+    "@neon-rs/load": "^0.0.74",
     "apache-arrow": "^12.0.0"
+  },
+  "os": [
+    "darwin",
+    "linux",
+    "win32"
+  ],
+  "cpu": [
+    "x64",
+    "arm64"
+  ],
+  "neon": {
+    "targets": {
+      "x86_64-apple-darwin": "vectordb-darwin-x64",
+      "aarch64-apple-darwin": "vectordb-darwin-arm64",
+      "x86_64-unknown-linux-gnu": "vectordb-linux-x64-gnu",
+      "aarch64-unknown-linux-gnu": "vectordb-linux-arm64-gnu",
+      "x86_64-pc-windows-msvc": "vectordb-win32-x64-msvc"
+    }
+  },
+  "optionalDependencies": {
+    "vectordb-darwin-arm64": "0.1.13",
+    "vectordb-darwin-x64": "0.1.13",
+    "vectordb-linux-arm64-gnu": "0.1.13",
+    "vectordb-linux-x64-gnu": "0.1.13",
+    "vectordb-win32-x64-msvc": "0.1.13"
   }
 }

node/src/index.ts

@@ -22,34 +22,145 @@ import { fromRecordsToBuffer } from './arrow'
 import type { EmbeddingFunction } from './embedding/embedding_function'
 
 // eslint-disable-next-line @typescript-eslint/no-var-requires
-const { databaseNew, databaseTableNames, databaseOpenTable, tableCreate, tableSearch, tableAdd, tableCreateVectorIndex } = require('../native.js')
+const { databaseNew, databaseTableNames, databaseOpenTable, databaseDropTable, tableCreate, tableSearch, tableAdd, tableCreateVectorIndex, tableCountRows, tableDelete } = require('../native.js')
 
 export type { EmbeddingFunction }
 export { OpenAIEmbeddingFunction } from './embedding/openai'
 
+export interface AwsCredentials {
+  accessKeyId: string
+
+  secretKey: string
+
+  sessionToken?: string
+}
+
+export interface ConnectionOptions {
+  uri: string
+  awsCredentials?: AwsCredentials
+}
+
 /**
  * Connect to a LanceDB instance at the given URI
  * @param uri The uri of the database.
  */
-export async function connect (uri: string): Promise<Connection> {
+export async function connect (uri: string): Promise<Connection>
-  const db = await databaseNew(uri)
+export async function connect (opts: Partial<ConnectionOptions>): Promise<Connection>
-  return new Connection(db, uri)
+export async function connect (arg: string | Partial<ConnectionOptions>): Promise<Connection> {
+  let opts: ConnectionOptions
+  if (typeof arg === 'string') {
+    opts = { uri: arg }
+  } else {
+    // opts = { uri: arg.uri, awsCredentials = arg.awsCredentials }
+    opts = Object.assign({
+      uri: '',
+      awsCredentials: undefined
+    }, arg)
+  }
+  const db = await databaseNew(opts.uri)
+  return new LocalConnection(db, opts)
+}
+
+/**
+ * A LanceDB Connection that allows you to open tables and create new ones.
+ *
+ * Connection could be local against filesystem or remote against a server.
+ */
+export interface Connection {
+  uri: string
+
+  tableNames(): Promise<string[]>
+
+  /**
+   * Open a table in the database.
+   *
+   * @param name The name of the table.
+   * @param embeddings An embedding function to use on this table
+   */
+  openTable<T>(name: string, embeddings?: EmbeddingFunction<T>): Promise<Table<T>>
+
+  /**
+   * Creates a new Table and initialize it with new data.
+   *
+   * @param {string} name - The name of the table.
+   * @param data - Non-empty Array of Records to be inserted into the table
+   * @param {WriteMode} mode - The write mode to use when creating the table.
+   * @param {EmbeddingFunction} embeddings - An embedding function to use on this table
+   */
+  createTable<T>(name: string, data: Array<Record<string, unknown>>, mode?: WriteMode, embeddings?: EmbeddingFunction<T>): Promise<Table<T>>
+
+  createTableArrow(name: string, table: ArrowTable): Promise<Table>
+
+  /**
+   * Drop an existing table.
+   * @param name The name of the table to drop.
+   */
+  dropTable(name: string): Promise<void>
+
+}
+
+/**
+ * A LanceDB Table is the collection of Records. Each Record has one or more vector fields.
+ */
+export interface Table<T = number[]> {
+  name: string
+
+  /**
+   * Creates a search query to find the nearest neighbors of the given search term
+   * @param query The query search term
+   */
+  search: (query: T) => Query<T>
+
+  /**
+   * Insert records into this Table.
+   *
+   * @param data Records to be inserted into the Table
+   * @return The number of rows added to the table
+   */
+  add: (data: Array<Record<string, unknown>>) => Promise<number>
+
+  /**
+   * Insert records into this Table, replacing its contents.
+   *
+   * @param data Records to be inserted into the Table
+   * @return The number of rows added to the table
+   */
+  overwrite: (data: Array<Record<string, unknown>>) => Promise<number>
+
+  /**
+   * Create an ANN index on this Table vector index.
+   *
+   * @param indexParams The parameters of this Index, @see VectorIndexParams.
+   */
+  createIndex: (indexParams: VectorIndexParams) => Promise<any>
+
+  /**
+   * Returns the number of rows in this table.
+   */
+  countRows: () => Promise<number>
+
+  /**
+   * Delete rows from this table.
+   *
+   * @param filter  A filter in the same format used by a sql WHERE clause.
+   */
+  delete: (filter: string) => Promise<void>
 }
 
 /**
  * A connection to a LanceDB database.
  */
-export class Connection {
+export class LocalConnection implements Connection {
-  private readonly _uri: string
+  private readonly _options: ConnectionOptions
   private readonly _db: any
 
-  constructor (db: any, uri: string) {
+  constructor (db: any, options: ConnectionOptions) {
-    this._uri = uri
+    this._options = options
     this._db = db
   }
 
   get uri (): string {
-    return this._uri
+    return this._options.uri
   }
 
   /**
@@ -72,12 +183,13 @@ export class Connection {
    * @param embeddings An embedding function to use on this Table
    */
   async openTable<T> (name: string, embeddings: EmbeddingFunction<T>): Promise<Table<T>>
+  async openTable<T> (name: string, embeddings?: EmbeddingFunction<T>): Promise<Table<T>>
   async openTable<T> (name: string, embeddings?: EmbeddingFunction<T>): Promise<Table<T>> {
     const tbl = await databaseOpenTable.call(this._db, name)
     if (embeddings !== undefined) {
-      return new Table(tbl, name, embeddings)
+      return new LocalTable(tbl, name, this._options, embeddings)
     } else {
-      return new Table(tbl, name)
+      return new LocalTable(tbl, name, this._options)
     }
   }
 
@@ -86,23 +198,41 @@ export class Connection {
    *
    * @param name The name of the table.
    * @param data Non-empty Array of Records to be inserted into the Table
+   * @param mode The write mode to use when creating the table.
    */
+  async createTable (name: string, data: Array<Record<string, unknown>>, mode?: WriteMode): Promise<Table>
+  async createTable (name: string, data: Array<Record<string, unknown>>, mode: WriteMode): Promise<Table>
 
-  async createTable (name: string, data: Array<Record<string, unknown>>): Promise<Table>
   /**
    * Creates a new Table and initialize it with new data.
    *
    * @param name The name of the table.
    * @param data Non-empty Array of Records to be inserted into the Table
+   * @param mode The write mode to use when creating the table.
    * @param embeddings An embedding function to use on this Table
    */
-  async createTable<T> (name: string, data: Array<Record<string, unknown>>, embeddings: EmbeddingFunction<T>): Promise<Table<T>>
+  async createTable<T> (name: string, data: Array<Record<string, unknown>>, mode: WriteMode, embeddings: EmbeddingFunction<T>): Promise<Table<T>>
-  async createTable<T> (name: string, data: Array<Record<string, unknown>>, embeddings?: EmbeddingFunction<T>): Promise<Table<T>> {
+  async createTable<T> (name: string, data: Array<Record<string, unknown>>, mode: WriteMode, embeddings?: EmbeddingFunction<T>): Promise<Table<T>>
-    const tbl = await tableCreate.call(this._db, name, await fromRecordsToBuffer(data, embeddings))
+  async createTable<T> (name: string, data: Array<Record<string, unknown>>, mode: WriteMode, embeddings?: EmbeddingFunction<T>): Promise<Table<T>> {
+    if (mode === undefined) {
+      mode = WriteMode.Create
+    }
+
+    const createArgs = [this._db, name, await fromRecordsToBuffer(data, embeddings), mode.toLowerCase()]
+    if (this._options.awsCredentials !== undefined) {
+      createArgs.push(this._options.awsCredentials.accessKeyId)
+      createArgs.push(this._options.awsCredentials.secretKey)
+      if (this._options.awsCredentials.sessionToken !== undefined) {
+        createArgs.push(this._options.awsCredentials.sessionToken)
+      }
+    }
+
+    const tbl = await tableCreate.call(...createArgs)
+
     if (embeddings !== undefined) {
-      return new Table(tbl, name, embeddings)
+      return new LocalTable(tbl, name, this._options, embeddings)
     } else {
-      return new Table(tbl, name)
+      return new LocalTable(tbl, name, this._options)
     }
   }
 
@@ -111,24 +241,35 @@ export class Connection {
     await tableCreate.call(this._db, name, Buffer.from(await writer.toUint8Array()))
     return await this.openTable(name)
   }
+
+  /**
+   * Drop an existing table.
+   * @param name The name of the table to drop.
+   */
+  async dropTable (name: string): Promise<void> {
+    await databaseDropTable.call(this._db, name)
+  }
 }
 
-export class Table<T = number[]> {
+export class LocalTable<T = number[]> implements Table<T> {
   private readonly _tbl: any
   private readonly _name: string
   private readonly _embeddings?: EmbeddingFunction<T>
+  private readonly _options: ConnectionOptions
 
-  constructor (tbl: any, name: string)
+  constructor (tbl: any, name: string, options: ConnectionOptions)
   /**
    * @param tbl
    * @param name
+   * @param options
    * @param embeddings An embedding function to use when interacting with this table
    */
-  constructor (tbl: any, name: string, embeddings: EmbeddingFunction<T>)
+  constructor (tbl: any, name: string, options: ConnectionOptions, embeddings: EmbeddingFunction<T>)
-  constructor (tbl: any, name: string, embeddings?: EmbeddingFunction<T>) {
+  constructor (tbl: any, name: string, options: ConnectionOptions, embeddings?: EmbeddingFunction<T>) {
     this._tbl = tbl
     this._name = name
     this._embeddings = embeddings
+    this._options = options
   }
 
   get name (): string {
@@ -150,7 +291,15 @@ export class Table<T = number[]> {
    * @return The number of rows added to the table
    */
   async add (data: Array<Record<string, unknown>>): Promise<number> {
-    return tableAdd.call(this._tbl, await fromRecordsToBuffer(data, this._embeddings), WriteMode.Append.toString())
+    const callArgs = [this._tbl, await fromRecordsToBuffer(data, this._embeddings), WriteMode.Append.toString()]
+    if (this._options.awsCredentials !== undefined) {
+      callArgs.push(this._options.awsCredentials.accessKeyId)
+      callArgs.push(this._options.awsCredentials.secretKey)
+      if (this._options.awsCredentials.sessionToken !== undefined) {
+        callArgs.push(this._options.awsCredentials.sessionToken)
+      }
+    }
+    return tableAdd.call(...callArgs)
   }
 
   /**
@@ -160,6 +309,14 @@ export class Table<T = number[]> {
    * @return The number of rows added to the table
    */
   async overwrite (data: Array<Record<string, unknown>>): Promise<number> {
+    const callArgs = [this._tbl, await fromRecordsToBuffer(data, this._embeddings), WriteMode.Overwrite.toString()]
+    if (this._options.awsCredentials !== undefined) {
+      callArgs.push(this._options.awsCredentials.accessKeyId)
+      callArgs.push(this._options.awsCredentials.secretKey)
+      if (this._options.awsCredentials.sessionToken !== undefined) {
+        callArgs.push(this._options.awsCredentials.sessionToken)
+      }
+    }
     return tableAdd.call(this._tbl, await fromRecordsToBuffer(data, this._embeddings), WriteMode.Overwrite.toString())
   }
 
@@ -173,14 +330,25 @@ export class Table<T = number[]> {
   }
 
   /**
-   * @deprecated Use [Table.createIndex]
+   * Returns the number of rows in this table.
    */
-  async create_index (indexParams: VectorIndexParams): Promise<any> {
+  async countRows (): Promise<number> {
-    return await this.createIndex(indexParams)
+    return tableCountRows.call(this._tbl)
+  }
+
+  /**
+   * Delete rows from this table.
+   *
+   * @param filter A filter in the same format used by a sql WHERE clause.
+   */
+  async delete (filter: string): Promise<void> {
+    return tableDelete.call(this._tbl, filter)
   }
 }
 
-interface IvfPQIndexConfig {
+/// Config to build IVF_PQ index.
+///
+export interface IvfPQIndexConfig {
   /**
    * The column to be indexed
    */
@@ -225,6 +393,11 @@ interface IvfPQIndexConfig {
    */
   max_opq_iters?: number
 
+  /**
+   * Replace an existing index with the same name if it exists.
+   */
+  replace?: boolean
+
   type: 'ivf_pq'
 }
 
@@ -293,6 +466,8 @@ export class Query<T = number[]> {
     return this
   }
 
+  where = this.filter
+
   /** Return only the specified columns.
    *
    * @param value Only select the specified columns. If not specified, all columns will be returned.
@@ -323,6 +498,7 @@ export class Query<T = number[]> {
 
     const buffer = await tableSearch.call(this._tbl, this)
     const data = tableFromIPC(buffer)
+
     return data.toArray().map((entry: Record<string, unknown>) => {
       const newObject: Record<string, unknown> = {}
       Object.keys(entry).forEach((key: string) => {
@@ -337,8 +513,15 @@ export class Query<T = number[]> {
   }
 }
 
+/**
+ * Write mode for writing a table.
+ */
 export enum WriteMode {
+  /** Create a new {@link Table}. */
+  Create = 'create',
+  /** Overwrite the existing {@link Table} if presented. */
   Overwrite = 'overwrite',
+  /** Append new data to the table. */
   Append = 'append'
 }
 
@@ -354,5 +537,10 @@ export enum MetricType {
   /**
    * Cosine distance
    */
-  Cosine = 'cosine'
+  Cosine = 'cosine',
+
+  /**
+   * Dot product
+   */
+  Dot = 'dot'
 }

node/src/test/io.ts

@@ -18,26 +18,48 @@ import { describe } from 'mocha'
 import { assert } from 'chai'
 
 import * as lancedb from '../index'
+import { type ConnectionOptions } from '../index'
 
 describe('LanceDB S3 client', function () {
   if (process.env.TEST_S3_BASE_URL != null) {
     const baseUri = process.env.TEST_S3_BASE_URL
     it('should have a valid url', async function () {
-      const uri = `${baseUri}/valid_url`
+      const opts = { uri: `${baseUri}/valid_url` }
-      const table = await createTestDB(uri, 2, 20)
+      const table = await createTestDB(opts, 2, 20)
-      const con = await lancedb.connect(uri)
+      const con = await lancedb.connect(opts)
-      assert.equal(con.uri, uri)
+      assert.equal(con.uri, opts.uri)
 
       const results = await table.search([0.1, 0.3]).limit(5).execute()
       assert.equal(results.length, 5)
-    })
+    }).timeout(10_000)
+  } else {
+    describe.skip('Skip S3 test', function () {})
+  }
+
+  if (process.env.TEST_S3_BASE_URL != null && process.env.TEST_AWS_ACCESS_KEY_ID != null && process.env.TEST_AWS_SECRET_ACCESS_KEY != null) {
+    const baseUri = process.env.TEST_S3_BASE_URL
+    it('use custom credentials', async function () {
+      const opts: ConnectionOptions = {
+        uri: `${baseUri}/custom_credentials`,
+        awsCredentials: {
+          accessKeyId: process.env.TEST_AWS_ACCESS_KEY_ID as string,
+          secretKey: process.env.TEST_AWS_SECRET_ACCESS_KEY as string
+        }
+      }
+      const table = await createTestDB(opts, 2, 20)
+      const con = await lancedb.connect(opts)
+      assert.equal(con.uri, opts.uri)
+
+      const results = await table.search([0.1, 0.3]).limit(5).execute()
+      assert.equal(results.length, 5)
+    }).timeout(10_000)
   } else {
     describe.skip('Skip S3 test', function () {})
   }
 })
 
-async function createTestDB (uri: string, numDimensions: number = 2, numRows: number = 2): Promise<lancedb.Table> {
+async function createTestDB (opts: ConnectionOptions, numDimensions: number = 2, numRows: number = 2): Promise<lancedb.Table> {
-  const con = await lancedb.connect(uri)
+  const con = await lancedb.connect(opts)
 
   const data = []
   for (let i = 0; i < numRows; i++) {

node/src/test/test.ts

@@ -1,4 +1,4 @@
-// Copyright 2023 Lance Developers.
+// Copyright 2023 LanceDB Developers.
 //
 // Licensed under the Apache License, Version 2.0 (the "License");
 // you may not use this file except in compliance with the License.
@@ -13,11 +13,16 @@
 // limitations under the License.
 
 import { describe } from 'mocha'
-import { assert } from 'chai'
 import { track } from 'temp'
+import * as chai from 'chai'
+import * as chaiAsPromised from 'chai-as-promised'
 
 import * as lancedb from '../index'
-import { type EmbeddingFunction, MetricType, Query } from '../index'
+import { type AwsCredentials, type EmbeddingFunction, MetricType, Query, WriteMode } from '../index'
+
+const expect = chai.expect
+const assert = chai.assert
+chai.use(chaiAsPromised)
 
 describe('LanceDB client', function () {
   describe('when creating a connection to lancedb', function () {
@@ -27,6 +32,22 @@ describe('LanceDB client', function () {
       assert.equal(con.uri, uri)
     })
 
+    it('should accept an options object', async function () {
+      const uri = await createTestDB()
+      const con = await lancedb.connect({ uri })
+      assert.equal(con.uri, uri)
+    })
+
+    it('should accept custom aws credentials', async function () {
+      const uri = await createTestDB()
+      const awsCredentials: AwsCredentials = {
+        accessKeyId: '',
+        secretKey: ''
+      }
+      const con = await lancedb.connect({ uri, awsCredentials })
+      assert.equal(con.uri, uri)
+    })
+
     it('should return the existing table names', async function () {
       const uri = await createTestDB()
       const con = await lancedb.connect(uri)
@@ -64,13 +85,20 @@ describe('LanceDB client', function () {
       assert.equal(results[0].id, 1)
     })
 
-    it('uses a filter', async function () {
+    it('uses a filter / where clause', async function () {
+      // eslint-disable-next-line @typescript-eslint/explicit-function-return-type
+      const assertResults = (results: Array<Record<string, unknown>>) => {
+        assert.equal(results.length, 1)
+        assert.equal(results[0].id, 2)
+      }
+
       const uri = await createTestDB()
       const con = await lancedb.connect(uri)
       const table = await con.openTable('vectors')
-      const results = await table.search([0.1, 0.1]).filter('id == 2').execute()
+      let results = await table.search([0.1, 0.1]).filter('id == 2').execute()
-      assert.equal(results.length, 1)
+      assertResults(results)
-      assert.equal(results[0].id, 2)
+      results = await table.search([0.1, 0.1]).where('id == 2').execute()
+      assertResults(results)
     })
 
     it('select only a subset of columns', async function () {
@@ -103,9 +131,32 @@ describe('LanceDB client', function () {
       const tableName = `vectors_${Math.floor(Math.random() * 100)}`
       const table = await con.createTable(tableName, data)
       assert.equal(table.name, tableName)
+      assert.equal(await table.countRows(), 2)
+    })
 
-      const results = await table.search([0.1, 0.3]).execute()
+    it('use overwrite flag to overwrite existing table', async function () {
-      assert.equal(results.length, 2)
+      const dir = await track().mkdir('lancejs')
+      const con = await lancedb.connect(dir)
+
+      const data = [
+        { id: 1, vector: [0.1, 0.2], price: 10 },
+        { id: 2, vector: [1.1, 1.2], price: 50 }
+      ]
+
+      const tableName = 'overwrite'
+      await con.createTable(tableName, data, WriteMode.Create)
+
+      const newData = [
+        { id: 1, vector: [0.1, 0.2], price: 10 },
+        { id: 2, vector: [1.1, 1.2], price: 50 },
+        { id: 3, vector: [1.1, 1.2], price: 50 }
+      ]
+
+      await expect(con.createTable(tableName, newData)).to.be.rejectedWith(Error, 'already exists')
+
+      const table = await con.createTable(tableName, newData, WriteMode.Overwrite)
+      assert.equal(table.name, tableName)
+      assert.equal(await table.countRows(), 3)
     })
 
     it('appends records to an existing table ', async function () {
@@ -118,16 +169,14 @@ describe('LanceDB client', function () {
       ]
 
       const table = await con.createTable('vectors', data)
-      const results = await table.search([0.1, 0.3]).execute()
+      assert.equal(await table.countRows(), 2)
-      assert.equal(results.length, 2)
 
       const dataAdd = [
         { id: 3, vector: [2.1, 2.2], price: 10, name: 'c' },
         { id: 4, vector: [3.1, 3.2], price: 50, name: 'd' }
       ]
       await table.add(dataAdd)
-      const resultsAdd = await table.search([0.1, 0.3]).execute()
+      assert.equal(await table.countRows(), 4)
-      assert.equal(resultsAdd.length, 4)
     })
 
     it('overwrite all records in a table', async function () {
@@ -135,16 +184,25 @@ describe('LanceDB client', function () {
       const con = await lancedb.connect(uri)
 
       const table = await con.openTable('vectors')
-      const results = await table.search([0.1, 0.3]).execute()
+      assert.equal(await table.countRows(), 2)
-      assert.equal(results.length, 2)
 
       const dataOver = [
         { vector: [2.1, 2.2], price: 10, name: 'foo' },
         { vector: [3.1, 3.2], price: 50, name: 'bar' }
       ]
       await table.overwrite(dataOver)
-      const resultsAdd = await table.search([0.1, 0.3]).execute()
+      assert.equal(await table.countRows(), 2)
-      assert.equal(resultsAdd.length, 2)
+    })
+
+    it('can delete records from a table', async function () {
+      const uri = await createTestDB()
+      const con = await lancedb.connect(uri)
+
+      const table = await con.openTable('vectors')
+      assert.equal(await table.countRows(), 2)
+
+      await table.delete('price = 10')
+      assert.equal(await table.countRows(), 1)
     })
   })
 
@@ -153,8 +211,25 @@ describe('LanceDB client', function () {
       const uri = await createTestDB(32, 300)
       const con = await lancedb.connect(uri)
       const table = await con.openTable('vectors')
-      await table.createIndex({ type: 'ivf_pq', column: 'vector', num_partitions: 2, max_iters: 2 })
+      await table.createIndex({ type: 'ivf_pq', column: 'vector', num_partitions: 2, max_iters: 2, num_sub_vectors: 2 })
     }).timeout(10_000) // Timeout is high partially because GH macos runner is pretty slow
+
+    it('replace an existing index', async function () {
+      const uri = await createTestDB(16, 300)
+      const con = await lancedb.connect(uri)
+      const table = await con.openTable('vectors')
+
+      await table.createIndex({ type: 'ivf_pq', column: 'vector', num_partitions: 2, max_iters: 2, num_sub_vectors: 2 })
+
+      // Replace should fail if the index already exists
+      await expect(table.createIndex({
+        type: 'ivf_pq', column: 'vector', num_partitions: 2, max_iters: 2, num_sub_vectors: 2, replace: false
+      })
+      ).to.be.rejectedWith('LanceError(Index)')
+
+      // Default replace = true
+      await table.createIndex({ type: 'ivf_pq', column: 'vector', num_partitions: 2, max_iters: 2, num_sub_vectors: 2 })
+    }).timeout(50_000)
   })
 
   describe('when using a custom embedding function', function () {
@@ -184,7 +259,7 @@ describe('LanceDB client', function () {
         { price: 10, name: 'foo' },
         { price: 50, name: 'bar' }
       ]
-      const table = await con.createTable('vectors', data, embeddings)
+      const table = await con.createTable('vectors', data, WriteMode.Create, embeddings)
       const results = await table.search('foo').execute()
       assert.equal(results.length, 2)
     })
@@ -223,3 +298,22 @@ async function createTestDB (numDimensions: number = 2, numRows: number = 2): Pr
   await con.createTable('vectors', data)
   return dir
 }
+
+describe('Drop table', function () {
+  it('drop a table', async function () {
+    const dir = await track().mkdir('lancejs')
+    const con = await lancedb.connect(dir)
+
+    const data = [
+      { price: 10, name: 'foo', vector: [1, 2, 3] },
+      { price: 50, name: 'bar', vector: [4, 5, 6] }
+    ]
+    await con.createTable('t1', data)
+    await con.createTable('t2', data)
+
+    assert.deepEqual(await con.tableNames(), ['t1', 't2'])
+
+    await con.dropTable('t1')
+    assert.deepEqual(await con.tableNames(), ['t2'])
+  })
+})

python/.bumpversion.cfg

@@ -1,5 +1,5 @@
 [bumpversion]
-current_version = 0.1.7
+current_version = 0.1.8
 commit = True
 message = [python] Bump version: {current_version} → {new_version}
 tag = True

python/README.md

@@ -0,0 +1,85 @@
+# LanceDB
+
+A Python library for [LanceDB](https://github.com/lancedb/lancedb).
+
+## Installation
+
+```bash
+pip install lancedb
+```
+
+## Usage
+
+### Basic Example
+
+```python
+import lancedb
+db = lancedb.connect('<PATH_TO_LANCEDB_DATASET>')
+table = db.open_table('my_table')
+results = table.search([0.1, 0.3]).limit(20).to_df()
+print(results)
+```
+
+
+## Development
+
+Create a virtual environment and activate it:
+
+```bash
+python -m venv venv
+. ./venv/bin/activate
+```
+
+Install the necessary packages:
+
+```bash
+python -m pip install .
+```
+
+To run the unit tests:
+
+```bash
+pytest
+```
+
+To run linter and automatically fix all errors:
+
+```bash
+black .
+isort .
+```
+
+If any packages are missing, install them with:
+
+```bash
+pip install <PACKAGE_NAME>
+```
+
+
+___
+For **Windows** users, there may be errors when installing packages, so these commands may be helpful:
+
+Activate the virtual environment:
+```bash
+. .\venv\Scripts\activate
+```
+
+You may need to run the installs separately:
+```bash
+pip install -e .[tests]
+pip install -e .[dev]
+```
+
+
+`tantivy` requires `rust` to be installed, so install it with `conda`, as it doesn't support windows installation:
+```bash
+pip install wheel
+pip install cargo
+conda install rust
+pip install tantivy
+```
+
+To run the unit tests:
+```bash
+pytest
+```

python/lancedb/__init__.py

@@ -11,16 +11,25 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 
-from .db import URI, LanceDBConnection
+from typing import Optional
+
+from .db import URI, DBConnection, LanceDBConnection
+from .remote.db import RemoteDBConnection
+from .schema import vector
 
 
-def connect(uri: URI) -> LanceDBConnection:
+def connect(
-    """Connect to a LanceDB instance at the given URI
+    uri: URI, *, api_key: Optional[str] = None, region: str = "us-west-2"
+) -> DBConnection:
+    """Connect to a LanceDB database.
 
     Parameters
     ----------
     uri: str or Path
         The uri of the database.
+    api_token: str, optional
+        If presented, connect to LanceDB cloud.
+        Otherwise, connect to a database on file system or cloud storage.
 
     Examples
     --------
@@ -34,9 +43,17 @@ def connect(uri: URI) -> LanceDBConnection:
 
     >>> db = lancedb.connect("s3://my-bucket/lancedb")
 
+    Connect to LancdDB cloud:
+
+    >>> db = lancedb.connect("db://my_database", api_key="ldb_...")
+
     Returns
     -------
-    conn : LanceDBConnection
+    conn : DBConnection
         A connection to a LanceDB database.
     """
+    if isinstance(uri, str) and uri.startswith("db://"):
+        if api_key is None:
+            raise ValueError(f"api_key is required to connected LanceDB cloud: {uri}")
+        return RemoteDBConnection(uri, api_key, region)
     return LanceDBConnection(uri)

python/lancedb/common.py

@@ -23,3 +23,13 @@ URI = Union[str, Path]
 # TODO support generator
 DATA = Union[List[dict], dict, pd.DataFrame]
 VECTOR_COLUMN_NAME = "vector"
+
+
+class Credential(str):
+    """Credential field"""
+
+    def __repr__(self) -> str:
+        return "********"
+
+    def __str__(self) -> str:
+        return "********"

python/lancedb/conftest.py

@@ -1,10 +1,8 @@
-import builtins
 import os
 
 import pytest
 
 # import lancedb so we don't have to in every example
-import lancedb
 
 
 @pytest.fixture(autouse=True)

python/lancedb/context.py

@@ -13,7 +13,8 @@
 from __future__ import annotations
 
 import pandas as pd
-from .exceptions import MissingValueError, MissingColumnError
+
+from .exceptions import MissingColumnError, MissingValueError
 
 
 def contextualize(raw_df: pd.DataFrame) -> Contextualizer:
@@ -52,14 +53,16 @@ def contextualize(raw_df: pd.DataFrame) -> Contextualizer:
     6        the lazy dog            1
     7          lazy dog I            1
     8          dog I love            1
-    >>> contextualize(data).window(7).stride(1).text_col('token').to_df()
+    9   I love sandwiches            2
+    10    love sandwiches            2
+    >>> contextualize(data).window(7).stride(1).min_window_size(7).text_col('token').to_df()
                                       token  document_id
     0   The quick brown fox jumped over the            1
     1  quick brown fox jumped over the lazy            1
     2    brown fox jumped over the lazy dog            1
     3        fox jumped over the lazy dog I            1
     4       jumped over the lazy dog I love            1
-
+    5   over the lazy dog I love sandwiches            1
 
     ``stride`` determines how many rows to skip between each window start. This can
     be used to reduce the total number of windows generated.
@@ -70,6 +73,8 @@ def contextualize(raw_df: pd.DataFrame) -> Contextualizer:
     2   brown fox jumped over            1
     4    jumped over the lazy            1
     6          the lazy dog I            1
+    8   dog I love sandwiches            1
+    10        love sandwiches            2
 
     ``groupby`` determines how to group the rows. For example, we would like to have
     context windows that don't cross document boundaries. In this case, we can
@@ -80,6 +85,25 @@ def contextualize(raw_df: pd.DataFrame) -> Contextualizer:
     0    The quick brown fox            1
     2  brown fox jumped over            1
     4   jumped over the lazy            1
+    6           the lazy dog            1
+    9      I love sandwiches            2
+
+    ``min_window_size`` determines the minimum size of the  context windows that are generated
+    This can be used to trim the last few context windows which have size less than
+    ``min_window_size``. By default context windows of size 1 are skipped.
+
+    >>> contextualize(data).window(6).stride(3).text_col('token').groupby('document_id').to_df()
+                                 token  document_id
+    0  The quick brown fox jumped over            1
+    3     fox jumped over the lazy dog            1
+    6                     the lazy dog            1
+    9                I love sandwiches            2
+
+    >>> contextualize(data).window(6).stride(3).min_window_size(4).text_col('token').groupby('document_id').to_df()
+                                 token  document_id
+    0  The quick brown fox jumped over            1
+    3     fox jumped over the lazy dog            1
+
     """
     return Contextualizer(raw_df)
 
@@ -92,6 +116,7 @@ class Contextualizer:
         self._groupby = None
         self._stride = None
         self._window = None
+        self._min_window_size = 2
         self._raw_df = raw_df
 
     def window(self, window: int) -> Contextualizer:
@@ -139,6 +164,17 @@ class Contextualizer:
         self._text_col = text_col
         return self
 
+    def min_window_size(self, min_window_size: int) -> Contextualizer:
+        """Set the (optional) min_window_size size for the context window.
+
+        Parameters
+        ----------
+        min_window_size: int
+            The min_window_size.
+        """
+        self._min_window_size = min_window_size
+        return self
+
     def to_df(self) -> pd.DataFrame:
         """Create the context windows and return a DataFrame."""
 
@@ -159,12 +195,19 @@ class Contextualizer:
 
         def process_group(grp):
             # For each group, create the text rolling window
+            # with values of size >= min_window_size
             text = grp[self._text_col].values
-            contexts = grp.iloc[: -self._window : self._stride, :].copy()
+            contexts = grp.iloc[:: self._stride, :].copy()
-            contexts[self._text_col] = [
+            windows = [
-                " ".join(text[start_i : start_i + self._window])
+                " ".join(text[start_i : min(start_i + self._window, len(grp))])
-                for start_i in range(0, len(grp) - self._window, self._stride)
+                for start_i in range(0, len(grp), self._stride)
+                if start_i + self._window <= len(grp)
+                or len(grp) - start_i >= self._min_window_size
             ]
+            # if last few rows dropped
+            if len(windows) < len(contexts):
+                contexts = contexts.iloc[: len(windows)]
+            contexts[self._text_col] = windows
             return contexts
 
         if self._groupby is None:

python/lancedb/db.py

@@ -14,128 +14,69 @@
 from __future__ import annotations
 
 import os
+from abc import ABC, abstractmethod
 from pathlib import Path
-import os
+from typing import Dict, Iterable, List, Optional, Tuple, Union
 
+import pandas as pd
 import pyarrow as pa
 from pyarrow import fs
 
 from .common import DATA, URI
-from .table import LanceTable
+from .table import LanceTable, Table
-from .util import get_uri_scheme, get_uri_location
+from .util import get_uri_location, get_uri_scheme
 
 
-class LanceDBConnection:
+class DBConnection(ABC):
-    """
+    """An active LanceDB connection interface."""
-    A connection to a LanceDB database.
-
-    Parameters
-    ----------
-    uri: str or Path
-        The root uri of the database.
-
-    Examples
-    --------
-    >>> import lancedb
-    >>> db = lancedb.connect("./.lancedb")
-    >>> db.create_table("my_table", data=[{"vector": [1.1, 1.2], "b": 2},
-    ...                                   {"vector": [0.5, 1.3], "b": 4}])
-    LanceTable(my_table)
-    >>> db.create_table("another_table", data=[{"vector": [0.4, 0.4], "b": 6}])
-    LanceTable(another_table)
-    >>> db.table_names()
-    ['another_table', 'my_table']
-    >>> len(db)
-    2
-    >>> db["my_table"]
-    LanceTable(my_table)
-    >>> "my_table" in db
-    True
-    >>> db.drop_table("my_table")
-    >>> db.drop_table("another_table")
-    """
-
-    def __init__(self, uri: URI):
-        is_local = isinstance(uri, Path) or get_uri_scheme(uri) == "file"
-        if is_local:
-            if isinstance(uri, str):
-                uri = Path(uri)
-            uri = uri.expanduser().absolute()
-            Path(uri).mkdir(parents=True, exist_ok=True)
-        self._uri = str(uri)
-
-    @property
-    def uri(self) -> str:
-        return self._uri
 
+    @abstractmethod
     def table_names(self) -> list[str]:
-        """Get the names of all tables in the database.
+        """List all table names in the database."""
-
+        pass
-        Returns
-        -------
-        list of str
-            A list of table names.
-        """
-        try:
-            filesystem, path = fs.FileSystem.from_uri(self.uri)
-        except pa.ArrowInvalid:
-            raise NotImplementedError("Unsupported scheme: " + self.uri)
-
-        try:
-            paths = filesystem.get_file_info(
-                fs.FileSelector(get_uri_location(self.uri))
-            )
-        except FileNotFoundError:
-            # It is ok if the file does not exist since it will be created
-            paths = []
-        tables = [
-            os.path.splitext(file_info.base_name)[0]
-            for file_info in paths
-            if file_info.extension == "lance"
-        ]
-        return tables
-
-    def __len__(self) -> int:
-        return len(self.table_names())
-
-    def __contains__(self, name: str) -> bool:
-        return name in self.table_names()
-
-    def __getitem__(self, name: str) -> LanceTable:
-        return self.open_table(name)
 
+    @abstractmethod
     def create_table(
         self,
         name: str,
-        data: DATA = None,
+        data: Optional[
-        schema: pa.Schema = None,
+            Union[List[dict], dict, pd.DataFrame, pa.Table, Iterable[pa.RecordBatch]],
+        ] = None,
+        schema: Optional[pa.Schema] = None,
         mode: str = "create",
-    ) -> LanceTable:
+        on_bad_vectors: str = "error",
-        """Create a table in the database.
+        fill_value: float = 0.0,
+    ) -> Table:
+        """Create a [Table][lancedb.table.Table] in the database.
 
         Parameters
         ----------
         name: str
             The name of the table.
         data: list, tuple, dict, pd.DataFrame; optional
-            The data to insert into the table.
+            The data to initialize the table. User must provide at least one of `data` or `schema`.
         schema: pyarrow.Schema; optional
             The schema of the table.
         mode: str; default "create"
-            The mode to use when creating the table.
+            The mode to use when creating the table. Can be either "create" or "overwrite".
             By default, if the table already exists, an exception is raised.
             If you want to overwrite the table, use mode="overwrite".
-
+        on_bad_vectors: str, default "error"
-        Note
+            What to do if any of the vectors are not the same size or contains NaNs.
-        ----
+            One of "error", "drop", "fill".
-        The vector index won't be created by default.
+        fill_value: float
-        To create the index, call the `create_index` method on the table.
+            The value to use when filling vectors. Only used if on_bad_vectors="fill".
 
         Returns
         -------
         LanceTable
             A reference to the newly created table.
 
+        !!! note
+
+            The vector index won't be created by default.
+            To create the index, call the `create_index` method on the table.
+
         Examples
         --------
 
@@ -181,7 +122,7 @@ class LanceDBConnection:
 
         Data is converted to Arrow before being written to disk. For maximum
         control over how data is saved, either provide the PyArrow schema to
-        convert to or else provide a PyArrow table directly.
+        convert to or else provide a [PyArrow Table](pyarrow.Table) directly.
 
         >>> custom_schema = pa.schema([
         ...   pa.field("vector", pa.list_(pa.float32(), 2)),
@@ -200,11 +141,168 @@ class LanceDBConnection:
         vector: [[[1.1,1.2],[0.2,1.8]]]
         lat: [[45.5,40.1]]
         long: [[-122.7,-74.1]]
+
+
+        It is also possible to create an table from `[Iterable[pa.RecordBatch]]`:
+
+
+        >>> import pyarrow as pa
+        >>> def make_batches():
+        ...     for i in range(5):
+        ...         yield pa.RecordBatch.from_arrays(
+        ...             [
+        ...                 pa.array([[3.1, 4.1], [5.9, 26.5]]),
+        ...                 pa.array(["foo", "bar"]),
+        ...                 pa.array([10.0, 20.0]),
+        ...             ],
+        ...             ["vector", "item", "price"],
+        ...         )
+        >>> schema=pa.schema([
+        ...     pa.field("vector", pa.list_(pa.float32())),
+        ...     pa.field("item", pa.utf8()),
+        ...     pa.field("price", pa.float32()),
+        ... ])
+        >>> db.create_table("table4", make_batches(), schema=schema)
+        LanceTable(table4)
+
         """
-        if data is not None:
+        raise NotImplementedError
-            tbl = LanceTable.create(self, name, data, schema, mode=mode)
+
-        else:
+    def __getitem__(self, name: str) -> LanceTable:
-            tbl = LanceTable(self, name)
+        return self.open_table(name)
+
+    def open_table(self, name: str) -> Table:
+        """Open a Lance Table in the database.
+
+        Parameters
+        ----------
+        name: str
+            The name of the table.
+
+        Returns
+        -------
+        A LanceTable object representing the table.
+        """
+        raise NotImplementedError
+
+    def drop_table(self, name: str):
+        """Drop a table from the database.
+
+        Parameters
+        ----------
+        name: str
+            The name of the table.
+        """
+        raise NotImplementedError
+
+
+class LanceDBConnection(DBConnection):
+    """
+    A connection to a LanceDB database.
+
+    Parameters
+    ----------
+    uri: str or Path
+        The root uri of the database.
+
+    Examples
+    --------
+    >>> import lancedb
+    >>> db = lancedb.connect("./.lancedb")
+    >>> db.create_table("my_table", data=[{"vector": [1.1, 1.2], "b": 2},
+    ...                                   {"vector": [0.5, 1.3], "b": 4}])
+    LanceTable(my_table)
+    >>> db.create_table("another_table", data=[{"vector": [0.4, 0.4], "b": 6}])
+    LanceTable(another_table)
+    >>> sorted(db.table_names())
+    ['another_table', 'my_table']
+    >>> len(db)
+    2
+    >>> db["my_table"]
+    LanceTable(my_table)
+    >>> "my_table" in db
+    True
+    >>> db.drop_table("my_table")
+    >>> db.drop_table("another_table")
+    """
+
+    def __init__(self, uri: URI):
+        if not isinstance(uri, Path):
+            scheme = get_uri_scheme(uri)
+        is_local = isinstance(uri, Path) or scheme == "file"
+        if is_local:
+            if isinstance(uri, str):
+                uri = Path(uri)
+            uri = uri.expanduser().absolute()
+            Path(uri).mkdir(parents=True, exist_ok=True)
+        self._uri = str(uri)
+
+        self._entered = False
+
+    @property
+    def uri(self) -> str:
+        return self._uri
+
+    def table_names(self) -> list[str]:
+        """Get the names of all tables in the database.
+
+        Returns
+        -------
+        list of str
+            A list of table names.
+        """
+        try:
+            filesystem, path = fs.FileSystem.from_uri(self.uri)
+        except pa.ArrowInvalid:
+            raise NotImplementedError("Unsupported scheme: " + self.uri)
+
+        try:
+            paths = filesystem.get_file_info(
+                fs.FileSelector(get_uri_location(self.uri))
+            )
+        except FileNotFoundError:
+            # It is ok if the file does not exist since it will be created
+            paths = []
+        tables = [
+            os.path.splitext(file_info.base_name)[0]
+            for file_info in paths
+            if file_info.extension == "lance"
+        ]
+        return tables
+
+    def __len__(self) -> int:
+        return len(self.table_names())
+
+    def __contains__(self, name: str) -> bool:
+        return name in self.table_names()
+
+    def create_table(
+        self,
+        name: str,
+        data: Optional[Union[List[dict], dict, pd.DataFrame]] = None,
+        schema: pa.Schema = None,
+        mode: str = "create",
+        on_bad_vectors: str = "error",
+        fill_value: float = 0.0,
+    ) -> LanceTable:
+        """Create a table in the database.
+
+        See
+        ---
+        DBConnection.create_table
+        """
+        if mode.lower() not in ["create", "overwrite"]:
+            raise ValueError("mode must be either 'create' or 'overwrite'")
+
+        tbl = LanceTable.create(
+            self,
+            name,
+            data,
+            schema,
+            mode=mode,
+            on_bad_vectors=on_bad_vectors,
+            fill_value=fill_value,
+        )
         return tbl
 
     def open_table(self, name: str) -> LanceTable:
@@ -219,7 +317,7 @@ class LanceDBConnection:
         -------
         A LanceTable object representing the table.
         """
-        return LanceTable(self, name)
+        return LanceTable.open(self, name)
 
     def drop_table(self, name: str):
         """Drop a table from the database.

python/lancedb/pydantic.py

@@ -0,0 +1,212 @@
+#  Copyright 2023 LanceDB Developers
+#
+#  Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+#  You may obtain a copy of the License at
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+#  Unless required by applicable law or agreed to in writing, software
+#  distributed under the License is distributed on an "AS IS" BASIS,
+#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#  See the License for the specific language governing permissions and
+#  limitations under the License.
+
+"""Pydantic adapter for LanceDB"""
+
+from __future__ import annotations
+
+import inspect
+import sys
+import types
+from abc import ABC, abstractmethod
+from typing import Any, List, Type, Union, _GenericAlias
+
+import pyarrow as pa
+import pydantic
+from pydantic_core import CoreSchema, core_schema
+
+
+class FixedSizeListMixin(ABC):
+    @staticmethod
+    @abstractmethod
+    def dim() -> int:
+        raise NotImplementedError
+
+    @staticmethod
+    @abstractmethod
+    def value_arrow_type() -> pa.DataType:
+        raise NotImplementedError
+
+
+def vector(
+    dim: int, value_type: pa.DataType = pa.float32()
+) -> Type[FixedSizeListMixin]:
+    """Pydantic Vector Type.
+
+    !!! warning
+        Experimental feature.
+
+    Parameters
+    ----------
+    dim : int
+        The dimension of the vector.
+    value_type : pyarrow.DataType, optional
+        The value type of the vector, by default pa.float32()
+
+    Examples
+    --------
+
+    >>> import pydantic
+    >>> from lancedb.pydantic import vector
+    ...
+    >>> class MyModel(pydantic.BaseModel):
+    ...     id: int
+    ...     url: str
+    ...     embeddings: vector(768)
+    >>> schema = pydantic_to_schema(MyModel)
+    >>> assert schema == pa.schema([
+    ...     pa.field("id", pa.int64(), False),
+    ...     pa.field("url", pa.utf8(), False),
+    ...     pa.field("embeddings", pa.list_(pa.float32(), 768), False)
+    ... ])
+    """
+
+    # TODO: make a public parameterized type.
+    class FixedSizeList(list, FixedSizeListMixin):
+        @staticmethod
+        def dim() -> int:
+            return dim
+
+        @staticmethod
+        def value_arrow_type() -> pa.DataType:
+            return value_type
+
+        @classmethod
+        def __get_pydantic_core_schema__(
+            cls, _source_type: Any, _handler: pydantic.GetCoreSchemaHandler
+        ) -> CoreSchema:
+            return core_schema.no_info_after_validator_function(
+                cls,
+                core_schema.list_schema(
+                    min_length=dim,
+                    max_length=dim,
+                    items_schema=core_schema.float_schema(),
+                ),
+            )
+
+    return FixedSizeList
+
+
+def _py_type_to_arrow_type(py_type: Type[Any]) -> pa.DataType:
+    """Convert Python Type to Arrow DataType.
+
+    Raises
+    ------
+    TypeError
+        If the type is not supported.
+    """
+    if py_type == int:
+        return pa.int64()
+    elif py_type == float:
+        return pa.float64()
+    elif py_type == str:
+        return pa.utf8()
+    elif py_type == bool:
+        return pa.bool_()
+    elif py_type == bytes:
+        return pa.binary()
+    raise TypeError(
+        f"Converting Pydantic type to Arrow Type: unsupported type {py_type}"
+    )
+
+
+def _pydantic_model_to_fields(model: pydantic.BaseModel) -> List[pa.Field]:
+    fields = []
+    for name, field in model.model_fields.items():
+        fields.append(_pydantic_to_field(name, field))
+    return fields
+
+
+def _pydantic_to_arrow_type(field: pydantic.fields.FieldInfo) -> pa.DataType:
+    """Convert a Pydantic FieldInfo to Arrow DataType"""
+    if isinstance(field.annotation, _GenericAlias) or (
+        sys.version_info > (3, 9) and isinstance(field.annotation, types.GenericAlias)
+    ):
+        origin = field.annotation.__origin__
+        args = field.annotation.__args__
+        if origin == list:
+            child = args[0]
+            return pa.list_(_py_type_to_arrow_type(child))
+        elif origin == Union:
+            if len(args) == 2 and args[1] == type(None):
+                return _py_type_to_arrow_type(args[0])
+    elif inspect.isclass(field.annotation):
+        if issubclass(field.annotation, pydantic.BaseModel):
+            # Struct
+            fields = _pydantic_model_to_fields(field.annotation)
+            return pa.struct(fields)
+        elif issubclass(field.annotation, FixedSizeListMixin):
+            return pa.list_(field.annotation.value_arrow_type(), field.annotation.dim())
+    return _py_type_to_arrow_type(field.annotation)
+
+
+def is_nullable(field: pydantic.fields.FieldInfo) -> bool:
+    """Check if a Pydantic FieldInfo is nullable."""
+    if isinstance(field.annotation, _GenericAlias):
+        origin = field.annotation.__origin__
+        args = field.annotation.__args__
+        if origin == Union:
+            if len(args) == 2 and args[1] == type(None):
+                return True
+    return False
+
+
+def _pydantic_to_field(name: str, field: pydantic.fields.FieldInfo) -> pa.Field:
+    """Convert a Pydantic field to a PyArrow Field."""
+    dt = _pydantic_to_arrow_type(field)
+    return pa.field(name, dt, is_nullable(field))
+
+
+def pydantic_to_schema(model: Type[pydantic.BaseModel]) -> pa.Schema:
+    """Convert a Pydantic model to a PyArrow Schema.
+
+    Parameters
+    ----------
+    model : Type[pydantic.BaseModel]
+        The Pydantic BaseModel to convert to Arrow Schema.
+
+    Returns
+    -------
+    pyarrow.Schema
+
+    Examples
+    --------
+
+    >>> from typing import List, Optional
+    >>> import pydantic
+    >>> from lancedb.pydantic import pydantic_to_schema
+    ...
+    >>> class InnerModel(pydantic.BaseModel):
+    ...     a: str
+    ...     b: Optional[float]
+    >>>
+    >>> class FooModel(pydantic.BaseModel):
+    ...     id: int
+    ...     s: Optional[str] = None
+    ...     vec: List[float]
+    ...     li: List[int]
+    ...     inner: InnerModel
+    >>> schema = pydantic_to_schema(FooModel)
+    >>> assert schema == pa.schema([
+    ...     pa.field("id", pa.int64(), False),
+    ...     pa.field("s", pa.utf8(), True),
+    ...     pa.field("vec", pa.list_(pa.float64()), False),
+    ...     pa.field("li", pa.list_(pa.int64()), False),
+    ...     pa.field("inner", pa.struct([
+    ...         pa.field("a", pa.utf8(), False),
+    ...         pa.field("b", pa.float64(), True),
+    ...     ]), False),
+    ... ])
+    """
+    fields = _pydantic_model_to_fields(model)
+    return pa.schema(fields)

python/lancedb/query.py

@@ -10,16 +10,47 @@
 #  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
+
 from __future__ import annotations
-from typing import Literal
+
+from typing import List, Literal, Optional, Union
 
 import numpy as np
 import pandas as pd
 import pyarrow as pa
+from pydantic import BaseModel
 
 from .common import VECTOR_COLUMN_NAME
 
 
+class Query(BaseModel):
+    """A Query"""
+
+    vector_column: str = VECTOR_COLUMN_NAME
+
+    # vector to search for
+    vector: List[float]
+
+    # sql filter to refine the query with
+    filter: Optional[str] = None
+
+    # top k results to return
+    k: int
+
+    # # metrics
+    metric: str = "L2"
+
+    # which columns to return in the results
+    columns: Optional[List[str]] = None
+
+    # optional query parameters for tuning the results,
+    # e.g. `{"nprobes": "10", "refine_factor": "10"}`
+    nprobes: int = 10
+
+    # Refine factor.
+    refine_factor: Optional[int] = None
+
+
 class LanceQueryBuilder:
     """
     A builder for nearest neighbor queries for LanceDB.
@@ -43,7 +74,12 @@ class LanceQueryBuilder:
     0  6  [0.4, 0.4]    0.0
     """
 
-    def __init__(self, table: "lancedb.table.LanceTable", query: np.ndarray):
+    def __init__(
+        self,
+        table: "lancedb.table.Table",
+        query: Union[np.ndarray, str],
+        vector_column: str = VECTOR_COLUMN_NAME,
+    ):
         self._metric = "L2"
         self._nprobes = 20
         self._refine_factor = None
@@ -52,6 +88,7 @@ class LanceQueryBuilder:
         self._limit = 10
         self._columns = None
         self._where = None
+        self._vector_column = vector_column
 
     def limit(self, limit: int) -> LanceQueryBuilder:
         """Set the maximum number of results to return.
@@ -168,24 +205,34 @@ class LanceQueryBuilder:
         and also the "score" column which is the distance between the query
         vector and the returned vector.
         """
-        ds = self._table.to_lance()
+
-        tbl = ds.to_table(
+        return self.to_arrow().to_pandas()
-            columns=self._columns,
+
+    def to_arrow(self) -> pa.Table:
+        """
+        Execute the query and return the results as an
+        [Apache Arrow Table](https://arrow.apache.org/docs/python/generated/pyarrow.Table.html#pyarrow.Table).
+
+        In addition to the selected columns, LanceDB also returns a vector
+        and also the "score" column which is the distance between the query
+        vector and the returned vectors.
+        """
+        vector = self._query if isinstance(self._query, list) else self._query.tolist()
+        query = Query(
+            vector=vector,
             filter=self._where,
-            nearest={
+            k=self._limit,
-                "column": VECTOR_COLUMN_NAME,
+            metric=self._metric,
-                "q": self._query,
+            columns=self._columns,
-                "k": self._limit,
+            nprobes=self._nprobes,
-                "metric": self._metric,
+            refine_factor=self._refine_factor,
-                "nprobes": self._nprobes,
+            vector_column=self._vector_column,
-                "refine_factor": self._refine_factor,
-            },
         )
-        return tbl.to_pandas()
+        return self._table._execute_query(query)
 
 
 class LanceFtsQueryBuilder(LanceQueryBuilder):
-    def to_df(self) -> pd.DataFrame:
+    def to_arrow(self) -> pd.Table:
         try:
             import tantivy
         except ImportError:
@@ -202,8 +249,9 @@ class LanceFtsQueryBuilder(LanceQueryBuilder):
         # get the scores and doc ids
         row_ids, scores = search_index(index, self._query, self._limit)
         if len(row_ids) == 0:
-            return pd.DataFrame()
+            empty_schema = pa.schema([pa.field("score", pa.float32())])
+            return pa.Table.from_pylist([], schema=empty_schema)
         scores = pa.array(scores)
         output_tbl = self._table.to_lance().take(row_ids, columns=self._columns)
         output_tbl = output_tbl.append_column("score", scores)
-        return output_tbl.to_pandas()
+        return output_tbl

python/lancedb/remote/__init__.py

@@ -0,0 +1,60 @@
+#  Copyright 2023 LanceDB Developers
+#
+#  Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+#  You may obtain a copy of the License at
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+#  Unless required by applicable law or agreed to in writing, software
+#  distributed under the License is distributed on an "AS IS" BASIS,
+#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#  See the License for the specific language governing permissions and
+#  limitations under the License.
+
+import abc
+from typing import List, Optional
+
+import attr
+import pyarrow as pa
+from pydantic import BaseModel
+
+__all__ = ["LanceDBClient", "VectorQuery", "VectorQueryResult"]
+
+
+class VectorQuery(BaseModel):
+    # vector to search for
+    vector: List[float]
+
+    # sql filter to refine the query with
+    filter: Optional[str] = None
+
+    # top k results to return
+    k: int
+
+    # # metrics
+    _metric: str = "L2"
+
+    # which columns to return in the results
+    columns: Optional[List[str]] = None
+
+    # optional query parameters for tuning the results,
+    # e.g. `{"nprobes": "10", "refine_factor": "10"}`
+    nprobes: int = 10
+
+    refine_factor: Optional[int] = None
+
+
+@attr.define
+class VectorQueryResult:
+    # for now the response is directly seralized into a pandas dataframe
+    tbl: pa.Table
+
+    def to_arrow(self) -> pa.Table:
+        return self.tbl
+
+
+class LanceDBClient(abc.ABC):
+    @abc.abstractmethod
+    def query(self, table_name: str, query: VectorQuery) -> VectorQueryResult:
+        """Query the LanceDB server for the given table and query."""
+        pass

python/lancedb/remote/arrow.py

@@ -0,0 +1,22 @@
+#  Copyright 2023 LanceDB Developers
+#
+#  Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+#  You may obtain a copy of the License at
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+#  Unless required by applicable law or agreed to in writing, software
+#  distributed under the License is distributed on an "AS IS" BASIS,
+#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#  See the License for the specific language governing permissions and
+#  limitations under the License.
+
+import pyarrow as pa
+
+
+def to_ipc_binary(table: pa.Table) -> bytes:
+    """Serialize a PyArrow Table to IPC binary."""
+    sink = pa.BufferOutputStream()
+    with pa.ipc.new_stream(sink, table.schema) as writer:
+        writer.write_table(table)
+    return sink.getvalue().to_pybytes()

python/lancedb/remote/client.py

@@ -0,0 +1,145 @@
+#  Copyright 2023 LanceDB Developers
+#
+#  Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+#  You may obtain a copy of the License at
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+#  Unless required by applicable law or agreed to in writing, software
+#  distributed under the License is distributed on an "AS IS" BASIS,
+#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#  See the License for the specific language governing permissions and
+#  limitations under the License.
+
+
+import functools
+from typing import Any, Callable, Dict, Optional, Union
+
+import aiohttp
+import attr
+import pyarrow as pa
+from pydantic import BaseModel
+
+from lancedb.common import Credential
+from lancedb.remote import VectorQuery, VectorQueryResult
+from lancedb.remote.errors import LanceDBClientError
+
+ARROW_STREAM_CONTENT_TYPE = "application/vnd.apache.arrow.stream"
+
+
+def _check_not_closed(f):
+    @functools.wraps(f)
+    def wrapped(self, *args, **kwargs):
+        if self.closed:
+            raise ValueError("Connection is closed")
+        return f(self, *args, **kwargs)
+
+    return wrapped
+
+
+async def _read_ipc(resp: aiohttp.ClientResponse) -> pa.Table:
+    resp_body = await resp.read()
+    with pa.ipc.open_file(pa.BufferReader(resp_body)) as reader:
+        return reader.read_all()
+
+
+@attr.define(slots=False)
+class RestfulLanceDBClient:
+    db_name: str
+    region: str
+    api_key: Credential
+    closed: bool = attr.field(default=False, init=False)
+
+    @functools.cached_property
+    def session(self) -> aiohttp.ClientSession:
+        url = f"https://{self.db_name}.{self.region}.api.lancedb.com"
+        return aiohttp.ClientSession(url)
+
+    async def close(self):
+        await self.session.close()
+        self.closed = True
+
+    @functools.cached_property
+    def headers(self) -> Dict[str, str]:
+        headers = {
+            "x-api-key": self.api_key,
+        }
+        if self.region == "local":  # Local test mode
+            headers["Host"] = f"{self.db_name}.{self.region}.api.lancedb.com"
+        return headers
+
+    @staticmethod
+    async def _check_status(resp: aiohttp.ClientResponse):
+        if resp.status == 404:
+            raise LanceDBClientError(f"Not found: {await resp.text()}")
+        elif 400 <= resp.status < 500:
+            raise LanceDBClientError(
+                f"Bad Request: {resp.status}, error: {await resp.text()}"
+            )
+        elif 500 <= resp.status < 600:
+            raise LanceDBClientError(
+                f"Internal Server Error: {resp.status}, error: {await resp.text()}"
+            )
+        elif resp.status != 200:
+            raise LanceDBClientError(
+                f"Unknown Error: {resp.status}, error: {await resp.text()}"
+            )
+
+    @_check_not_closed
+    async def get(self, uri: str, params: Union[Dict[str, Any], BaseModel] = None):
+        """Send a GET request and returns the deserialized response payload."""
+        if isinstance(params, BaseModel):
+            params: Dict[str, Any] = params.dict(exclude_none=True)
+        async with self.session.get(uri, params=params, headers=self.headers) as resp:
+            await self._check_status(resp)
+            return await resp.json()
+
+    @_check_not_closed
+    async def post(
+        self,
+        uri: str,
+        data: Union[Dict[str, Any], BaseModel, bytes],
+        params: Optional[Dict[str, Any]] = None,
+        content_type: Optional[str] = None,
+        deserialize: Callable = lambda resp: resp.json(),
+    ) -> Dict[str, Any]:
+        """Send a POST request and returns the deserialized response payload.
+
+        Parameters
+        ----------
+        uri : str
+            The uri to send the POST request to.
+        data: Union[Dict[str, Any], BaseModel]
+
+        """
+        if isinstance(data, BaseModel):
+            data: Dict[str, Any] = data.dict(exclude_none=True)
+        if isinstance(data, bytes):
+            req_kwargs = {"data": data}
+        else:
+            req_kwargs = {"json": data}
+
+        headers = self.headers.copy()
+        if content_type is not None:
+            headers["content-type"] = content_type
+        async with self.session.post(
+            uri,
+            headers=headers,
+            params=params,
+            **req_kwargs,
+        ) as resp:
+            resp: aiohttp.ClientResponse = resp
+            await self._check_status(resp)
+            return await deserialize(resp)
+
+    @_check_not_closed
+    async def list_tables(self):
+        """List all tables in the database."""
+        json = await self.get("/v1/table/", {})
+        return json["tables"]
+
+    @_check_not_closed
+    async def query(self, table_name: str, query: VectorQuery) -> VectorQueryResult:
+        """Query a table."""
+        tbl = await self.post(f"/v1/table/{table_name}/", query, deserialize=_read_ipc)
+        return VectorQueryResult(tbl)

python/lancedb/remote/db.py

@@ -0,0 +1,104 @@
+#  Copyright 2023 LanceDB Developers
+#
+#  Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+#  You may obtain a copy of the License at
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+#  Unless required by applicable law or agreed to in writing, software
+#  distributed under the License is distributed on an "AS IS" BASIS,
+#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#  See the License for the specific language governing permissions and
+#  limitations under the License.
+
+import asyncio
+import uuid
+from typing import List
+from urllib.parse import urlparse
+
+import pyarrow as pa
+
+from lancedb.common import DATA
+from lancedb.db import DBConnection
+from lancedb.schema import schema_to_json
+from lancedb.table import Table, _sanitize_data
+
+from .arrow import to_ipc_binary
+from .client import ARROW_STREAM_CONTENT_TYPE, RestfulLanceDBClient
+
+
+class RemoteDBConnection(DBConnection):
+    """A connection to a remote LanceDB database."""
+
+    def __init__(self, db_url: str, api_key: str, region: str):
+        """Connect to a remote LanceDB database."""
+        parsed = urlparse(db_url)
+        if parsed.scheme != "db":
+            raise ValueError(f"Invalid scheme: {parsed.scheme}, only accepts db://")
+        self.db_name = parsed.netloc
+        self.api_key = api_key
+        self._client = RestfulLanceDBClient(self.db_name, region, api_key)
+        try:
+            self._loop = asyncio.get_running_loop()
+        except RuntimeError:
+            self._loop = asyncio.get_event_loop()
+
+    def __repr__(self) -> str:
+        return f"RemoveConnect(name={self.db_name})"
+
+    def table_names(self) -> List[str]:
+        """List the names of all tables in the database."""
+        result = self._loop.run_until_complete(self._client.list_tables())
+        return result
+
+    def open_table(self, name: str) -> Table:
+        """Open a Lance Table in the database.
+
+        Parameters
+        ----------
+        name: str
+            The name of the table.
+
+        Returns
+        -------
+        A LanceTable object representing the table.
+        """
+        from .table import RemoteTable
+
+        # TODO: check if table exists
+
+        return RemoteTable(self, name)
+
+    def create_table(
+        self,
+        name: str,
+        data: DATA = None,
+        schema: pa.Schema = None,
+        on_bad_vectors: str = "error",
+        fill_value: float = 0.0,
+    ) -> Table:
+        if data is None and schema is None:
+            raise ValueError("Either data or schema must be provided.")
+        if data is not None:
+            data = _sanitize_data(
+                data, schema, on_bad_vectors=on_bad_vectors, fill_value=fill_value
+            )
+        else:
+            if schema is None:
+                raise ValueError("Either data or schema must be provided")
+            data = pa.Table.from_pylist([], schema=schema)
+
+        from .table import RemoteTable
+
+        data = to_ipc_binary(data)
+        request_id = uuid.uuid4().hex
+
+        self._loop.run_until_complete(
+            self._client.post(
+                f"/v1/table/{name}/create",
+                data=data,
+                params={"request_id": request_id},
+                content_type=ARROW_STREAM_CONTENT_TYPE,
+            )
+        )
+        return RemoteTable(self, name)

python/lancedb/remote/errors.py

@@ -0,0 +1,16 @@
+#  Copyright 2023 LanceDB Developers
+#
+#  Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+#  You may obtain a copy of the License at
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+#  Unless required by applicable law or agreed to in writing, software
+#  distributed under the License is distributed on an "AS IS" BASIS,
+#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#  See the License for the specific language governing permissions and
+#  limitations under the License.
+
+
+class LanceDBClientError(RuntimeError):
+    pass

python/lancedb/remote/table.py

@@ -0,0 +1,91 @@
+#  Copyright 2023 LanceDB Developers
+#
+#  Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+#  You may obtain a copy of the License at
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+#  Unless required by applicable law or agreed to in writing, software
+#  distributed under the License is distributed on an "AS IS" BASIS,
+#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#  See the License for the specific language governing permissions and
+#  limitations under the License.
+
+import uuid
+from functools import cached_property
+from typing import Union
+
+import pyarrow as pa
+
+from lancedb.common import DATA, VEC, VECTOR_COLUMN_NAME
+
+from ..query import LanceQueryBuilder, Query
+from ..schema import json_to_schema
+from ..table import Query, Table, _sanitize_data
+from .arrow import to_ipc_binary
+from .client import ARROW_STREAM_CONTENT_TYPE
+from .db import RemoteDBConnection
+
+
+class RemoteTable(Table):
+    def __init__(self, conn: RemoteDBConnection, name: str):
+        self._conn = conn
+        self._name = name
+
+    def __repr__(self) -> str:
+        return f"RemoteTable({self._conn.db_name}.{self.name})"
+
+    @cached_property
+    def schema(self) -> pa.Schema:
+        """Return the schema of the table."""
+        resp = self._conn._loop.run_until_complete(
+            self._conn._client.get(f"/v1/table/{self._name}/describe")
+        )
+        schema = json_to_schema(resp["schema"])
+        return schema
+
+    def to_arrow(self) -> pa.Table:
+        raise NotImplementedError
+
+    def create_index(
+        self,
+        metric="L2",
+        num_partitions=256,
+        num_sub_vectors=96,
+        vector_column_name: str = VECTOR_COLUMN_NAME,
+        replace: bool = True,
+    ):
+        raise NotImplementedError
+
+    def add(
+        self,
+        data: DATA,
+        mode: str = "append",
+        on_bad_vectors: str = "error",
+        fill_value: float = 0.0,
+    ) -> int:
+        data = _sanitize_data(
+            data, self.schema, on_bad_vectors=on_bad_vectors, fill_value=fill_value
+        )
+        payload = to_ipc_binary(data)
+
+        request_id = uuid.uuid4().hex
+
+        self._conn._loop.run_until_complete(
+            self._conn._client.post(
+                f"/v1/table/{self._name}/insert",
+                data=payload,
+                params={"request_id": request_id, "mode": mode},
+                content_type=ARROW_STREAM_CONTENT_TYPE,
+            )
+        )
+        return len(data)
+
+    def search(
+        self, query: Union[VEC, str], vector_column: str = VECTOR_COLUMN_NAME
+    ) -> LanceQueryBuilder:
+        return LanceQueryBuilder(self, query, vector_column)
+
+    def _execute_query(self, query: Query) -> pa.Table:
+        result = self._conn._client.query(self._name, query)
+        return self._conn._loop.run_until_complete(result).to_arrow()

python/lancedb/schema.py

@@ -0,0 +1,45 @@
+#  Copyright 2023 LanceDB Developers
+#
+#  Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+#  You may obtain a copy of the License at
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+#  Unless required by applicable law or agreed to in writing, software
+#  distributed under the License is distributed on an "AS IS" BASIS,
+#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#  See the License for the specific language governing permissions and
+#  limitations under the License.
+
+"""Schema related utilities."""
+
+from typing import Any, Dict, Type
+
+import pyarrow as pa
+from lance import json_to_schema, schema_to_json
+
+
+def vector(dimension: int, value_type: pa.DataType = pa.float32()) -> pa.DataType:
+    """A help function to create a vector type.
+
+    Parameters
+    ----------
+    dimension: The dimension of the vector.
+    value_type: pa.DataType, optional
+        The type of the value in the vector.
+
+    Returns
+    -------
+    A PyArrow DataType for vectors.
+
+    Examples
+    --------
+
+    >>> import pyarrow as pa
+    >>> import lancedb
+    >>> schema = pa.schema([
+    ...     pa.field("id", pa.int64()),
+    ...     pa.field("vector", lancedb.vector(756)),
+    ... ])
+    """
+    return pa.list_(value_type, dimension)

python/lancedb/table.py

@@ -14,44 +14,49 @@
 from __future__ import annotations
 
 import os
-import shutil
+from abc import ABC, abstractmethod
 from functools import cached_property
-from typing import List, Union
+from typing import Iterable, List, Union
 
 import lance
 import numpy as np
 import pandas as pd
 import pyarrow as pa
+import pyarrow.compute as pc
+import pyarrow.fs
 from lance import LanceDataset
 from lance.vector import vec_to_table
 
 from .common import DATA, VEC, VECTOR_COLUMN_NAME
-from .query import LanceFtsQueryBuilder, LanceQueryBuilder
+from .query import LanceFtsQueryBuilder, LanceQueryBuilder, Query
-from .util import get_uri_scheme
 
 
-def _sanitize_data(data, schema):
+def _sanitize_data(data, schema, on_bad_vectors, fill_value):
     if isinstance(data, list):
         data = pa.Table.from_pylist(data)
-        data = _sanitize_schema(data, schema=schema)
+        data = _sanitize_schema(
+            data, schema=schema, on_bad_vectors=on_bad_vectors, fill_value=fill_value
+        )
     if isinstance(data, dict):
         data = vec_to_table(data)
     if isinstance(data, pd.DataFrame):
         data = pa.Table.from_pandas(data)
-        data = _sanitize_schema(data, schema=schema)
+        data = _sanitize_schema(
-    if not isinstance(data, pa.Table):
+            data, schema=schema, on_bad_vectors=on_bad_vectors, fill_value=fill_value
+        )
+    if not isinstance(data, (pa.Table, Iterable)):
         raise TypeError(f"Unsupported data type: {type(data)}")
     return data
 
 
-class LanceTable:
+class Table(ABC):
     """
-    A table in a LanceDB database.
+    A [Table](Table) is a collection of Records in a LanceDB [Database](Database).
 
     Examples
     --------
 
-    Create using [LanceDBConnection.create_table][lancedb.LanceDBConnection.create_table]
+    Create using [DBConnection.create_table][lancedb.DBConnection.create_table]
     (more examples in that method's documentation).
 
     >>> import lancedb
@@ -66,12 +71,12 @@ class LanceTable:
     vector: [[[1.1,1.2]]]
     b: [[2]]
 
-    Can append new data with [LanceTable.add][lancedb.table.LanceTable.add].
+    Can append new data with [Table.add()][lancedb.table.Table.add].
 
     >>> table.add([{"vector": [0.5, 1.3], "b": 4}])
     2
 
-    Can query the table with [LanceTable.search][lancedb.table.LanceTable.search].
+    Can query the table with [Table.search][lancedb.table.Table.search].
 
     >>> table.search([0.4, 0.4]).select(["b"]).to_df()
        b      vector  score
@@ -79,8 +84,128 @@ class LanceTable:
     1  2  [1.1, 1.2]   1.13
 
     Search queries are much faster when an index is created. See
-    [LanceTable.create_index][lancedb.table.LanceTable.create_index].
+    [Table.create_index][lancedb.table.Table.create_index].
+    """
 
+    @abstractmethod
+    def schema(self) -> pa.Schema:
+        """Return the [Arrow Schema](https://arrow.apache.org/docs/python/api/datatypes.html#) of
+        this [Table](Table)
+
+        """
+        raise NotImplementedError
+
+    def to_pandas(self) -> pd.DataFrame:
+        """Return the table as a pandas DataFrame.
+
+        Returns
+        -------
+        pd.DataFrame
+        """
+        return self.to_arrow().to_pandas()
+
+    @abstractmethod
+    def to_arrow(self) -> pa.Table:
+        """Return the table as a pyarrow Table.
+
+        Returns
+        -------
+        pa.Table
+        """
+        raise NotImplementedError
+
+    def create_index(
+        self,
+        metric="L2",
+        num_partitions=256,
+        num_sub_vectors=96,
+        vector_column_name: str = VECTOR_COLUMN_NAME,
+        replace: bool = True,
+    ):
+        """Create an index on the table.
+
+        Parameters
+        ----------
+        metric: str, default "L2"
+            The distance metric to use when creating the index.
+            Valid values are "L2", "cosine", or "dot".
+            L2 is euclidean distance.
+        num_partitions: int
+            The number of IVF partitions to use when creating the index.
+            Default is 256.
+        num_sub_vectors: int
+            The number of PQ sub-vectors to use when creating the index.
+            Default is 96.
+        vector_column_name: str, default "vector"
+            The vector column name to create the index.
+        replace: bool, default True
+            If True, replace the existing index if it exists.
+            If False, raise an error if duplicate index exists.
+        """
+        raise NotImplementedError
+
+    @abstractmethod
+    def add(
+        self,
+        data: DATA,
+        mode: str = "append",
+        on_bad_vectors: str = "error",
+        fill_value: float = 0.0,
+    ) -> int:
+        """Add more data to the [Table](Table).
+
+        Parameters
+        ----------
+        data: list-of-dict, dict, pd.DataFrame
+            The data to insert into the table.
+        mode: str
+            The mode to use when writing the data. Valid values are
+            "append" and "overwrite".
+        on_bad_vectors: str, default "error"
+            What to do if any of the vectors are not the same size or contains NaNs.
+            One of "error", "drop", "fill".
+        fill_value: float, default 0.
+            The value to use when filling vectors. Only used if on_bad_vectors="fill".
+
+        Returns
+        -------
+        int
+            The number of vectors in the table.
+        """
+        raise NotImplementedError
+
+    @abstractmethod
+    def search(
+        self, query: Union[VEC, str], vector_column: str = VECTOR_COLUMN_NAME
+    ) -> LanceQueryBuilder:
+        """Create a search query to find the nearest neighbors
+        of the given query vector.
+
+        Parameters
+        ----------
+        query: list, np.ndarray
+            The query vector.
+        vector_column: str, default "vector"
+            The name of the vector column to search.
+
+        Returns
+        -------
+        LanceQueryBuilder
+            A query builder object representing the query.
+            Once executed, the query returns selected columns, the vector,
+            and also the "score" column which is the distance between the query
+            vector and the returned vector.
+        """
+        raise NotImplementedError
+
+    @abstractmethod
+    def _execute_query(self, query: Query) -> pa.Table:
+        pass
+
+
+class LanceTable(Table):
+    """
+    A table in a LanceDB database.
     """
 
     def __init__(
@@ -92,6 +217,7 @@ class LanceTable:
 
     def _reset_dataset(self):
         try:
+            if "_dataset" in self.__dict__:
                 del self.__dict__["_dataset"]
         except AttributeError:
             pass
@@ -184,27 +310,22 @@ class LanceTable:
     def _dataset_uri(self) -> str:
         return os.path.join(self._conn.uri, f"{self.name}.lance")
 
-    def create_index(self, metric="L2", num_partitions=256, num_sub_vectors=96):
+    def create_index(
-        """Create an index on the table.
+        self,
-
+        metric="L2",
-        Parameters
+        num_partitions=256,
-        ----------
+        num_sub_vectors=96,
-        metric: str, default "L2"
+        vector_column_name=VECTOR_COLUMN_NAME,
-            The distance metric to use when creating the index. Valid values are "L2" or "cosine".
+        replace: bool = True,
-            L2 is euclidean distance.
+    ):
-        num_partitions: int
+        """Create an index on the table."""
-            The number of IVF partitions to use when creating the index.
-            Default is 256.
-        num_sub_vectors: int
-            The number of PQ sub-vectors to use when creating the index.
-            Default is 96.
-        """
         self._dataset.create_index(
-            column=VECTOR_COLUMN_NAME,
+            column=vector_column_name,
             index_type="IVF_PQ",
             metric=metric,
             num_partitions=num_partitions,
             num_sub_vectors=num_sub_vectors,
+            replace=replace,
         )
         self._reset_dataset()
 
@@ -237,7 +358,13 @@ class LanceTable:
         """Return the LanceDataset backing this table."""
         return self._dataset
 
-    def add(self, data: DATA, mode: str = "append") -> int:
+    def add(
+        self,
+        data: DATA,
+        mode: str = "append",
+        on_bad_vectors: str = "error",
+        fill_value: float = 0.0,
+    ) -> int:
         """Add data to the table.
 
         Parameters
@@ -247,18 +374,28 @@ class LanceTable:
         mode: str
             The mode to use when writing the data. Valid values are
             "append" and "overwrite".
+        on_bad_vectors: str, default "error"
+            What to do if any of the vectors are not the same size or contains NaNs.
+            One of "error", "drop", "fill".
+        fill_value: float, default 0.
+            The value to use when filling vectors. Only used if on_bad_vectors="fill".
 
         Returns
         -------
         int
             The number of vectors in the table.
         """
-        data = _sanitize_data(data, self.schema)
+        # TODO: manage table listing and metadata separately
+        data = _sanitize_data(
+            data, self.schema, on_bad_vectors=on_bad_vectors, fill_value=fill_value
+        )
         lance.write_dataset(data, self._dataset_uri, mode=mode)
         self._reset_dataset()
         return len(self)
 
-    def search(self, query: Union[VEC, str]) -> LanceQueryBuilder:
+    def search(
+        self, query: Union[VEC, str], vector_column_name=VECTOR_COLUMN_NAME
+    ) -> LanceQueryBuilder:
         """Create a search query to find the nearest neighbors
         of the given query vector.
 
@@ -266,6 +403,8 @@ class LanceTable:
         ----------
         query: list, np.ndarray
             The query vector.
+        vector_column_name: str, default "vector"
+            The name of the vector column to search.
 
         Returns
         -------
@@ -277,7 +416,7 @@ class LanceTable:
         """
         if isinstance(query, str):
             # fts
-            return LanceFtsQueryBuilder(self, query)
+            return LanceFtsQueryBuilder(self, query, vector_column_name)
 
         if isinstance(query, list):
             query = np.array(query)
@@ -285,17 +424,127 @@ class LanceTable:
             query = query.astype(np.float32)
         else:
             raise TypeError(f"Unsupported query type: {type(query)}")
-        return LanceQueryBuilder(self, query)
+        return LanceQueryBuilder(self, query, vector_column_name)
 
     @classmethod
-    def create(cls, db, name, data, schema=None, mode="create"):
+    def create(
+        cls,
+        db,
+        name,
+        data=None,
+        schema=None,
+        mode="create",
+        on_bad_vectors: str = "error",
+        fill_value: float = 0.0,
+    ):
+        """
+        Create a new table.
+
+        Examples
+        --------
+        >>> import lancedb
+        >>> import pandas as pd
+        >>> data = pd.DataFrame({"x": [1, 2, 3], "vector": [[1, 2], [3, 4], [5, 6]]})
+        >>> db = lancedb.connect("./.lancedb")
+        >>> table = db.create_table("my_table", data)
+        >>> table.to_pandas()
+           x      vector
+        0  1  [1.0, 2.0]
+        1  2  [3.0, 4.0]
+        2  3  [5.0, 6.0]
+
+        Parameters
+        ----------
+        db: LanceDB
+            The LanceDB instance to create the table in.
+        name: str
+            The name of the table to create.
+        data: list-of-dict, dict, pd.DataFrame, default None
+            The data to insert into the table.
+            At least one of `data` or `schema` must be provided.
+        schema: dict, optional
+            The schema of the table. If not provided, the schema is inferred from the data.
+            At least one of `data` or `schema` must be provided.
+        mode: str, default "create"
+            The mode to use when writing the data. Valid values are
+            "create", "overwrite", and "append".
+        on_bad_vectors: str, default "error"
+            What to do if any of the vectors are not the same size or contains NaNs.
+            One of "error", "drop", "fill".
+        fill_value: float, default 0.
+            The value to use when filling vectors. Only used if on_bad_vectors="fill".
+        """
         tbl = LanceTable(db, name)
-        data = _sanitize_data(data, schema)
+        if data is not None:
-        lance.write_dataset(data, tbl._dataset_uri, mode=mode)
+            data = _sanitize_data(
+                data, schema, on_bad_vectors=on_bad_vectors, fill_value=fill_value
+            )
+        else:
+            if schema is None:
+                raise ValueError("Either data or schema must be provided")
+            data = pa.Table.from_pylist([], schema=schema)
+        lance.write_dataset(data, tbl._dataset_uri, schema=schema, mode=mode)
+        return LanceTable(db, name)
+
+    @classmethod
+    def open(cls, db, name):
+        tbl = cls(db, name)
+        if not os.path.exists(tbl._dataset_uri):
+            raise FileNotFoundError(
+                f"Table {name} does not exist. Please first call db.create_table({name}, data)"
+            )
         return tbl
 
+    def delete(self, where: str):
+        """Delete rows from the table.
 
-def _sanitize_schema(data: pa.Table, schema: pa.Schema = None) -> pa.Table:
+        Parameters
+        ----------
+        where: str
+            The SQL where clause to use when deleting rows.
+
+        Examples
+        --------
+        >>> import lancedb
+        >>> import pandas as pd
+        >>> data = pd.DataFrame({"x": [1, 2, 3], "vector": [[1, 2], [3, 4], [5, 6]]})
+        >>> db = lancedb.connect("./.lancedb")
+        >>> table = db.create_table("my_table", data)
+        >>> table.to_pandas()
+           x      vector
+        0  1  [1.0, 2.0]
+        1  2  [3.0, 4.0]
+        2  3  [5.0, 6.0]
+        >>> table.delete("x = 2")
+        >>> table.to_pandas()
+           x      vector
+        0  1  [1.0, 2.0]
+        1  3  [5.0, 6.0]
+        """
+        self._dataset.delete(where)
+
+    def _execute_query(self, query: Query) -> pa.Table:
+        ds = self.to_lance()
+        return ds.to_table(
+            columns=query.columns,
+            filter=query.filter,
+            nearest={
+                "column": query.vector_column,
+                "q": query.vector,
+                "k": query.k,
+                "metric": query.metric,
+                "nprobes": query.nprobes,
+                "refine_factor": query.refine_factor,
+            },
+        )
+
+
+def _sanitize_schema(
+    data: pa.Table,
+    schema: pa.Schema = None,
+    on_bad_vectors: str = "error",
+    fill_value: float = 0.0,
+) -> pa.Table:
     """Ensure that the table has the expected schema.
 
     Parameters
@@ -305,21 +554,41 @@ def _sanitize_schema(data: pa.Table, schema: pa.Schema = None) -> pa.Table:
     schema: pa.Schema; optional
         The expected schema. If not provided, this just converts the
         vector column to fixed_size_list(float32) if necessary.
+    on_bad_vectors: str, default "error"
+        What to do if any of the vectors are not the same size or contains NaNs.
+        One of "error", "drop", "fill".
+    fill_value: float, default 0.
+        The value to use when filling vectors. Only used if on_bad_vectors="fill".
     """
     if schema is not None:
         if data.schema == schema:
             return data
         # cast the columns to the expected types
         data = data.combine_chunks()
-        data = _sanitize_vector_column(data, vector_column_name=VECTOR_COLUMN_NAME)
+        data = _sanitize_vector_column(
+            data,
+            vector_column_name=VECTOR_COLUMN_NAME,
+            on_bad_vectors=on_bad_vectors,
+            fill_value=fill_value,
+        )
         return pa.Table.from_arrays(
             [data[name] for name in schema.names], schema=schema
         )
     # just check the vector column
-    return _sanitize_vector_column(data, vector_column_name=VECTOR_COLUMN_NAME)
+    return _sanitize_vector_column(
+        data,
+        vector_column_name=VECTOR_COLUMN_NAME,
+        on_bad_vectors=on_bad_vectors,
+        fill_value=fill_value,
+    )
 
 
-def _sanitize_vector_column(data: pa.Table, vector_column_name: str) -> pa.Table:
+def _sanitize_vector_column(
+    data: pa.Table,
+    vector_column_name: str,
+    on_bad_vectors: str = "error",
+    fill_value: float = 0.0,
+) -> pa.Table:
     """
     Ensure that the vector column exists and has type fixed_size_list(float32)
 
@@ -329,19 +598,103 @@ def _sanitize_vector_column(data: pa.Table, vector_column_name: str) -> pa.Table
         The table to sanitize.
     vector_column_name: str
         The name of the vector column.
+    on_bad_vectors: str, default "error"
+        What to do if any of the vectors are not the same size or contains NaNs.
+        One of "error", "drop", "fill".
+    fill_value: float, default 0.0
+        The value to use when filling vectors. Only used if on_bad_vectors="fill".
     """
     if vector_column_name not in data.column_names:
         raise ValueError(f"Missing vector column: {vector_column_name}")
+    # ChunkedArray is annoying to work with, so we combine chunks here
     vec_arr = data[vector_column_name].combine_chunks()
-    if pa.types.is_fixed_size_list(vec_arr.type):
+    if pa.types.is_list(data[vector_column_name].type):
-        return data
+        # if it's a variable size list array we make sure the dimensions are all the same
-    if not pa.types.is_list(vec_arr.type):
+        has_jagged_ndims = len(vec_arr.values) % len(data) != 0
+        if has_jagged_ndims:
+            data = _sanitize_jagged(
+                data, fill_value, on_bad_vectors, vec_arr, vector_column_name
+            )
+            vec_arr = data[vector_column_name].combine_chunks()
+    elif not pa.types.is_fixed_size_list(vec_arr.type):
         raise TypeError(f"Unsupported vector column type: {vec_arr.type}")
+
+    vec_arr = ensure_fixed_size_list_of_f32(vec_arr)
+    data = data.set_column(
+        data.column_names.index(vector_column_name), vector_column_name, vec_arr
+    )
+
+    has_nans = pc.any(pc.is_nan(vec_arr.values)).as_py()
+    if has_nans:
+        data = _sanitize_nans(
+            data, fill_value, on_bad_vectors, vec_arr, vector_column_name
+        )
+
+    return data
+
+
+def ensure_fixed_size_list_of_f32(vec_arr):
     values = vec_arr.values
     if not pa.types.is_float32(values.type):
         values = values.cast(pa.float32())
-    list_size = len(values) / len(data)
+    if pa.types.is_fixed_size_list(vec_arr.type):
+        list_size = vec_arr.type.list_size
+    else:
+        list_size = len(values) / len(vec_arr)
     vec_arr = pa.FixedSizeListArray.from_arrays(values, list_size)
-    return data.set_column(
+    return vec_arr
+
+
+def _sanitize_jagged(data, fill_value, on_bad_vectors, vec_arr, vector_column_name):
+    """Sanitize jagged vectors."""
+    if on_bad_vectors == "error":
+        raise ValueError(
+            f"Vector column {vector_column_name} has variable length vectors "
+            "Set on_bad_vectors='drop' to remove them, or "
+            "set on_bad_vectors='fill' and fill_value=<value> to replace them."
+        )
+
+    lst_lengths = pc.list_value_length(vec_arr)
+    ndims = pc.max(lst_lengths).as_py()
+    correct_ndims = pc.equal(lst_lengths, ndims)
+
+    if on_bad_vectors == "fill":
+        if fill_value is None:
+            raise ValueError(
+                "`fill_value` must not be None if `on_bad_vectors` is 'fill'"
+            )
+        fill_arr = pa.scalar([float(fill_value)] * ndims)
+        vec_arr = pc.if_else(correct_ndims, vec_arr, fill_arr)
+        data = data.set_column(
             data.column_names.index(vector_column_name), vector_column_name, vec_arr
         )
+    elif on_bad_vectors == "drop":
+        data = data.filter(correct_ndims)
+    return data
+
+
+def _sanitize_nans(data, fill_value, on_bad_vectors, vec_arr, vector_column_name):
+    """Sanitize NaNs in vectors"""
+    if on_bad_vectors == "error":
+        raise ValueError(
+            f"Vector column {vector_column_name} has NaNs. "
+            "Set on_bad_vectors='drop' to remove them, or "
+            "set on_bad_vectors='fill' and fill_value=<value> to replace them."
+        )
+    elif on_bad_vectors == "fill":
+        if fill_value is None:
+            raise ValueError(
+                "`fill_value` must not be None if `on_bad_vectors` is 'fill'"
+            )
+        fill_value = float(fill_value)
+        values = pc.if_else(pc.is_nan(vec_arr.values), fill_value, vec_arr.values)
+        ndims = len(vec_arr[0])
+        vec_arr = pa.FixedSizeListArray.from_arrays(values, ndims)
+        data = data.set_column(
+            data.column_names.index(vector_column_name), vector_column_name, vec_arr
+        )
+    elif on_bad_vectors == "drop":
+        is_value_nan = pc.is_nan(vec_arr.values).to_numpy(zero_copy_only=False)
+        is_full = np.any(~is_value_nan.reshape(-1, vec_arr.type.list_size), axis=1)
+        data = data.filter(is_full)
+    return data

python/lancedb/util.py

@@ -11,9 +11,7 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 
-from urllib.parse import ParseResult, urlparse
+from urllib.parse import urlparse
-
-from pyarrow import fs
 
 
 def get_uri_scheme(uri: str) -> str:

python/pyproject.toml

@@ -1,7 +1,7 @@
 [project]
 name = "lancedb"
-version = "0.1.7"
+version = "0.1.11"
-dependencies = ["pylance>=0.4.17", "ratelimiter", "retry", "tqdm"]
+dependencies = ["pylance~=0.5.8", "ratelimiter", "retry", "tqdm", "aiohttp", "pydantic>=2", "attr"]
 description = "lancedb"
 authors = [
     { name = "LanceDB Devs", email = "dev@lancedb.com" },
@@ -37,7 +37,7 @@ repository = "https://github.com/lancedb/lancedb"
 
 [project.optional-dependencies]
 tests = [
-    "pytest", "pytest-mock", "doctest"
+    "pytest", "pytest-mock", "pytest-asyncio"
 ]
 dev = [
     "ruff", "pre-commit", "black"

python/tests/test_context.py

@@ -0,0 +1,77 @@
+#  Copyright 2023 LanceDB Developers
+#
+#  Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+#  You may obtain a copy of the License at
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+#  Unless required by applicable law or agreed to in writing, software
+#  distributed under the License is distributed on an "AS IS" BASIS,
+#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#  See the License for the specific language governing permissions and
+#  limitations under the License.
+
+import pandas as pd
+import pytest
+
+from lancedb.context import contextualize
+
+
+@pytest.fixture
+def raw_df() -> pd.DataFrame:
+    return pd.DataFrame(
+        {
+            "token": [
+                "The",
+                "quick",
+                "brown",
+                "fox",
+                "jumped",
+                "over",
+                "the",
+                "lazy",
+                "dog",
+                "I",
+                "love",
+                "sandwiches",
+            ],
+            "document_id": [1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2],
+        }
+    )
+
+
+def test_contextualizer(raw_df: pd.DataFrame):
+    result = (
+        contextualize(raw_df)
+        .window(6)
+        .stride(3)
+        .text_col("token")
+        .groupby("document_id")
+        .to_df()["token"]
+        .to_list()
+    )
+
+    assert result == [
+        "The quick brown fox jumped over",
+        "fox jumped over the lazy dog",
+        "the lazy dog",
+        "I love sandwiches",
+    ]
+
+
+def test_contextualizer_with_threshold(raw_df: pd.DataFrame):
+    result = (
+        contextualize(raw_df)
+        .window(6)
+        .stride(3)
+        .text_col("token")
+        .groupby("document_id")
+        .min_window_size(4)
+        .to_df()["token"]
+        .to_list()
+    )
+
+    assert result == [
+        "The quick brown fox jumped over",
+        "fox jumped over the lazy dog",
+    ]

python/tests/test_db.py

@@ -11,7 +11,9 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 
+import numpy as np
 import pandas as pd
+import pyarrow as pa
 import pytest
 
 import lancedb
@@ -74,6 +76,32 @@ def test_ingest_pd(tmp_path):
     assert db.open_table("test").name == db["test"].name
 
 
+def test_ingest_record_batch_iterator(tmp_path):
+    def batch_reader():
+        for i in range(5):
+            yield pa.RecordBatch.from_arrays(
+                [
+                    pa.array([[3.1, 4.1], [5.9, 26.5]]),
+                    pa.array(["foo", "bar"]),
+                    pa.array([10.0, 20.0]),
+                ],
+                ["vector", "item", "price"],
+            )
+
+    db = lancedb.connect(tmp_path)
+    tbl = db.create_table(
+        "test",
+        batch_reader(),
+        schema=pa.schema(
+            [
+                pa.field("vector", pa.list_(pa.float32())),
+                pa.field("item", pa.utf8()),
+                pa.field("price", pa.float32()),
+            ]
+        ),
+    )
+
+
 def test_create_mode(tmp_path):
     db = lancedb.connect(tmp_path)
     data = pd.DataFrame(
@@ -120,3 +148,43 @@ def test_delete_table(tmp_path):
 
     db.create_table("test", data=data)
     assert db.table_names() == ["test"]
+
+
+def test_empty_or_nonexistent_table(tmp_path):
+    db = lancedb.connect(tmp_path)
+    with pytest.raises(Exception):
+        db.create_table("test_with_no_data")
+
+    with pytest.raises(Exception):
+        db.open_table("does_not_exist")
+
+    schema = pa.schema([pa.field("a", pa.int32())])
+    db.create_table("test", schema=schema)
+
+
+def test_replace_index(tmp_path):
+    db = lancedb.connect(uri=tmp_path)
+    table = db.create_table(
+        "test",
+        [
+            {"vector": np.random.rand(128), "item": "foo", "price": float(i)}
+            for i in range(1000)
+        ],
+    )
+    table.create_index(
+        num_partitions=2,
+        num_sub_vectors=4,
+    )
+
+    with pytest.raises(Exception):
+        table.create_index(
+            num_partitions=2,
+            num_sub_vectors=4,
+            replace=False,
+        )
+
+    table.create_index(
+        num_partitions=2,
+        num_sub_vectors=4,
+        replace=True,
+    )

python/tests/test_e2e_remote_db.py

@@ -0,0 +1,27 @@
+#  Copyright 2023 LanceDB Developers
+#
+#  Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+#  You may obtain a copy of the License at
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+#  Unless required by applicable law or agreed to in writing, software
+#  distributed under the License is distributed on an "AS IS" BASIS,
+#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#  See the License for the specific language governing permissions and
+#  limitations under the License.
+
+import numpy as np
+import pytest
+
+from lancedb import LanceDBConnection
+
+# TODO: setup integ test mark and script
+
+
+@pytest.mark.skip(reason="Need to set up a local server")
+def test_against_local_server():
+    conn = LanceDBConnection("lancedb+http://localhost:10024")
+    table = conn.open_table("sift1m_ivf1024_pq16")
+    df = table.search(np.random.rand(128)).to_df()
+    assert len(df) == 10

python/tests/test_embeddings.py

@@ -14,6 +14,7 @@ import sys
 
 import numpy as np
 import pyarrow as pa
+
 from lancedb.embeddings import with_embeddings
 
 

python/tests/test_fts.py

@@ -13,13 +13,13 @@
 import os
 import random
 
-import lancedb.fts
 import numpy as np
 import pandas as pd
 import pytest
 import tantivy
 
 import lancedb as ldb
+import lancedb.fts
 
 
 @pytest.fixture

python/tests/test_io.py

@@ -12,6 +12,7 @@
 #  limitations under the License.
 
 import os
+
 import pytest
 
 import lancedb

python/tests/test_pydantic.py

@@ -0,0 +1,155 @@
+#  Copyright 2023 LanceDB Developers
+#
+#  Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+#  You may obtain a copy of the License at
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+#  Unless required by applicable law or agreed to in writing, software
+#  distributed under the License is distributed on an "AS IS" BASIS,
+#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#  See the License for the specific language governing permissions and
+#  limitations under the License.
+
+
+import json
+import sys
+from typing import List, Optional
+
+import pyarrow as pa
+import pydantic
+import pytest
+
+from lancedb.pydantic import pydantic_to_schema, vector
+
+
+@pytest.mark.skipif(
+    sys.version_info < (3, 9),
+    reason="using native type alias requires python3.9 or higher",
+)
+def test_pydantic_to_arrow():
+    class StructModel(pydantic.BaseModel):
+        a: str
+        b: Optional[float]
+
+    class TestModel(pydantic.BaseModel):
+        id: int
+        s: str
+        vec: list[float]
+        li: List[int]
+        opt: Optional[str] = None
+        st: StructModel
+        # d: dict
+
+    m = TestModel(
+        id=1, s="hello", vec=[1.0, 2.0, 3.0], li=[2, 3, 4], st=StructModel(a="a", b=1.0)
+    )
+
+    schema = pydantic_to_schema(TestModel)
+
+    expect_schema = pa.schema(
+        [
+            pa.field("id", pa.int64(), False),
+            pa.field("s", pa.utf8(), False),
+            pa.field("vec", pa.list_(pa.float64()), False),
+            pa.field("li", pa.list_(pa.int64()), False),
+            pa.field("opt", pa.utf8(), True),
+            pa.field(
+                "st",
+                pa.struct(
+                    [pa.field("a", pa.utf8(), False), pa.field("b", pa.float64(), True)]
+                ),
+                False,
+            ),
+        ]
+    )
+    assert schema == expect_schema
+
+
+def test_pydantic_to_arrow_py38():
+    class StructModel(pydantic.BaseModel):
+        a: str
+        b: Optional[float]
+
+    class TestModel(pydantic.BaseModel):
+        id: int
+        s: str
+        vec: List[float]
+        li: List[int]
+        opt: Optional[str] = None
+        st: StructModel
+        # d: dict
+
+    m = TestModel(
+        id=1, s="hello", vec=[1.0, 2.0, 3.0], li=[2, 3, 4], st=StructModel(a="a", b=1.0)
+    )
+
+    schema = pydantic_to_schema(TestModel)
+
+    expect_schema = pa.schema(
+        [
+            pa.field("id", pa.int64(), False),
+            pa.field("s", pa.utf8(), False),
+            pa.field("vec", pa.list_(pa.float64()), False),
+            pa.field("li", pa.list_(pa.int64()), False),
+            pa.field("opt", pa.utf8(), True),
+            pa.field(
+                "st",
+                pa.struct(
+                    [pa.field("a", pa.utf8(), False), pa.field("b", pa.float64(), True)]
+                ),
+                False,
+            ),
+        ]
+    )
+    assert schema == expect_schema
+
+
+def test_fixed_size_list_field():
+    class TestModel(pydantic.BaseModel):
+        vec: vector(16)
+        li: List[int]
+
+    data = TestModel(vec=list(range(16)), li=[1, 2, 3])
+    assert json.loads(data.model_dump_json()) == {
+        "vec": list(range(16)),
+        "li": [1, 2, 3],
+    }
+
+    schema = pydantic_to_schema(TestModel)
+    assert schema == pa.schema(
+        [
+            pa.field("vec", pa.list_(pa.float32(), 16), False),
+            pa.field("li", pa.list_(pa.int64()), False),
+        ]
+    )
+
+    json_schema = TestModel.model_json_schema()
+    assert json_schema == {
+        "properties": {
+            "vec": {
+                "items": {"type": "number"},
+                "maxItems": 16,
+                "minItems": 16,
+                "title": "Vec",
+                "type": "array",
+            },
+            "li": {"items": {"type": "integer"}, "title": "Li", "type": "array"},
+        },
+        "required": ["vec", "li"],
+        "title": "TestModel",
+        "type": "object",
+    }
+
+
+def test_fixed_size_list_validation():
+    class TestModel(pydantic.BaseModel):
+        vec: vector(8)
+
+    with pytest.raises(pydantic.ValidationError):
+        TestModel(vec=range(9))
+
+    with pytest.raises(pydantic.ValidationError):
+        TestModel(vec=range(7))
+
+    TestModel(vec=range(8))

python/tests/test_query.py

@@ -11,22 +11,42 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 
+import unittest.mock as mock
+
 import lance
 import numpy as np
-import pandas as pd
 import pandas.testing as tm
 import pyarrow as pa
 import pytest
-from lancedb.query import LanceQueryBuilder
+
+from lancedb.db import LanceDBConnection
+from lancedb.query import LanceQueryBuilder, Query
+from lancedb.table import LanceTable
 
 
 class MockTable:
     def __init__(self, tmp_path):
         self.uri = tmp_path
+        self._conn = LanceDBConnection(self.uri)
 
     def to_lance(self):
         return lance.dataset(self.uri)
 
+    def _execute_query(self, query):
+        ds = self.to_lance()
+        return ds.to_table(
+            columns=query.columns,
+            filter=query.filter,
+            nearest={
+                "column": query.vector_column,
+                "q": query.vector,
+                "k": query.k,
+                "metric": query.metric,
+                "nprobes": query.nprobes,
+                "refine_factor": query.refine_factor,
+            },
+        )
+
 
 @pytest.fixture
 def table(tmp_path) -> MockTable:
@@ -45,24 +65,30 @@ def table(tmp_path) -> MockTable:
 
 
 def test_query_builder(table):
-    df = LanceQueryBuilder(table, [0, 0]).limit(1).select(["id"]).to_df()
+    df = LanceQueryBuilder(table, [0, 0], "vector").limit(1).select(["id"]).to_df()
     assert df["id"].values[0] == 1
     assert all(df["vector"].values[0] == [1, 2])
 
 
 def test_query_builder_with_filter(table):
-    df = LanceQueryBuilder(table, [0, 0]).where("id = 2").to_df()
+    df = LanceQueryBuilder(table, [0, 0], "vector").where("id = 2").to_df()
     assert df["id"].values[0] == 2
     assert all(df["vector"].values[0] == [3, 4])
 
 
 def test_query_builder_with_metric(table):
     query = [4, 8]
-    df_default = LanceQueryBuilder(table, query).to_df()
+    vector_column_name = "vector"
-    df_l2 = LanceQueryBuilder(table, query).metric("L2").to_df()
+    df_default = LanceQueryBuilder(table, query, vector_column_name).to_df()
+    df_l2 = LanceQueryBuilder(table, query, vector_column_name).metric("L2").to_df()
     tm.assert_frame_equal(df_default, df_l2)
 
-    df_cosine = LanceQueryBuilder(table, query).metric("cosine").limit(1).to_df()
+    df_cosine = (
+        LanceQueryBuilder(table, query, vector_column_name)
+        .metric("cosine")
+        .limit(1)
+        .to_df()
+    )
     assert df_cosine.score[0] == pytest.approx(
         cosine_distance(query, df_cosine.vector[0]),
         abs=1e-6,
@@ -70,5 +96,33 @@ def test_query_builder_with_metric(table):
     assert 0 <= df_cosine.score[0] <= 1
 
 
+def test_query_builder_with_different_vector_column():
+    table = mock.MagicMock(spec=LanceTable)
+    query = [4, 8]
+    vector_column_name = "foo_vector"
+    builder = (
+        LanceQueryBuilder(table, query, vector_column_name)
+        .metric("cosine")
+        .where("b < 10")
+        .select(["b"])
+        .limit(2)
+    )
+    ds = mock.Mock()
+    table.to_lance.return_value = ds
+    builder.to_arrow()
+    table._execute_query.assert_called_once_with(
+        Query(
+            vector=query,
+            filter="b < 10",
+            k=2,
+            metric="cosine",
+            columns=["b"],
+            nprobes=20,
+            refine_factor=None,
+            vector_column="foo_vector",
+        )
+    )
+
+
 def cosine_distance(vec1, vec2):
     return 1 - np.dot(vec1, vec2) / (np.linalg.norm(vec1) * np.linalg.norm(vec2))

python/tests/test_remote_client.py

@@ -0,0 +1,95 @@
+#  Copyright 2023 LanceDB Developers
+#
+#  Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+#  You may obtain a copy of the License at
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+#  Unless required by applicable law or agreed to in writing, software
+#  distributed under the License is distributed on an "AS IS" BASIS,
+#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#  See the License for the specific language governing permissions and
+#  limitations under the License.
+
+import attr
+import numpy as np
+import pandas as pd
+import pyarrow as pa
+import pytest
+from aiohttp import web
+
+from lancedb.remote.client import RestfulLanceDBClient, VectorQuery
+
+
+@attr.define
+class MockLanceDBServer:
+    runner: web.AppRunner = attr.field(init=False)
+    site: web.TCPSite = attr.field(init=False)
+
+    async def query_handler(self, request: web.Request) -> web.Response:
+        table_name = request.match_info["table_name"]
+        assert table_name == "test_table"
+
+        await request.json()
+        # TODO: do some matching
+
+        vecs = pd.Series([np.random.rand(128) for x in range(10)], name="vector")
+        ids = pd.Series(range(10), name="id")
+        df = pd.DataFrame([vecs, ids]).T
+
+        batch = pa.RecordBatch.from_pandas(
+            df,
+            schema=pa.schema(
+                [
+                    pa.field("vector", pa.list_(pa.float32(), 128)),
+                    pa.field("id", pa.int64()),
+                ]
+            ),
+        )
+
+        sink = pa.BufferOutputStream()
+        with pa.ipc.new_file(sink, batch.schema) as writer:
+            writer.write_batch(batch)
+
+        return web.Response(body=sink.getvalue().to_pybytes())
+
+    async def setup(self):
+        app = web.Application()
+        app.add_routes([web.post("/table/{table_name}", self.query_handler)])
+        self.runner = web.AppRunner(app)
+        await self.runner.setup()
+        self.site = web.TCPSite(self.runner, "localhost", 8111)
+
+    async def start(self):
+        await self.site.start()
+
+    async def stop(self):
+        await self.runner.cleanup()
+
+
+@pytest.mark.skip(reason="flaky somehow, fix later")
+@pytest.mark.asyncio
+async def test_e2e_with_mock_server():
+    mock_server = MockLanceDBServer()
+    await mock_server.setup()
+    await mock_server.start()
+
+    try:
+        client = RestfulLanceDBClient("lancedb+http://localhost:8111")
+        df = (
+            await client.query(
+                "test_table",
+                VectorQuery(
+                    vector=np.random.rand(128).tolist(),
+                    k=10,
+                    _metric="L2",
+                    columns=["id", "vector"],
+                ),
+            )
+        ).to_df()
+
+        assert "vector" in df.columns
+        assert "id" in df.columns
+    finally:
+        # make sure we don't leak resources
+        await mock_server.stop()

python/tests/test_remote_db.py

@@ -0,0 +1,35 @@
+#  Copyright 2023 LanceDB Developers
+#
+#  Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+#  You may obtain a copy of the License at
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+#  Unless required by applicable law or agreed to in writing, software
+#  distributed under the License is distributed on an "AS IS" BASIS,
+#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#  See the License for the specific language governing permissions and
+#  limitations under the License.
+
+import pyarrow as pa
+
+import lancedb
+from lancedb.remote.client import VectorQuery, VectorQueryResult
+
+
+class FakeLanceDBClient:
+    async def close(self):
+        pass
+
+    async def query(self, table_name: str, query: VectorQuery) -> VectorQueryResult:
+        assert table_name == "test"
+        t = pa.schema([]).empty_table()
+        return VectorQueryResult(t)
+
+
+def test_remote_db():
+    conn = lancedb.connect("db://client-will-be-injected", api_key="fake")
+    setattr(conn, "_client", FakeLanceDBClient())
+
+    table = conn["test"]
+    table.search([1.0, 2.0]).to_df()

python/tests/test_table.py

@@ -11,11 +11,17 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 
+import functools
 from pathlib import Path
+from unittest.mock import PropertyMock, patch
 
+import numpy as np
 import pandas as pd
 import pyarrow as pa
 import pytest
+from lance.vector import vec_to_table
+
+from lancedb.db import LanceDBConnection
 from lancedb.table import LanceTable
 
 
@@ -23,6 +29,10 @@ class MockDB:
     def __init__(self, uri: Path):
         self.uri = uri
 
+    @functools.cached_property
+    def is_managed_remote(self) -> bool:
+        return False
+
 
 @pytest.fixture
 def db(tmp_path) -> MockDB:
@@ -80,7 +90,31 @@ def test_create_table(db):
         assert expected == tbl
 
 
+def test_empty_table(db):
+    schema = pa.schema(
+        [
+            pa.field("vector", pa.list_(pa.float32(), 2)),
+            pa.field("item", pa.string()),
+            pa.field("price", pa.float32()),
+        ]
+    )
+    tbl = LanceTable.create(db, "test", schema=schema)
+    data = [
+        {"vector": [3.1, 4.1], "item": "foo", "price": 10.0},
+        {"vector": [5.9, 26.5], "item": "bar", "price": 20.0},
+    ]
+    tbl.add(data=data)
+
+
 def test_add(db):
+    schema = pa.schema(
+        [
+            pa.field("vector", pa.list_(pa.float32(), 2)),
+            pa.field("item", pa.string()),
+            pa.field("price", pa.float64()),
+        ]
+    )
+
     table = LanceTable.create(
         db,
         "test",
@@ -89,7 +123,19 @@ def test_add(db):
             {"vector": [5.9, 26.5], "item": "bar", "price": 20.0},
         ],
     )
+    _add(table, schema)
 
+    table = LanceTable.create(db, "test2", schema=schema)
+    table.add(
+        data=[
+            {"vector": [3.1, 4.1], "item": "foo", "price": 10.0},
+            {"vector": [5.9, 26.5], "item": "bar", "price": 20.0},
+        ],
+    )
+    _add(table, schema)
+
+
+def _add(table, schema):
     # table = LanceTable(db, "test")
     assert len(table) == 2
 
@@ -104,13 +150,7 @@ def test_add(db):
             pa.array(["foo", "bar", "new"]),
             pa.array([10.0, 20.0, 30.0]),
         ],
-        schema=pa.schema(
+        schema=schema,
-            [
-                pa.field("vector", pa.list_(pa.float32(), 2)),
-                pa.field("item", pa.string()),
-                pa.field("price", pa.float64()),
-            ]
-        ),
     )
     assert expected == table.to_arrow()
 
@@ -136,3 +176,83 @@ def test_versioning(db):
     table.checkout(1)
     assert table.version == 1
     assert len(table) == 2
+
+
+def test_create_index_method():
+    with patch.object(LanceTable, "_reset_dataset", return_value=None):
+        with patch.object(
+            LanceTable, "_dataset", new_callable=PropertyMock
+        ) as mock_dataset:
+            # Setup mock responses
+            mock_dataset.return_value.create_index.return_value = None
+
+            # Create a LanceTable object
+            connection = LanceDBConnection(uri="mock.uri")
+            table = LanceTable(connection, "test_table")
+
+            # Call the create_index method
+            table.create_index(
+                metric="L2",
+                num_partitions=256,
+                num_sub_vectors=96,
+                vector_column_name="vector",
+                replace=True,
+            )
+
+            # Check that the _dataset.create_index method was called
+            # with the right parameters
+            mock_dataset.return_value.create_index.assert_called_once_with(
+                column="vector",
+                index_type="IVF_PQ",
+                metric="L2",
+                num_partitions=256,
+                num_sub_vectors=96,
+                replace=True,
+            )
+
+
+def test_add_with_nans(db):
+    # by default we raise an error on bad input vectors
+    bad_data = [
+        {"vector": [np.nan], "item": "bar", "price": 20.0},
+        {"vector": [5], "item": "bar", "price": 20.0},
+        {"vector": [np.nan, np.nan], "item": "bar", "price": 20.0},
+        {"vector": [np.nan, 5.0], "item": "bar", "price": 20.0},
+    ]
+    for row in bad_data:
+        with pytest.raises(ValueError):
+            LanceTable.create(
+                db,
+                "error_test",
+                data=[{"vector": [3.1, 4.1], "item": "foo", "price": 10.0}, row],
+            )
+
+    table = LanceTable.create(
+        db,
+        "drop_test",
+        data=[
+            {"vector": [3.1, 4.1], "item": "foo", "price": 10.0},
+            {"vector": [np.nan], "item": "bar", "price": 20.0},
+            {"vector": [5], "item": "bar", "price": 20.0},
+            {"vector": [np.nan, np.nan], "item": "bar", "price": 20.0},
+        ],
+        on_bad_vectors="drop",
+    )
+    assert len(table) == 1
+
+    # We can fill bad input with some value
+    table = LanceTable.create(
+        db,
+        "fill_test",
+        data=[
+            {"vector": [3.1, 4.1], "item": "foo", "price": 10.0},
+            {"vector": [np.nan], "item": "bar", "price": 20.0},
+            {"vector": [np.nan, np.nan], "item": "bar", "price": 20.0},
+        ],
+        on_bad_vectors="fill",
+        fill_value=0.0,
+    )
+    assert len(table) == 3
+    arrow_tbl = table.to_lance().to_table(filter="item == 'bar'")
+    v = arrow_tbl["vector"].to_pylist()[0]
+    assert np.allclose(v, np.array([0.0, 0.0]))

rust/ffi/node/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "vectordb-node"
-version = "0.1.0"
+version = "0.1.13"
 description = "Serverless, low-latency vector database for AI applications"
 license = "Apache-2.0"
 edition = "2018"
@@ -10,12 +10,16 @@ exclude = ["index.node"]
 crate-type = ["cdylib"]
 
 [dependencies]
-arrow-array = "37.0"
+arrow-array = { workspace = true }
-arrow-ipc = "37.0"
+arrow-ipc = { workspace = true }
-arrow-schema = "37.0"
+arrow-schema = { workspace = true }
 once_cell = "1"
 futures = "0.3"
-lance = "0.4.17"
+half = { workspace = true }
+lance = { workspace = true }
 vectordb = { path = "../../vectordb" }
 tokio = { version = "1.23", features = ["rt-multi-thread"] }
 neon = {version = "0.10.1", default-features = false, features = ["channel-api", "napi-6", "promise-api", "task-api"] }
+object_store = { workspace = true, features = ["aws"] }
+async-trait = "0"
+env_logger = "0"

rust/ffi/node/src/arrow.rs

@@ -13,7 +13,6 @@
 // limitations under the License.
 
 use std::io::Cursor;
-use std::ops::Deref;
 use std::sync::Arc;
 
 use arrow_array::cast::as_list_array;
@@ -25,10 +24,13 @@ use lance::arrow::{FixedSizeListArrayExt, RecordBatchExt};
 pub(crate) fn convert_record_batch(record_batch: RecordBatch) -> RecordBatch {
     let column = record_batch
         .column_by_name("vector")
+        .cloned()
         .expect("vector column is missing");
-    let arr = as_list_array(column.deref());
+    // TODO: we should just consume the underlaying js buffer in the future instead of this arrow around a bunch of times
+    let arr = as_list_array(column.as_ref());
     let list_size = arr.values().len() / record_batch.num_rows();
-    let r = FixedSizeListArray::try_new(arr.values(), list_size as i32).unwrap();
+    let r =
+        FixedSizeListArray::try_new_from_values(arr.values().to_owned(), list_size as i32).unwrap();
 
     let schema = Arc::new(Schema::new(vec![Field::new(
         "vector",

rust/ffi/node/src/index/vector.rs

@@ -97,6 +97,7 @@ fn get_index_params_builder(
                 let ivf_params = IvfBuildParams {
                     num_partitions: np,
                     max_iters,
+                    centroids: None,
                 };
                 index_builder.ivf_params(ivf_params)
             });
@@ -121,6 +122,10 @@ fn get_index_params_builder(
                 .map_err(|t| t.to_string())?
                 .map(|s| pq_params.max_opq_iters = s.value(cx) as usize);
 
+            obj.get_opt::<JsBoolean, _, _>(cx, "replace")
+                .map_err(|t| t.to_string())?
+                .map(|s| index_builder.replace(s.value(cx)));
+
             Ok(index_builder)
         }
         t => Err(format!("{} is not a valid index type", t).to_string()),

rust/ffi/node/src/lib.rs

@@ -17,20 +17,23 @@ use std::convert::TryFrom;
 use std::ops::Deref;
 use std::sync::{Arc, Mutex};
 
-use arrow_array::{Float32Array, RecordBatchReader};
+use arrow_array::{Float32Array, RecordBatchIterator};
 use arrow_ipc::writer::FileWriter;
+use async_trait::async_trait;
 use futures::{TryFutureExt, TryStreamExt};
-use lance::arrow::RecordBatchBuffer;
+use lance::dataset::{ReadParams, WriteMode, WriteParams};
-use lance::dataset::WriteMode;
 use lance::index::vector::MetricType;
+use lance::io::object_store::ObjectStoreParams;
 use neon::prelude::*;
 use neon::types::buffer::TypedArray;
+use object_store::aws::{AwsCredential, AwsCredentialProvider};
+use object_store::CredentialProvider;
 use once_cell::sync::OnceCell;
 use tokio::runtime::Runtime;
 
 use vectordb::database::Database;
 use vectordb::error::Error;
-use vectordb::table::Table;
+use vectordb::table::{OpenTableParams, Table};
 
 use crate::arrow::arrow_buffer_to_record_batch;
 
@@ -50,8 +53,38 @@ struct JsTable {
 
 impl Finalize for JsTable {}
 
+// TODO: object_store didn't export this type so I copied it.
+// Make a requiest to object_store to export this type
+#[derive(Debug)]
+pub struct StaticCredentialProvider<T> {
+    credential: Arc<T>,
+}
+
+impl<T> StaticCredentialProvider<T> {
+    pub fn new(credential: T) -> Self {
+        Self {
+            credential: Arc::new(credential),
+        }
+    }
+}
+
+#[async_trait]
+impl<T> CredentialProvider for StaticCredentialProvider<T>
+where
+    T: std::fmt::Debug + Send + Sync,
+{
+    type Credential = T;
+
+    async fn get_credential(&self) -> object_store::Result<Arc<T>> {
+        Ok(Arc::clone(&self.credential))
+    }
+}
+
 fn runtime<'a, C: Context<'a>>(cx: &mut C) -> NeonResult<&'static Runtime> {
     static RUNTIME: OnceCell<Runtime> = OnceCell::new();
+    static LOG: OnceCell<()> = OnceCell::new();
+
+    LOG.get_or_init(|| env_logger::init());
 
     RUNTIME.get_or_try_init(|| Runtime::new().or_else(|err| cx.throw_error(err.to_string())))
 }
@@ -98,25 +131,101 @@ fn database_table_names(mut cx: FunctionContext) -> JsResult<JsPromise> {
     Ok(promise)
 }
 
+fn get_aws_creds<T>(
+    cx: &mut FunctionContext,
+    arg_starting_location: i32,
+) -> Result<Option<AwsCredentialProvider>, NeonResult<T>> {
+    let secret_key_id = cx
+        .argument_opt(arg_starting_location)
+        .map(|arg| arg.downcast_or_throw::<JsString, FunctionContext>(cx).ok())
+        .flatten()
+        .map(|v| v.value(cx));
+
+    let secret_key = cx
+        .argument_opt(arg_starting_location + 1)
+        .map(|arg| arg.downcast_or_throw::<JsString, FunctionContext>(cx).ok())
+        .flatten()
+        .map(|v| v.value(cx));
+
+    let temp_token = cx
+        .argument_opt(arg_starting_location + 2)
+        .map(|arg| arg.downcast_or_throw::<JsString, FunctionContext>(cx).ok())
+        .flatten()
+        .map(|v| v.value(cx));
+
+    match (secret_key_id, secret_key, temp_token) {
+        (Some(key_id), Some(key), optional_token) => Ok(Some(Arc::new(
+            StaticCredentialProvider::new(AwsCredential {
+                key_id: key_id,
+                secret_key: key,
+                token: optional_token,
+            }),
+        ))),
+        (None, None, None) => Ok(None),
+        _ => Err(cx.throw_error("Invalid credentials configuration")),
+    }
+}
+
 fn database_open_table(mut cx: FunctionContext) -> JsResult<JsPromise> {
     let db = cx
         .this()
         .downcast_or_throw::<JsBox<JsDatabase>, _>(&mut cx)?;
     let table_name = cx.argument::<JsString>(0)?.value(&mut cx);
 
+    let aws_creds = match get_aws_creds(&mut cx, 1) {
+        Ok(creds) => creds,
+        Err(err) => return err,
+    };
+
+    let param = ReadParams {
+        store_options: Some(ObjectStoreParams {
+            aws_credentials: aws_creds,
+            ..ObjectStoreParams::default()
+        }),
+        ..ReadParams::default()
+    };
+
+    let rt = runtime(&mut cx)?;
+    let channel = cx.channel();
+    let database = db.database.clone();
+
+    let (deferred, promise) = cx.promise();
+    rt.spawn(async move {
+        let table_rst = database
+            .open_table_with_params(
+                &table_name,
+                OpenTableParams {
+                    open_table_params: param,
+                },
+            )
+            .await;
+
+        deferred.settle_with(&channel, move |mut cx| {
+            let table = Arc::new(Mutex::new(
+                table_rst.or_else(|err| cx.throw_error(err.to_string()))?,
+            ));
+            Ok(cx.boxed(JsTable { table }))
+        });
+    });
+    Ok(promise)
+}
+
+fn database_drop_table(mut cx: FunctionContext) -> JsResult<JsPromise> {
+    let db = cx
+        .this()
+        .downcast_or_throw::<JsBox<JsDatabase>, _>(&mut cx)?;
+    let table_name = cx.argument::<JsString>(0)?.value(&mut cx);
+
     let rt = runtime(&mut cx)?;
     let channel = cx.channel();
     let database = db.database.clone();
 
     let (deferred, promise) = cx.promise();
     rt.spawn(async move {
-        let table_rst = database.open_table(&table_name).await;
+        let result = database.drop_table(&table_name).await;
-
         deferred.settle_with(&channel, move |mut cx| {
-            let table = Arc::new(Mutex::new(
+            result.or_else(|err| cx.throw_error(err.to_string()))?;
-                table_rst.or_else(|err| cx.throw_error(err.to_string()))?,
+            Ok(cx.null())
-            ));
-            Ok(cx.boxed(JsTable { table }))
         });
     });
     Ok(promise)
@@ -212,6 +321,15 @@ fn table_create(mut cx: FunctionContext) -> JsResult<JsPromise> {
     let table_name = cx.argument::<JsString>(0)?.value(&mut cx);
     let buffer = cx.argument::<JsBuffer>(1)?;
     let batches = arrow_buffer_to_record_batch(buffer.as_slice(&mut cx));
+    let schema = batches[0].schema();
+
+    // Write mode
+    let mode = match cx.argument::<JsString>(2)?.value(&mut cx).as_str() {
+        "overwrite" => WriteMode::Overwrite,
+        "append" => WriteMode::Append,
+        "create" => WriteMode::Create,
+        _ => return cx.throw_error("Table::create only supports 'overwrite' and 'create' modes"),
+    };
 
     let rt = runtime(&mut cx)?;
     let channel = cx.channel();
@@ -219,9 +337,25 @@ fn table_create(mut cx: FunctionContext) -> JsResult<JsPromise> {
     let (deferred, promise) = cx.promise();
     let database = db.database.clone();
 
+    let aws_creds = match get_aws_creds(&mut cx, 3) {
+        Ok(creds) => creds,
+        Err(err) => return err,
+    };
+
+    let params = WriteParams {
+        store_params: Some(ObjectStoreParams {
+            aws_credentials: aws_creds,
+            ..ObjectStoreParams::default()
+        }),
+        mode: mode,
+        ..WriteParams::default()
+    };
+
     rt.block_on(async move {
-        let batch_reader: Box<dyn RecordBatchReader> = Box::new(RecordBatchBuffer::new(batches));
+        let batch_reader = RecordBatchIterator::new(batches.into_iter().map(Ok), schema);
-        let table_rst = database.create_table(&table_name, batch_reader).await;
+        let table_rst = database
+            .create_table(&table_name, batch_reader, Some(params))
+            .await;
 
         deferred.settle_with(&channel, move |mut cx| {
             let table = Arc::new(Mutex::new(
@@ -244,6 +378,7 @@ fn table_add(mut cx: FunctionContext) -> JsResult<JsPromise> {
     let buffer = cx.argument::<JsBuffer>(0)?;
     let write_mode = cx.argument::<JsString>(1)?.value(&mut cx);
     let batches = arrow_buffer_to_record_batch(buffer.as_slice(&mut cx));
+    let schema = batches[0].schema();
 
     let rt = runtime(&mut cx)?;
     let channel = cx.channel();
@@ -252,26 +387,82 @@ fn table_add(mut cx: FunctionContext) -> JsResult<JsPromise> {
     let table = js_table.table.clone();
     let write_mode = write_mode_map.get(write_mode.as_str()).cloned();
 
+    let aws_creds = match get_aws_creds(&mut cx, 2) {
+        Ok(creds) => creds,
+        Err(err) => return err,
+    };
+
+    let params = WriteParams {
+        store_params: Some(ObjectStoreParams {
+            aws_credentials: aws_creds,
+            ..ObjectStoreParams::default()
+        }),
+        mode: write_mode.unwrap_or(WriteMode::Append),
+        ..WriteParams::default()
+    };
+
     rt.block_on(async move {
-        let batch_reader: Box<dyn RecordBatchReader> = Box::new(RecordBatchBuffer::new(batches));
+        let batch_reader = RecordBatchIterator::new(batches.into_iter().map(Ok), schema);
-        let add_result = table.lock().unwrap().add(batch_reader, write_mode).await;
+        let add_result = table.lock().unwrap().add(batch_reader, Some(params)).await;
 
         deferred.settle_with(&channel, move |mut cx| {
-            let added = add_result.or_else(|err| cx.throw_error(err.to_string()))?;
+            let _added = add_result.or_else(|err| cx.throw_error(err.to_string()))?;
-            Ok(cx.number(added as f64))
+            Ok(cx.boolean(true))
         });
     });
     Ok(promise)
 }
 
+fn table_count_rows(mut cx: FunctionContext) -> JsResult<JsPromise> {
+    let js_table = cx.this().downcast_or_throw::<JsBox<JsTable>, _>(&mut cx)?;
+    let rt = runtime(&mut cx)?;
+    let channel = cx.channel();
+
+    let (deferred, promise) = cx.promise();
+    let table = js_table.table.clone();
+
+    rt.block_on(async move {
+        let num_rows_result = table.lock().unwrap().count_rows().await;
+
+        deferred.settle_with(&channel, move |mut cx| {
+            let num_rows = num_rows_result.or_else(|err| cx.throw_error(err.to_string()))?;
+            Ok(cx.number(num_rows as f64))
+        });
+    });
+    Ok(promise)
+}
+
+fn table_delete(mut cx: FunctionContext) -> JsResult<JsPromise> {
+    let js_table = cx.this().downcast_or_throw::<JsBox<JsTable>, _>(&mut cx)?;
+    let rt = runtime(&mut cx)?;
+    let channel = cx.channel();
+
+    let (deferred, promise) = cx.promise();
+    let table = js_table.table.clone();
+
+    let predicate = cx.argument::<JsString>(0)?.value(&mut cx);
+
+    let delete_result = rt.block_on(async move { table.lock().unwrap().delete(&predicate).await });
+
+    deferred.settle_with(&channel, move |mut cx| {
+        delete_result.or_else(|err| cx.throw_error(err.to_string()))?;
+        Ok(cx.undefined())
+    });
+
+    Ok(promise)
+}
+
 #[neon::main]
 fn main(mut cx: ModuleContext) -> NeonResult<()> {
     cx.export_function("databaseNew", database_new)?;
     cx.export_function("databaseTableNames", database_table_names)?;
     cx.export_function("databaseOpenTable", database_open_table)?;
+    cx.export_function("databaseDropTable", database_drop_table)?;
     cx.export_function("tableSearch", table_search)?;
     cx.export_function("tableCreate", table_create)?;
     cx.export_function("tableAdd", table_add)?;
+    cx.export_function("tableCountRows", table_count_rows)?;
+    cx.export_function("tableDelete", table_delete)?;
     cx.export_function(
         "tableCreateVectorIndex",
         index::vector::table_create_vector_index,

rust/vectordb/Cargo.toml

@@ -1,20 +1,20 @@
 [package]
 name = "vectordb"
-version = "0.0.1"
+version = "0.1.13"
 edition = "2021"
 description = "Serverless, low-latency vector database for AI applications"
 license = "Apache-2.0"
 repository = "https://github.com/lancedb/lancedb"
 
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
-
 [dependencies]
-arrow-array = "37.0"
+arrow-array = { workspace = true }
-arrow-data = "37.0"
+arrow-data = { workspace = true }
-arrow-schema = "37.0"
+arrow-schema = { workspace = true }
-object_store = "0.5.6"
+object_store = { workspace = true }
 snafu = "0.7.4"
-lance = "0.4.17"
+half = { workspace = true }
+lance = { workspace = true }
 tokio = { version = "1.23", features = ["rt-multi-thread"] }
 
 [dev-dependencies]