Bump version: 0.21.3-beta.0 → 0.22.0-beta.0

Fix restore to always checkout latest version, following local restore
api implementation

a1d1833a40/rust/lancedb/src/table.rs (L1910)
Otherwise
table.create_table -> version 1
table.add_table -> version 2
table.checkout(1), table.restore() -> the version remains at 1 (should
checkout_latest inside restore method to update version to latest
version and allow write operation)
table.checkout_latest() -> version is 3
can do write operations

.bumpversion.toml

@@ -1,5 +1,5 @@
 [tool.bumpversion]
-current_version = "0.18.0"
+current_version = "0.18.3-beta.0"
 parse = """(?x)
     (?P<major>0|[1-9]\\d*)\\.
     (?P<minor>0|[1-9]\\d*)\\.
@@ -87,26 +87,11 @@ glob = "node/package.json"
 replace = "\"@lancedb/vectordb-linux-x64-gnu\": \"{new_version}\""
 search = "\"@lancedb/vectordb-linux-x64-gnu\": \"{current_version}\""
 
-[[tool.bumpversion.files]]
-glob = "node/package.json"
-replace = "\"@lancedb/vectordb-linux-arm64-musl\": \"{new_version}\""
-search = "\"@lancedb/vectordb-linux-arm64-musl\": \"{current_version}\""
-
-[[tool.bumpversion.files]]
-glob = "node/package.json"
-replace = "\"@lancedb/vectordb-linux-x64-musl\": \"{new_version}\""
-search = "\"@lancedb/vectordb-linux-x64-musl\": \"{current_version}\""
-
 [[tool.bumpversion.files]]
 glob = "node/package.json"
 replace = "\"@lancedb/vectordb-win32-x64-msvc\": \"{new_version}\""
 search = "\"@lancedb/vectordb-win32-x64-msvc\": \"{current_version}\""
 
-[[tool.bumpversion.files]]
-glob = "node/package.json"
-replace = "\"@lancedb/vectordb-win32-arm64-msvc\": \"{new_version}\""
-search = "\"@lancedb/vectordb-win32-arm64-msvc\": \"{current_version}\""
-
 # Cargo files
 # ------------
 [[tool.bumpversion.files]]

.cargo/config.toml

@@ -34,6 +34,10 @@ rustflags = ["-C", "target-cpu=haswell", "-C", "target-feature=+avx2,+fma,+f16c"
 [target.x86_64-unknown-linux-musl]
 rustflags = ["-C", "target-cpu=haswell", "-C", "target-feature=-crt-static,+avx2,+fma,+f16c"]
 
+[target.aarch64-unknown-linux-musl]
+linker = "aarch64-linux-musl-gcc"
+rustflags = ["-C", "target-feature=-crt-static"]
+
 [target.aarch64-apple-darwin]
 rustflags = ["-C", "target-cpu=apple-m1", "-C", "target-feature=+neon,+fp16,+fhm,+dotprod"]
 
@@ -44,4 +48,4 @@ rustflags = ["-Ctarget-feature=+crt-static"]
 
 # Experimental target for Arm64 Windows
 [target.aarch64-pc-windows-msvc]
-rustflags = ["-Ctarget-feature=+crt-static"]
+rustflags = ["-Ctarget-feature=+crt-static"]

.github/workflows/build_linux_wheel/action.yml

@@ -36,8 +36,7 @@ runs:
         args: ${{ inputs.args }}
         before-script-linux: |
           set -e
-          yum install -y openssl-devel \
-            && curl -L https://github.com/protocolbuffers/protobuf/releases/download/v24.4/protoc-24.4-linux-$(uname -m).zip > /tmp/protoc.zip \
+          curl -L https://github.com/protocolbuffers/protobuf/releases/download/v24.4/protoc-24.4-linux-$(uname -m).zip > /tmp/protoc.zip \
             && unzip /tmp/protoc.zip -d /usr/local \
             && rm /tmp/protoc.zip
     - name: Build Arm Manylinux Wheel
@@ -52,7 +51,7 @@ runs:
         args: ${{ inputs.args }}
         before-script-linux: |
           set -e
-          yum install -y openssl-devel clang \
+          yum install -y clang \
             && curl -L https://github.com/protocolbuffers/protobuf/releases/download/v24.4/protoc-24.4-linux-aarch_64.zip > /tmp/protoc.zip \
             && unzip /tmp/protoc.zip -d /usr/local \
             && rm /tmp/protoc.zip

.github/workflows/java-publish.yml

@@ -43,7 +43,7 @@ jobs:
       - uses: Swatinem/rust-cache@v2
       - uses: actions-rust-lang/setup-rust-toolchain@v1
         with:
-          toolchain: "1.79.0"
+          toolchain: "1.81.0"
           cache-workspaces: "./java/core/lancedb-jni"
           # 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.
@@ -97,7 +97,7 @@ jobs:
       - name: Dry run
         if: github.event_name == 'pull_request'
         run: |
-          mvn --batch-mode -DskipTests package
+          mvn --batch-mode -DskipTests -Drust.release.build=true package
       - name: Set github
         run: |
           git config --global user.email "LanceDB Github Runner"
@@ -108,7 +108,7 @@ jobs:
           echo "use-agent" >> ~/.gnupg/gpg.conf
           echo "pinentry-mode loopback" >> ~/.gnupg/gpg.conf
           export GPG_TTY=$(tty)
-          mvn --batch-mode -DskipTests -DpushChanges=false -Dgpg.passphrase=${{ secrets.GPG_PASSPHRASE }} deploy -P deploy-to-ossrh
+          mvn --batch-mode -DskipTests -Drust.release.build=true -DpushChanges=false -Dgpg.passphrase=${{ secrets.GPG_PASSPHRASE }} deploy -P deploy-to-ossrh
         env:
           SONATYPE_USER: ${{ secrets.SONATYPE_USER }}
           SONATYPE_TOKEN: ${{ secrets.SONATYPE_TOKEN }}

.github/workflows/pypi-publish.yml

@@ -4,6 +4,10 @@ on:
   push:
     tags:
       - 'python-v*'
+  pull_request:
+    # This should trigger a dry run (we skip the final publish step)
+    paths:
+      - .github/workflows/pypi-publish.yml
 
 jobs:
   linux:
@@ -46,6 +50,7 @@ jobs:
           arm-build: ${{ matrix.config.platform == 'aarch64' }}
           manylinux: ${{ matrix.config.manylinux }}
       - uses: ./.github/workflows/upload_wheel
+        if: startsWith(github.ref, 'refs/tags/python-v')
         with:
           pypi_token: ${{ secrets.LANCEDB_PYPI_API_TOKEN }}
           fury_token: ${{ secrets.FURY_TOKEN }}
@@ -75,6 +80,7 @@ jobs:
           python-minor-version: 8
           args: "--release --strip --target ${{ matrix.config.target }} --features fp16kernels"
       - uses: ./.github/workflows/upload_wheel
+        if: startsWith(github.ref, 'refs/tags/python-v')
         with:
           pypi_token: ${{ secrets.LANCEDB_PYPI_API_TOKEN }}
           fury_token: ${{ secrets.FURY_TOKEN }}
@@ -96,10 +102,12 @@ jobs:
           args: "--release --strip"
           vcpkg_token: ${{ secrets.VCPKG_GITHUB_PACKAGES }}
       - uses: ./.github/workflows/upload_wheel
+        if: startsWith(github.ref, 'refs/tags/python-v')
         with:
           pypi_token: ${{ secrets.LANCEDB_PYPI_API_TOKEN }}
           fury_token: ${{ secrets.FURY_TOKEN }}
   gh-release:
+    if: startsWith(github.ref, 'refs/tags/python-v')
     runs-on: ubuntu-latest
     permissions:
       contents: write

.github/workflows/python.yml

@@ -13,6 +13,11 @@ concurrency:
   group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }}
   cancel-in-progress: true
 
+env:
+  # Color output for pytest is off by default.
+  PYTEST_ADDOPTS: "--color=yes"
+  FORCE_COLOR: "1"
+
 jobs:
   lint:
     name: "Lint"
@@ -131,6 +136,10 @@ jobs:
       - uses: ./.github/workflows/run_tests
         with:
           integration: true
+      - name: Test without pylance
+        run: |
+          pip uninstall -y pylance
+          pytest -vv python/tests/test_table.py
       # Make sure wheels are not included in the Rust cache
       - name: Delete wheels
         run: rm -rf target/wheels

.github/workflows/rust.yml

@@ -157,153 +157,33 @@ jobs:
 
   windows:
     runs-on: windows-2022
+    strategy:
+      matrix:
+        target:
+        - x86_64-pc-windows-msvc
+        - aarch64-pc-windows-msvc
+    defaults:
+      run:
+        working-directory: rust/lancedb
     steps:
       - uses: actions/checkout@v4
       - uses: Swatinem/rust-cache@v2
         with:
           workspaces: rust
       - name: Install Protoc v21.12
-        working-directory: C:\
+        run: choco install --no-progress protoc
+      - name: Build
         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
+          rustup target add ${{ matrix.target }}
+          $env:VCPKG_ROOT = $env:VCPKG_INSTALLATION_ROOT
+          cargo build --features remote --tests --locked --target ${{ matrix.target }}
       - name: Run tests
+        # Can only run tests when target matches host
+        if: ${{ matrix.target == 'x86_64-pc-windows-msvc' }}
         run: |
           $env:VCPKG_ROOT = $env:VCPKG_INSTALLATION_ROOT
           cargo test --features remote --locked
 
-  windows-arm64-cross:
-    # We cross compile in Node releases, so we want to make sure
-    # this can run successfully.
-    runs-on: ubuntu-latest
-    container: alpine:edge
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v4
-      - name: Install dependencies (part 1)
-        run: |
-          set -e
-          apk add protobuf-dev curl clang lld llvm19 grep npm bash msitools sed
-      - name: Install rust
-        uses: actions-rust-lang/setup-rust-toolchain@v1
-        with:
-          target: aarch64-pc-windows-msvc
-      - name: Install dependencies (part 2)
-        run: |
-          set -e
-          mkdir -p sysroot
-          cd sysroot
-          sh ../ci/sysroot-aarch64-pc-windows-msvc.sh
-      - name: Check
-        env:
-          CC: clang
-          AR: llvm-ar
-          C_INCLUDE_PATH: /usr/aarch64-pc-windows-msvc/usr/include
-          CARGO_BUILD_TARGET: aarch64-pc-windows-msvc
-          RUSTFLAGS: -Ctarget-feature=+crt-static,+neon,+fp16,+fhm,+dotprod -Clinker=lld -Clink-arg=/LIBPATH:/usr/aarch64-pc-windows-msvc/usr/lib -Clink-arg=arm64rt.lib
-        run: |
-          source $HOME/.cargo/env
-          cargo check --features remote --locked
-
-  windows-arm64:
-    runs-on: windows-4x-arm
-    steps:
-      - name: Install Git
-        run: |
-          Invoke-WebRequest -Uri "https://github.com/git-for-windows/git/releases/download/v2.44.0.windows.1/Git-2.44.0-64-bit.exe" -OutFile "git-installer.exe"
-          Start-Process -FilePath "git-installer.exe" -ArgumentList "/VERYSILENT", "/NORESTART" -Wait
-        shell: powershell
-      - name: Add Git to PATH
-        run: |
-          Add-Content $env:GITHUB_PATH "C:\Program Files\Git\bin"
-          $env:Path = [System.Environment]::GetEnvironmentVariable("Path","Machine") + ";" + [System.Environment]::GetEnvironmentVariable("Path","User")
-        shell: powershell
-      - name: Configure Git symlinks
-        run: git config --global core.symlinks true
-      - uses: actions/checkout@v4
-      - uses: actions/setup-python@v5
-        with:
-          python-version: "3.13"
-      - name: Install Visual Studio Build Tools
-        run: |
-          Invoke-WebRequest -Uri "https://aka.ms/vs/17/release/vs_buildtools.exe" -OutFile "vs_buildtools.exe"
-          Start-Process -FilePath "vs_buildtools.exe" -ArgumentList "--quiet", "--wait", "--norestart", "--nocache", `
-            "--installPath", "C:\BuildTools", `
-            "--add", "Microsoft.VisualStudio.Component.VC.Tools.ARM64", `
-            "--add", "Microsoft.VisualStudio.Component.VC.Tools.x86.x64", `
-            "--add", "Microsoft.VisualStudio.Component.Windows11SDK.22621", `
-            "--add", "Microsoft.VisualStudio.Component.VC.ATL", `
-            "--add", "Microsoft.VisualStudio.Component.VC.ATLMFC", `
-            "--add", "Microsoft.VisualStudio.Component.VC.Llvm.Clang" -Wait
-        shell: powershell
-      - name: Add Visual Studio Build Tools to PATH
-        run: |
-          $vsPath = "C:\BuildTools\VC\Tools\MSVC"
-          $latestVersion = (Get-ChildItem $vsPath | Sort-Object {[version]$_.Name} -Descending)[0].Name
-          Add-Content $env:GITHUB_PATH "C:\BuildTools\VC\Tools\MSVC\$latestVersion\bin\Hostx64\arm64"
-          Add-Content $env:GITHUB_PATH "C:\BuildTools\VC\Tools\MSVC\$latestVersion\bin\Hostx64\x64"
-          Add-Content $env:GITHUB_PATH "C:\Program Files (x86)\Windows Kits\10\bin\10.0.22621.0\arm64"
-          Add-Content $env:GITHUB_PATH "C:\Program Files (x86)\Windows Kits\10\bin\10.0.22621.0\x64"
-          Add-Content $env:GITHUB_PATH "C:\BuildTools\VC\Tools\Llvm\x64\bin"
-
-          # Add MSVC runtime libraries to LIB
-          $env:LIB = "C:\BuildTools\VC\Tools\MSVC\$latestVersion\lib\arm64;" +
-                     "C:\Program Files (x86)\Windows Kits\10\Lib\10.0.22621.0\um\arm64;" +
-                     "C:\Program Files (x86)\Windows Kits\10\Lib\10.0.22621.0\ucrt\arm64"
-          Add-Content $env:GITHUB_ENV "LIB=$env:LIB"
-
-          # Add INCLUDE paths
-          $env:INCLUDE = "C:\BuildTools\VC\Tools\MSVC\$latestVersion\include;" +
-                        "C:\Program Files (x86)\Windows Kits\10\Include\10.0.22621.0\ucrt;" +
-                        "C:\Program Files (x86)\Windows Kits\10\Include\10.0.22621.0\um;" +
-                        "C:\Program Files (x86)\Windows Kits\10\Include\10.0.22621.0\shared"
-          Add-Content $env:GITHUB_ENV "INCLUDE=$env:INCLUDE"
-        shell: powershell
-      - name: Install Rust
-        run: |
-          Invoke-WebRequest https://win.rustup.rs/x86_64 -OutFile rustup-init.exe
-          .\rustup-init.exe -y --default-host aarch64-pc-windows-msvc --default-toolchain 1.83.0
-        shell: powershell
-      - name: Add Rust to PATH
-        run: |
-          Add-Content $env:GITHUB_PATH "$env:USERPROFILE\.cargo\bin"
-        shell: powershell
-      - uses: Swatinem/rust-cache@v2
-        with:
-          workspaces: rust
-      - name: Install 7-Zip ARM
-        run: |
-          New-Item -Path 'C:\7zip' -ItemType Directory
-          Invoke-WebRequest https://7-zip.org/a/7z2408-arm64.exe -OutFile C:\7zip\7z-installer.exe
-          Start-Process -FilePath C:\7zip\7z-installer.exe -ArgumentList '/S' -Wait
-        shell: powershell
-      - name: Add 7-Zip to PATH
-        run: Add-Content $env:GITHUB_PATH "C:\Program Files\7-Zip"
-        shell: powershell
-      - name: Install Protoc v21.12
-        working-directory: C:\
-        run: |
-          if (Test-Path 'C:\protoc') {
-            Write-Host "Protoc directory exists, skipping installation"
-            return
-          }
-          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
-          & 'C:\Program Files\7-Zip\7z.exe' x protoc.zip
-        shell: powershell
-      - name: Add Protoc to PATH
-        run: Add-Content $env:GITHUB_PATH "C:\protoc\bin"
-        shell: powershell
-      - name: Run tests
-        run: |
-          $env:VCPKG_ROOT = $env:VCPKG_INSTALLATION_ROOT
-          cargo test --target aarch64-pc-windows-msvc --features remote --locked
-
   msrv:
     # Check the minimum supported Rust version
     name: MSRV Check - Rust v${{ matrix.msrv }}

Cargo.toml

@@ -21,14 +21,16 @@ categories = ["database-implementations"]
 rust-version = "1.78.0"
 
 [workspace.dependencies]
-lance = { "version" = "=0.24.1", "features" = ["dynamodb"] }
-lance-io = { version = "=0.24.1" }
-lance-index = { version = "=0.24.1" }
-lance-linalg = { version = "=0.24.1" }
-lance-table = { version = "=0.24.1" }
-lance-testing = { version = "=0.24.1" }
-lance-datafusion = { version = "=0.24.1" }
-lance-encoding = { version = "=0.24.1" }
+lance = { "version" = "=0.25.1", "features" = [
+    "dynamodb",
+], tag = "v0.25.1-beta.3", git = "https://github.com/lancedb/lance.git" }
+lance-io = { version = "=0.25.1", tag = "v0.25.1-beta.3", git = "https://github.com/lancedb/lance.git" }
+lance-index = { version = "=0.25.1", tag = "v0.25.1-beta.3", git = "https://github.com/lancedb/lance.git" }
+lance-linalg = { version = "=0.25.1", tag = "v0.25.1-beta.3", git = "https://github.com/lancedb/lance.git" }
+lance-table = { version = "=0.25.1", tag = "v0.25.1-beta.3", git = "https://github.com/lancedb/lance.git" }
+lance-testing = { version = "=0.25.1", tag = "v0.25.1-beta.3", git = "https://github.com/lancedb/lance.git" }
+lance-datafusion = { version = "=0.25.1", tag = "v0.25.1-beta.3", git = "https://github.com/lancedb/lance.git" }
+lance-encoding = { version = "=0.25.1", tag = "v0.25.1-beta.3", git = "https://github.com/lancedb/lance.git" }
 # Note that this one does not include pyarrow
 arrow = { version = "54.1", optional = false }
 arrow-array = "54.1"
@@ -70,3 +72,6 @@ base64ct = "=1.6.0"
 
 # Workaround for: https://github.com/eira-fransham/crunchy/issues/13
 crunchy = "=0.2.2"
+
+# Workaround for: https://github.com/Lokathor/bytemuck/issues/306
+bytemuck_derive = ">=1.8.1, <1.9.0"

README.md

@@ -1,9 +1,17 @@
+<a href="https://cloud.lancedb.com" target="_blank">
+  <img src="https://github.com/user-attachments/assets/92dad0a2-2a37-4ce1-b783-0d1b4f30a00c" alt="LanceDB Cloud Public Beta" width="100%" style="max-width: 100%;">
+</a>
+
 <div align="center">
 <p align="center">
 
-<img width="275" alt="LanceDB Logo" src="https://github.com/lancedb/lancedb/assets/5846846/37d7c7ad-c2fd-4f56-9f16-fffb0d17c73a">
+<picture>
+  <source media="(prefers-color-scheme: dark)" srcset="https://github.com/user-attachments/assets/ac270358-333e-4bea-a132-acefaa94040e">
+  <source media="(prefers-color-scheme: light)" srcset="https://github.com/user-attachments/assets/b864d814-0d29-4784-8fd9-807297c758c0">
+  <img alt="LanceDB Logo" src="https://github.com/user-attachments/assets/b864d814-0d29-4784-8fd9-807297c758c0" width=300>
+</picture>
 
-**Developer-friendly, database for multimodal AI**
+**Search More, Manage Less**
 
 <a href='https://github.com/lancedb/vectordb-recipes/tree/main' target="_blank"><img alt='LanceDB' src='https://img.shields.io/badge/VectorDB_Recipes-100000?style=for-the-badge&logo=LanceDB&logoColor=white&labelColor=645cfb&color=645cfb'/></a>
 <a href='https://lancedb.github.io/lancedb/' target="_blank"><img alt='lancdb' src='https://img.shields.io/badge/DOCS-100000?style=for-the-badge&logo=lancdb&logoColor=white&labelColor=645cfb&color=645cfb'/></a>

ci/build_linux_artifacts_nodejs.sh

@@ -1,21 +0,0 @@
-#!/bin/bash
-set -e
-ARCH=${1:-x86_64}
-
-# We pass down the current user so that when we later mount the local files
-# into the container, the files are accessible by the current user.
-pushd ci/manylinux_node
-docker build \
-    -t lancedb-node-manylinux-$ARCH \
-    --build-arg="ARCH=$ARCH" \
-    --build-arg="DOCKER_USER=$(id -u)" \
-    --progress=plain \
-    .
-popd
-
-# We turn on memory swap to avoid OOM killer
-docker run \
-    -v $(pwd):/io -w /io \
-    --memory-swap=-1 \
-    lancedb-node-manylinux-$ARCH \
-    bash ci/manylinux_node/build_lancedb.sh $ARCH

ci/build_macos_artifacts_nodejs.sh

@@ -1,34 +0,0 @@
-# Builds the macOS artifacts (nodejs binaries).
-# Usage: ./ci/build_macos_artifacts_nodejs.sh [target]
-# Targets supported: x86_64-apple-darwin aarch64-apple-darwin
-set -e
-
-prebuild_rust() {
-    # Building here for the sake of easier debugging.
-    pushd rust/lancedb
-    echo "Building rust library for $1"
-    export RUST_BACKTRACE=1
-    cargo build --release --target $1
-    popd
-}
-
-build_node_binaries() {
-    pushd nodejs
-    echo "Building nodejs library for $1"
-    export RUST_TARGET=$1
-    npm run build-release
-    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/manylinux_node/Dockerfile

@@ -1,5 +1,5 @@
 # Many linux dockerfile with Rust, Node, and Lance dependencies installed.
-# This container allows building the node modules native libraries in an 
+# This container allows building the node modules native libraries in an
 # environment with a very old glibc, so that we are compatible with a wide
 # range of linux distributions.
 ARG ARCH=x86_64
@@ -9,10 +9,6 @@ FROM quay.io/pypa/manylinux_2_28_${ARCH}
 ARG ARCH=x86_64
 ARG DOCKER_USER=default_user
 
-# Install static openssl
-COPY install_openssl.sh install_openssl.sh
-RUN ./install_openssl.sh ${ARCH} > /dev/null
-
 # Protobuf is also installed as root.
 COPY install_protobuf.sh install_protobuf.sh
 RUN ./install_protobuf.sh ${ARCH}
@@ -21,7 +17,7 @@ ENV DOCKER_USER=${DOCKER_USER}
 # Create a group and user, but only if it doesn't exist
 RUN echo ${ARCH} && id -u ${DOCKER_USER} >/dev/null 2>&1 || adduser --user-group --create-home --uid ${DOCKER_USER} build_user
 
-# We switch to the user to install Rust and Node, since those like to be 
+# We switch to the user to install Rust and Node, since those like to be
 # installed at the user level.
 USER ${DOCKER_USER}
 

ci/manylinux_node/build_lancedb.sh

@@ -1,19 +0,0 @@
-#!/bin/bash
-# Builds the nodejs module for manylinux. Invoked by ci/build_linux_artifacts_nodejs.sh.
-set -e
-ARCH=${1:-x86_64}
-
-if [ "$ARCH" = "x86_64" ]; then
-    export OPENSSL_LIB_DIR=/usr/local/lib64/
-else
-    export OPENSSL_LIB_DIR=/usr/local/lib/
-fi
-export OPENSSL_STATIC=1
-export OPENSSL_INCLUDE_DIR=/usr/local/include/openssl
-
-#Alpine doesn't have .bashrc
-FILE=$HOME/.bashrc && test -f $FILE && source $FILE
-
-cd nodejs
-npm ci
-npm run build-release

ci/manylinux_node/build_vectordb.sh

@@ -4,14 +4,6 @@ set -e
 ARCH=${1:-x86_64}
 TARGET_TRIPLE=${2:-x86_64-unknown-linux-gnu}
 
-if [ "$ARCH" = "x86_64" ]; then
-    export OPENSSL_LIB_DIR=/usr/local/lib64/
-else
-    export OPENSSL_LIB_DIR=/usr/local/lib/
-fi
-export OPENSSL_STATIC=1
-export OPENSSL_INCLUDE_DIR=/usr/local/include/openssl
-
 #Alpine doesn't have .bashrc
 FILE=$HOME/.bashrc && test -f $FILE && source $FILE
 

ci/manylinux_node/install_openssl.sh

@@ -1,26 +0,0 @@
-#!/bin/bash
-# Builds openssl from source so we can statically link to it
-
-# this is to avoid the error we get with the system installation:
-# /usr/bin/ld: <library>: version node not found for symbol SSLeay@@OPENSSL_1.0.1
-# /usr/bin/ld: failed to set dynamic section sizes: Bad value
-set -e
-
-git clone -b OpenSSL_1_1_1v \
-    --single-branch \
-    https://github.com/openssl/openssl.git
-
-pushd openssl
-
-if [[ $1 == x86_64* ]]; then
-    ARCH=linux-x86_64
-else
-    # gnu target
-    ARCH=linux-aarch64
-fi
-
-./Configure no-shared $ARCH
-
-make
-
-make install

docs/mkdocs.yml

@@ -124,6 +124,9 @@ nav:
               - Overview: hybrid_search/hybrid_search.md
               - Comparing Rerankers: hybrid_search/eval.md
               - Airbnb financial data example: notebooks/hybrid_search.ipynb
+          - Late interaction with MultiVector search:
+              - Overview: guides/multi-vector.md
+              - Example: notebooks/Multivector_on_LanceDB.ipynb
           - RAG:
               - Vanilla RAG: rag/vanilla_rag.md
               - Multi-head RAG: rag/multi_head_rag.md
@@ -233,13 +236,6 @@ nav:
           - 👾 JavaScript (vectordb): javascript/modules.md
           - 👾 JavaScript (lancedb): js/globals.md
           - 🦀 Rust: https://docs.rs/lancedb/latest/lancedb/
-      - ☁️ LanceDB Cloud:
-          - Overview: cloud/index.md
-          - API reference:
-              - 🐍 Python: python/saas-python.md
-              - 👾 JavaScript: javascript/modules.md
-              - REST API: cloud/rest.md
-          - FAQs: cloud/cloud_faq.md
 
   - Quick start: basic.md
   - Concepts:
@@ -260,6 +256,9 @@ nav:
           - Overview: hybrid_search/hybrid_search.md
           - Comparing Rerankers: hybrid_search/eval.md
           - Airbnb financial data example: notebooks/hybrid_search.ipynb
+      - Late interaction with MultiVector search:
+          - Overview: guides/multi-vector.md
+          - Document search Example: notebooks/Multivector_on_LanceDB.ipynb
       - RAG:
           - Vanilla RAG: rag/vanilla_rag.md
           - Multi-head RAG: rag/multi_head_rag.md
@@ -363,13 +362,6 @@ nav:
       - Javascript (vectordb): javascript/modules.md
       - Javascript (lancedb): js/globals.md
       - Rust: https://docs.rs/lancedb/latest/lancedb/index.html
-  - LanceDB Cloud:
-      - Overview: cloud/index.md
-      - API reference:
-          - 🐍 Python: python/saas-python.md
-          - 👾 JavaScript: javascript/modules.md
-          - REST API: cloud/rest.md
-      - FAQs: cloud/cloud_faq.md
 
 extra_css:
   - styles/global.css

docs/openapi.yml

@@ -171,7 +171,7 @@ paths:
                 distance_type:
                   type: string
                   description: |
-                    The distance metric to use for search. L2, Cosine, Dot and Hamming are supported. Default is L2.
+                    The distance metric to use for search. l2, Cosine, Dot and Hamming are supported. Default is l2.
                 bypass_vector_index:
                   type: boolean
                   description: |
@@ -450,7 +450,7 @@ paths:
                   type: string
                   nullable: false
                   description: |
-                    The metric type to use for the index. L2, Cosine, Dot are supported.
+                    The metric type to use for the index. l2, Cosine, Dot are supported.
                 index_type:
                   type: string
       responses:

docs/src/ann_indexes.md

@@ -69,7 +69,7 @@ Lance supports `IVF_PQ` index type by default.
 
 The following IVF_PQ paramters can be specified:
 
-- **distance_type**: The distance metric to use. By default it uses euclidean distance "`L2`".
+- **distance_type**: The distance metric to use. By default it uses euclidean distance "`l2`".
   We also support "cosine" and "dot" distance as well.
 - **num_partitions**: The number of partitions in the index. The default is the square root
   of the number of rows.

docs/src/cloud/index.md

@@ -2,7 +2,7 @@
 
 LanceDB Cloud is a SaaS (software-as-a-service) solution that runs serverless in the cloud, clearly separating storage from compute. It's designed to be highly scalable without breaking the bank. LanceDB Cloud is currently in private beta with general availability coming soon, but you can apply for early access with the private beta release by signing up below.
 
-[Try out LanceDB Cloud](https://noteforms.com/forms/lancedb-mailing-list-cloud-kty1o5?notionforms=1&utm_source=notionforms){ .md-button .md-button--primary }
+[Try out LanceDB Cloud (Public Beta)](https://cloud.lancedb.com){ .md-button .md-button--primary }
 
 ## Architecture
 

docs/src/concepts/index_hnsw.md

@@ -59,7 +59,7 @@ Then the greedy search routine operates as follows:
 
 There are three key parameters to set when constructing an HNSW index:
 
-* `metric`: Use an `L2` euclidean distance metric. We also support `dot` and `cosine` distance.
+* `metric`: Use an `l2` euclidean distance metric. We also support `dot` and `cosine` distance.
 * `m`: The number of neighbors to select for each vector in the HNSW graph.
 * `ef_construction`: The number of candidates to evaluate during the construction of the HNSW graph.
 

docs/src/concepts/index_ivfpq.md

@@ -47,7 +47,7 @@ We can combine the above concepts to understand how to build and query an IVF-PQ
 
 There are three key parameters to set when constructing an IVF-PQ index:
 
-* `metric`: Use an `L2` euclidean distance metric. We also support `dot` and `cosine` distance.
+* `metric`: Use an `l2` euclidean distance metric. We also support `dot` and `cosine` distance.
 * `num_partitions`: The number of partitions in the IVF portion of the index.
 * `num_sub_vectors`: The number of sub-vectors that will be created during Product Quantization (PQ).
 
@@ -56,7 +56,7 @@ In Python, the index can be created as follows:
 ```python
 # Create and train the index for a 1536-dimensional vector
 # Make sure you have enough data in the table for an effective training step
-tbl.create_index(metric="L2", num_partitions=256, num_sub_vectors=96)
+tbl.create_index(metric="l2", num_partitions=256, num_sub_vectors=96)
 ```
 !!! note
     `num_partitions`=256 and `num_sub_vectors`=96 does not work for every dataset. Those values needs to be adjusted for your particular dataset.

docs/src/embeddings/understanding_embeddings.md

@@ -54,7 +54,7 @@ As mentioned, after creating embedding, each data point is represented as a vect
 
 Points that are close to each other in vector space are considered similar (or appear in similar contexts), and points that are far away are considered dissimilar. To quantify this closeness, we use distance as a metric which can be measured in the  following way - 
 
-1. **Euclidean Distance (L2)**: It calculates the straight-line distance between two points (vectors) in a multidimensional space.
+1. **Euclidean Distance (l2)**: It calculates the straight-line distance between two points (vectors) in a multidimensional space.
 2. **Cosine Similarity**: It measures the cosine of the angle between two vectors, providing a normalized measure of similarity based on their direction.
 3. **Dot product**: It is calculated as the sum of the products of their corresponding components. To measure relatedness it considers both the magnitude and direction of the vectors.
 

docs/src/examples/index.md

@@ -8,15 +8,5 @@ LanceDB provides language APIs, allowing you to embed a database in your languag
 * 👾 [JavaScript](examples_js.md) examples
 * 🦀 Rust examples (coming soon)
 
-## Python Applications powered by LanceDB
-
-| Project Name | Description |
-| --- | --- |
-| **Ultralytics Explorer 🚀**<br>[![Ultralytics](https://img.shields.io/badge/Ultralytics-Docs-green?labelColor=0f3bc4&style=flat-square&logo=https://cdn.prod.website-files.com/646dd1f1a3703e451ba81ecc/64994922cf2a6385a4bf4489_UltralyticsYOLO_mark_blue.svg&link=https://docs.ultralytics.com/datasets/explorer/)](https://docs.ultralytics.com/datasets/explorer/)<br>[![Open In Collab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/ultralytics/ultralytics/blob/main/docs/en/datasets/explorer/explorer.ipynb) |  - 🔍 **Explore CV Datasets**: Semantic search, SQL queries, vector similarity, natural language.<br>- 🖥️ **GUI & Python API**: Seamless dataset interaction.<br>- ⚡ **Efficient & Scalable**: Leverages LanceDB for large datasets.<br>- 📊 **Detailed Analysis**: Easily analyze data patterns.<br>- 🌐 **Browser GUI Demo**: Create embeddings, search images, run queries. |
-| **Website Chatbot🤖**<br>[![GitHub](https://img.shields.io/badge/github-%23121011.svg?style=for-the-badge&logo=github&logoColor=white)](https://github.com/lancedb/lancedb-vercel-chatbot)<br>[![Deploy with Vercel](https://vercel.com/button)](https://vercel.com/new/clone?repository-url=https%3A%2F%2Fgithub.com%2Flancedb%2Flancedb-vercel-chatbot&amp;env=OPENAI_API_KEY&amp;envDescription=OpenAI%20API%20Key%20for%20chat%20completion.&amp;project-name=lancedb-vercel-chatbot&amp;repository-name=lancedb-vercel-chatbot&amp;demo-title=LanceDB%20Chatbot%20Demo&amp;demo-description=Demo%20website%20chatbot%20with%20LanceDB.&amp;demo-url=https%3A%2F%2Flancedb.vercel.app&amp;demo-image=https%3A%2F%2Fi.imgur.com%2FazVJtvr.png) | - 🌐 **Chatbot from Sitemap/Docs**: Create a chatbot using site or document context.<br>- 🚀 **Embed LanceDB in Next.js**: Lightweight, on-prem storage.<br>- 🧠 **AI-Powered Context Retrieval**: Efficiently access relevant data.<br>- 🔧 **Serverless & Native JS**: Seamless integration with Next.js.<br>- ⚡ **One-Click Deploy on Vercel**: Quick and easy setup.. |
-
-## Nodejs Applications powered by LanceDB
-
-| Project Name | Description |
-| --- | --- |
-| **Langchain Writing Assistant✍️ **<br>[![Github](../assets/github.svg)](https://github.com/lancedb/vectordb-recipes/tree/main/applications/node/lanchain_writing_assistant) |  - **📂 Data Source Integration**:  Use your own data by specifying data source file, and the app instantly processes it to provide insights. <br>- **🧠 Intelligent Suggestions**:  Powered by LangChain.js and LanceDB, it improves writing productivity and accuracy.  <br>- **💡 Enhanced Writing Experience**: It delivers real-time contextual insights and factual suggestions while the user writes. |
+!!! tip "Hosted LanceDB"
+    If you want S3 cost-efficiency and local performance via a simple serverless API, checkout **LanceDB Cloud**. For private deployments, high performance at extreme scale, or if you have strict security requirements, talk to us about **LanceDB Enterprise**. [Learn more](https://docs.lancedb.com/)

docs/src/guides/multi-vector.md

@@ -0,0 +1,85 @@
+# Late interaction & MultiVector embedding type
+Late interaction is a technique used in retrieval that calculates the relevance of a query to a document by comparing their multi-vector representations. The key difference between late interaction and other popular methods:
+
+![late interaction vs other methods](https://raw.githubusercontent.com/lancedb/assets/b035a0ceb2c237734e0d393054c146d289792339/docs/assets/integration/colbert-blog-interaction.svg)
+
+
+[ Illustration from https://jina.ai/news/what-is-colbert-and-late-interaction-and-why-they-matter-in-search/]
+
+<b>No interaction:</b> Refers to independently embedding the query and document, that are compared to calcualte similarity without any interaction between them. This is typically used in vector search operations.
+
+<b>Partial interaction</b> Refers to a specific approach where the similarity computation happens primarily between query vectors and document vectors, without extensive interaction between individual components of each. An example of this is dual-encoder models like BERT.
+
+<b>Early full interaction</b> Refers to techniques like cross-encoders that process query and docs in pairs with full interaction across various stages of encoding. This is a powerful, but relatively slower technique. Because it requires processing query and docs in pairs, doc embeddings can't be pre-computed for fast retrieval. This is why cross encoders are typically used as reranking models combined with vector search. Learn more about [LanceDB Reranking support](https://lancedb.github.io/lancedb/reranking/).
+
+<b>Late interaction</b> Late interaction is a technique that calculates the doc and query similarity independently and then the interaction or evaluation happens during the retrieval process. This is typically used in retrieval models like ColBERT. Unlike early interaction, It allows speeding up the retrieval process without compromising the depth of semantic analysis.
+
+## Internals of ColBERT 
+Let's take a look at the steps involved in performing late interaction based retrieval using ColBERT:
+
+• ColBERT employs BERT-based encoders for both queries `(fQ)` and documents `(fD)`
+• A single BERT model is shared between query and document encoders and special tokens distinguish input types: `[Q]` for queries and `[D]` for documents
+
+**Query Encoder (fQ):**
+• Query q is tokenized into WordPiece tokens: `q1, q2, ..., ql`. `[Q]` token is prepended right after BERT's `[CLS]` token
+• If query length < Nq, it's padded with [MASK] tokens up to Nq.
+• The padded sequence goes through BERT's transformer architecture
+• Final embeddings are L2-normalized.
+
+**Document Encoder (fD):**
+• Document d is tokenized into tokens `d1, d2, ..., dm`. `[D]` token is prepended after `[CLS]` token
+• Unlike queries, documents are NOT padded with `[MASK]` tokens
+• Document tokens are processed through BERT and the same linear layer
+
+**Late Interaction:**
+• Late interaction estimates relevance score `S(q,d)` using embedding `Eq` and `Ed`. Late interaction happens after independent encoding
+• For each query embedding, maximum similarity is computed against all document embeddings
+• The similarity measure can be cosine similarity or squared L2 distance
+
+**MaxSim Calculation:**
+```
+S(q,d) := Σ max(Eqi⋅EdjT)
+          i∈|Eq| j∈|Ed|
+```
+• This finds the best matching document embedding for each query embedding
+• Captures relevance based on strongest local matches between contextual embeddings
+
+## LanceDB MultiVector type
+LanceDB supports multivector type, this is useful when you have multiple vectors for a single item (e.g. with ColBert and ColPali).
+
+You can index on a column with multivector type and search on it, the query can be single vector or multiple vectors. For now, only cosine metric is supported for multivector search. The vector value type can be float16, float32 or float64. LanceDB integrateds [ConteXtualized Token Retriever(XTR)](https://arxiv.org/abs/2304.01982), which introduces a simple, yet novel, objective function that encourages the model to retrieve the most important document tokens first. 
+
+```python
+import lancedb
+import numpy as np
+import pyarrow as pa
+
+db = lancedb.connect("data/multivector_demo")
+schema = pa.schema(
+    [
+        pa.field("id", pa.int64()),
+        # float16, float32, and float64 are supported
+        pa.field("vector", pa.list_(pa.list_(pa.float32(), 256))),
+    ]
+)
+data = [
+    {
+        "id": i,
+        "vector": np.random.random(size=(2, 256)).tolist(),
+    }
+    for i in range(1024)
+]
+tbl = db.create_table("my_table", data=data, schema=schema)
+
+# only cosine similarity is supported for multi-vectors
+tbl.create_index(metric="cosine")
+
+# query with single vector
+query = np.random.random(256).astype(np.float16)
+tbl.search(query).to_arrow()
+
+# query with multiple vectors
+query = np.random.random(size=(2, 256))
+tbl.search(query).to_arrow()
+```
+Find more about vector search in LanceDB [here](https://lancedb.github.io/lancedb/search/#multivector-type).

docs/src/guides/tables.md

@@ -1001,9 +1001,11 @@ In LanceDB OSS, users can set the `read_consistency_interval` parameter on conne
 
 There are three possible settings for `read_consistency_interval`:
 
-1. **Unset (default)**: The database does not check for updates to tables made by other processes. This provides the best query performance, but means that clients may not see the most up-to-date data. This setting is suitable for applications where the data does not change during the lifetime of the table reference.
-2. **Zero seconds (Strong consistency)**: The database checks for updates on every read. This provides the strongest consistency guarantees, ensuring that all clients see the latest committed data. However, it has the most overhead. This setting is suitable when consistency matters more than having high QPS.
-3. **Custom interval (Eventual consistency)**: The database checks for updates at a custom interval, such as every 5 seconds. This provides eventual consistency, allowing for some lag between write and read operations. Performance wise, this is a middle ground between strong consistency and no consistency check. This setting is suitable for applications where immediate consistency is not critical, but clients should see updated data eventually.
+1. **Unset**: The database does not check for updates to tables made by other processes. This setting is suitable for applications where the data does not change during the lifetime of the table reference.
+2. **Zero seconds (Strong consistency)**: The database checks for updates on every read. This provides the strongest consistency guarantees, ensuring that all clients see the latest committed data. However, it has the most overhead. This setting is suitable when consistency matters more than having high QPS. For best performance, combine this setting with the storage option `new_table_enable_v2_manifest_paths` set to `true`.
+3. **Custom interval (Eventual consistency, the default)**: The database checks for updates at a custom interval. By default, this is every 5 seconds. This provides eventual consistency, allowing for some lag between write and read operations. Performance wise, this is a middle ground between strong consistency and no consistency check. This setting is suitable for applications where immediate consistency is not critical, but clients should see updated data eventually.
+
+You can always force a synchronization by calling `checkout_latest()` / `checkoutLatest()` on a table.
 
 !!! tip "Consistency in LanceDB Cloud"
 
@@ -1041,7 +1043,21 @@ There are three possible settings for `read_consistency_interval`:
         --8<-- "python/python/tests/docs/test_guide_tables.py:table_async_eventual_consistency"
         ```
 
-    By default, a `Table` will never check for updates from other writers. To manually check for updates you can use `checkout_latest`:
+    For no consistency, use `None`:
+
+    === "Sync API"
+
+        ```python
+        --8<-- "python/python/tests/docs/test_guide_tables.py:table_no_consistency"
+        ```
+
+    === "Async API"
+
+        ```python
+        --8<-- "python/python/tests/docs/test_guide_tables.py:table_async_no_consistency"
+        ```
+
+    To manually check for updates you can use `checkout_latest`:
 
     === "Sync API"
 
@@ -1059,15 +1075,25 @@ There are three possible settings for `read_consistency_interval`:
     To set strong consistency, use `0`:
 
     ```ts
-    const db = await lancedb.connect({ uri: "./.lancedb", readConsistencyInterval: 0 });
-    const tbl = await db.openTable("my_table");
+    --8<-- "nodejs/examples/basic.test.ts:table_strong_consistency"
     ```
 
     For eventual consistency, specify the update interval as seconds:
 
     ```ts
-    const db = await lancedb.connect({ uri: "./.lancedb", readConsistencyInterval: 5 });
-    const tbl = await db.openTable("my_table");
+    --8<-- "nodejs/examples/basic.test.ts:table_eventual_consistency"
+    ```
+
+    For no consistency, use `null`:
+
+    ```ts
+    --8<-- "nodejs/examples/basic.test.ts:table_no_consistency"
+    ```
+
+    To manually check for updates you can use `checkoutLatest`:
+
+    ```ts
+    --8<-- "nodejs/examples/basic.test.ts:table_checkout_latest"
     ```
 
 <!-- Node doesn't yet support the version time travel: https://github.com/lancedb/lancedb/issues/1007

docs/src/index.md

@@ -4,6 +4,9 @@ LanceDB is an open-source vector database for AI that's designed to store, manag
 
 Both the database and the underlying data format are designed from the ground up to be **easy-to-use**, **scalable** and **cost-effective**.
 
+!!! tip "Hosted LanceDB"
+    If you want S3 cost-efficiency and local performance via a simple serverless API, checkout **LanceDB Cloud**. For private deployments, high performance at extreme scale, or if you have strict security requirements, talk to us about **LanceDB Enterprise**. [Learn more](https://docs.lancedb.com/)
+
 ![](assets/lancedb_and_lance.png)
 
 ## Truly multi-modal
@@ -20,7 +23,7 @@ LanceDB **OSS** is an **open-source**, batteries-included embedded vector databa
 
 LanceDB **Cloud** is a SaaS (software-as-a-service) solution that runs serverless in the cloud, making the storage clearly separated from compute. It's designed to be cost-effective and highly scalable without breaking the bank. LanceDB Cloud is currently in private beta with general availability coming soon, but you can apply for early access with the private beta release by signing up below.
 
-[Try out LanceDB Cloud](https://noteforms.com/forms/lancedb-mailing-list-cloud-kty1o5?notionforms=1&utm_source=notionforms){ .md-button .md-button--primary }
+[Try out LanceDB Cloud (Public Beta) Now](https://cloud.lancedb.com){ .md-button .md-button--primary }
 
 ## Why use LanceDB?
 

docs/src/integrations/langchain.md

@@ -108,7 +108,7 @@ This method creates a scalar(for non-vector cols) or a vector index on a table.
 |:---|:---|:---|:---|
 |`vector_col`|`Optional[str]`| Provide if you want to create index on a vector column. |`None`|
 |`col_name`|`Optional[str]`| Provide if you want to create index on a non-vector column. |`None`|
-|`metric`|`Optional[str]` |Provide the metric to use for vector index. choice of metrics: 'L2', 'dot', 'cosine'. |`L2`|
+|`metric`|`Optional[str]` |Provide the metric to use for vector index. choice of metrics: 'l2', 'dot', 'cosine'. |`l2`|
 |`num_partitions`|`Optional[int]`|Number of partitions to use for the index.|`256`|
 |`num_sub_vectors`|`Optional[int]` |Number of sub-vectors to use for the index.|`96`|
 |`index_cache_size`|`Optional[int]` |Size of the index cache.|`None`|

docs/src/integrations/llamaIndex.md

@@ -125,7 +125,7 @@ The exhaustive list of parameters for `LanceDBVectorStore` vector store are :
         ```               
 - **_table_exists(self, tbl_name: `Optional[str]` = `None`) -> `bool`** : Returns `True` if `tbl_name` exists in database.
 - __create_index(  
-  self, scalar: `Optional[bool]` = False, col_name: `Optional[str]` = None, num_partitions: `Optional[int]` = 256, num_sub_vectors: `Optional[int]` = 96, index_cache_size: `Optional[int]` = None, metric: `Optional[str]` = "L2",  
+  self, scalar: `Optional[bool]` = False, col_name: `Optional[str]` = None, num_partitions: `Optional[int]` = 256, num_sub_vectors: `Optional[int]` = 96, index_cache_size: `Optional[int]` = None, metric: `Optional[str]` = "l2",  
 ) -> `None`__ : Creates a scalar(for non-vector cols) or a vector index on a table.
         Make sure your vector column has enough data before creating an index on it.
 

docs/src/javascript/enums/MetricType.md

@@ -10,7 +10,7 @@ Distance metrics type.
 
 - [Cosine](MetricType.md#cosine)
 - [Dot](MetricType.md#dot)
-- [L2](MetricType.md#l2)
+- [l2](MetricType.md#l2)
 
 ## Enumeration Members
 

docs/src/javascript/interfaces/IvfPQIndexConfig.md

@@ -85,7 +85,7 @@ ___
 
 • `Optional` **metric\_type**: [`MetricType`](../enums/MetricType.md)
 
-Metric type, L2 or Cosine
+Metric type, l2 or Cosine
 
 #### Defined in
 

docs/src/js/README.md

@@ -15,11 +15,9 @@ npm install @lancedb/lancedb
 This will download the appropriate native library for your platform. We currently
 support:
 
-- Linux (x86_64 and aarch64)
+- Linux (x86_64 and aarch64 on glibc and musl)
 - MacOS (Intel and ARM/M1/M2)
-- Windows (x86_64 only)
-
-We do not yet support musl-based Linux (such as Alpine Linux) or aarch64 Windows.
+- Windows (x86_64 and aarch64)
 
 ## Usage
 

docs/src/js/classes/Index.md

@@ -126,6 +126,37 @@ the vectors.
 
 ***
 
+### ivfFlat()
+
+```ts
+static ivfFlat(options?): Index
+```
+
+Create an IvfFlat index
+
+This index groups vectors into partitions of similar vectors.  Each partition keeps track of
+a centroid which is the average value of all vectors in the group.
+
+During a query the centroids are compared with the query vector to find the closest
+partitions.  The vectors in these partitions are then searched to find
+the closest vectors.
+
+The partitioning process is called IVF and the `num_partitions` parameter controls how
+many groups to create.
+
+Note that training an IVF FLAT index on a large dataset is a slow operation and
+currently is also a memory intensive operation.
+
+#### Parameters
+
+* **options?**: `Partial`<[`IvfFlatOptions`](../interfaces/IvfFlatOptions.md)>
+
+#### Returns
+
+[`Index`](Index.md)
+
+***
+
 ### ivfPq()
 
 ```ts

docs/src/js/classes/Query.md

@@ -30,6 +30,53 @@ protected inner: Query | Promise<Query>;
 
 ## Methods
 
+### analyzePlan()
+
+```ts
+analyzePlan(): Promise<string>
+```
+
+Executes the query and returns the physical query plan annotated with runtime metrics.
+
+This is useful for debugging and performance analysis, as it shows how the query was executed
+and includes metrics such as elapsed time, rows processed, and I/O statistics.
+
+#### Returns
+
+`Promise`&lt;`string`&gt;
+
+A query execution plan with runtime metrics for each step.
+
+#### Example
+
+```ts
+import * as lancedb from "@lancedb/lancedb"
+
+const db = await lancedb.connect("./.lancedb");
+const table = await db.createTable("my_table", [
+  { vector: [1.1, 0.9], id: "1" },
+]);
+
+const plan = await table.query().nearestTo([0.5, 0.2]).analyzePlan();
+
+Example output (with runtime metrics inlined):
+AnalyzeExec verbose=true, metrics=[]
+ ProjectionExec: expr=[id@3 as id, vector@0 as vector, _distance@2 as _distance], metrics=[output_rows=1, elapsed_compute=3.292µs]
+  Take: columns="vector, _rowid, _distance, (id)", metrics=[output_rows=1, elapsed_compute=66.001µs, batches_processed=1, bytes_read=8, iops=1, requests=1]
+   CoalesceBatchesExec: target_batch_size=1024, metrics=[output_rows=1, elapsed_compute=3.333µs]
+    GlobalLimitExec: skip=0, fetch=10, metrics=[output_rows=1, elapsed_compute=167ns]
+     FilterExec: _distance@2 IS NOT NULL, metrics=[output_rows=1, elapsed_compute=8.542µs]
+      SortExec: TopK(fetch=10), expr=[_distance@2 ASC NULLS LAST], metrics=[output_rows=1, elapsed_compute=63.25µs, row_replacements=1]
+       KNNVectorDistance: metric=l2, metrics=[output_rows=1, elapsed_compute=114.333µs, output_batches=1]
+        LanceScan: uri=/path/to/data, projection=[vector], row_id=true, row_addr=false, ordered=false, metrics=[output_rows=1, elapsed_compute=103.626µs, bytes_read=549, iops=2, requests=2]
+```
+
+#### Inherited from
+
+[`QueryBase`](QueryBase.md).[`analyzePlan`](QueryBase.md#analyzeplan)
+
+***
+
 ### execute()
 
 ```ts

docs/src/js/classes/QueryBase.md

@@ -36,6 +36,49 @@ protected inner: NativeQueryType | Promise<NativeQueryType>;
 
 ## Methods
 
+### analyzePlan()
+
+```ts
+analyzePlan(): Promise<string>
+```
+
+Executes the query and returns the physical query plan annotated with runtime metrics.
+
+This is useful for debugging and performance analysis, as it shows how the query was executed
+and includes metrics such as elapsed time, rows processed, and I/O statistics.
+
+#### Returns
+
+`Promise`&lt;`string`&gt;
+
+A query execution plan with runtime metrics for each step.
+
+#### Example
+
+```ts
+import * as lancedb from "@lancedb/lancedb"
+
+const db = await lancedb.connect("./.lancedb");
+const table = await db.createTable("my_table", [
+  { vector: [1.1, 0.9], id: "1" },
+]);
+
+const plan = await table.query().nearestTo([0.5, 0.2]).analyzePlan();
+
+Example output (with runtime metrics inlined):
+AnalyzeExec verbose=true, metrics=[]
+ ProjectionExec: expr=[id@3 as id, vector@0 as vector, _distance@2 as _distance], metrics=[output_rows=1, elapsed_compute=3.292µs]
+  Take: columns="vector, _rowid, _distance, (id)", metrics=[output_rows=1, elapsed_compute=66.001µs, batches_processed=1, bytes_read=8, iops=1, requests=1]
+   CoalesceBatchesExec: target_batch_size=1024, metrics=[output_rows=1, elapsed_compute=3.333µs]
+    GlobalLimitExec: skip=0, fetch=10, metrics=[output_rows=1, elapsed_compute=167ns]
+     FilterExec: _distance@2 IS NOT NULL, metrics=[output_rows=1, elapsed_compute=8.542µs]
+      SortExec: TopK(fetch=10), expr=[_distance@2 ASC NULLS LAST], metrics=[output_rows=1, elapsed_compute=63.25µs, row_replacements=1]
+       KNNVectorDistance: metric=l2, metrics=[output_rows=1, elapsed_compute=114.333µs, output_batches=1]
+        LanceScan: uri=/path/to/data, projection=[vector], row_id=true, row_addr=false, ordered=false, metrics=[output_rows=1, elapsed_compute=103.626µs, bytes_read=549, iops=2, requests=2]
+```
+
+***
+
 ### execute()
 
 ```ts

docs/src/js/classes/VectorQuery.md

@@ -48,6 +48,53 @@ addQueryVector(vector): VectorQuery
 
 ***
 
+### analyzePlan()
+
+```ts
+analyzePlan(): Promise<string>
+```
+
+Executes the query and returns the physical query plan annotated with runtime metrics.
+
+This is useful for debugging and performance analysis, as it shows how the query was executed
+and includes metrics such as elapsed time, rows processed, and I/O statistics.
+
+#### Returns
+
+`Promise`&lt;`string`&gt;
+
+A query execution plan with runtime metrics for each step.
+
+#### Example
+
+```ts
+import * as lancedb from "@lancedb/lancedb"
+
+const db = await lancedb.connect("./.lancedb");
+const table = await db.createTable("my_table", [
+  { vector: [1.1, 0.9], id: "1" },
+]);
+
+const plan = await table.query().nearestTo([0.5, 0.2]).analyzePlan();
+
+Example output (with runtime metrics inlined):
+AnalyzeExec verbose=true, metrics=[]
+ ProjectionExec: expr=[id@3 as id, vector@0 as vector, _distance@2 as _distance], metrics=[output_rows=1, elapsed_compute=3.292µs]
+  Take: columns="vector, _rowid, _distance, (id)", metrics=[output_rows=1, elapsed_compute=66.001µs, batches_processed=1, bytes_read=8, iops=1, requests=1]
+   CoalesceBatchesExec: target_batch_size=1024, metrics=[output_rows=1, elapsed_compute=3.333µs]
+    GlobalLimitExec: skip=0, fetch=10, metrics=[output_rows=1, elapsed_compute=167ns]
+     FilterExec: _distance@2 IS NOT NULL, metrics=[output_rows=1, elapsed_compute=8.542µs]
+      SortExec: TopK(fetch=10), expr=[_distance@2 ASC NULLS LAST], metrics=[output_rows=1, elapsed_compute=63.25µs, row_replacements=1]
+       KNNVectorDistance: metric=l2, metrics=[output_rows=1, elapsed_compute=114.333µs, output_batches=1]
+        LanceScan: uri=/path/to/data, projection=[vector], row_id=true, row_addr=false, ordered=false, metrics=[output_rows=1, elapsed_compute=103.626µs, bytes_read=549, iops=2, requests=2]
+```
+
+#### Inherited from
+
+[`QueryBase`](QueryBase.md).[`analyzePlan`](QueryBase.md#analyzeplan)
+
+***
+
 ### bypassVectorIndex()
 
 ```ts

docs/src/js/functions/packBits.md

@@ -0,0 +1,19 @@
+[**@lancedb/lancedb**](../README.md) • **Docs**
+
+***
+
+[@lancedb/lancedb](../globals.md) / packBits
+
+# Function: packBits()
+
+```ts
+function packBits(data): number[]
+```
+
+## Parameters
+
+* **data**: `number`[]
+
+## Returns
+
+`number`[]

docs/src/js/globals.md

@@ -39,6 +39,7 @@
 - [IndexConfig](interfaces/IndexConfig.md)
 - [IndexOptions](interfaces/IndexOptions.md)
 - [IndexStatistics](interfaces/IndexStatistics.md)
+- [IvfFlatOptions](interfaces/IvfFlatOptions.md)
 - [IvfPqOptions](interfaces/IvfPqOptions.md)
 - [OpenTableOptions](interfaces/OpenTableOptions.md)
 - [OptimizeOptions](interfaces/OptimizeOptions.md)
@@ -66,3 +67,4 @@
 
 - [connect](functions/connect.md)
 - [makeArrowTable](functions/makeArrowTable.md)
+- [packBits](functions/packBits.md)

docs/src/js/interfaces/ColumnAlteration.md

@@ -16,7 +16,7 @@ must be provided.
 ### dataType?
 
 ```ts
-optional dataType: string;
+optional dataType: string | DataType<Type, any>;
 ```
 
 A new data type for the column. If not provided then the data type will not be changed.

docs/src/js/interfaces/ConnectionOptions.md

@@ -44,7 +44,7 @@ for testing purposes.
 ### readConsistencyInterval?
 
 ```ts
-optional readConsistencyInterval: number;
+optional readConsistencyInterval: null | number;
 ```
 
 (For LanceDB OSS only): The interval, in seconds, at which to check for

docs/src/js/interfaces/HnswPqOptions.md

@@ -24,18 +24,18 @@ The following distance types are available:
 
 "l2" - Euclidean distance. This is a very common distance metric that
 accounts for both magnitude and direction when determining the distance
-between vectors. L2 distance has a range of [0, ∞).
+between vectors. l2 distance has a range of [0, ∞).
 
 "cosine" - Cosine distance.  Cosine distance is a distance metric
 calculated from the cosine similarity between two vectors. Cosine
 similarity is a measure of similarity between two non-zero vectors of an
 inner product space. It is defined to equal the cosine of the angle
-between them.  Unlike L2, the cosine distance is not affected by the
+between them.  Unlike l2, the cosine distance is not affected by the
 magnitude of the vectors.  Cosine distance has a range of [0, 2].
 
 "dot" - Dot product. Dot distance is the dot product of two vectors. Dot
 distance has a range of (-∞, ∞). If the vectors are normalized (i.e. their
-L2 norm is 1), then dot distance is equivalent to the cosine distance.
+l2 norm is 1), then dot distance is equivalent to the cosine distance.
 
 ***
 

docs/src/js/interfaces/HnswSqOptions.md

@@ -24,18 +24,18 @@ The following distance types are available:
 
 "l2" - Euclidean distance. This is a very common distance metric that
 accounts for both magnitude and direction when determining the distance
-between vectors. L2 distance has a range of [0, ∞).
+between vectors. l2 distance has a range of [0, ∞).
 
 "cosine" - Cosine distance.  Cosine distance is a distance metric
 calculated from the cosine similarity between two vectors. Cosine
 similarity is a measure of similarity between two non-zero vectors of an
 inner product space. It is defined to equal the cosine of the angle
-between them.  Unlike L2, the cosine distance is not affected by the
+between them.  Unlike l2, the cosine distance is not affected by the
 magnitude of the vectors.  Cosine distance has a range of [0, 2].
 
 "dot" - Dot product. Dot distance is the dot product of two vectors. Dot
 distance has a range of (-∞, ∞). If the vectors are normalized (i.e. their
-L2 norm is 1), then dot distance is equivalent to the cosine distance.
+l2 norm is 1), then dot distance is equivalent to the cosine distance.
 
 ***
 

docs/src/js/interfaces/IndexStatistics.md

@@ -30,6 +30,17 @@ The type of the index
 
 ***
 
+### loss?
+
+```ts
+optional loss: number;
+```
+
+The KMeans loss value of the index,
+it is only present for vector indices.
+
+***
+
 ### numIndexedRows
 
 ```ts

docs/src/js/interfaces/IvfFlatOptions.md

@@ -0,0 +1,112 @@
+[**@lancedb/lancedb**](../README.md) • **Docs**
+
+***
+
+[@lancedb/lancedb](../globals.md) / IvfFlatOptions
+
+# Interface: IvfFlatOptions
+
+Options to create an `IVF_FLAT` index
+
+## Properties
+
+### distanceType?
+
+```ts
+optional distanceType: "l2" | "cosine" | "dot" | "hamming";
+```
+
+Distance type to use to build the index.
+
+Default value is "l2".
+
+This is used when training the index to calculate the IVF partitions
+(vectors are grouped in partitions with similar vectors according to this
+distance type).
+
+The distance type used to train an index MUST match the distance type used
+to search the index.  Failure to do so will yield inaccurate results.
+
+The following distance types are available:
+
+"l2" - Euclidean distance. This is a very common distance metric that
+accounts for both magnitude and direction when determining the distance
+between vectors. l2 distance has a range of [0, ∞).
+
+"cosine" - Cosine distance.  Cosine distance is a distance metric
+calculated from the cosine similarity between two vectors. Cosine
+similarity is a measure of similarity between two non-zero vectors of an
+inner product space. It is defined to equal the cosine of the angle
+between them.  Unlike l2, the cosine distance is not affected by the
+magnitude of the vectors.  Cosine distance has a range of [0, 2].
+
+Note: the cosine distance is undefined when one (or both) of the vectors
+are all zeros (there is no direction).  These vectors are invalid and may
+never be returned from a vector search.
+
+"dot" - Dot product. Dot distance is the dot product of two vectors. Dot
+distance has a range of (-∞, ∞). If the vectors are normalized (i.e. their
+l2 norm is 1), then dot distance is equivalent to the cosine distance.
+
+"hamming" - Hamming distance. Hamming distance is a distance metric
+calculated from the number of bits that are different between two vectors.
+Hamming distance has a range of [0, dimension]. Note that the hamming distance
+is only valid for binary vectors.
+
+***
+
+### maxIterations?
+
+```ts
+optional maxIterations: number;
+```
+
+Max iteration to train IVF kmeans.
+
+When training an IVF FLAT index we use kmeans to calculate the partitions.  This parameter
+controls how many iterations of kmeans to run.
+
+Increasing this might improve the quality of the index but in most cases these extra
+iterations have diminishing returns.
+
+The default value is 50.
+
+***
+
+### numPartitions?
+
+```ts
+optional numPartitions: number;
+```
+
+The number of IVF partitions to create.
+
+This value should generally scale with the number of rows in the dataset.
+By default the number of partitions is the square root of the number of
+rows.
+
+If this value is too large then the first part of the search (picking the
+right partition) will be slow.  If this value is too small then the second
+part of the search (searching within a partition) will be slow.
+
+***
+
+### sampleRate?
+
+```ts
+optional sampleRate: number;
+```
+
+The number of vectors, per partition, to sample when training IVF kmeans.
+
+When an IVF FLAT index is trained, we need to calculate partitions.  These are groups
+of vectors that are similar to each other.  To do this we use an algorithm called kmeans.
+
+Running kmeans on a large dataset can be slow.  To speed this up we run kmeans on a
+random sample of the data.  This parameter controls the size of the sample.  The total
+number of vectors used to train the index is `sample_rate * num_partitions`.
+
+Increasing this value might improve the quality of the index but in most cases the
+default should be sufficient.
+
+The default value is 256.

docs/src/js/interfaces/IvfPqOptions.md

@@ -31,13 +31,13 @@ The following distance types are available:
 
 "l2" - Euclidean distance. This is a very common distance metric that
 accounts for both magnitude and direction when determining the distance
-between vectors. L2 distance has a range of [0, ∞).
+between vectors. l2 distance has a range of [0, ∞).
 
 "cosine" - Cosine distance.  Cosine distance is a distance metric
 calculated from the cosine similarity between two vectors. Cosine
 similarity is a measure of similarity between two non-zero vectors of an
 inner product space. It is defined to equal the cosine of the angle
-between them.  Unlike L2, the cosine distance is not affected by the
+between them.  Unlike l2, the cosine distance is not affected by the
 magnitude of the vectors.  Cosine distance has a range of [0, 2].
 
 Note: the cosine distance is undefined when one (or both) of the vectors
@@ -46,7 +46,7 @@ never be returned from a vector search.
 
 "dot" - Dot product. Dot distance is the dot product of two vectors. Dot
 distance has a range of (-∞, ∞). If the vectors are normalized (i.e. their
-L2 norm is 1), then dot distance is equivalent to the cosine distance.
+l2 norm is 1), then dot distance is equivalent to the cosine distance.
 
 ***
 

docs/src/python/python.md

@@ -59,8 +59,6 @@ is also an [asynchronous API client](#connections-asynchronous).
 
 ::: lancedb.embeddings.open_clip.OpenClipEmbeddings
 
-::: lancedb.embeddings.utils.with_embeddings
-
 ## Context
 
 ::: lancedb.context.contextualize

docs/src/search.md

@@ -15,7 +15,7 @@ Currently, LanceDB supports the following metrics:
 
 | Metric    | Description                                                                 |
 | --------- | --------------------------------------------------------------------------- |
-| `l2`      | [Euclidean / L2 distance](https://en.wikipedia.org/wiki/Euclidean_distance) |
+| `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)                 |
 | `hamming` | [Hamming Distance](https://en.wikipedia.org/wiki/Hamming_distance)          |
@@ -138,6 +138,19 @@ LanceDB supports binary vectors as a data type, and has the ability to search bi
         --8<-- "python/python/tests/docs/test_binary_vector.py:async_binary_vector"
         ```
 
+    === "TypeScript"
+
+        ```ts
+        --8<-- "nodejs/examples/search.test.ts:import"
+
+        --8<-- "nodejs/examples/search.test.ts:import_bin_util"
+
+        --8<-- "nodejs/examples/search.test.ts:ingest_binary_data"
+
+        --8<-- "nodejs/examples/search.test.ts:search_binary_data"
+        ```
+
+
 ## Multivector type
 
 LanceDB supports multivector type, this is useful when you have multiple vectors for a single item (e.g. with ColBert and ColPali).

docs/src/sql.md

@@ -7,7 +7,7 @@ performed on the top-k results returned by the vector search. However, pre-filte
 option that performs the filter prior to vector search. This can be useful to narrow down
 the search space of a very large dataset to reduce query latency.
 
-Note that both pre-filtering and post-filtering can yield false positives. For pre-filtering, if the filter is too selective, it might eliminate relevant items that the vector search would have otherwise identified as a good match. In this case, increasing `nprobes` parameter will help reduce such false positives. It is recommended to set `use_index=false` if you know that the filter is highly selective.
+Note that both pre-filtering and post-filtering can yield false positives. For pre-filtering, if the filter is too selective, it might eliminate relevant items that the vector search would have otherwise identified as a good match. In this case, increasing `nprobes` parameter will help reduce such false positives. It is recommended to call `bypass_vector_index()` if you know that the filter is highly selective.
 
 Similarly, a highly selective post-filter can lead to false positives. Increasing both `nprobes` and `refine_factor` can mitigate this issue. When deciding between pre-filtering and post-filtering, pre-filtering is generally the safer choice if you're uncertain.
 

docs/src/troubleshooting.md

@@ -8,6 +8,11 @@ For trouble shooting, the best place to ask is in our Discord, under the relevan
 language channel. By asking in the language-specific channel, it makes it more
 likely that someone who knows the answer will see your question.
 
+## Common issues
+
+* Multiprocessing with `fork` is not supported. You should use `spawn` instead.
+* Data returned by queries may not reflect the most recent writes, depending on configuration. LanceDB uses eventual consistency by default. See [consistency](/docs/src/guides/tables.md#consistency) for more information.
+
 ## Enabling logging
 
 To provide more information, especially for LanceDB Cloud related issues, enable
@@ -31,3 +36,9 @@ print the resolved query plan. You can use the `explain_plan` method to do this:
 * Python Sync: [LanceQueryBuilder.explain_plan][lancedb.query.LanceQueryBuilder.explain_plan]
 * Python Async: [AsyncQueryBase.explain_plan][lancedb.query.AsyncQueryBase.explain_plan]
 * Node @lancedb/lancedb: [LanceQueryBuilder.explainPlan](/lancedb/js/classes/QueryBase/#explainplan)
+
+To understand how a query was actually executed—including metrics like execution time, number of rows processed, I/O stats, and more—use the analyze_plan method. This executes the query and returns a physical execution plan annotated with runtime metrics, making it especially helpful for performance tuning and debugging.
+
+* Python Sync: [LanceQueryBuilder.analyze_plan][lancedb.query.LanceQueryBuilder.analyze_plan]
+* Python Async: [AsyncQueryBase.analyze_plan][lancedb.query.AsyncQueryBase.analyze_plan]
+* Node @lancedb/lancedb: [LanceQueryBuilder.analyzePlan](/lancedb/js/classes/QueryBase/#analyzePlan)

java/.gitignore

@@ -0,0 +1,3 @@
+*.iml
+.java-version
+

java/core/pom.xml

@@ -8,13 +8,16 @@
     <parent>
         <groupId>com.lancedb</groupId>
         <artifactId>lancedb-parent</artifactId>
-        <version>0.18.0-final.0</version>
+        <version>0.18.3-beta.0</version>
         <relativePath>../pom.xml</relativePath>
     </parent>
 
     <artifactId>lancedb-core</artifactId>
     <name>LanceDB Core</name>
     <packaging>jar</packaging>
+    <properties>
+        <rust.release.build>false</rust.release.build>
+    </properties>
 
     <dependencies>
         <dependency>
@@ -68,7 +71,7 @@
                                 </goals>
                                 <configuration>
                                     <path>lancedb-jni</path>
-                                    <release>true</release>
+                                    <release>${rust.release.build}</release>
                                     <!-- Copy native libraries to target/classes for runtime access -->
                                     <copyTo>${project.build.directory}/classes/nativelib</copyTo>
                                     <copyWithPlatformDir>true</copyWithPlatformDir>

java/core/src/main/java/com/lancedb/lancedb/Connection.java

@@ -1,16 +1,25 @@
-// SPDX-License-Identifier: Apache-2.0
-// SPDX-FileCopyrightText: Copyright The LanceDB Authors
-
+/*
+ * 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.
+ */
 package com.lancedb.lancedb;
 
 import io.questdb.jar.jni.JarJniLoader;
+
 import java.io.Closeable;
 import java.util.List;
 import java.util.Optional;
 
-/**
- * Represents LanceDB database.
- */
+/** Represents LanceDB database. */
 public class Connection implements Closeable {
   static {
     JarJniLoader.loadLib(Connection.class, "/nativelib", "lancedb_jni");
@@ -18,14 +27,11 @@ public class Connection implements Closeable {
 
   private long nativeConnectionHandle;
 
-  /**
-   * Connect to a LanceDB instance.
-   */
+  /** Connect to a LanceDB instance. */
   public static native Connection connect(String uri);
 
   /**
-   * Get the names of all tables in the database. The names are sorted in
-   * ascending order.
+   * Get the names of all tables in the database. The names are sorted in ascending order.
    *
    * @return the table names
    */
@@ -34,8 +40,7 @@ public class Connection implements Closeable {
   }
 
   /**
-   * Get the names of filtered tables in the database. The names are sorted in
-   * ascending order.
+   * Get the names of filtered tables in the database. The names are sorted in ascending order.
    *
    * @param limit The number of results to return.
    * @return the table names
@@ -45,12 +50,11 @@ public class Connection implements Closeable {
   }
 
   /**
-   * Get the names of filtered tables in the database. The names are sorted in
-   * ascending order.
+   * Get the names of filtered tables in the database. The names are sorted in ascending order.
    *
    * @param startAfter If present, only return names that come lexicographically after the supplied
-   *                   value. This can be combined with limit to implement pagination
-   *                   by setting this to the last table name from the previous page.
+   *     value. This can be combined with limit to implement pagination by setting this to the last
+   *     table name from the previous page.
    * @return the table names
    */
   public List<String> tableNames(String startAfter) {
@@ -58,12 +62,11 @@ public class Connection implements Closeable {
   }
 
   /**
-   * Get the names of filtered tables in the database. The names are sorted in
-   * ascending order.
+   * Get the names of filtered tables in the database. The names are sorted in ascending order.
    *
    * @param startAfter If present, only return names that come lexicographically after the supplied
-   *                   value. This can be combined with limit to implement pagination
-   *                   by setting this to the last table name from the previous page.
+   *     value. This can be combined with limit to implement pagination by setting this to the last
+   *     table name from the previous page.
    * @param limit The number of results to return.
    * @return the table names
    */
@@ -72,22 +75,19 @@ public class Connection implements Closeable {
   }
 
   /**
-   * Get the names of filtered tables in the database. The names are sorted in
-   * ascending order.
+   * Get the names of filtered tables in the database. The names are sorted in ascending order.
    *
    * @param startAfter If present, only return names that come lexicographically after the supplied
-   *                   value. This can be combined with limit to implement pagination
-   *                   by setting this to the last table name from the previous page.
+   *     value. This can be combined with limit to implement pagination by setting this to the last
+   *     table name from the previous page.
    * @param limit The number of results to return.
    * @return the table names
    */
-  public native List<String> tableNames(
-      Optional<String> startAfter, Optional<Integer> limit);
+  public native List<String> tableNames(Optional<String> startAfter, Optional<Integer> limit);
 
   /**
-   * Closes this connection and releases any system resources associated with it. If
-   * the connection is
-   * already closed, then invoking this method has no effect.
+   * Closes this connection and releases any system resources associated with it. If the connection
+   * is already closed, then invoking this method has no effect.
    */
   @Override
   public void close() {
@@ -98,8 +98,7 @@ public class Connection implements Closeable {
   }
 
   /**
-   * Native method to release the Lance connection resources associated with the
-   * given handle.
+   * Native method to release the Lance connection resources associated with the given handle.
    *
    * @param handle The native handle to the connection resource.
    */

java/core/src/test/java/com/lancedb/lancedb/ConnectionTest.java

@@ -1,27 +1,35 @@
-// SPDX-License-Identifier: Apache-2.0
-// SPDX-FileCopyrightText: Copyright The LanceDB Authors
+/*
+ * 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.
+ */
 package com.lancedb.lancedb;
 
-import static org.junit.jupiter.api.Assertions.assertEquals;
-import static org.junit.jupiter.api.Assertions.assertTrue;
-
-import java.nio.file.Path;
-import java.util.List;
-import java.net.URL;
 import org.junit.jupiter.api.BeforeAll;
 import org.junit.jupiter.api.Test;
 import org.junit.jupiter.api.io.TempDir;
 
+import java.net.URL;
+import java.nio.file.Path;
+import java.util.List;
+
+import static org.junit.jupiter.api.Assertions.assertEquals;
+import static org.junit.jupiter.api.Assertions.assertTrue;
+
 public class ConnectionTest {
   private static final String[] TABLE_NAMES = {
-      "dataset_version",
-      "new_empty_dataset",
-      "test",
-      "write_stream"
+    "dataset_version", "new_empty_dataset", "test", "write_stream"
   };
 
-  @TempDir
-  static Path tempDir; // Temporary directory for the tests
+  @TempDir static Path tempDir; // Temporary directory for the tests
   private static URL lanceDbURL;
 
   @BeforeAll
@@ -53,18 +61,21 @@ public class ConnectionTest {
   @Test
   void tableNamesStartAfter() {
     try (Connection conn = Connection.connect(lanceDbURL.toString())) {
-      assertTableNamesStartAfter(conn, TABLE_NAMES[0], 3, TABLE_NAMES[1], TABLE_NAMES[2], TABLE_NAMES[3]);
+      assertTableNamesStartAfter(
+          conn, TABLE_NAMES[0], 3, TABLE_NAMES[1], TABLE_NAMES[2], TABLE_NAMES[3]);
       assertTableNamesStartAfter(conn, TABLE_NAMES[1], 2, TABLE_NAMES[2], TABLE_NAMES[3]);
       assertTableNamesStartAfter(conn, TABLE_NAMES[2], 1, TABLE_NAMES[3]);
       assertTableNamesStartAfter(conn, TABLE_NAMES[3], 0);
-      assertTableNamesStartAfter(conn, "a_dataset", 4, TABLE_NAMES[0], TABLE_NAMES[1], TABLE_NAMES[2], TABLE_NAMES[3]);
+      assertTableNamesStartAfter(
+          conn, "a_dataset", 4, TABLE_NAMES[0], TABLE_NAMES[1], TABLE_NAMES[2], TABLE_NAMES[3]);
       assertTableNamesStartAfter(conn, "o_dataset", 2, TABLE_NAMES[2], TABLE_NAMES[3]);
       assertTableNamesStartAfter(conn, "v_dataset", 1, TABLE_NAMES[3]);
       assertTableNamesStartAfter(conn, "z_dataset", 0);
     }
   }
 
-  private void assertTableNamesStartAfter(Connection conn, String startAfter, int expectedSize, String... expectedNames) {
+  private void assertTableNamesStartAfter(
+      Connection conn, String startAfter, int expectedSize, String... expectedNames) {
     List<String> tableNames = conn.tableNames(startAfter);
     assertEquals(expectedSize, tableNames.size());
     for (int i = 0; i < expectedNames.length; i++) {
@@ -74,7 +85,7 @@ public class ConnectionTest {
 
   @Test
   void tableNamesLimit() {
-      try (Connection conn = Connection.connect(lanceDbURL.toString())) {
+    try (Connection conn = Connection.connect(lanceDbURL.toString())) {
       for (int i = 0; i <= TABLE_NAMES.length; i++) {
         List<String> tableNames = conn.tableNames(i);
         assertEquals(i, tableNames.size());

java/pom.xml

@@ -6,7 +6,7 @@
 
     <groupId>com.lancedb</groupId>
     <artifactId>lancedb-parent</artifactId>
-    <version>0.18.0-final.0</version>
+    <version>0.18.3-beta.0</version>
     <packaging>pom</packaging>
 
     <name>LanceDB Parent</name>
@@ -29,6 +29,25 @@
     <properties>
         <project.build.sourceEncoding>UTF-8</project.build.sourceEncoding>
         <arrow.version>15.0.0</arrow.version>
+        <spotless.skip>false</spotless.skip>
+        <spotless.version>2.30.0</spotless.version>
+        <spotless.java.googlejavaformat.version>1.7</spotless.java.googlejavaformat.version>
+        <spotless.delimiter>package</spotless.delimiter>
+        <spotless.license.header>
+            /*
+            * 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.
+            */
+        </spotless.license.header>
     </properties>
 
     <modules>
@@ -127,7 +146,8 @@
                 <configuration>
                     <configLocation>google_checks.xml</configLocation>
                     <consoleOutput>true</consoleOutput>
-                    <failsOnError>true</failsOnError>
+                    <failsOnError>false</failsOnError>
+                    <failOnViolation>false</failOnViolation>
                     <violationSeverity>warning</violationSeverity>
                     <linkXRef>false</linkXRef>
                 </configuration>
@@ -141,6 +161,10 @@
                     </execution>
                 </executions>
             </plugin>
+            <plugin>
+                <groupId>com.diffplug.spotless</groupId>
+                <artifactId>spotless-maven-plugin</artifactId>
+            </plugin>
         </plugins>
         <pluginManagement>
             <plugins>
@@ -166,7 +190,6 @@
                     <artifactId>maven-surefire-plugin</artifactId>
                     <version>3.2.5</version>
                     <configuration>
-                        <argLine>--add-opens=java.base/java.nio=ALL-UNNAMED</argLine>
                         <forkNode
                             implementation="org.apache.maven.plugin.surefire.extensions.SurefireForkNodeFactory" />
                         <useSystemClassLoader>false</useSystemClassLoader>
@@ -180,6 +203,54 @@
                     <artifactId>maven-install-plugin</artifactId>
                     <version>2.5.2</version>
                 </plugin>
+                <plugin>
+                    <groupId>com.diffplug.spotless</groupId>
+                    <artifactId>spotless-maven-plugin</artifactId>
+                    <version>${spotless.version}</version>
+                    <configuration>
+                        <skip>${spotless.skip}</skip>
+                        <upToDateChecking>
+                            <enabled>true</enabled>
+                        </upToDateChecking>
+                        <java>
+                            <includes>
+                                <include>src/main/java/**/*.java</include>
+                                <include>src/test/java/**/*.java</include>
+                            </includes>
+                            <googleJavaFormat>
+                                <version>${spotless.java.googlejavaformat.version}</version>
+                                <style>GOOGLE</style>
+                            </googleJavaFormat>
+
+                            <importOrder>
+                                <order>com.lancedb.lance,,javax,java,\#</order>
+                            </importOrder>
+
+                            <removeUnusedImports />
+                        </java>
+                        <scala>
+                            <includes>
+                                <include>src/main/scala/**/*.scala</include>
+                                <include>src/main/scala-*/**/*.scala</include>
+                                <include>src/test/scala/**/*.scala</include>
+                                <include>src/test/scala-*/**/*.scala</include>
+                            </includes>
+                        </scala>
+                        <licenseHeader>
+                            <content>${spotless.license.header}</content>
+                            <delimiter>${spotless.delimiter}</delimiter>
+                        </licenseHeader>
+                    </configuration>
+                    <executions>
+                        <execution>
+                            <id>spotless-check</id>
+                            <phase>validate</phase>
+                            <goals>
+                                <goal>apply</goal>
+                            </goals>
+                        </execution>
+                    </executions>
+                </plugin>
             </plugins>
         </pluginManagement>
     </build>

node/package-lock.json

@@ -1,12 +1,12 @@
 {
   "name": "vectordb",
-  "version": "0.18.0",
+  "version": "0.18.3-beta.0",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {
     "": {
       "name": "vectordb",
-      "version": "0.18.0",
+      "version": "0.18.3-beta.0",
       "cpu": [
         "x64",
         "arm64"
@@ -52,14 +52,11 @@
         "uuid": "^9.0.0"
       },
       "optionalDependencies": {
-        "@lancedb/vectordb-darwin-arm64": "0.18.0",
-        "@lancedb/vectordb-darwin-x64": "0.18.0",
-        "@lancedb/vectordb-linux-arm64-gnu": "0.18.0",
-        "@lancedb/vectordb-linux-arm64-musl": "0.18.0",
-        "@lancedb/vectordb-linux-x64-gnu": "0.18.0",
-        "@lancedb/vectordb-linux-x64-musl": "0.18.0",
-        "@lancedb/vectordb-win32-arm64-msvc": "0.18.0",
-        "@lancedb/vectordb-win32-x64-msvc": "0.18.0"
+        "@lancedb/vectordb-darwin-arm64": "0.18.3-beta.0",
+        "@lancedb/vectordb-darwin-x64": "0.18.3-beta.0",
+        "@lancedb/vectordb-linux-arm64-gnu": "0.18.3-beta.0",
+        "@lancedb/vectordb-linux-x64-gnu": "0.18.3-beta.0",
+        "@lancedb/vectordb-win32-x64-msvc": "0.18.3-beta.0"
       },
       "peerDependencies": {
         "@apache-arrow/ts": "^14.0.2",
@@ -330,9 +327,9 @@
       }
     },
     "node_modules/@lancedb/vectordb-darwin-arm64": {
-      "version": "0.18.0",
-      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-darwin-arm64/-/vectordb-darwin-arm64-0.18.0.tgz",
-      "integrity": "sha512-ormNCmny1j64aSZRrZeUQ1Zs8cOFKrW14NgTmW3AehDuru+Ep+8AriHA5Pmyi6raBOZfNzDSiZs/LTzzyVaa7g==",
+      "version": "0.18.3-beta.0",
+      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-darwin-arm64/-/vectordb-darwin-arm64-0.18.3-beta.0.tgz",
+      "integrity": "sha512-dhJ5VlXV2N/L67mIpTSePhb8krX0FyQgpuz3I+4T4vYuU5JEF3cmedQ5TF5+3cGJhZim4PHRYLkfgCyTlxcqUg==",
       "cpu": [
         "arm64"
       ],
@@ -343,9 +340,9 @@
       ]
     },
     "node_modules/@lancedb/vectordb-darwin-x64": {
-      "version": "0.18.0",
-      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-darwin-x64/-/vectordb-darwin-x64-0.18.0.tgz",
-      "integrity": "sha512-S4skQ1RXXQJciq40s84Kyy7v3YC+nao8pX4xUyxDcKRx+90Qg9eH+tehs6XLN1IjrQT/9CWKaE5wxZmv6Oys4g==",
+      "version": "0.18.3-beta.0",
+      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-darwin-x64/-/vectordb-darwin-x64-0.18.3-beta.0.tgz",
+      "integrity": "sha512-SHqPkuyfe87d5skf9GERzdeu6AKvVIbXMUwl5N+dVrE7HH6qiuP2HvOmiyHS2lJFgo0Ph8jSBVzPDxxtjF36Dg==",
       "cpu": [
         "x64"
       ],
@@ -356,22 +353,9 @@
       ]
     },
     "node_modules/@lancedb/vectordb-linux-arm64-gnu": {
-      "version": "0.18.0",
-      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-linux-arm64-gnu/-/vectordb-linux-arm64-gnu-0.18.0.tgz",
-      "integrity": "sha512-1txr4tasVdxy321/4Fw8GJPjzrf84F02L9ffN8JebHmmR0S8uk2MKf2WsyLaSVRPd4YHIvvf3qmG0RGaUsb2sw==",
-      "cpu": [
-        "arm64"
-      ],
-      "license": "Apache-2.0",
-      "optional": true,
-      "os": [
-        "linux"
-      ]
-    },
-    "node_modules/@lancedb/vectordb-linux-arm64-musl": {
-      "version": "0.18.0",
-      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-linux-arm64-musl/-/vectordb-linux-arm64-musl-0.18.0.tgz",
-      "integrity": "sha512-8xS1xaoJeFDx6WmDBcfueWvIbdNX/ptQXfoC7hYICwNHizjlyt4O3Nxz8uG9URMF1y9saUYUditIHLzLVZc76g==",
+      "version": "0.18.3-beta.0",
+      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-linux-arm64-gnu/-/vectordb-linux-arm64-gnu-0.18.3-beta.0.tgz",
+      "integrity": "sha512-ohnWsV1n9cxL5ik/GGL4FdQ04Ff9REELcNb1zgmJYyEfwyc6TH9m5HdySO/1ACPZJiLbML4gSvZ10J0Zyb+2SA==",
       "cpu": [
         "arm64"
       ],
@@ -382,9 +366,9 @@
       ]
     },
     "node_modules/@lancedb/vectordb-linux-x64-gnu": {
-      "version": "0.18.0",
-      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-linux-x64-gnu/-/vectordb-linux-x64-gnu-0.18.0.tgz",
-      "integrity": "sha512-8XUc2UnEV3awv0DGJS5gRA7yTkicX6oPN7GudXXxycCKL33FJ2ah7hkeDia9Bhk9MmvTonvsEDvUSqnglcpqfA==",
+      "version": "0.18.3-beta.0",
+      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-linux-x64-gnu/-/vectordb-linux-x64-gnu-0.18.3-beta.0.tgz",
+      "integrity": "sha512-nhbW2CKaBSUesiYCPBd9fAsDYIJLadlGsrb2gfjODlFy+2Lpnbz6T9SuV7dNqj6KBw+KHhaRhLqta7tyMZm/EA==",
       "cpu": [
         "x64"
       ],
@@ -394,36 +378,10 @@
         "linux"
       ]
     },
-    "node_modules/@lancedb/vectordb-linux-x64-musl": {
-      "version": "0.18.0",
-      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-linux-x64-musl/-/vectordb-linux-x64-musl-0.18.0.tgz",
-      "integrity": "sha512-LV7TuWgLcL82Wdq+EH2Xs3+apqeLohwYLlVIauVAwKEHvdwyNxTOW9TaNLvHXcbylIh7agl2xXvASCNhYZAyzA==",
-      "cpu": [
-        "x64"
-      ],
-      "license": "Apache-2.0",
-      "optional": true,
-      "os": [
-        "linux"
-      ]
-    },
-    "node_modules/@lancedb/vectordb-win32-arm64-msvc": {
-      "version": "0.18.0",
-      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-win32-arm64-msvc/-/vectordb-win32-arm64-msvc-0.18.0.tgz",
-      "integrity": "sha512-kxdCnKfvnuDKoKZRUBbreMBpimHb+k9/pFR48GN6JSrIuzUDx5G1VjHKBmaFhbveZCOBjjtYlg/8ohnWQHZfeA==",
-      "cpu": [
-        "arm64"
-      ],
-      "license": "Apache-2.0",
-      "optional": true,
-      "os": [
-        "win32"
-      ]
-    },
     "node_modules/@lancedb/vectordb-win32-x64-msvc": {
-      "version": "0.18.0",
-      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-win32-x64-msvc/-/vectordb-win32-x64-msvc-0.18.0.tgz",
-      "integrity": "sha512-uAE80q50cAp4gHoGvclxJqZGqj3/9oN9kz8iXgNIxiPngqnN01kVyaj4ulm4Qk/nauWUhHJ3tjTh/+CpkhSc2Q==",
+      "version": "0.18.3-beta.0",
+      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-win32-x64-msvc/-/vectordb-win32-x64-msvc-0.18.3-beta.0.tgz",
+      "integrity": "sha512-VE4TvMdZ7DIrTC8VYylGxEcH4h2UEejSwGX4PxRzrN9QsCQ4m4pOh3L/UguSO3g+Y1QEaGE20iWQoX6wgSEUhA==",
       "cpu": [
         "x64"
       ],

node/package.json

@@ -1,6 +1,6 @@
 {
   "name": "vectordb",
-  "version": "0.18.0",
+  "version": "0.18.3-beta.0",
   "description": " Serverless, low-latency vector database for AI applications",
   "private": false,
   "main": "dist/index.js",
@@ -85,20 +85,14 @@
       "aarch64-apple-darwin": "@lancedb/vectordb-darwin-arm64",
       "x86_64-unknown-linux-gnu": "@lancedb/vectordb-linux-x64-gnu",
       "aarch64-unknown-linux-gnu": "@lancedb/vectordb-linux-arm64-gnu",
-      "x86_64-unknown-linux-musl": "@lancedb/vectordb-linux-x64-musl",
-      "aarch64-unknown-linux-musl": "@lancedb/vectordb-linux-arm64-musl",
-      "x86_64-pc-windows-msvc": "@lancedb/vectordb-win32-x64-msvc",
-      "aarch64-pc-windows-msvc": "@lancedb/vectordb-win32-arm64-msvc"
+      "x86_64-pc-windows-msvc": "@lancedb/vectordb-win32-x64-msvc"
     }
   },
   "optionalDependencies": {
-    "@lancedb/vectordb-darwin-x64": "0.18.0",
-    "@lancedb/vectordb-darwin-arm64": "0.18.0",
-    "@lancedb/vectordb-linux-x64-gnu": "0.18.0",
-    "@lancedb/vectordb-linux-arm64-gnu": "0.18.0",
-    "@lancedb/vectordb-linux-x64-musl": "0.18.0",
-    "@lancedb/vectordb-linux-arm64-musl": "0.18.0",
-    "@lancedb/vectordb-win32-x64-msvc": "0.18.0",
-    "@lancedb/vectordb-win32-arm64-msvc": "0.18.0"
+    "@lancedb/vectordb-darwin-x64": "0.18.3-beta.0",
+    "@lancedb/vectordb-darwin-arm64": "0.18.3-beta.0",
+    "@lancedb/vectordb-linux-x64-gnu": "0.18.3-beta.0",
+    "@lancedb/vectordb-linux-arm64-gnu": "0.18.3-beta.0",
+    "@lancedb/vectordb-win32-x64-msvc": "0.18.3-beta.0"
   }
 }

node/src/index.ts

@@ -1299,7 +1299,7 @@ export interface IvfPQIndexConfig {
   index_name?: string
 
   /**
-   * Metric type, L2 or Cosine
+   * Metric type, l2 or Cosine
    */
   metric_type?: MetricType
 

node/src/integration_test/test.ts

@@ -110,7 +110,7 @@ describe('LanceDB Mirrored Store Integration test', function () {
 
       fs.readdir(path.join(mirroredPath, 'data'), { withFileTypes: true }, (err, files) => {
         if (err != null) throw err
-        assert.equal(files.length, 1)
+        assert.equal(files.length, 1, `Found files: ${files.map(f => f.name)}`)
         assert.isTrue(files[0].name.endsWith('.lance'))
       })
 

nodejs/.npmignore

@@ -22,3 +22,4 @@ build.rs
 jest.config.js
 tsconfig.json
 typedoc.json
+typedoc_post_process.js

nodejs/Cargo.toml

@@ -1,7 +1,7 @@
 [package]
 name = "lancedb-nodejs"
 edition.workspace = true
-version = "0.18.0"
+version = "0.18.3-beta.0"
 license.workspace = true
 description.workspace = true
 repository.workspace = true
@@ -18,7 +18,7 @@ arrow-array.workspace = true
 arrow-schema.workspace = true
 env_logger.workspace = true
 futures.workspace = true
-lancedb = { path = "../rust/lancedb", features = ["remote"] }
+lancedb = { path = "../rust/lancedb" }
 napi = { version = "2.16.8", default-features = false, features = [
     "napi9",
     "async"
@@ -30,3 +30,8 @@ log.workspace = true
 
 [build-dependencies]
 napi-build = "2.1"
+
+[features]
+default = ["remote"]
+fp16kernels = ["lancedb/fp16kernels"]
+remote = ["lancedb/remote"]

nodejs/README.md

@@ -11,11 +11,9 @@ npm install @lancedb/lancedb
 This will download the appropriate native library for your platform. We currently
 support:
 
-- Linux (x86_64 and aarch64)
+- Linux (x86_64 and aarch64 on glibc and musl)
 - MacOS (Intel and ARM/M1/M2)
-- Windows (x86_64 only)
-
-We do not yet support musl-based Linux (such as Alpine Linux) or aarch64 Windows.
+- Windows (x86_64 and aarch64)
 
 ## Usage
 

nodejs/__test__/connection.test.ts

@@ -17,7 +17,7 @@ describe("when connecting", () => {
   it("should connect", async () => {
     const db = await connect(tmpDir.name);
     expect(db.display()).toBe(
-      `ListingDatabase(uri=${tmpDir.name}, read_consistency_interval=None)`,
+      `ListingDatabase(uri=${tmpDir.name}, read_consistency_interval=5s)`,
     );
   });
 

nodejs/__test__/table.test.ts

@@ -21,9 +21,11 @@ import {
   Int64,
   List,
   Schema,
+  Uint8,
   Utf8,
   makeArrowTable,
 } from "../lancedb/arrow";
+import * as arrow from "../lancedb/arrow";
 import {
   EmbeddingFunction,
   LanceSchema,
@@ -56,7 +58,7 @@ describe.each([arrow15, arrow16, arrow17, arrow18])(
 
     it("be displayable", async () => {
       expect(table.display()).toMatch(
-        /NativeTable\(some_table, uri=.*, read_consistency_interval=None\)/,
+        /NativeTable\(some_table, uri=.*, read_consistency_interval=5s\)/,
       );
       table.close();
       expect(table.display()).toBe("ClosedTable(some_table)");
@@ -278,6 +280,15 @@ describe.each([arrow15, arrow16, arrow17, arrow18])(
       expect(res.getChild("y")?.toJSON()).toEqual([2, null, null, null]);
       expect(res.getChild("z")?.toJSON()).toEqual([null, null, 3n, 5n]);
     });
+
+    it("should handle null vectors at end of data", async () => {
+      // https://github.com/lancedb/lancedb/issues/2240
+      const data = [{ vector: [1, 2, 3] }, { vector: null }];
+      const db = await connect("memory://");
+
+      const table = await db.createTable("my_table", data);
+      expect(await table.countRows()).toEqual(2);
+    });
   },
 );
 
@@ -460,6 +471,8 @@ describe("When creating an index", () => {
       indexType: "IvfPq",
       columns: ["vec"],
     });
+    const stats = await tbl.indexStats("vec_idx");
+    expect(stats?.loss).toBeDefined();
 
     // Search without specifying the column
     let rst = await tbl
@@ -620,6 +633,23 @@ describe("When creating an index", () => {
     expect(plan2).not.toMatch("LanceScan");
   });
 
+  it("should be able to run analyze plan", async () => {
+    await tbl.createIndex("vec");
+    await tbl.add([
+      {
+        id: 300,
+        vec: Array(32)
+          .fill(1)
+          .map(() => Math.random()),
+        tags: [],
+      },
+    ]);
+
+    const plan = await tbl.query().nearestTo(queryVec).analyzePlan();
+    expect(plan).toMatch("AnalyzeExec");
+    expect(plan).toMatch("metrics=");
+  });
+
   it("should be able to query with row id", async () => {
     const results = await tbl
       .query()
@@ -720,6 +750,7 @@ describe("When creating an index", () => {
     expect(stats?.distanceType).toBeUndefined();
     expect(stats?.indexType).toEqual("BTREE");
     expect(stats?.numIndices).toEqual(1);
+    expect(stats?.loss).toBeUndefined();
   });
 
   test("when getting stats on non-existent index", async () => {
@@ -727,6 +758,38 @@ describe("When creating an index", () => {
     expect(stats).toBeUndefined();
   });
 
+  test("create ivf_flat with binary vectors", async () => {
+    const db = await connect(tmpDir.name);
+    const binarySchema = new Schema([
+      new Field("id", new Int32(), true),
+      new Field("vec", new FixedSizeList(32, new Field("item", new Uint8()))),
+    ]);
+    const tbl = await db.createTable(
+      "binary",
+      makeArrowTable(
+        Array(300)
+          .fill(1)
+          .map((_, i) => ({
+            id: i,
+            vec: Array(32)
+              .fill(1)
+              .map(() => Math.floor(Math.random() * 255)),
+          })),
+        { schema: binarySchema },
+      ),
+    );
+    await tbl.createIndex("vec", {
+      config: Index.ivfFlat({ numPartitions: 10, distanceType: "hamming" }),
+    });
+
+    // query with binary vectors
+    const queryVec = Array(32)
+      .fill(1)
+      .map(() => Math.floor(Math.random() * 255));
+    const rst = await tbl.query().limit(5).nearestTo(queryVec).toArrow();
+    expect(rst.numRows).toBe(5);
+  });
+
   // TODO: Move this test to the query API test (making sure we can reject queries
   // when the dimension is incorrect)
   test("two columns with different dimensions", async () => {
@@ -920,6 +983,93 @@ describe("schema evolution", function () {
       new Field("price", new Float64(), true),
     ]);
     expect(await table.schema()).toEqual(expectedSchema2);
+
+    await table.alterColumns([
+      {
+        path: "vector",
+        dataType: new FixedSizeList(2, new Field("item", new Float64(), true)),
+      },
+    ]);
+    const expectedSchema3 = new Schema([
+      new Field("new_id", new Int32(), true),
+      new Field(
+        "vector",
+        new FixedSizeList(2, new Field("item", new Float64(), true)),
+        true,
+      ),
+      new Field("price", new Float64(), true),
+    ]);
+    expect(await table.schema()).toEqual(expectedSchema3);
+  });
+
+  it("can cast to various types", async function () {
+    const con = await connect(tmpDir.name);
+
+    // integers
+    const intTypes = [
+      new arrow.Int8(),
+      new arrow.Int16(),
+      new arrow.Int32(),
+      new arrow.Int64(),
+      new arrow.Uint8(),
+      new arrow.Uint16(),
+      new arrow.Uint32(),
+      new arrow.Uint64(),
+    ];
+    const tableInts = await con.createTable("ints", [{ id: 1n }], {
+      schema: new Schema([new Field("id", new Int64(), true)]),
+    });
+    for (const intType of intTypes) {
+      await tableInts.alterColumns([{ path: "id", dataType: intType }]);
+      const schema = new Schema([new Field("id", intType, true)]);
+      expect(await tableInts.schema()).toEqual(schema);
+    }
+
+    // floats
+    const floatTypes = [
+      new arrow.Float16(),
+      new arrow.Float32(),
+      new arrow.Float64(),
+    ];
+    const tableFloats = await con.createTable("floats", [{ val: 2.1 }], {
+      schema: new Schema([new Field("val", new Float32(), true)]),
+    });
+    for (const floatType of floatTypes) {
+      await tableFloats.alterColumns([{ path: "val", dataType: floatType }]);
+      const schema = new Schema([new Field("val", floatType, true)]);
+      expect(await tableFloats.schema()).toEqual(schema);
+    }
+
+    // Lists of floats
+    const listTypes = [
+      new arrow.List(new arrow.Field("item", new arrow.Float32(), true)),
+      new arrow.FixedSizeList(
+        2,
+        new arrow.Field("item", new arrow.Float64(), true),
+      ),
+      new arrow.FixedSizeList(
+        2,
+        new arrow.Field("item", new arrow.Float16(), true),
+      ),
+      new arrow.FixedSizeList(
+        2,
+        new arrow.Field("item", new arrow.Float32(), true),
+      ),
+    ];
+    const tableLists = await con.createTable("lists", [{ val: [2.1, 3.2] }], {
+      schema: new Schema([
+        new Field(
+          "val",
+          new FixedSizeList(2, new arrow.Field("item", new Float32())),
+          true,
+        ),
+      ]),
+    });
+    for (const listType of listTypes) {
+      await tableLists.alterColumns([{ path: "val", dataType: listType }]);
+      const schema = new Schema([new Field("val", listType, true)]);
+      expect(await tableLists.schema()).toEqual(schema);
+    }
   });
 
   it("can drop a column from the schema", async function () {
@@ -1213,6 +1363,30 @@ describe("when calling explainPlan", () => {
   });
 });
 
+describe("when calling analyzePlan", () => {
+  let tmpDir: tmp.DirResult;
+  let table: Table;
+  let queryVec: number[];
+  beforeEach(async () => {
+    tmpDir = tmp.dirSync({ unsafeCleanup: true });
+    const con = await connect(tmpDir.name);
+    table = await con.createTable("vectors", [{ id: 1, vector: [1.1, 0.9] }]);
+  });
+
+  afterEach(() => {
+    tmpDir.removeCallback();
+  });
+
+  it("retrieves runtime metrics", async () => {
+    queryVec = Array(2)
+      .fill(1)
+      .map(() => Math.random());
+    const plan = await table.query().nearestTo(queryVec).analyzePlan();
+    console.log("Query Plan:\n", plan); // <--- Print the plan
+    expect(plan).toMatch("AnalyzeExec");
+  });
+});
+
 describe("column name options", () => {
   let tmpDir: tmp.DirResult;
   let table: Table;

nodejs/examples/basic.test.ts

@@ -132,6 +132,17 @@ test("basic table examples", async () => {
       },
     ]);
     // --8<-- [end:alter_columns]
+    // --8<-- [start:alter_columns_vector]
+    await tbl.alterColumns([
+      {
+        path: "vector",
+        dataType: new arrow.FixedSizeList(
+          2,
+          new arrow.Field("item", new arrow.Float16(), false),
+        ),
+      },
+    ]);
+    // --8<-- [end:alter_columns_vector]
     // --8<-- [start:drop_columns]
     await tbl.dropColumns(["dbl_price"]);
     // --8<-- [end:drop_columns]
@@ -191,5 +202,35 @@ test("basic table examples", async () => {
       // --8<-- [end:create_f16_table]
       await db.dropTable("f16_tbl");
     }
+    const uri = databaseDir;
+    await db.createTable("my_table", [{ id: 1 }, { id: 2 }]);
+    {
+      // --8<-- [start:table_strong_consistency]
+      const db = await lancedb.connect({ uri, readConsistencyInterval: 0 });
+      const tbl = await db.openTable("my_table");
+      // --8<-- [end:table_strong_consistency]
+    }
+    {
+      // --8<-- [start:table_eventual_consistency]
+      const db = await lancedb.connect({ uri, readConsistencyInterval: 5 });
+      const tbl = await db.openTable("my_table");
+      // --8<-- [end:table_eventual_consistency]
+    }
+    {
+      // --8<-- [start:table_no_consistency]
+      const db = await lancedb.connect({ uri, readConsistencyInterval: null });
+      const tbl = await db.openTable("my_table");
+      // --8<-- [end:table_no_consistency]
+    }
+    {
+      // --8<-- [start:table_checkout_latest]
+      const tbl = await db.openTable("my_table");
+
+      // (Other writes happen to test_table_async from another process)
+
+      // Check for updates
+      tbl.checkoutLatest();
+      // --8<-- [end:table_checkout_latest]
+    }
   });
 });

nodejs/examples/search.test.ts

@@ -4,9 +4,12 @@ import { expect, test } from "@jest/globals";
 // --8<-- [start:import]
 import * as lancedb from "@lancedb/lancedb";
 // --8<-- [end:import]
+// --8<-- [start:import_bin_util]
+import { Field, FixedSizeList, Int32, Schema, Uint8 } from "apache-arrow";
+// --8<-- [end:import_bin_util]
 import { withTempDirectory } from "./util.ts";
 
-test("full text search", async () => {
+test("vector search", async () => {
   await withTempDirectory(async (databaseDir) => {
     {
       const db = await lancedb.connect(databaseDir);
@@ -14,8 +17,6 @@ test("full text search", async () => {
       const data = Array.from({ length: 10_000 }, (_, i) => ({
         vector: Array(128).fill(i),
         id: `${i}`,
-        content: "",
-        longId: `${i}`,
       }));
 
       await db.createTable("my_vectors", data);
@@ -52,5 +53,41 @@ test("full text search", async () => {
       expect(r.distance).toBeGreaterThanOrEqual(0.1);
       expect(r.distance).toBeLessThan(0.2);
     }
+
+    {
+      // --8<-- [start:ingest_binary_data]
+      const schema = new Schema([
+        new Field("id", new Int32(), true),
+        new Field("vec", new FixedSizeList(32, new Field("item", new Uint8()))),
+      ]);
+      const data = lancedb.makeArrowTable(
+        Array(1_000)
+          .fill(0)
+          .map((_, i) => ({
+            // the 256 bits would be store in 32 bytes,
+            // if your data is already in this format, you can skip the packBits step
+            id: i,
+            vec: lancedb.packBits(Array(256).fill(i % 2)),
+          })),
+        { schema: schema },
+      );
+
+      const tbl = await db.createTable("binary_table", data);
+      await tbl.createIndex("vec", {
+        config: lancedb.Index.ivfFlat({
+          numPartitions: 10,
+          distanceType: "hamming",
+        }),
+      });
+      // --8<-- [end:ingest_binary_data]
+
+      // --8<-- [start:search_binary_data]
+      const query = Array(32)
+        .fill(1)
+        .map(() => Math.floor(Math.random() * 255));
+      const results = await tbl.query().nearestTo(query).limit(10).toArrow();
+      // --8<-- [end:search_binary_data
+      expect(results.numRows).toBe(10);
+    }
   });
 });

nodejs/lancedb/arrow.ts

@@ -8,7 +8,11 @@ import {
   Bool,
   BufferType,
   DataType,
+  DateUnit,
+  Date_,
+  Decimal,
   Dictionary,
+  Duration,
   Field,
   FixedSizeBinary,
   FixedSizeList,
@@ -21,19 +25,22 @@ import {
   LargeBinary,
   List,
   Null,
+  Precision,
   RecordBatch,
   RecordBatchFileReader,
   RecordBatchFileWriter,
   RecordBatchStreamWriter,
   Schema,
   Struct,
+  Timestamp,
+  Type,
   Utf8,
   Vector,
   makeVector as arrowMakeVector,
+  vectorFromArray as badVectorFromArray,
   makeBuilder,
   makeData,
   makeTable,
-  vectorFromArray,
 } from "apache-arrow";
 import { Buffers } from "apache-arrow/data";
 import { type EmbeddingFunction } from "./embedding/embedding_function";
@@ -179,6 +186,21 @@ export class VectorColumnOptions {
   }
 }
 
+// biome-ignore lint/suspicious/noExplicitAny: skip
+function vectorFromArray(data: any, type?: DataType) {
+  // Workaround for: https://github.com/apache/arrow/issues/45862
+  // If FSL type with float
+  if (DataType.isFixedSizeList(type) && DataType.isFloat(type.valueType)) {
+    const extendedData = [...data, new Array(type.listSize).fill(0.0)];
+    const array = badVectorFromArray(extendedData, type);
+    return array.slice(0, data.length);
+  } else if (type === undefined) {
+    return badVectorFromArray(data);
+  } else {
+    return badVectorFromArray(data, type);
+  }
+}
+
 /** Options to control the makeArrowTable call. */
 export class MakeArrowTableOptions {
   /*
@@ -1170,3 +1192,137 @@ function validateSchemaEmbeddings(
 
   return new Schema(fields, schema.metadata);
 }
+
+interface JsonDataType {
+  type: string;
+  fields?: JsonField[];
+  length?: number;
+}
+
+interface JsonField {
+  name: string;
+  type: JsonDataType;
+  nullable: boolean;
+  metadata: Map<string, string>;
+}
+
+// Matches format of https://github.com/lancedb/lance/blob/main/rust/lance/src/arrow/json.rs
+export function dataTypeToJson(dataType: DataType): JsonDataType {
+  switch (dataType.typeId) {
+    // For primitives, matches https://github.com/lancedb/lance/blob/e12bb9eff2a52f753668d4b62c52e4d72b10d294/rust/lance-core/src/datatypes.rs#L185
+    case Type.Null:
+      return { type: "null" };
+    case Type.Bool:
+      return { type: "bool" };
+    case Type.Int8:
+      return { type: "int8" };
+    case Type.Int16:
+      return { type: "int16" };
+    case Type.Int32:
+      return { type: "int32" };
+    case Type.Int64:
+      return { type: "int64" };
+    case Type.Uint8:
+      return { type: "uint8" };
+    case Type.Uint16:
+      return { type: "uint16" };
+    case Type.Uint32:
+      return { type: "uint32" };
+    case Type.Uint64:
+      return { type: "uint64" };
+    case Type.Int: {
+      const bitWidth = (dataType as Int).bitWidth;
+      const signed = (dataType as Int).isSigned;
+      const prefix = signed ? "" : "u";
+      return { type: `${prefix}int${bitWidth}` };
+    }
+    case Type.Float: {
+      switch ((dataType as Float).precision) {
+        case Precision.HALF:
+          return { type: "halffloat" };
+        case Precision.SINGLE:
+          return { type: "float" };
+        case Precision.DOUBLE:
+          return { type: "double" };
+      }
+      throw Error("Unsupported float precision");
+    }
+    case Type.Float16:
+      return { type: "halffloat" };
+    case Type.Float32:
+      return { type: "float" };
+    case Type.Float64:
+      return { type: "double" };
+    case Type.Utf8:
+      return { type: "string" };
+    case Type.Binary:
+      return { type: "binary" };
+    case Type.LargeUtf8:
+      return { type: "large_string" };
+    case Type.LargeBinary:
+      return { type: "large_binary" };
+    case Type.List:
+      return {
+        type: "list",
+        fields: [fieldToJson((dataType as List).children[0])],
+      };
+    case Type.FixedSizeList: {
+      const fixedSizeList = dataType as FixedSizeList;
+      return {
+        type: "fixed_size_list",
+        fields: [fieldToJson(fixedSizeList.children[0])],
+        length: fixedSizeList.listSize,
+      };
+    }
+    case Type.Struct:
+      return {
+        type: "struct",
+        fields: (dataType as Struct).children.map(fieldToJson),
+      };
+    case Type.Date: {
+      const unit = (dataType as Date_).unit;
+      return {
+        type: unit === DateUnit.DAY ? "date32:day" : "date64:ms",
+      };
+    }
+    case Type.Timestamp: {
+      const timestamp = dataType as Timestamp;
+      const timezone = timestamp.timezone || "-";
+      return {
+        type: `timestamp:${timestamp.unit}:${timezone}`,
+      };
+    }
+    case Type.Decimal: {
+      const decimal = dataType as Decimal;
+      return {
+        type: `decimal:${decimal.bitWidth}:${decimal.precision}:${decimal.scale}`,
+      };
+    }
+    case Type.Duration: {
+      const duration = dataType as Duration;
+      return { type: `duration:${duration.unit}` };
+    }
+    case Type.FixedSizeBinary: {
+      const byteWidth = (dataType as FixedSizeBinary).byteWidth;
+      return { type: `fixed_size_binary:${byteWidth}` };
+    }
+    case Type.Dictionary: {
+      const dict = dataType as Dictionary;
+      const indexType = dataTypeToJson(dict.indices);
+      const valueType = dataTypeToJson(dict.valueType);
+      return {
+        type: `dict:${valueType.type}:${indexType.type}:false`,
+      };
+    }
+  }
+  throw new Error("Unsupported data type");
+}
+
+function fieldToJson(field: Field): JsonField {
+  return {
+    name: field.name,
+    type: dataTypeToJson(field.type),
+    nullable: field.nullable,
+    metadata: field.metadata,
+  };
+}

nodejs/lancedb/index.ts

@@ -14,7 +14,6 @@ import {
 
 export {
   AddColumnsSql,
-  ColumnAlteration,
   ConnectionOptions,
   IndexStatistics,
   IndexConfig,
@@ -54,6 +53,7 @@ export {
   Index,
   IndexOptions,
   IvfPqOptions,
+  IvfFlatOptions,
   HnswPqOptions,
   HnswSqOptions,
   FtsOptions,
@@ -65,6 +65,7 @@ export {
   UpdateOptions,
   OptimizeOptions,
   Version,
+  ColumnAlteration,
 } from "./table";
 
 export { MergeInsertBuilder } from "./merge";
@@ -79,7 +80,7 @@ export {
   DataLike,
   IntoVector,
 } from "./arrow";
-export { IntoSql } from "./util";
+export { IntoSql, packBits } from "./util";
 
 /**
  * Connect to a LanceDB instance at the given URI.

nodejs/lancedb/indices.ts

@@ -62,13 +62,13 @@ export interface IvfPqOptions {
    *
    * "l2" - Euclidean distance. This is a very common distance metric that
    * accounts for both magnitude and direction when determining the distance
-   * between vectors. L2 distance has a range of [0, ∞).
+   * between vectors. l2 distance has a range of [0, ∞).
    *
    * "cosine" - Cosine distance.  Cosine distance is a distance metric
    * calculated from the cosine similarity between two vectors. Cosine
    * similarity is a measure of similarity between two non-zero vectors of an
    * inner product space. It is defined to equal the cosine of the angle
-   * between them.  Unlike L2, the cosine distance is not affected by the
+   * between them.  Unlike l2, the cosine distance is not affected by the
    * magnitude of the vectors.  Cosine distance has a range of [0, 2].
    *
    * Note: the cosine distance is undefined when one (or both) of the vectors
@@ -77,7 +77,7 @@ export interface IvfPqOptions {
    *
    * "dot" - Dot product. Dot distance is the dot product of two vectors. Dot
    * distance has a range of (-∞, ∞). If the vectors are normalized (i.e. their
-   * L2 norm is 1), then dot distance is equivalent to the cosine distance.
+   * l2 norm is 1), then dot distance is equivalent to the cosine distance.
    */
   distanceType?: "l2" | "cosine" | "dot";
 
@@ -125,18 +125,18 @@ export interface HnswPqOptions {
    *
    * "l2" - Euclidean distance. This is a very common distance metric that
    * accounts for both magnitude and direction when determining the distance
-   * between vectors. L2 distance has a range of [0, ∞).
+   * between vectors. l2 distance has a range of [0, ∞).
    *
    * "cosine" - Cosine distance.  Cosine distance is a distance metric
    * calculated from the cosine similarity between two vectors. Cosine
    * similarity is a measure of similarity between two non-zero vectors of an
    * inner product space. It is defined to equal the cosine of the angle
-   * between them.  Unlike L2, the cosine distance is not affected by the
+   * between them.  Unlike l2, the cosine distance is not affected by the
    * magnitude of the vectors.  Cosine distance has a range of [0, 2].
    *
    * "dot" - Dot product. Dot distance is the dot product of two vectors. Dot
    * distance has a range of (-∞, ∞). If the vectors are normalized (i.e. their
-   * L2 norm is 1), then dot distance is equivalent to the cosine distance.
+   * l2 norm is 1), then dot distance is equivalent to the cosine distance.
    */
   distanceType?: "l2" | "cosine" | "dot";
 
@@ -241,18 +241,18 @@ export interface HnswSqOptions {
    *
    * "l2" - Euclidean distance. This is a very common distance metric that
    * accounts for both magnitude and direction when determining the distance
-   * between vectors. L2 distance has a range of [0, ∞).
+   * between vectors. l2 distance has a range of [0, ∞).
    *
    * "cosine" - Cosine distance.  Cosine distance is a distance metric
    * calculated from the cosine similarity between two vectors. Cosine
    * similarity is a measure of similarity between two non-zero vectors of an
    * inner product space. It is defined to equal the cosine of the angle
-   * between them.  Unlike L2, the cosine distance is not affected by the
+   * between them.  Unlike l2, the cosine distance is not affected by the
    * magnitude of the vectors.  Cosine distance has a range of [0, 2].
    *
    * "dot" - Dot product. Dot distance is the dot product of two vectors. Dot
    * distance has a range of (-∞, ∞). If the vectors are normalized (i.e. their
-   * L2 norm is 1), then dot distance is equivalent to the cosine distance.
+   * l2 norm is 1), then dot distance is equivalent to the cosine distance.
    */
   distanceType?: "l2" | "cosine" | "dot";
 
@@ -327,6 +327,94 @@ export interface HnswSqOptions {
   efConstruction?: number;
 }
 
+/**
+ * Options to create an `IVF_FLAT` index
+ */
+export interface IvfFlatOptions {
+  /**
+   * The number of IVF partitions to create.
+   *
+   * This value should generally scale with the number of rows in the dataset.
+   * By default the number of partitions is the square root of the number of
+   * rows.
+   *
+   * If this value is too large then the first part of the search (picking the
+   * right partition) will be slow.  If this value is too small then the second
+   * part of the search (searching within a partition) will be slow.
+   */
+  numPartitions?: number;
+
+  /**
+   * Distance type to use to build the index.
+   *
+   * Default value is "l2".
+   *
+   * This is used when training the index to calculate the IVF partitions
+   * (vectors are grouped in partitions with similar vectors according to this
+   * distance type).
+   *
+   * The distance type used to train an index MUST match the distance type used
+   * to search the index.  Failure to do so will yield inaccurate results.
+   *
+   * The following distance types are available:
+   *
+   * "l2" - Euclidean distance. This is a very common distance metric that
+   * accounts for both magnitude and direction when determining the distance
+   * between vectors. l2 distance has a range of [0, ∞).
+   *
+   * "cosine" - Cosine distance.  Cosine distance is a distance metric
+   * calculated from the cosine similarity between two vectors. Cosine
+   * similarity is a measure of similarity between two non-zero vectors of an
+   * inner product space. It is defined to equal the cosine of the angle
+   * between them.  Unlike l2, the cosine distance is not affected by the
+   * magnitude of the vectors.  Cosine distance has a range of [0, 2].
+   *
+   * Note: the cosine distance is undefined when one (or both) of the vectors
+   * are all zeros (there is no direction).  These vectors are invalid and may
+   * never be returned from a vector search.
+   *
+   * "dot" - Dot product. Dot distance is the dot product of two vectors. Dot
+   * distance has a range of (-∞, ∞). If the vectors are normalized (i.e. their
+   * l2 norm is 1), then dot distance is equivalent to the cosine distance.
+   *
+   * "hamming" - Hamming distance. Hamming distance is a distance metric
+   * calculated from the number of bits that are different between two vectors.
+   * Hamming distance has a range of [0, dimension]. Note that the hamming distance
+   * is only valid for binary vectors.
+   */
+  distanceType?: "l2" | "cosine" | "dot" | "hamming";
+
+  /**
+   * Max iteration to train IVF kmeans.
+   *
+   * When training an IVF FLAT index we use kmeans to calculate the partitions.  This parameter
+   * controls how many iterations of kmeans to run.
+   *
+   * Increasing this might improve the quality of the index but in most cases these extra
+   * iterations have diminishing returns.
+   *
+   * The default value is 50.
+   */
+  maxIterations?: number;
+
+  /**
+   * The number of vectors, per partition, to sample when training IVF kmeans.
+   *
+   * When an IVF FLAT index is trained, we need to calculate partitions.  These are groups
+   * of vectors that are similar to each other.  To do this we use an algorithm called kmeans.
+   *
+   * Running kmeans on a large dataset can be slow.  To speed this up we run kmeans on a
+   * random sample of the data.  This parameter controls the size of the sample.  The total
+   * number of vectors used to train the index is `sample_rate * num_partitions`.
+   *
+   * Increasing this value might improve the quality of the index but in most cases the
+   * default should be sufficient.
+   *
+   * The default value is 256.
+   */
+  sampleRate?: number;
+}
+
 /**
  * Options to create a full text search index
  */
@@ -426,6 +514,33 @@ export class Index {
     );
   }
 
+  /**
+   * Create an IvfFlat index
+   *
+   * This index groups vectors into partitions of similar vectors.  Each partition keeps track of
+   * a centroid which is the average value of all vectors in the group.
+   *
+   * During a query the centroids are compared with the query vector to find the closest
+   * partitions.  The vectors in these partitions are then searched to find
+   * the closest vectors.
+   *
+   * The partitioning process is called IVF and the `num_partitions` parameter controls how
+   * many groups to create.
+   *
+   * Note that training an IVF FLAT index on a large dataset is a slow operation and
+   * currently is also a memory intensive operation.
+   */
+  static ivfFlat(options?: Partial<IvfFlatOptions>) {
+    return new Index(
+      LanceDbIndex.ivfFlat(
+        options?.distanceType,
+        options?.numPartitions,
+        options?.maxIterations,
+        options?.sampleRate,
+      ),
+    );
+  }
+
   /**
    * Create a btree index
    *

nodejs/lancedb/query.ts

@@ -348,6 +348,43 @@ export class QueryBase<NativeQueryType extends NativeQuery | NativeVectorQuery>
       return this.inner.explainPlan(verbose);
     }
   }
+
+  /**
+   * Executes the query and returns the physical query plan annotated with runtime metrics.
+   *
+   * This is useful for debugging and performance analysis, as it shows how the query was executed
+   * and includes metrics such as elapsed time, rows processed, and I/O statistics.
+   *
+   * @example
+   * import * as lancedb from "@lancedb/lancedb"
+   *
+   * const db = await lancedb.connect("./.lancedb");
+   * const table = await db.createTable("my_table", [
+   *   { vector: [1.1, 0.9], id: "1" },
+   * ]);
+   *
+   * const plan = await table.query().nearestTo([0.5, 0.2]).analyzePlan();
+   *
+   * Example output (with runtime metrics inlined):
+   * AnalyzeExec verbose=true, metrics=[]
+   *  ProjectionExec: expr=[id@3 as id, vector@0 as vector, _distance@2 as _distance], metrics=[output_rows=1, elapsed_compute=3.292µs]
+   *   Take: columns="vector, _rowid, _distance, (id)", metrics=[output_rows=1, elapsed_compute=66.001µs, batches_processed=1, bytes_read=8, iops=1, requests=1]
+   *    CoalesceBatchesExec: target_batch_size=1024, metrics=[output_rows=1, elapsed_compute=3.333µs]
+   *     GlobalLimitExec: skip=0, fetch=10, metrics=[output_rows=1, elapsed_compute=167ns]
+   *      FilterExec: _distance@2 IS NOT NULL, metrics=[output_rows=1, elapsed_compute=8.542µs]
+   *       SortExec: TopK(fetch=10), expr=[_distance@2 ASC NULLS LAST], metrics=[output_rows=1, elapsed_compute=63.25µs, row_replacements=1]
+   *        KNNVectorDistance: metric=l2, metrics=[output_rows=1, elapsed_compute=114.333µs, output_batches=1]
+   *         LanceScan: uri=/path/to/data, projection=[vector], row_id=true, row_addr=false, ordered=false, metrics=[output_rows=1, elapsed_compute=103.626µs, bytes_read=549, iops=2, requests=2]
+   *
+   * @returns A query execution plan with runtime metrics for each step.
+   */
+  async analyzePlan(): Promise<string> {
+    if (this.inner instanceof Promise) {
+      return this.inner.then((inner) => inner.analyzePlan());
+    } else {
+      return this.inner.analyzePlan();
+    }
+  }
 }
 
 /**

nodejs/lancedb/table.ts

@@ -4,8 +4,10 @@
 import {
   Table as ArrowTable,
   Data,
+  DataType,
   IntoVector,
   Schema,
+  dataTypeToJson,
   fromDataToBuffer,
   tableFromIPC,
 } from "./arrow";
@@ -15,13 +17,13 @@ import { IndexOptions } from "./indices";
 import { MergeInsertBuilder } from "./merge";
 import {
   AddColumnsSql,
-  ColumnAlteration,
   IndexConfig,
   IndexStatistics,
   OptimizeStats,
   Table as _NativeTable,
 } from "./native";
 import { Query, VectorQuery } from "./query";
+import { sanitizeType } from "./sanitize";
 import { IntoSql, toSQL } from "./util";
 export { IndexConfig } from "./native";
 
@@ -618,7 +620,27 @@ export class LocalTable extends Table {
   }
 
   async alterColumns(columnAlterations: ColumnAlteration[]): Promise<void> {
-    await this.inner.alterColumns(columnAlterations);
+    const processedAlterations = columnAlterations.map((alteration) => {
+      if (typeof alteration.dataType === "string") {
+        return {
+          ...alteration,
+          dataType: JSON.stringify({ type: alteration.dataType }),
+        };
+      } else if (alteration.dataType === undefined) {
+        return {
+          ...alteration,
+          dataType: undefined,
+        };
+      } else {
+        const dataType = sanitizeType(alteration.dataType);
+        return {
+          ...alteration,
+          dataType: JSON.stringify(dataTypeToJson(dataType)),
+        };
+      }
+    });
+
+    await this.inner.alterColumns(processedAlterations);
   }
 
   async dropColumns(columnNames: string[]): Promise<void> {
@@ -711,3 +733,38 @@ export class LocalTable extends Table {
     await this.inner.migrateManifestPathsV2();
   }
 }
+
+/**
+ *  A definition of a column alteration. The alteration changes the column at
+ * `path` to have the new name `name`, to be nullable if `nullable` is true,
+ * and to have the data type `data_type`. At least one of `rename` or `nullable`
+ * must be provided.
+ */
+export interface ColumnAlteration {
+  /**
+   * The path to the column to alter. This is a dot-separated path to the column.
+   * If it is a top-level column then it is just the name of the column. If it is
+   * a nested column then it is the path to the column, e.g. "a.b.c" for a column
+   * `c` nested inside a column `b` nested inside a column `a`.
+   */
+  path: string;
+  /**
+   * The new name of the column. If not provided then the name will not be changed.
+   * This must be distinct from the names of all other columns in the table.
+   */
+  rename?: string;
+  /**
+   * A new data type for the column. If not provided then the data type will not be changed.
+   * Changing data types is limited to casting to the same general type. For example, these
+   * changes are valid:
+   * * `int32` -> `int64` (integers)
+   * * `double` -> `float` (floats)
+   * * `string` -> `large_string` (strings)
+   * But these changes are not:
+   * * `int32` -> `double` (mix integers and floats)
+   * * `string` -> `int32` (mix strings and integers)
+   */
+  dataType?: string | DataType;
+  /** Set the new nullability. Note that a nullable column cannot be made non-nullable. */
+  nullable?: boolean;
+}

nodejs/lancedb/util.ts

@@ -35,6 +35,16 @@ export function toSQL(value: IntoSql): string {
   }
 }
 
+export function packBits(data: Array<number>): Array<number> {
+  const packed = Array(data.length >> 3).fill(0);
+  for (let i = 0; i < data.length; i++) {
+    const byte = i >> 3;
+    const bit = i & 7;
+    packed[byte] |= data[i] << bit;
+  }
+  return packed;
+}
+
 export class TTLCache {
   // biome-ignore lint/suspicious/noExplicitAny: <explanation>
   private readonly cache: Map<string, { value: any; expires: number }>;

nodejs/npm/darwin-arm64/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-darwin-arm64",
-	"version": "0.18.0",
+	"version": "0.18.3-beta.0",
 	"os": ["darwin"],
 	"cpu": ["arm64"],
 	"main": "lancedb.darwin-arm64.node",

nodejs/npm/darwin-x64/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-darwin-x64",
-	"version": "0.18.0",
+	"version": "0.18.3-beta.0",
 	"os": ["darwin"],
 	"cpu": ["x64"],
 	"main": "lancedb.darwin-x64.node",

nodejs/npm/linux-arm64-gnu/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-linux-arm64-gnu",
-	"version": "0.18.0",
+	"version": "0.18.3-beta.0",
 	"os": ["linux"],
 	"cpu": ["arm64"],
 	"main": "lancedb.linux-arm64-gnu.node",

nodejs/npm/linux-arm64-musl/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-linux-arm64-musl",
-	"version": "0.18.0",
+	"version": "0.18.3-beta.0",
 	"os": ["linux"],
 	"cpu": ["arm64"],
 	"main": "lancedb.linux-arm64-musl.node",

nodejs/npm/linux-x64-gnu/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-linux-x64-gnu",
-	"version": "0.18.0",
+	"version": "0.18.3-beta.0",
 	"os": ["linux"],
 	"cpu": ["x64"],
 	"main": "lancedb.linux-x64-gnu.node",

nodejs/npm/linux-x64-musl/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-linux-x64-musl",
-	"version": "0.18.0",
+	"version": "0.18.3-beta.0",
 	"os": ["linux"],
 	"cpu": ["x64"],
 	"main": "lancedb.linux-x64-musl.node",

nodejs/npm/win32-arm64-msvc/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@lancedb/lancedb-win32-arm64-msvc",
-  "version": "0.18.0",
+  "version": "0.18.3-beta.0",
   "os": [
     "win32"
   ],

nodejs/npm/win32-x64-msvc/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-win32-x64-msvc",
-	"version": "0.18.0",
+	"version": "0.18.3-beta.0",
 	"os": ["win32"],
 	"cpu": ["x64"],
 	"main": "lancedb.win32-x64-msvc.node",

nodejs/package-lock.json

@@ -1,12 +1,12 @@
 {
   "name": "@lancedb/lancedb",
-  "version": "0.18.0",
+  "version": "0.18.3-beta.0",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {
     "": {
       "name": "@lancedb/lancedb",
-      "version": "0.18.0",
+      "version": "0.18.3-beta.0",
       "cpu": [
         "x64",
         "arm64"

nodejs/package.json

@@ -11,7 +11,7 @@
     "ann"
   ],
   "private": false,
-  "version": "0.18.0",
+  "version": "0.18.3-beta.0",
   "main": "dist/index.js",
   "exports": {
     ".": "./dist/index.js",
@@ -29,7 +29,6 @@
         "aarch64-apple-darwin",
         "x86_64-unknown-linux-gnu",
         "aarch64-unknown-linux-gnu",
-        "x86_64-unknown-linux-musl",
         "aarch64-unknown-linux-musl",
         "x86_64-pc-windows-msvc",
         "aarch64-pc-windows-msvc"
@@ -74,8 +73,10 @@
     "artifacts": "napi artifacts",
     "build:debug": "napi build --platform --no-const-enum --dts ../lancedb/native.d.ts --js ../lancedb/native.js lancedb",
     "build:release": "napi build --platform --no-const-enum --release --dts ../lancedb/native.d.ts --js ../lancedb/native.js dist/",
-    "build": "npm run build:debug && tsc -b && shx cp lancedb/native.d.ts dist/native.d.ts && shx cp lancedb/*.node dist/",
-    "build-release": "npm run build:release && tsc -b && shx cp lancedb/native.d.ts dist/native.d.ts",
+    "build": "npm run build:debug && npm run tsc && shx cp lancedb/*.node dist/",
+    "build-release": "npm run build:release && npm run tsc",
+    "tsc": "tsc -b",
+    "posttsc": "shx cp lancedb/native.d.ts dist/native.d.ts",
     "lint-ci": "biome ci .",
     "docs": "typedoc --plugin typedoc-plugin-markdown --treatWarningsAsErrors --out ../docs/src/js lancedb/index.ts",
     "postdocs": "node typedoc_post_process.js",

nodejs/src/connection.rs

@@ -48,8 +48,16 @@ impl Connection {
     pub async fn new(uri: String, options: ConnectionOptions) -> napi::Result<Self> {
         let mut builder = ConnectBuilder::new(&uri);
         if let Some(interval) = options.read_consistency_interval {
-            builder =
-                builder.read_consistency_interval(std::time::Duration::from_secs_f64(interval));
+            match interval {
+                Either::A(seconds) => {
+                    builder = builder.read_consistency_interval(Some(
+                        std::time::Duration::from_secs_f64(seconds),
+                    ));
+                }
+                Either::B(_) => {
+                    builder = builder.read_consistency_interval(None);
+                }
+            }
         }
         if let Some(storage_options) = options.storage_options {
             for (key, value) in storage_options {

nodejs/src/index.rs

@@ -4,7 +4,9 @@
 use std::sync::Mutex;
 
 use lancedb::index::scalar::{BTreeIndexBuilder, FtsIndexBuilder};
-use lancedb::index::vector::{IvfHnswPqIndexBuilder, IvfHnswSqIndexBuilder, IvfPqIndexBuilder};
+use lancedb::index::vector::{
+    IvfFlatIndexBuilder, IvfHnswPqIndexBuilder, IvfHnswSqIndexBuilder, IvfPqIndexBuilder,
+};
 use lancedb::index::Index as LanceDbIndex;
 use napi_derive::napi;
 
@@ -63,6 +65,32 @@ impl Index {
         })
     }
 
+    #[napi(factory)]
+    pub fn ivf_flat(
+        distance_type: Option<String>,
+        num_partitions: Option<u32>,
+        max_iterations: Option<u32>,
+        sample_rate: Option<u32>,
+    ) -> napi::Result<Self> {
+        let mut ivf_flat_builder = IvfFlatIndexBuilder::default();
+        if let Some(distance_type) = distance_type {
+            let distance_type = parse_distance_type(distance_type)?;
+            ivf_flat_builder = ivf_flat_builder.distance_type(distance_type);
+        }
+        if let Some(num_partitions) = num_partitions {
+            ivf_flat_builder = ivf_flat_builder.num_partitions(num_partitions);
+        }
+        if let Some(max_iterations) = max_iterations {
+            ivf_flat_builder = ivf_flat_builder.max_iterations(max_iterations);
+        }
+        if let Some(sample_rate) = sample_rate {
+            ivf_flat_builder = ivf_flat_builder.sample_rate(sample_rate);
+        }
+        Ok(Self {
+            inner: Mutex::new(Some(LanceDbIndex::IvfFlat(ivf_flat_builder))),
+        })
+    }
+
     #[napi(factory)]
     pub fn btree() -> Self {
         Self {

nodejs/src/lib.rs

@@ -4,6 +4,7 @@
 use std::collections::HashMap;
 
 use env_logger::Env;
+use napi::{bindgen_prelude::Null, Either};
 use napi_derive::*;
 
 mod connection;
@@ -18,7 +19,6 @@ mod table;
 mod util;
 
 #[napi(object)]
-#[derive(Debug)]
 pub struct ConnectionOptions {
     /// (For LanceDB OSS only): The interval, in seconds, at which to check for
     /// updates to the table from other processes. If None, then consistency is not
@@ -29,7 +29,7 @@ pub struct ConnectionOptions {
     /// has passed since the last check, then the table will be checked for updates.
     /// Note: this consistency only applies to read operations. Write operations are
     /// always consistent.
-    pub read_consistency_interval: Option<f64>,
+    pub read_consistency_interval: Option<Either<f64, Null>>,
     /// (For LanceDB OSS only): configuration for object storage.
     ///
     /// The available options are described at https://lancedb.github.io/lancedb/guides/storage/

nodejs/src/query.rs

@@ -114,6 +114,16 @@ impl Query {
             ))
         })
     }
+
+    #[napi(catch_unwind)]
+    pub async fn analyze_plan(&self) -> napi::Result<String> {
+        self.inner.analyze_plan().await.map_err(|e| {
+            napi::Error::from_reason(format!(
+                "Failed to execute analyze plan: {}",
+                convert_error(&e)
+            ))
+        })
+    }
 }
 
 #[napi]
@@ -259,4 +269,14 @@ impl VectorQuery {
             ))
         })
     }
+
+    #[napi(catch_unwind)]
+    pub async fn analyze_plan(&self) -> napi::Result<String> {
+        self.inner.analyze_plan().await.map_err(|e| {
+            napi::Error::from_reason(format!(
+                "Failed to execute analyze plan: {}",
+                convert_error(&e)
+            ))
+        })
+    }
 }

nodejs/src/table.rs

@@ -498,6 +498,9 @@ pub struct IndexStatistics {
     pub distance_type: Option<String>,
     /// The number of parts this index is split into.
     pub num_indices: Option<u32>,
+    /// The KMeans loss value of the index,
+    /// it is only present for vector indices.
+    pub loss: Option<f64>,
 }
 impl From<lancedb::index::IndexStatistics> for IndexStatistics {
     fn from(value: lancedb::index::IndexStatistics) -> Self {
@@ -507,6 +510,7 @@ impl From<lancedb::index::IndexStatistics> for IndexStatistics {
             index_type: value.index_type.to_string(),
             distance_type: value.distance_type.map(|d| d.to_string()),
             num_indices: value.num_indices,
+            loss: value.loss,
         }
     }
 }

python/.bumpversion.toml

@@ -1,5 +1,5 @@
 [tool.bumpversion]
-current_version = "0.21.1"
+current_version = "0.22.0-beta.0"
 parse = """(?x)
     (?P<major>0|[1-9]\\d*)\\.
     (?P<minor>0|[1-9]\\d*)\\.

python/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "lancedb-python"
-version = "0.21.1"
+version = "0.22.0-beta.0"
 edition.workspace = true
 description = "Python bindings for LanceDB"
 license.workspace = true
@@ -33,10 +33,6 @@ pyo3-build-config = { version = "0.23", features = [
 ] }
 
 [features]
-default = ["default-tls", "remote"]
+default = ["remote"]
 fp16kernels = ["lancedb/fp16kernels"]
 remote = ["lancedb/remote"]
-# TLS
-default-tls = ["lancedb/default-tls"]
-native-tls = ["lancedb/native-tls"]
-rustls-tls = ["lancedb/rustls-tls"]

python/pyproject.toml

@@ -9,7 +9,6 @@ dependencies = [
     "pydantic>=1.10",
     "packaging",
     "overrides>=0.7",
-    "pylance>=0.23.2",
 ]
 description = "lancedb"
 authors = [{ name = "LanceDB Devs", email = "dev@lancedb.com" }]
@@ -55,6 +54,7 @@ tests = [
     "polars>=0.19, <=1.3.0",
     "tantivy",
     "pyarrow-stubs",
+    "pylance>=0.23.2",
 ]
 dev = [
     "ruff",
@@ -63,7 +63,7 @@ dev = [
     'typing-extensions>=4.0.0; python_version < "3.11"',
 ]
 docs = ["mkdocs", "mkdocs-jupyter", "mkdocs-material", "mkdocstrings[python]"]
-clip = ["torch", "pillow", "open-clip"]
+clip = ["torch", "pillow", "open-clip-torch"]
 embeddings = [
     "requests>=2.31.0",
     "openai>=1.6.1",

python/python/lancedb/__init__.py

@@ -7,6 +7,7 @@ import os
 from concurrent.futures import ThreadPoolExecutor
 from datetime import timedelta
 from typing import Dict, Optional, Union, Any
+import warnings
 
 __version__ = importlib.metadata.version("lancedb")
 
@@ -25,7 +26,7 @@ def connect(
     api_key: Optional[str] = None,
     region: str = "us-east-1",
     host_override: Optional[str] = None,
-    read_consistency_interval: Optional[timedelta] = None,
+    read_consistency_interval: Optional[timedelta] = timedelta(seconds=5),
     request_thread_pool: Optional[Union[int, ThreadPoolExecutor]] = None,
     client_config: Union[ClientConfig, Dict[str, Any], None] = None,
     storage_options: Optional[Dict[str, str]] = None,
@@ -48,9 +49,8 @@ def connect(
     read_consistency_interval: timedelta, default None
         (For LanceDB OSS only)
         The interval at which to check for updates to the table from other
-        processes. If None, then consistency is not checked. For performance
-        reasons, this is the default. For strong consistency, set this to
-        zero seconds. Then every read will check for updates from other
+        processes. If None, then consistency is not checked. For strong consistency,
+        set this to zero seconds. Then every read will check for updates from other
         processes. As a compromise, you can set this to a non-zero timedelta
         for eventual consistency. If more than that interval has passed since
         the last check, then the table will be checked for updates. Note: this
@@ -121,7 +121,7 @@ async def connect_async(
     api_key: Optional[str] = None,
     region: str = "us-east-1",
     host_override: Optional[str] = None,
-    read_consistency_interval: Optional[timedelta] = None,
+    read_consistency_interval: Optional[timedelta] = timedelta(seconds=5),
     client_config: Optional[Union[ClientConfig, Dict[str, Any]]] = None,
     storage_options: Optional[Dict[str, str]] = None,
 ) -> AsyncConnection:
@@ -142,9 +142,8 @@ async def connect_async(
     read_consistency_interval: timedelta, default None
         (For LanceDB OSS only)
         The interval at which to check for updates to the table from other
-        processes. If None, then consistency is not checked. For performance
-        reasons, this is the default. For strong consistency, set this to
-        zero seconds. Then every read will check for updates from other
+        processes. If None, then consistency is not checked. For strong consistency,
+        set this to zero seconds. Then every read will check for updates from other
         processes. As a compromise, you can set this to a non-zero timedelta
         for eventual consistency. If more than that interval has passed since
         the last check, then the table will be checked for updates. Note: this
@@ -213,3 +212,13 @@ __all__ = [
     "RemoteDBConnection",
     "__version__",
 ]
+
+
+def __warn_on_fork():
+    warnings.warn(
+        "lance is not fork-safe. If you are using multiprocessing, use spawn instead.",
+    )
+
+
+if hasattr(os, "register_at_fork"):
+    os.register_at_fork(before=__warn_on_fork)

python/python/lancedb/_lancedb.pyi

@@ -48,10 +48,11 @@ class Table:
     async def version(self) -> int: ...
     async def checkout(self, version: int): ...
     async def checkout_latest(self): ...
-    async def restore(self): ...
+    async def restore(self, version: Optional[int] = None): ...
     async def list_indices(self) -> list[IndexConfig]: ...
     async def delete(self, filter: str): ...
     async def add_columns(self, columns: list[tuple[str, str]]) -> None: ...
+    async def add_columns_with_schema(self, schema: pa.Schema) -> None: ...
     async def alter_columns(self, columns: list[dict[str, Any]]) -> None: ...
     async def optimize(
         self,
@@ -94,6 +95,9 @@ class Query:
     def nearest_to(self, query_vec: pa.Array) -> VectorQuery: ...
     def nearest_to_text(self, query: dict) -> FTSQuery: ...
     async def execute(self, max_batch_length: Optional[int]) -> RecordBatchStream: ...
+    async def explain_plan(self, verbose: Optional[bool]) -> str: ...
+    async def analyze_plan(self) -> str: ...
+    def to_query_request(self) -> PyQueryRequest: ...
 
 class FTSQuery:
     def where(self, filter: str): ...
@@ -107,7 +111,7 @@ class FTSQuery:
     def add_query_vector(self, query_vec: pa.Array) -> None: ...
     def nearest_to(self, query_vec: pa.Array) -> HybridQuery: ...
     async def execute(self, max_batch_length: Optional[int]) -> RecordBatchStream: ...
-    async def explain_plan(self) -> str: ...
+    def to_query_request(self) -> PyQueryRequest: ...
 
 class VectorQuery:
     async def execute(self) -> RecordBatchStream: ...
@@ -123,6 +127,7 @@ class VectorQuery:
     def nprobes(self, nprobes: int): ...
     def bypass_vector_index(self): ...
     def nearest_to_text(self, query: dict) -> HybridQuery: ...
+    def to_query_request(self) -> PyQueryRequest: ...
 
 class HybridQuery:
     def where(self, filter: str): ...
@@ -140,6 +145,33 @@ class HybridQuery:
     def to_fts_query(self) -> FTSQuery: ...
     def get_limit(self) -> int: ...
     def get_with_row_id(self) -> bool: ...
+    def to_query_request(self) -> PyQueryRequest: ...
+
+class PyFullTextSearchQuery:
+    columns: Optional[List[str]]
+    query: str
+    limit: Optional[int]
+    wand_factor: Optional[float]
+
+class PyQueryRequest:
+    limit: Optional[int]
+    offset: Optional[int]
+    filter: Optional[Union[str, bytes]]
+    full_text_search: Optional[PyFullTextSearchQuery]
+    select: Optional[Union[str, List[str]]]
+    fast_search: Optional[bool]
+    with_row_id: Optional[bool]
+    column: Optional[str]
+    query_vector: Optional[List[pa.Array]]
+    nprobes: Optional[int]
+    lower_bound: Optional[float]
+    upper_bound: Optional[float]
+    ef: Optional[int]
+    refine_factor: Optional[int]
+    distance_type: Optional[str]
+    bypass_vector_index: Optional[bool]
+    postfilter: Optional[bool]
+    norm: Optional[str]
 
 class CompactionStats:
     fragments_removed: int

python/python/lancedb/common.py

@@ -7,10 +7,9 @@ from typing import Iterable, List, Optional, Union
 
 import numpy as np
 import pyarrow as pa
+import pyarrow.dataset
 
-from .util import safe_import_pandas
-
-pd = safe_import_pandas()
+from .dependencies import pandas as pd
 
 DATA = Union[List[dict], "pd.DataFrame", pa.Table, Iterable[pa.RecordBatch]]
 VEC = Union[list, np.ndarray, pa.Array, pa.ChunkedArray]

python/python/lancedb/context.py

@@ -8,9 +8,7 @@ import deprecation
 
 from . import __version__
 from .exceptions import MissingColumnError, MissingValueError
-from .util import safe_import_pandas
-
-pd = safe_import_pandas()
+from .dependencies import pandas as pd
 
 
 def contextualize(raw_df: "pd.DataFrame") -> Contextualizer:

python/python/lancedb/db.py

@@ -6,6 +6,7 @@ from __future__ import annotations
 
 from abc import abstractmethod
 from pathlib import Path
+from datetime import timedelta
 from typing import TYPE_CHECKING, Dict, Iterable, List, Literal, Optional, Union
 
 from lancedb.embeddings.registry import EmbeddingFunctionRegistry
@@ -32,7 +33,6 @@ import deprecation
 if TYPE_CHECKING:
     import pyarrow as pa
     from .pydantic import LanceModel
-    from datetime import timedelta
 
     from ._lancedb import Connection as LanceDbConnection
     from .common import DATA, URI
@@ -318,9 +318,8 @@ class LanceDBConnection(DBConnection):
         The root uri of the database.
     read_consistency_interval: timedelta, default None
         The interval at which to check for updates to the table from other
-        processes. If None, then consistency is not checked. For performance
-        reasons, this is the default. For strong consistency, set this to
-        zero seconds. Then every read will check for updates from other
+        processes. If None, then consistency is not checked. For strong consistency,
+        set this to zero seconds. Then every read will check for updates from other
         processes. As a compromise, you can set this to a non-zero timedelta
         for eventual consistency. If more than that interval has passed since
         the last check, then the table will be checked for updates. Note: this
@@ -352,7 +351,7 @@ class LanceDBConnection(DBConnection):
         self,
         uri: URI,
         *,
-        read_consistency_interval: Optional[timedelta] = None,
+        read_consistency_interval: Optional[timedelta] = timedelta(seconds=5),
         storage_options: Optional[Dict[str, str]] = None,
     ):
         if not isinstance(uri, Path):

python/python/lancedb/dependencies.py

@@ -30,6 +30,7 @@ _TORCH_AVAILABLE = True
 _HUGGING_FACE_AVAILABLE = True
 _TENSORFLOW_AVAILABLE = True
 _RAY_AVAILABLE = True
+_LANCE_AVAILABLE = True
 
 
 class _LazyModule(ModuleType):
@@ -53,6 +54,7 @@ class _LazyModule(ModuleType):
         "torch": "torch.",
         "tensorflow": "tf.",
         "ray": "ray.",
+        "lance": "lance.",
     }
 
     def __init__(
@@ -169,6 +171,7 @@ if TYPE_CHECKING:
     import ray
     import tensorflow
     import torch
+    import lance
 else:
     # heavy/optional third party libs
     numpy, _NUMPY_AVAILABLE = _lazy_import("numpy")
@@ -178,6 +181,7 @@ else:
     datasets, _HUGGING_FACE_AVAILABLE = _lazy_import("datasets")
     tensorflow, _TENSORFLOW_AVAILABLE = _lazy_import("tensorflow")
     ray, _RAY_AVAILABLE = _lazy_import("ray")
+    lance, _LANCE_AVAILABLE = _lazy_import("lance")
 
 
 @lru_cache(maxsize=None)
@@ -232,6 +236,12 @@ def _check_for_ray(obj: Any, *, check_type: bool = True) -> bool:
     )
 
 
+def _check_for_lance(obj: Any, *, check_type: bool = True) -> bool:
+    return _LANCE_AVAILABLE and _might_be(
+        cast(Hashable, type(obj) if check_type else obj), "lance"
+    )
+
+
 __all__ = [
     # lazy-load third party libs
     "datasets",
@@ -241,6 +251,7 @@ __all__ = [
     "ray",
     "tensorflow",
     "torch",
+    "lance",
     # lazy utilities
     "_check_for_hugging_face",
     "_check_for_numpy",
@@ -249,6 +260,7 @@ __all__ = [
     "_check_for_tensorflow",
     "_check_for_torch",
     "_check_for_ray",
+    "_check_for_lance",
     "_LazyModule",
     # exported flags/guards
     "_NUMPY_AVAILABLE",
@@ -258,4 +270,5 @@ __all__ = [
     "_HUGGING_FACE_AVAILABLE",
     "_TENSORFLOW_AVAILABLE",
     "_RAY_AVAILABLE",
+    "_LANCE_AVAILABLE",
 ]

python/python/lancedb/embeddings/__init__.py

@@ -16,7 +16,6 @@ from .sentence_transformers import SentenceTransformerEmbeddings
 from .gte import GteEmbeddings
 from .transformers import TransformersEmbeddingFunction, ColbertEmbeddings
 from .imagebind import ImageBindEmbeddings
-from .utils import with_embeddings
 from .jinaai import JinaEmbeddings
 from .watsonx import WatsonxEmbeddings
 from .voyageai import VoyageAIEmbeddingFunction

python/python/lancedb/embeddings/base.py

@@ -17,12 +17,13 @@ class EmbeddingFunction(BaseModel, ABC):
     """
     An ABC for embedding functions.
 
-    All concrete embedding functions must implement the following:
+    All concrete embedding functions must implement the following methods:
     1. compute_query_embeddings() which takes a query and returns a list of embeddings
-    2. get_source_embeddings() which returns a list of embeddings for the source column
+    2. compute_source_embeddings() which returns a list of embeddings for
+       the source column
     For text data, the two will be the same. For multi-modal data, the source column
     might be images and the vector column might be text.
-    3. ndims method which returns the number of dimensions of the vector column
+    3. ndims() which returns the number of dimensions of the vector column
     """
 
     __slots__ = ("__weakref__",)  # pydantic 1.x compatibility

python/python/lancedb/embeddings/utils.py

@@ -16,9 +16,8 @@ from functools import wraps
 from typing import Callable, List, Union
 import numpy as np
 import pyarrow as pa
-from lance.vector import vec_to_table
 
-from ..util import deprecated, safe_import_pandas
+from ..dependencies import pandas as pd
 
 
 # ruff: noqa: PERF203
@@ -41,8 +40,6 @@ def retry(tries=10, delay=1, max_delay=30, backoff=3, jitter=1):
     return wrapper
 
 
-pd = safe_import_pandas()
-
 DATA = Union[pa.Table, "pd.DataFrame"]
 TEXT = Union[str, List[str], pa.Array, pa.ChunkedArray, np.ndarray]
 IMAGES = Union[
@@ -87,52 +84,6 @@ class RateLimiter:
         return wrapper
 
 
-@deprecated
-def with_embeddings(
-    func: Callable,
-    data: DATA,
-    column: str = "text",
-    wrap_api: bool = True,
-    show_progress: bool = False,
-    batch_size: int = 1000,
-) -> pa.Table:
-    """Add a vector column to a table using the given embedding function.
-
-    The new columns will be called "vector".
-
-    Parameters
-    ----------
-    func : Callable
-        A function that takes a list of strings and returns a list of vectors.
-    data : pa.Table or pd.DataFrame
-        The data to add an embedding column to.
-    column : str, default "text"
-        The name of the column to use as input to the embedding function.
-    wrap_api : bool, default True
-        Whether to wrap the embedding function in a retry and rate limiter.
-    show_progress : bool, default False
-        Whether to show a progress bar.
-    batch_size : int, default 1000
-        The number of row values to pass to each call of the embedding function.
-
-    Returns
-    -------
-    pa.Table
-        The input table with a new column called "vector" containing the embeddings.
-    """
-    func = FunctionWrapper(func)
-    if wrap_api:
-        func = func.retry().rate_limit()
-    func = func.batch_size(batch_size)
-    if show_progress:
-        func = func.show_progress()
-    if pd is not None and isinstance(data, pd.DataFrame):
-        data = pa.Table.from_pandas(data, preserve_index=False)
-    embeddings = func(data[column].to_numpy())
-    table = vec_to_table(np.array(embeddings))
-    return data.append_column("vector", table["vector"])
-
-
 class FunctionWrapper:
     """
     A wrapper for embedding functions that adds rate limiting, retries, and batching.