Bump version: 0.24.1-beta.2 → 0.24.1

This is preparation for a stable release

.bumpversion.toml

@@ -1,5 +1,5 @@
 [tool.bumpversion]
-current_version = "0.20.1-beta.1"
+current_version = "0.21.1-beta.1"
 parse = """(?x)
     (?P<major>0|[1-9]\\d*)\\.
     (?P<minor>0|[1-9]\\d*)\\.

.github/workflows/npm-publish.yml

@@ -550,6 +550,9 @@ jobs:
           bash ci/update_lockfiles.sh
       - name: Push new commit
         uses: ad-m/github-push-action@master
+        with:
+          github_token: ${{ secrets.LANCEDB_RELEASE_TOKEN }}
+          branch: main
       - name: Notify Slack Action
         uses: ravsamhq/notify-slack-action@2.3.0
         if: ${{ always() }}

Cargo.toml

@@ -21,14 +21,14 @@ categories = ["database-implementations"]
 rust-version = "1.78.0"
 
 [workspace.dependencies]
-lance = { "version" = "=0.30.0", "features" = ["dynamodb"], tag = "v0.30.0-beta.1", git="https://github.com/lancedb/lance.git" }
+lance = { "version" = "=0.31.1", features = ["dynamodb"] }
-lance-io = { version = "=0.30.0", tag = "v0.30.0-beta.1", git="https://github.com/lancedb/lance.git" }
+lance-io = { "version" = "=0.31.1" }
-lance-index = { version = "=0.30.0", tag = "v0.30.0-beta.1", git="https://github.com/lancedb/lance.git" }
+lance-index = { "version" = "=0.31.1" }
-lance-linalg = { version = "=0.30.0", tag = "v0.30.0-beta.1", git="https://github.com/lancedb/lance.git" }
+lance-linalg = { "version" = "=0.31.1" }
-lance-table = { version = "=0.30.0", tag = "v0.30.0-beta.1", git="https://github.com/lancedb/lance.git" }
+lance-table = { "version" = "=0.31.1" }
-lance-testing = { version = "=0.30.0", tag = "v0.30.0-beta.1", git="https://github.com/lancedb/lance.git" }
+lance-testing = { "version" = "=0.31.1" }
-lance-datafusion = { version = "=0.30.0", tag = "v0.30.0-beta.1", git="https://github.com/lancedb/lance.git" }
+lance-datafusion = { "version" = "=0.31.1" }
-lance-encoding = { version = "=0.30.0", tag = "v0.30.0-beta.1", git="https://github.com/lancedb/lance.git" }
+lance-encoding = { "version" = "=0.31.1" }
 # Note that this one does not include pyarrow
 arrow = { version = "55.1", optional = false }
 arrow-array = "55.1"
@@ -39,20 +39,20 @@ arrow-schema = "55.1"
 arrow-arith = "55.1"
 arrow-cast = "55.1"
 async-trait = "0"
-datafusion = { version = "47.0", default-features = false }
+datafusion = { version = "48.0", default-features = false }
-datafusion-catalog = "47.0"
+datafusion-catalog = "48.0"
-datafusion-common = { version = "47.0", default-features = false }
+datafusion-common = { version = "48.0", default-features = false }
-datafusion-execution = "47.0"
+datafusion-execution = "48.0"
-datafusion-expr = "47.0"
+datafusion-expr = "48.0"
-datafusion-physical-plan = "47.0"
+datafusion-physical-plan = "48.0"
 env_logger = "0.11"
-half = { "version" = "=2.5.0", default-features = false, features = [
+half = { "version" = "2.6.0", default-features = false, features = [
     "num-traits",
 ] }
 futures = "0"
 log = "0.4"
 moka = { version = "0.12", features = ["future"] }
-object_store = "0.11.0"
+object_store = "0.12.0"
 pin-project = "1.0.7"
 snafu = "0.8"
 url = "2"

docs/src/notebooks/Multivector_on_LanceDB.ipynb

@@ -428,7 +428,7 @@
         "\n",
         "**Why?**  \n",
         "Embedding the UFO dataset and ingesting it into LanceDB takes **~2 hours on a T4 GPU**. To save time:  \n",
-        "- **Use the pre-prepared table with index created ** (provided below) to proceed directly to step7: search.  \n",
+        "- **Use the pre-prepared table with index created** (provided below) to proceed directly to **Step 7**: search.  \n",
         "- **Step 5a** contains the full ingestion code for reference (run it only if necessary).  \n",
         "- **Step 6** contains the details on creating the index on the multivector column"
       ]

java/core/pom.xml

@@ -8,7 +8,7 @@
     <parent>
         <groupId>com.lancedb</groupId>
         <artifactId>lancedb-parent</artifactId>
-        <version>0.20.1-beta.1</version>
+        <version>0.21.1-beta.1</version>
         <relativePath>../pom.xml</relativePath>
     </parent>
 

java/pom.xml

@@ -6,7 +6,7 @@
 
     <groupId>com.lancedb</groupId>
     <artifactId>lancedb-parent</artifactId>
-    <version>0.20.1-beta.1</version>
+    <version>0.21.1-beta.1</version>
     <packaging>pom</packaging>
 
     <name>LanceDB Parent</name>

node/package-lock.json

@@ -1,12 +1,12 @@
 {
   "name": "vectordb",
-  "version": "0.20.1-beta.1",
+  "version": "0.21.1-beta.1",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {
     "": {
       "name": "vectordb",
-      "version": "0.20.1-beta.1",
+      "version": "0.21.1-beta.1",
       "cpu": [
         "x64",
         "arm64"
@@ -52,11 +52,11 @@
         "uuid": "^9.0.0"
       },
       "optionalDependencies": {
-        "@lancedb/vectordb-darwin-arm64": "0.20.1-beta.1",
+        "@lancedb/vectordb-darwin-arm64": "0.21.1-beta.1",
-        "@lancedb/vectordb-darwin-x64": "0.20.1-beta.1",
+        "@lancedb/vectordb-darwin-x64": "0.21.1-beta.1",
-        "@lancedb/vectordb-linux-arm64-gnu": "0.20.1-beta.1",
+        "@lancedb/vectordb-linux-arm64-gnu": "0.21.1-beta.1",
-        "@lancedb/vectordb-linux-x64-gnu": "0.20.1-beta.1",
+        "@lancedb/vectordb-linux-x64-gnu": "0.21.1-beta.1",
-        "@lancedb/vectordb-win32-x64-msvc": "0.20.1-beta.1"
+        "@lancedb/vectordb-win32-x64-msvc": "0.21.1-beta.1"
       },
       "peerDependencies": {
         "@apache-arrow/ts": "^14.0.2",
@@ -327,9 +327,9 @@
       }
     },
     "node_modules/@lancedb/vectordb-darwin-arm64": {
-      "version": "0.20.1-beta.1",
+      "version": "0.21.1-beta.1",
-      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-darwin-arm64/-/vectordb-darwin-arm64-0.20.1-beta.1.tgz",
+      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-darwin-arm64/-/vectordb-darwin-arm64-0.21.1-beta.1.tgz",
-      "integrity": "sha512-DPD8gwFQz5aENYYbTFS/l3YX/rqzS6Kj2B4IZERccVFULQsdR5YwtaAfFwTMp7NSnsjWKwJAknohiMZlJr4njQ==",
+      "integrity": "sha512-D9SOLFb/40E2/9bt82xOti3jogRAaR1UkT2LfGZJw/0wBu8d8/xKjWgfm3d26S5K6in6DWsX1njLxevrFqD5HA==",
       "cpu": [
         "arm64"
       ],
@@ -339,9 +339,9 @@
       ]
     },
     "node_modules/@lancedb/vectordb-darwin-x64": {
-      "version": "0.20.1-beta.1",
+      "version": "0.21.1-beta.1",
-      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-darwin-x64/-/vectordb-darwin-x64-0.20.1-beta.1.tgz",
+      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-darwin-x64/-/vectordb-darwin-x64-0.21.1-beta.1.tgz",
-      "integrity": "sha512-lTPtlRSTC08UgQW5Bv8WYhdbogAgUJ+9ejg+UE+fwP9gEsgEKXL/SHBm+9gmAlTo7LbrxJjg0CtCde/mW68UTw==",
+      "integrity": "sha512-JnZ41aDOJs6LWfI9t/+MnpqsK/Fj9r/hDdZSOjcQquLOcm2eP3NnvEnDvn+1pqWBN6ceqf1avTatPBGnD/yhNA==",
       "cpu": [
         "x64"
       ],
@@ -351,9 +351,9 @@
       ]
     },
     "node_modules/@lancedb/vectordb-linux-arm64-gnu": {
-      "version": "0.20.1-beta.1",
+      "version": "0.21.1-beta.1",
-      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-linux-arm64-gnu/-/vectordb-linux-arm64-gnu-0.20.1-beta.1.tgz",
+      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-linux-arm64-gnu/-/vectordb-linux-arm64-gnu-0.21.1-beta.1.tgz",
-      "integrity": "sha512-w/3O9FvwQiGegYsM21yZ0FezfOFVsW7HttYwwPzZMZaCpK3/i+LvZVSqwO4qXHHJBtHgKevonINyvVlg5487aQ==",
+      "integrity": "sha512-Xnw0wYtnfzVUr4DzppJCSx+HZdAHr6sqMC8SdaYNQ9XEjBZE20n5SO2AdBYjejbmONJ7lpGs3ydnLIZ6N40dAQ==",
       "cpu": [
         "arm64"
       ],
@@ -363,9 +363,9 @@
       ]
     },
     "node_modules/@lancedb/vectordb-linux-x64-gnu": {
-      "version": "0.20.1-beta.1",
+      "version": "0.21.1-beta.1",
-      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-linux-x64-gnu/-/vectordb-linux-x64-gnu-0.20.1-beta.1.tgz",
+      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-linux-x64-gnu/-/vectordb-linux-x64-gnu-0.21.1-beta.1.tgz",
-      "integrity": "sha512-rq7Q6Lq9kJmBcgwplYQVJmRbyeP+xPVmXyyQfAO3IjekqeSsyjj1HoCZYqZIfBZyN5ELiSvIJB0731aKf9pr1A==",
+      "integrity": "sha512-7S7gV13hv9Ho5W1Jat3FYiaMJOjRAwZOol7lKvOhU+sR/tJMEfZIOWAgymoqhAowbMtf+wwLoeKacfybXGET/w==",
       "cpu": [
         "x64"
       ],
@@ -375,9 +375,9 @@
       ]
     },
     "node_modules/@lancedb/vectordb-win32-x64-msvc": {
-      "version": "0.20.1-beta.1",
+      "version": "0.21.1-beta.1",
-      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-win32-x64-msvc/-/vectordb-win32-x64-msvc-0.20.1-beta.1.tgz",
+      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-win32-x64-msvc/-/vectordb-win32-x64-msvc-0.21.1-beta.1.tgz",
-      "integrity": "sha512-kHra0SEXeMKdgqi5h0igsqHcBr73hKBhEVJBa8VTv1DUv6Jvazwl4B4ueqllcyD4k3vvOTb2XzZomm7dhQ9QnA==",
+      "integrity": "sha512-w6fEQA9IquvJ/GUYfiawRQvvdFD6OU44UW9JWm+FoscUFzdLiV7qmH4QjYEeEXQD7ob83ikFaxXGPTksYXpNOA==",
       "cpu": [
         "x64"
       ],

node/package.json

@@ -1,6 +1,6 @@
 {
   "name": "vectordb",
-  "version": "0.20.1-beta.1",
+  "version": "0.21.1-beta.1",
   "description": " Serverless, low-latency vector database for AI applications",
   "private": false,
   "main": "dist/index.js",
@@ -89,10 +89,10 @@
     }
   },
   "optionalDependencies": {
-    "@lancedb/vectordb-darwin-x64": "0.20.1-beta.1",
+    "@lancedb/vectordb-darwin-x64": "0.21.1-beta.1",
-    "@lancedb/vectordb-darwin-arm64": "0.20.1-beta.1",
+    "@lancedb/vectordb-darwin-arm64": "0.21.1-beta.1",
-    "@lancedb/vectordb-linux-x64-gnu": "0.20.1-beta.1",
+    "@lancedb/vectordb-linux-x64-gnu": "0.21.1-beta.1",
-    "@lancedb/vectordb-linux-arm64-gnu": "0.20.1-beta.1",
+    "@lancedb/vectordb-linux-arm64-gnu": "0.21.1-beta.1",
-    "@lancedb/vectordb-win32-x64-msvc": "0.20.1-beta.1"
+    "@lancedb/vectordb-win32-x64-msvc": "0.21.1-beta.1"
   }
 }

nodejs/Cargo.toml

@@ -1,7 +1,7 @@
 [package]
 name = "lancedb-nodejs"
 edition.workspace = true
-version = "0.20.1-beta.1"
+version = "0.21.1-beta.1"
 license.workspace = true
 description.workspace = true
 repository.workspace = true

nodejs/__test__/table.test.ts

@@ -368,9 +368,9 @@ describe("merge insert", () => {
       { a: 4, b: "z" },
     ];
 
-    expect(
+    const result = (await table.toArrow()).toArray().sort((a, b) => a.a - b.a);
-      JSON.parse(JSON.stringify((await table.toArrow()).toArray())),
+
-    ).toEqual(expected);
+    expect(result.map((row) => ({ ...row }))).toEqual(expected);
   });
   test("conditional update", async () => {
     const newData = [
@@ -1650,13 +1650,25 @@ describe.each([arrow15, arrow16, arrow17, arrow18])(
       expect(resultSet.has("fob")).toBe(true);
       expect(resultSet.has("fo")).toBe(true);
       expect(resultSet.has("food")).toBe(true);
+
+      const prefixResults = await table
+        .search(
+          new MatchQuery("foo", "text", { fuzziness: 3, prefixLength: 3 }),
+        )
+        .toArray();
+      expect(prefixResults.length).toBe(2);
+      const resultSet2 = new Set(prefixResults.map((r) => r.text));
+      expect(resultSet2.has("foo")).toBe(true);
+      expect(resultSet2.has("food")).toBe(true);
     });
 
     test("full text search boolean query", async () => {
       const db = await connect(tmpDir.name);
       const data = [
-        { text: "hello world", vector: [0.1, 0.2, 0.3] },
+        { text: "The cat and dog are playing" },
-        { text: "goodbye world", vector: [0.4, 0.5, 0.6] },
+        { text: "The cat is sleeping" },
+        { text: "The dog is barking" },
+        { text: "The dog chases the cat" },
       ];
       const table = await db.createTable("test", data);
       await table.createIndex("text", {
@@ -1666,22 +1678,32 @@ describe.each([arrow15, arrow16, arrow17, arrow18])(
       const shouldResults = await table
         .search(
           new BooleanQuery([
-            [Occur.Should, new MatchQuery("hello", "text")],
+            [Occur.Should, new MatchQuery("cat", "text")],
-            [Occur.Should, new MatchQuery("goodbye", "text")],
+            [Occur.Should, new MatchQuery("dog", "text")],
           ]),
         )
         .toArray();
-      expect(shouldResults.length).toBe(2);
+      expect(shouldResults.length).toBe(4);
 
       const mustResults = await table
         .search(
           new BooleanQuery([
-            [Occur.Must, new MatchQuery("hello", "text")],
+            [Occur.Must, new MatchQuery("cat", "text")],
-            [Occur.Must, new MatchQuery("world", "text")],
+            [Occur.Must, new MatchQuery("dog", "text")],
           ]),
         )
         .toArray();
-      expect(mustResults.length).toBe(1);
+      expect(mustResults.length).toBe(2);
+
+      const mustNotResults = await table
+        .search(
+          new BooleanQuery([
+            [Occur.Must, new MatchQuery("cat", "text")],
+            [Occur.MustNot, new MatchQuery("dog", "text")],
+          ]),
+        )
+        .toArray();
+      expect(mustNotResults.length).toBe(1);
     });
 
     test.each([

nodejs/lancedb/query.ts

@@ -812,10 +812,12 @@ export enum Operator {
  *
  * - `Must`: The term must be present in the document.
  * - `Should`: The term should contribute to the document score, but is not required.
+ * - `MustNot`: The term must not be present in the document.
  */
 export enum Occur {
-  Must = "MUST",
   Should = "SHOULD",
+  Must = "MUST",
+  MustNot = "MUST_NOT",
 }
 
 /**
@@ -856,6 +858,7 @@ export class MatchQuery implements FullTextQuery {
    *   - `fuzziness`: The fuzziness level for the query (default is 0).
    *   - `maxExpansions`: The maximum number of terms to consider for fuzzy matching (default is 50).
    *   - `operator`: The logical operator to use for combining terms in the query (default is "OR").
+   *   - `prefixLength`: The number of beginning characters being unchanged for fuzzy matching.
    */
   constructor(
     query: string,
@@ -865,6 +868,7 @@ export class MatchQuery implements FullTextQuery {
       fuzziness?: number;
       maxExpansions?: number;
       operator?: Operator;
+      prefixLength?: number;
     },
   ) {
     let fuzziness = options?.fuzziness;
@@ -878,6 +882,7 @@ export class MatchQuery implements FullTextQuery {
       fuzziness,
       options?.maxExpansions ?? 50,
       options?.operator ?? Operator.Or,
+      options?.prefixLength ?? 0,
     );
   }
 

nodejs/npm/darwin-arm64/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-darwin-arm64",
-	"version": "0.20.1-beta.1",
+	"version": "0.21.1-beta.1",
 	"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": "0.20.1-beta.1",
+	"version": "0.21.1-beta.1",
 	"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": "0.20.1-beta.1",
+	"version": "0.21.1-beta.1",
 	"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": "0.20.1-beta.1",
+	"version": "0.21.1-beta.1",
 	"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": "0.20.1-beta.1",
+	"version": "0.21.1-beta.1",
 	"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": "0.20.1-beta.1",
+	"version": "0.21.1-beta.1",
 	"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": "0.20.1-beta.1",
+  "version": "0.21.1-beta.1",
   "os": [
     "win32"
   ],

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

@@ -1,6 +1,6 @@
 {
 	"name": "@lancedb/lancedb-win32-x64-msvc",
-	"version": "0.20.1-beta.1",
+	"version": "0.21.1-beta.1",
 	"os": ["win32"],
 	"cpu": ["x64"],
 	"main": "lancedb.win32-x64-msvc.node",

nodejs/package-lock.json

@@ -1,12 +1,12 @@
 {
   "name": "@lancedb/lancedb",
-  "version": "0.20.1-beta.1",
+  "version": "0.21.1-beta.1",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {
     "": {
       "name": "@lancedb/lancedb",
-      "version": "0.20.1-beta.1",
+      "version": "0.21.1-beta.1",
       "cpu": [
         "x64",
         "arm64"

nodejs/package.json

@@ -11,7 +11,7 @@
     "ann"
   ],
   "private": false,
-  "version": "0.20.1-beta.1",
+  "version": "0.21.1-beta.1",
   "main": "dist/index.js",
   "exports": {
     ".": "./dist/index.js",

nodejs/src/query.rs

@@ -335,6 +335,7 @@ impl JsFullTextQuery {
         fuzziness: Option<u32>,
         max_expansions: u32,
         operator: String,
+        prefix_length: u32,
     ) -> napi::Result<Self> {
         Ok(Self {
             inner: MatchQuery::new(query)
@@ -347,6 +348,7 @@ impl JsFullTextQuery {
                         napi::Error::from_reason(format!("Invalid operator: {}", e))
                     })?,
                 )
+                .with_prefix_length(prefix_length)
                 .into(),
         })
     }

python/.bumpversion.toml

@@ -1,5 +1,5 @@
 [tool.bumpversion]
-current_version = "0.23.1-beta.2"
+current_version = "0.24.1"
 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 = "0.23.1-beta.2"
+version = "0.24.1"
 edition.workspace = true
 description = "Python bindings for LanceDB"
 license.workspace = true

python/pyproject.toml

@@ -85,7 +85,7 @@ embeddings = [
     "boto3>=1.28.57",
     "awscli>=1.29.57",
     "botocore>=1.31.57",
-    "ollama",
+    "ollama>=0.3.0",
     "ibm-watsonx-ai>=1.1.2",
 ]
 azure = ["adlfs>=2024.2.0"]

python/python/lancedb/embeddings/ollama.py

@@ -2,14 +2,15 @@
 # SPDX-FileCopyrightText: Copyright The LanceDB Authors
 
 from functools import cached_property
-from typing import TYPE_CHECKING, List, Optional, Union
+from typing import TYPE_CHECKING, List, Optional, Sequence, Union
+
+import numpy as np
 
 from ..util import attempt_import_or_raise
 from .base import TextEmbeddingFunction
 from .registry import register
 
 if TYPE_CHECKING:
-    import numpy as np
     import ollama
 
 
@@ -28,23 +29,21 @@ class OllamaEmbeddings(TextEmbeddingFunction):
     keep_alive: Optional[Union[float, str]] = None
     ollama_client_kwargs: Optional[dict] = {}
 
-    def ndims(self):
+    def ndims(self) -> int:
         return len(self.generate_embeddings(["foo"])[0])
 
-    def _compute_embedding(self, text) -> Union["np.array", None]:
+    def _compute_embedding(self, text: Sequence[str]) -> Sequence[Sequence[float]]:
-        return (
+        response = self._ollama_client.embed(
-            self._ollama_client.embeddings(
+            model=self.name,
-                model=self.name,
+            input=text,
-                prompt=text,
+            options=self.options,
-                options=self.options,
+            keep_alive=self.keep_alive,
-                keep_alive=self.keep_alive,
-            )["embedding"]
-            or None
         )
+        return response.embeddings
 
     def generate_embeddings(
-        self, texts: Union[List[str], "np.ndarray"]
+        self, texts: Union[List[str], np.ndarray]
-    ) -> list[Union["np.array", None]]:
+    ) -> list[Union[np.array, None]]:
         """
         Get the embeddings for the given texts
 
@@ -54,8 +53,8 @@ class OllamaEmbeddings(TextEmbeddingFunction):
             The texts to embed
         """
         # TODO retry, rate limit, token limit
-        embeddings = [self._compute_embedding(text) for text in texts]
+        embeddings = self._compute_embedding(texts)
-        return embeddings
+        return list(embeddings)
 
     @cached_property
     def _ollama_client(self) -> "ollama.Client":

python/python/lancedb/query.py

@@ -101,8 +101,9 @@ class FullTextOperator(str, Enum):
 
 
 class Occur(str, Enum):
-    MUST = "MUST"
     SHOULD = "SHOULD"
+    MUST = "MUST"
+    MUST_NOT = "MUST_NOT"
 
 
 @pydantic.dataclasses.dataclass
@@ -181,6 +182,9 @@ class MatchQuery(FullTextQuery):
         Can be either `AND` or `OR`.
         If `AND`, all terms in the query must match.
         If `OR`, at least one term in the query must match.
+    prefix_length : int, optional
+        The number of beginning characters being unchanged for fuzzy matching.
+        This is useful to achieve prefix matching.
     """
 
     query: str
@@ -189,6 +193,7 @@ class MatchQuery(FullTextQuery):
     fuzziness: int = pydantic.Field(0, kw_only=True)
     max_expansions: int = pydantic.Field(50, kw_only=True)
     operator: FullTextOperator = pydantic.Field(FullTextOperator.OR, kw_only=True)
+    prefix_length: int = pydantic.Field(0, kw_only=True)
 
     def query_type(self) -> FullTextQueryType:
         return FullTextQueryType.MATCH
@@ -1446,10 +1451,13 @@ class LanceFtsQueryBuilder(LanceQueryBuilder):
 
         query = self._query
         if self._phrase_query:
-            raise NotImplementedError(
+            if isinstance(query, str):
-                "Phrase query is not yet supported in Lance FTS. "
+                if not query.startswith('"') or not query.endswith('"'):
-                "Use tantivy-based index instead for now."
+                    query = f'"{query}"'
-            )
+            elif isinstance(query, FullTextQuery) and not isinstance(
+                query, PhraseQuery
+            ):
+                raise TypeError("Please use PhraseQuery for phrase queries.")
         query = self.to_query_object()
         results = self._table._execute_query(query, timeout=timeout)
         results = results.read_all()
@@ -3034,15 +3042,21 @@ class AsyncHybridQuery(AsyncQueryBase, AsyncVectorQueryBase):
         >>> asyncio.run(doctest_example()) # doctest: +ELLIPSIS, +NORMALIZE_WHITESPACE
         Vector Search Plan:
         ProjectionExec: expr=[vector@0 as vector, text@3 as text, _distance@2 as _distance]
-            Take: columns="vector, _rowid, _distance, (text)"
+          Take: columns="vector, _rowid, _distance, (text)"
-                CoalesceBatchesExec: target_batch_size=1024
+            CoalesceBatchesExec: target_batch_size=1024
-                GlobalLimitExec: skip=0, fetch=10
+              GlobalLimitExec: skip=0, fetch=10
-                    FilterExec: _distance@2 IS NOT NULL
+                FilterExec: _distance@2 IS NOT NULL
-                    SortExec: TopK(fetch=10), expr=[_distance@2 ASC NULLS LAST], preserve_partitioning=[false]
+                  SortExec: TopK(fetch=10), expr=[_distance@2 ASC NULLS LAST], preserve_partitioning=[false]
-                        KNNVectorDistance: metric=l2
+                    KNNVectorDistance: metric=l2
-                        LanceScan: uri=..., projection=[vector], row_id=true, row_addr=false, ordered=false
+                      LanceScan: uri=..., projection=[vector], row_id=true, row_addr=false, ordered=false
+        <BLANKLINE>
         FTS Search Plan:
-        LanceScan: uri=..., projection=[vector, text], row_id=false, row_addr=false, ordered=true
+        ProjectionExec: expr=[vector@2 as vector, text@3 as text, _score@1 as _score]
+          Take: columns="_rowid, _score, (vector), (text)"
+            CoalesceBatchesExec: target_batch_size=1024
+              GlobalLimitExec: skip=0, fetch=10
+                MatchQuery: query=hello
+        <BLANKLINE>
 
         Parameters
         ----------

python/python/lancedb/remote/table.py

@@ -18,7 +18,7 @@ from lancedb._lancedb import (
     UpdateResult,
 )
 from lancedb.embeddings.base import EmbeddingFunctionConfig
-from lancedb.index import FTS, BTree, Bitmap, HnswPq, HnswSq, IvfFlat, IvfPq, LabelList
+from lancedb.index import FTS, BTree, Bitmap, HnswSq, IvfFlat, IvfPq, LabelList
 from lancedb.remote.db import LOOP
 import pyarrow as pa
 
@@ -89,7 +89,7 @@ class RemoteTable(Table):
 
     def to_pandas(self):
         """to_pandas() is not yet supported on LanceDB cloud."""
-        return NotImplementedError("to_pandas() is not yet supported on LanceDB cloud.")
+        raise NotImplementedError("to_pandas() is not yet supported on LanceDB cloud.")
 
     def checkout(self, version: Union[int, str]):
         return LOOP.run(self._table.checkout(version))
@@ -186,6 +186,8 @@ class RemoteTable(Table):
         accelerator: Optional[str] = None,
         index_type="vector",
         wait_timeout: Optional[timedelta] = None,
+        *,
+        num_bits: int = 8,
     ):
         """Create an index on the table.
         Currently, the only parameters that matter are
@@ -220,11 +222,6 @@ class RemoteTable(Table):
         >>> table.create_index("l2", "vector") # doctest: +SKIP
         """
 
-        if num_partitions is not None:
-            logging.warning(
-                "num_partitions is not supported on LanceDB cloud."
-                "This parameter will be tuned automatically."
-            )
         if num_sub_vectors is not None:
             logging.warning(
                 "num_sub_vectors is not supported on LanceDB cloud."
@@ -244,13 +241,21 @@ class RemoteTable(Table):
 
         index_type = index_type.upper()
         if index_type == "VECTOR" or index_type == "IVF_PQ":
-            config = IvfPq(distance_type=metric)
+            config = IvfPq(
+                distance_type=metric,
+                num_partitions=num_partitions,
+                num_sub_vectors=num_sub_vectors,
+                num_bits=num_bits,
+            )
         elif index_type == "IVF_HNSW_PQ":
-            config = HnswPq(distance_type=metric)
+            raise ValueError(
+                "IVF_HNSW_PQ is not supported on LanceDB cloud."
+                "Please use IVF_HNSW_SQ instead."
+            )
         elif index_type == "IVF_HNSW_SQ":
-            config = HnswSq(distance_type=metric)
+            config = HnswSq(distance_type=metric, num_partitions=num_partitions)
         elif index_type == "IVF_FLAT":
-            config = IvfFlat(distance_type=metric)
+            config = IvfFlat(distance_type=metric, num_partitions=num_partitions)
         else:
             raise ValueError(
                 f"Unknown vector index type: {index_type}. Valid options are"

python/python/lancedb/table.py

@@ -827,7 +827,7 @@ class Table(ABC):
         ordering_field_names: Optional[Union[str, List[str]]] = None,
         replace: bool = False,
         writer_heap_size: Optional[int] = 1024 * 1024 * 1024,
-        use_tantivy: bool = True,
+        use_tantivy: bool = False,
         tokenizer_name: Optional[str] = None,
         with_position: bool = False,
         # tokenizer configs:
@@ -864,7 +864,7 @@ class Table(ABC):
             The tokenizer to use for the index. Can be "raw", "default" or the 2 letter
             language code followed by "_stem". So for english it would be "en_stem".
             For available languages see: https://docs.rs/tantivy/latest/tantivy/tokenizer/enum.Language.html
-        use_tantivy: bool, default True
+        use_tantivy: bool, default False
             If True, use the legacy full-text search implementation based on tantivy.
             If False, use the new full-text search implementation based on lance-index.
         with_position: bool, default False
@@ -1970,7 +1970,7 @@ class LanceTable(Table):
         ordering_field_names: Optional[Union[str, List[str]]] = None,
         replace: bool = False,
         writer_heap_size: Optional[int] = 1024 * 1024 * 1024,
-        use_tantivy: bool = True,
+        use_tantivy: bool = False,
         tokenizer_name: Optional[str] = None,
         with_position: bool = False,
         # tokenizer configs:

python/python/tests/docs/test_search.py

@@ -6,7 +6,7 @@ import lancedb
 
 # --8<-- [end:import-lancedb]
 # --8<-- [start:import-numpy]
-from lancedb.query import BoostQuery, MatchQuery
+from lancedb.query import BooleanQuery, BoostQuery, MatchQuery, Occur
 import numpy as np
 import pyarrow as pa
 
@@ -191,6 +191,15 @@ def test_fts_fuzzy_query():
         "food",  # 1 insertion
     }
 
+    results = table.search(
+        MatchQuery("foo", "text", fuzziness=1, prefix_length=3)
+    ).to_pandas()
+    assert len(results) == 2
+    assert set(results["text"].to_list()) == {
+        "foo",
+        "food",
+    }
+
 
 @pytest.mark.skipif(
     os.name == "nt", reason="Need to fix https://github.com/lancedb/lance/issues/3905"
@@ -240,6 +249,60 @@ def test_fts_boost_query():
     )
 
 
+@pytest.mark.skipif(
+    os.name == "nt", reason="Need to fix https://github.com/lancedb/lance/issues/3905"
+)
+def test_fts_boolean_query(tmp_path):
+    uri = tmp_path / "boolean-example"
+    db = lancedb.connect(uri)
+    table = db.create_table(
+        "my_table_fts_boolean",
+        data=[
+            {"text": "The cat and dog are playing"},
+            {"text": "The cat is sleeping"},
+            {"text": "The dog is barking"},
+            {"text": "The dog chases the cat"},
+        ],
+        mode="overwrite",
+    )
+    table.create_fts_index("text", use_tantivy=False, replace=True)
+
+    # SHOULD
+    results = table.search(
+        MatchQuery("cat", "text") | MatchQuery("dog", "text")
+    ).to_pandas()
+    assert len(results) == 4
+    assert set(results["text"].to_list()) == {
+        "The cat and dog are playing",
+        "The cat is sleeping",
+        "The dog is barking",
+        "The dog chases the cat",
+    }
+    # MUST
+    results = table.search(
+        MatchQuery("cat", "text") & MatchQuery("dog", "text")
+    ).to_pandas()
+    assert len(results) == 2
+    assert set(results["text"].to_list()) == {
+        "The cat and dog are playing",
+        "The dog chases the cat",
+    }
+
+    # MUST NOT
+    results = table.search(
+        BooleanQuery(
+            [
+                (Occur.MUST, MatchQuery("cat", "text")),
+                (Occur.MUST_NOT, MatchQuery("dog", "text")),
+            ]
+        )
+    ).to_pandas()
+    assert len(results) == 1
+    assert set(results["text"].to_list()) == {
+        "The cat is sleeping",
+    }
+
+
 @pytest.mark.skipif(
     os.name == "nt", reason="Need to fix https://github.com/lancedb/lance/issues/3905"
 )

python/python/tests/test_query.py

@@ -775,6 +775,82 @@ async def test_explain_plan_async(table_async: AsyncTable):
     assert "KNN" in plan
 
 
+@pytest.mark.asyncio
+async def test_explain_plan_fts(table_async: AsyncTable):
+    """Test explain plan for FTS queries"""
+    # Create FTS index
+    from lancedb.index import FTS
+
+    await table_async.create_index("text", config=FTS())
+
+    # Test pure FTS query
+    query = await table_async.search("dog", query_type="fts", fts_columns="text")
+    plan = await query.explain_plan()
+    # Should show FTS details (issue #2465 is now fixed)
+    assert "MatchQuery: query=dog" in plan
+    assert "GlobalLimitExec" in plan  # Default limit
+
+    # Test FTS query with limit
+    query_with_limit = await table_async.search(
+        "dog", query_type="fts", fts_columns="text"
+    )
+    plan_with_limit = await query_with_limit.limit(1).explain_plan()
+    assert "MatchQuery: query=dog" in plan_with_limit
+    assert "GlobalLimitExec: skip=0, fetch=1" in plan_with_limit
+
+    # Test FTS query with offset and limit
+    query_with_offset = await table_async.search(
+        "dog", query_type="fts", fts_columns="text"
+    )
+    plan_with_offset = await query_with_offset.offset(1).limit(1).explain_plan()
+    assert "MatchQuery: query=dog" in plan_with_offset
+    assert "GlobalLimitExec: skip=1, fetch=1" in plan_with_offset
+
+
+@pytest.mark.asyncio
+async def test_explain_plan_vector_with_limit_offset(table_async: AsyncTable):
+    """Test explain plan for vector queries with limit and offset"""
+    # Test vector query with limit
+    plan_with_limit = await (
+        table_async.query().nearest_to(pa.array([1, 2])).limit(1).explain_plan()
+    )
+    assert "KNN" in plan_with_limit
+    assert "GlobalLimitExec: skip=0, fetch=1" in plan_with_limit
+
+    # Test vector query with offset and limit
+    plan_with_offset = await (
+        table_async.query()
+        .nearest_to(pa.array([1, 2]))
+        .offset(1)
+        .limit(1)
+        .explain_plan()
+    )
+    assert "KNN" in plan_with_offset
+    assert "GlobalLimitExec: skip=1, fetch=1" in plan_with_offset
+
+
+@pytest.mark.asyncio
+async def test_explain_plan_with_filters(table_async: AsyncTable):
+    """Test explain plan for queries with filters"""
+    # Test vector query with filter
+    plan_with_filter = await (
+        table_async.query().nearest_to(pa.array([1, 2])).where("id = 1").explain_plan()
+    )
+    assert "KNN" in plan_with_filter
+    assert "FilterExec" in plan_with_filter
+
+    # Test FTS query with filter
+    from lancedb.index import FTS
+
+    await table_async.create_index("text", config=FTS())
+    query_fts_filter = await table_async.search(
+        "dog", query_type="fts", fts_columns="text"
+    )
+    plan_fts_filter = await query_fts_filter.where("id = 1").explain_plan()
+    assert "MatchQuery: query=dog" in plan_fts_filter
+    assert "FilterExec: id@" in plan_fts_filter  # Should show filter details
+
+
 @pytest.mark.asyncio
 async def test_query_camelcase_async(tmp_path):
     db = await lancedb.connect_async(tmp_path)

python/python/tests/test_remote_db.py

@@ -210,6 +210,25 @@ async def test_retry_error():
         assert cause.status_code == 429
 
 
+def test_table_unimplemented_functions():
+    def handler(request):
+        if request.path == "/v1/table/test/create/?mode=create":
+            request.send_response(200)
+            request.send_header("Content-Type", "application/json")
+            request.end_headers()
+            request.wfile.write(b"{}")
+        else:
+            request.send_response(404)
+            request.end_headers()
+
+    with mock_lancedb_connection(handler) as db:
+        table = db.create_table("test", [{"id": 1}])
+        with pytest.raises(NotImplementedError):
+            table.to_arrow()
+        with pytest.raises(NotImplementedError):
+            table.to_pandas()
+
+
 def test_table_add_in_threadpool():
     def handler(request):
         if request.path == "/v1/table/test/insert/":

python/python/tests/test_s3.py

@@ -245,7 +245,7 @@ def test_s3_dynamodb_sync(s3_bucket: str, commit_table: str, monkeypatch):
         NotImplementedError,
         match="Full-text search is only supported on the local filesystem",
     ):
-        table.create_fts_index("x")
+        table.create_fts_index("x", use_tantivy=True)
 
     # make sure list tables still works
     assert db.table_names() == ["test_ddb_sync"]

python/src/query.rs

@@ -50,8 +50,9 @@ impl FromPyObject<'_> for PyLanceDB<FtsQuery> {
                 let fuzziness = ob.getattr("fuzziness")?.extract()?;
                 let max_expansions = ob.getattr("max_expansions")?.extract()?;
                 let operator = ob.getattr("operator")?.extract::<String>()?;
+                let prefix_length = ob.getattr("prefix_length")?.extract()?;
 
-                Ok(PyLanceDB(
+                Ok(Self(
                     MatchQuery::new(query)
                         .with_column(Some(column))
                         .with_boost(boost)
@@ -60,6 +61,7 @@ impl FromPyObject<'_> for PyLanceDB<FtsQuery> {
                         .with_operator(Operator::try_from(operator.as_str()).map_err(|e| {
                             PyValueError::new_err(format!("Invalid operator: {}", e))
                         })?)
+                        .with_prefix_length(prefix_length)
                         .into(),
                 ))
             }
@@ -68,7 +70,7 @@ impl FromPyObject<'_> for PyLanceDB<FtsQuery> {
                 let column = ob.getattr("column")?.extract()?;
                 let slop = ob.getattr("slop")?.extract()?;
 
-                Ok(PyLanceDB(
+                Ok(Self(
                     PhraseQuery::new(query)
                         .with_column(Some(column))
                         .with_slop(slop)
@@ -76,10 +78,10 @@ impl FromPyObject<'_> for PyLanceDB<FtsQuery> {
                 ))
             }
             "BoostQuery" => {
-                let positive: PyLanceDB<FtsQuery> = ob.getattr("positive")?.extract()?;
+                let positive: Self = ob.getattr("positive")?.extract()?;
-                let negative: PyLanceDB<FtsQuery> = ob.getattr("negative")?.extract()?;
+                let negative: Self = ob.getattr("negative")?.extract()?;
                 let negative_boost = ob.getattr("negative_boost")?.extract()?;
-                Ok(PyLanceDB(
+                Ok(Self(
                     BoostQuery::new(positive.0, negative.0, negative_boost).into(),
                 ))
             }
@@ -101,18 +103,17 @@ impl FromPyObject<'_> for PyLanceDB<FtsQuery> {
                 let op = Operator::try_from(operator.as_str())
                     .map_err(|e| PyValueError::new_err(format!("Invalid operator: {}", e)))?;
 
-                Ok(PyLanceDB(q.with_operator(op).into()))
+                Ok(Self(q.with_operator(op).into()))
             }
             "BooleanQuery" => {
-                let queries: Vec<(String, PyLanceDB<FtsQuery>)> =
+                let queries: Vec<(String, Self)> = ob.getattr("queries")?.extract()?;
-                    ob.getattr("queries")?.extract()?;
                 let mut sub_queries = Vec::with_capacity(queries.len());
                 for (occur, q) in queries {
                     let occur = Occur::try_from(occur.as_str())
                         .map_err(|e| PyValueError::new_err(e.to_string()))?;
                     sub_queries.push((occur, q.0));
                 }
-                Ok(PyLanceDB(BooleanQuery::new(sub_queries).into()))
+                Ok(Self(BooleanQuery::new(sub_queries).into()))
             }
             name => Err(PyValueError::new_err(format!(
                 "Unsupported FTS query type: {}",
@@ -139,7 +140,8 @@ impl<'py> IntoPyObject<'py> for PyLanceDB<FtsQuery> {
                 kwargs.set_item("boost", query.boost)?;
                 kwargs.set_item("fuzziness", query.fuzziness)?;
                 kwargs.set_item("max_expansions", query.max_expansions)?;
-                kwargs.set_item("operator", operator_to_str(query.operator))?;
+                kwargs.set_item::<_, &str>("operator", query.operator.into())?;
+                kwargs.set_item("prefix_length", query.prefix_length)?;
                 namespace
                     .getattr(intern!(py, "MatchQuery"))?
                     .call((query.terms, query.column.unwrap()), Some(&kwargs))
@@ -152,8 +154,8 @@ impl<'py> IntoPyObject<'py> for PyLanceDB<FtsQuery> {
                     .call((query.terms, query.column.unwrap()), Some(&kwargs))
             }
             FtsQuery::Boost(query) => {
-                let positive = PyLanceDB(query.positive.as_ref().clone()).into_pyobject(py)?;
+                let positive = Self(query.positive.as_ref().clone()).into_pyobject(py)?;
-                let negative = PyLanceDB(query.negative.as_ref().clone()).into_pyobject(py)?;
+                let negative = Self(query.negative.as_ref().clone()).into_pyobject(py)?;
                 let kwargs = PyDict::new(py);
                 kwargs.set_item("negative_boost", query.negative_boost)?;
                 namespace
@@ -169,19 +171,25 @@ impl<'py> IntoPyObject<'py> for PyLanceDB<FtsQuery> {
                     .unzip();
                 let kwargs = PyDict::new(py);
                 kwargs.set_item("boosts", boosts)?;
-                kwargs.set_item("operator", operator_to_str(first.operator))?;
+                kwargs.set_item::<_, &str>("operator", first.operator.into())?;
                 namespace
                     .getattr(intern!(py, "MultiMatchQuery"))?
                     .call((first.terms.clone(), columns), Some(&kwargs))
             }
             FtsQuery::Boolean(query) => {
-                let mut queries = Vec::with_capacity(query.must.len() + query.should.len());
+                let mut queries: Vec<(&str, Bound<'py, PyAny>)> = Vec::with_capacity(
-                for q in query.must {
+                    query.should.len() + query.must.len() + query.must_not.len(),
-                    queries.push((occur_to_str(Occur::Must), PyLanceDB(q).into_pyobject(py)?));
+                );
-                }
                 for q in query.should {
-                    queries.push((occur_to_str(Occur::Should), PyLanceDB(q).into_pyobject(py)?));
+                    queries.push((Occur::Should.into(), Self(q).into_pyobject(py)?));
                 }
+                for q in query.must {
+                    queries.push((Occur::Must.into(), Self(q).into_pyobject(py)?));
+                }
+                for q in query.must_not {
+                    queries.push((Occur::MustNot.into(), Self(q).into_pyobject(py)?));
+                }
+
                 namespace
                     .getattr(intern!(py, "BooleanQuery"))?
                     .call1((queries,))
@@ -190,21 +198,6 @@ impl<'py> IntoPyObject<'py> for PyLanceDB<FtsQuery> {
     }
 }
 
-fn operator_to_str(op: Operator) -> &'static str {
-    match op {
-        Operator::And => "AND",
-        Operator::Or => "OR",
-    }
-}
-
-fn occur_to_str(occur: Occur) -> &'static str {
-    match occur {
-        Occur::Must => "MUST",
-        Occur::Should => "SHOULD",
-        Occur::MustNot => "MUST NOT",
-    }
-}
-
 // Python representation of query vector(s)
 #[derive(Clone)]
 pub struct PyQueryVectors(Vec<Arc<dyn Array>>);
@@ -569,7 +562,10 @@ impl FTSQuery {
     }
 
     pub fn explain_plan(self_: PyRef<'_, Self>, verbose: bool) -> PyResult<Bound<'_, PyAny>> {
-        let inner = self_.inner.clone();
+        let inner = self_
+            .inner
+            .clone()
+            .full_text_search(self_.fts_query.clone());
         future_into_py(self_.py(), async move {
             inner
                 .explain_plan(verbose)
@@ -579,7 +575,10 @@ impl FTSQuery {
     }
 
     pub fn analyze_plan(self_: PyRef<'_, Self>) -> PyResult<Bound<'_, PyAny>> {
-        let inner = self_.inner.clone();
+        let inner = self_
+            .inner
+            .clone()
+            .full_text_search(self_.fts_query.clone());
         future_into_py(self_.py(), async move {
             inner
                 .analyze_plan()

rust/ffi/node/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "lancedb-node"
-version = "0.20.1-beta.1"
+version = "0.21.1-beta.1"
 description = "Serverless, low-latency vector database for AI applications"
 license.workspace = true
 edition.workspace = true

rust/lancedb/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "lancedb"
-version = "0.20.1-beta.1"
+version = "0.21.1-beta.1"
 edition.workspace = true
 description = "LanceDB: A serverless, low-latency vector database for AI applications"
 license.workspace = true

rust/lancedb/src/catalog/listing.rs

@@ -105,7 +105,7 @@ impl ListingCatalog {
     }
 
     async fn open_path(path: &str) -> Result<Self> {
-        let (object_store, base_path) = ObjectStore::from_uri(path).await.unwrap();
+        let (object_store, base_path) = ObjectStore::from_uri(path).await?;
         if object_store.is_local() {
             Self::try_create_dir(path).context(CreateDirSnafu { path })?;
         }

rust/lancedb/src/io/object_store.rs

@@ -107,7 +107,7 @@ impl ObjectStore for MirroringObjectStore {
         self.primary.delete(location).await
     }
 
-    fn list(&self, prefix: Option<&Path>) -> BoxStream<'_, Result<ObjectMeta>> {
+    fn list(&self, prefix: Option<&Path>) -> BoxStream<'static, Result<ObjectMeta>> {
         self.primary.list(prefix)
     }
 

rust/lancedb/src/io/object_store/io_tracking.rs

@@ -119,7 +119,7 @@ impl ObjectStore for IoTrackingStore {
         let result = self.target.get(location).await;
         if let Ok(result) = &result {
             let num_bytes = result.range.end - result.range.start;
-            self.record_read(num_bytes as u64);
+            self.record_read(num_bytes);
         }
         result
     }
@@ -128,12 +128,12 @@ impl ObjectStore for IoTrackingStore {
         let result = self.target.get_opts(location, options).await;
         if let Ok(result) = &result {
             let num_bytes = result.range.end - result.range.start;
-            self.record_read(num_bytes as u64);
+            self.record_read(num_bytes);
         }
         result
     }
 
-    async fn get_range(&self, location: &Path, range: std::ops::Range<usize>) -> OSResult<Bytes> {
+    async fn get_range(&self, location: &Path, range: std::ops::Range<u64>) -> OSResult<Bytes> {
         let result = self.target.get_range(location, range).await;
         if let Ok(result) = &result {
             self.record_read(result.len() as u64);
@@ -144,7 +144,7 @@ impl ObjectStore for IoTrackingStore {
     async fn get_ranges(
         &self,
         location: &Path,
-        ranges: &[std::ops::Range<usize>],
+        ranges: &[std::ops::Range<u64>],
     ) -> OSResult<Vec<Bytes>> {
         let result = self.target.get_ranges(location, ranges).await;
         if let Ok(result) = &result {
@@ -170,7 +170,7 @@ impl ObjectStore for IoTrackingStore {
         self.target.delete_stream(locations)
     }
 
-    fn list(&self, prefix: Option<&Path>) -> BoxStream<'_, OSResult<ObjectMeta>> {
+    fn list(&self, prefix: Option<&Path>) -> BoxStream<'static, OSResult<ObjectMeta>> {
         self.record_read(0);
         self.target.list(prefix)
     }
@@ -179,7 +179,7 @@ impl ObjectStore for IoTrackingStore {
         &self,
         prefix: Option<&Path>,
         offset: &Path,
-    ) -> BoxStream<'_, OSResult<ObjectMeta>> {
+    ) -> BoxStream<'static, OSResult<ObjectMeta>> {
         self.record_read(0);
         self.target.list_with_offset(prefix, offset)
     }

rust/lancedb/src/remote/table.rs

@@ -57,6 +57,8 @@ use crate::{
 };
 
 const REQUEST_TIMEOUT_HEADER: HeaderName = HeaderName::from_static("x-request-timeout-ms");
+const METRIC_TYPE_KEY: &str = "metric_type";
+const INDEX_TYPE_KEY: &str = "index_type";
 
 pub struct RemoteTags<'a, S: HttpSend = Sender> {
     inner: &'a RemoteTable<S>,
@@ -997,23 +999,53 @@ impl<S: HttpSend> BaseTable for RemoteTable<S> {
             "column": column
         });
 
-        let (index_type, distance_type) = match index.index {
+        match index.index {
             // TODO: Should we pass the actual index parameters? SaaS does not
             // yet support them.
-            Index::IvfFlat(index) => ("IVF_FLAT", Some(index.distance_type)),
+            Index::IvfFlat(index) => {
-            Index::IvfPq(index) => ("IVF_PQ", Some(index.distance_type)),
+                body[INDEX_TYPE_KEY] = serde_json::Value::String("IVF_FLAT".to_string());
-            Index::IvfHnswSq(index) => ("IVF_HNSW_SQ", Some(index.distance_type)),
+                body[METRIC_TYPE_KEY] =
-            Index::BTree(_) => ("BTREE", None),
+                    serde_json::Value::String(index.distance_type.to_string().to_lowercase());
-            Index::Bitmap(_) => ("BITMAP", None),
+                if let Some(num_partitions) = index.num_partitions {
-            Index::LabelList(_) => ("LABEL_LIST", None),
+                    body["num_partitions"] = serde_json::Value::Number(num_partitions.into());
+                }
+            }
+            Index::IvfPq(index) => {
+                body[INDEX_TYPE_KEY] = serde_json::Value::String("IVF_PQ".to_string());
+                body[METRIC_TYPE_KEY] =
+                    serde_json::Value::String(index.distance_type.to_string().to_lowercase());
+                if let Some(num_partitions) = index.num_partitions {
+                    body["num_partitions"] = serde_json::Value::Number(num_partitions.into());
+                }
+                if let Some(num_bits) = index.num_bits {
+                    body["num_bits"] = serde_json::Value::Number(num_bits.into());
+                }
+            }
+            Index::IvfHnswSq(index) => {
+                body[INDEX_TYPE_KEY] = serde_json::Value::String("IVF_HNSW_SQ".to_string());
+                body[METRIC_TYPE_KEY] =
+                    serde_json::Value::String(index.distance_type.to_string().to_lowercase());
+                if let Some(num_partitions) = index.num_partitions {
+                    body["num_partitions"] = serde_json::Value::Number(num_partitions.into());
+                }
+            }
+            Index::BTree(_) => {
+                body[INDEX_TYPE_KEY] = serde_json::Value::String("BTREE".to_string());
+            }
+            Index::Bitmap(_) => {
+                body[INDEX_TYPE_KEY] = serde_json::Value::String("BITMAP".to_string());
+            }
+            Index::LabelList(_) => {
+                body[INDEX_TYPE_KEY] = serde_json::Value::String("LABEL_LIST".to_string());
+            }
             Index::FTS(fts) => {
+                body[INDEX_TYPE_KEY] = serde_json::Value::String("FTS".to_string());
                 let params = serde_json::to_value(&fts).map_err(|e| Error::InvalidInput {
                     message: format!("failed to serialize FTS index params {:?}", e),
                 })?;
                 for (key, value) in params.as_object().unwrap() {
                     body[key] = value.clone();
                 }
-                ("FTS", None)
             }
             Index::Auto => {
                 let schema = self.schema().await?;
@@ -1023,9 +1055,11 @@ impl<S: HttpSend> BaseTable for RemoteTable<S> {
                         message: format!("Column {} not found in schema", column),
                     })?;
                 if supported_vector_data_type(field.data_type()) {
-                    ("IVF_PQ", Some(DistanceType::L2))
+                    body[INDEX_TYPE_KEY] = serde_json::Value::String("IVF_PQ".to_string());
+                    body[METRIC_TYPE_KEY] =
+                        serde_json::Value::String(DistanceType::L2.to_string().to_lowercase());
                 } else if supported_btree_data_type(field.data_type()) {
-                    ("BTREE", None)
+                    body[INDEX_TYPE_KEY] = serde_json::Value::String("BTREE".to_string());
                 } else {
                     return Err(Error::NotSupported {
                         message: format!(
@@ -1042,12 +1076,6 @@ impl<S: HttpSend> BaseTable for RemoteTable<S> {
                 })
             }
         };
-        body["index_type"] = serde_json::Value::String(index_type.into());
-        if let Some(distance_type) = distance_type {
-            // Phalanx expects this to be lowercase right now.
-            body["metric_type"] =
-                serde_json::Value::String(distance_type.to_string().to_lowercase());
-        }
 
         let request = request.json(&body);
 
@@ -1429,11 +1457,12 @@ mod tests {
     use chrono::{DateTime, Utc};
     use futures::{future::BoxFuture, StreamExt, TryFutureExt};
     use lance_index::scalar::inverted::query::MatchQuery;
-    use lance_index::scalar::FullTextSearchQuery;
+    use lance_index::scalar::{FullTextSearchQuery, InvertedIndexParams};
     use reqwest::Body;
     use rstest::rstest;
+    use serde_json::json;
 
-    use crate::index::vector::IvfFlatIndexBuilder;
+    use crate::index::vector::{IvfFlatIndexBuilder, IvfHnswSqIndexBuilder};
     use crate::remote::db::DEFAULT_SERVER_VERSION;
     use crate::remote::JSON_CONTENT_TYPE;
     use crate::{
@@ -2433,29 +2462,79 @@ mod tests {
         let cases = [
             (
                 "IVF_FLAT",
-                Some("hamming"),
+                json!({
+                    "metric_type": "hamming",
+                }),
                 Index::IvfFlat(IvfFlatIndexBuilder::default().distance_type(DistanceType::Hamming)),
             ),
-            ("IVF_PQ", Some("l2"), Index::IvfPq(Default::default())),
+            (
+                "IVF_FLAT",
+                json!({
+                    "metric_type": "hamming",
+                    "num_partitions": 128,
+                }),
+                Index::IvfFlat(
+                    IvfFlatIndexBuilder::default()
+                        .distance_type(DistanceType::Hamming)
+                        .num_partitions(128),
+                ),
+            ),
             (
                 "IVF_PQ",
-                Some("cosine"),
+                json!({
-                Index::IvfPq(IvfPqIndexBuilder::default().distance_type(DistanceType::Cosine)),
+                    "metric_type": "l2",
+                }),
+                Index::IvfPq(Default::default()),
+            ),
+            (
+                "IVF_PQ",
+                json!({
+                    "metric_type": "cosine",
+                    "num_partitions": 128,
+                    "num_bits": 4,
+                }),
+                Index::IvfPq(
+                    IvfPqIndexBuilder::default()
+                        .distance_type(DistanceType::Cosine)
+                        .num_partitions(128)
+                        .num_bits(4),
+                ),
             ),
             (
                 "IVF_HNSW_SQ",
-                Some("l2"),
+                json!({
+                    "metric_type": "l2",
+                }),
                 Index::IvfHnswSq(Default::default()),
             ),
+            (
+                "IVF_HNSW_SQ",
+                json!({
+                    "metric_type": "l2",
+                    "num_partitions": 128,
+                }),
+                Index::IvfHnswSq(
+                    IvfHnswSqIndexBuilder::default()
+                        .distance_type(DistanceType::L2)
+                        .num_partitions(128),
+                ),
+            ),
             // HNSW_PQ isn't yet supported on SaaS
-            ("BTREE", None, Index::BTree(Default::default())),
+            ("BTREE", json!({}), Index::BTree(Default::default())),
-            ("BITMAP", None, Index::Bitmap(Default::default())),
+            ("BITMAP", json!({}), Index::Bitmap(Default::default())),
-            ("LABEL_LIST", None, Index::LabelList(Default::default())),
+            (
-            ("FTS", None, Index::FTS(Default::default())),
+                "LABEL_LIST",
+                json!({}),
+                Index::LabelList(Default::default()),
+            ),
+            (
+                "FTS",
+                serde_json::to_value(InvertedIndexParams::default()).unwrap(),
+                Index::FTS(Default::default()),
+            ),
         ];
 
-        for (index_type, distance_type, index) in cases {
+        for (index_type, expected_body, index) in cases {
-            let params = index.clone();
             let table = Table::new_with_handler("my_table", move |request| {
                 assert_eq!(request.method(), "POST");
                 assert_eq!(request.url().path(), "/v1/table/my_table/create_index/");
@@ -2465,19 +2544,9 @@ mod tests {
                 );
                 let body = request.body().unwrap().as_bytes().unwrap();
                 let body: serde_json::Value = serde_json::from_slice(body).unwrap();
-                let mut expected_body = serde_json::json!({
+                let mut expected_body = expected_body.clone();
-                    "column": "a",
+                expected_body["column"] = "a".into();
-                    "index_type": index_type,
+                expected_body[INDEX_TYPE_KEY] = index_type.into();
-                });
-                if let Some(distance_type) = distance_type {
-                    expected_body["metric_type"] = distance_type.to_lowercase().into();
-                }
-                if let Index::FTS(fts) = &params {
-                    let params = serde_json::to_value(fts).unwrap();
-                    for (key, value) in params.as_object().unwrap() {
-                        expected_body[key] = value.clone();
-                    }
-                }
 
                 assert_eq!(body, expected_body);
 

rust/lancedb/src/table/datafusion.rs

@@ -392,9 +392,18 @@ pub mod tests {
                 } else {
                     expected_line.trim()
                 };
-                assert_eq!(&actual_trimmed[..expected_trimmed.len()], expected_trimmed);
+                assert_eq!(
+                    &actual_trimmed[..expected_trimmed.len()],
+                    expected_trimmed,
+                    "\nactual:\n{physical_plan}\nexpected:\n{expected}"
+                );
             }
-            assert_eq!(lines_checked, expected.lines().count());
+            assert_eq!(
+                lines_checked,
+                expected.lines().count(),
+                "\nlines_checked:\n{lines_checked}\nexpected:\n{}",
+                expected.lines().count()
+            );
         }
     }
 
@@ -477,9 +486,9 @@ pub mod tests {
         TestFixture::check_plan(
             plan,
             "MetadataEraserExec
-             RepartitionExec:...
              CoalesceBatchesExec:...
              FilterExec: i@0 >= 5
+             RepartitionExec:...
              ProjectionExec:...
              LanceScan:...",
         )

rust/lancedb/src/table/dataset.rs

@@ -129,7 +129,9 @@ impl DatasetRef {
                 dataset: ref mut ds,
                 ..
             } => {
-                *ds = dataset;
+                if dataset.manifest().version > ds.manifest().version {
+                    *ds = dataset;
+                }
             }
             _ => unreachable!("Dataset should be in latest mode at this point"),
         }