[python] Bump version: 0.3.0 → 0.3.1

#488

Cargo.toml

@@ -17,6 +17,7 @@ arrow-ord = "43.0"
 arrow-schema = "43.0"
 arrow-arith = "43.0"
 arrow-cast = "43.0"
+chrono = "0.4.23"
 half = { "version" = "=2.2.1", default-features = false, features = [
     "num-traits"
 ] }

README.md

@@ -54,8 +54,7 @@ const table = await db.createTable('vectors',
       [{ id: 1, vector: [0.1, 0.2], item: "foo", price: 10 },
        { id: 2, vector: [1.1, 1.2], item: "bar", price: 50 }])
 
-const query = table.search([0.1, 0.3]);
+const query = table.search([0.1, 0.3]).limit(2);
-query.limit = 20;
 const results = await query.execute();
 ```
 
@@ -72,7 +71,7 @@ db = lancedb.connect(uri)
 table = db.create_table("my_table",
                          data=[{"vector": [3.1, 4.1], "item": "foo", "price": 10.0},
                                {"vector": [5.9, 26.5], "item": "bar", "price": 20.0}])
-result = table.search([100, 100]).limit(2).to_df()
+result = table.search([100, 100]).limit(2).to_pandas()
 ```
 
 ## Blogs, Tutorials & Videos

docs/mkdocs.yml

@@ -21,6 +21,7 @@ theme:
     - navigation.tracking
     - navigation.instant
     - navigation.indexes
+    - navigation.expand
   icon:
     repo: fontawesome/brands/github
   custom_dir: overrides
@@ -68,7 +69,7 @@ nav:
   - 🏢 Home: index.md
   - 💡 Basics: basic.md
   - 📚 Guides:
-    - Tables: guides/tables.md
+    - Create Ingest Update Delete: guides/tables.md
     - Vector Search: search.md
     - SQL filters: sql.md
     - Indexing: ann_indexes.md
@@ -96,9 +97,11 @@ nav:
     - Serverless Website Chatbot: examples/serverless_website_chatbot.md
     - YouTube Transcript Search: examples/youtube_transcript_bot_with_nodejs.md
     - TransformersJS Embedding Search: examples/transformerjs_embedding_search_nodejs.md
+  - ⚙️ CLI & Config: cli_config.md
+
 - Basics: basic.md
 - Guides:
-  - Tables: guides/tables.md
+  - Create Ingest Update Delete: guides/tables.md
   - Vector Search: search.md
   - SQL filters: sql.md
   - Indexing: ann_indexes.md

docs/src/ann_indexes.md

@@ -97,7 +97,7 @@ There are a couple of parameters that can be used to fine-tune the search:
          .limit(2) \
          .nprobes(20) \
          .refine_factor(10) \
-         .to_df()
+         .to_pandas()
      ```
      ```
                                               vector       item       _distance
@@ -124,7 +124,7 @@ You can further filter the elements returned by a search using a where clause.
 
 === "Python"
      ```python
-     tbl.search(np.random.random((1536))).where("item != 'item 1141'").to_df()
+     tbl.search(np.random.random((1536))).where("item != 'item 1141'").to_pandas()
      ```
 
 === "Javascript"
@@ -141,7 +141,7 @@ You can select the columns returned by the query using a select clause.
 
 === "Python"
      ```python
-     tbl.search(np.random.random((1536))).select(["vector"]).to_df()
+     tbl.search(np.random.random((1536))).select(["vector"]).to_pandas()
      ```
      ```
         vector                                             _distance

docs/src/basic.md

@@ -146,7 +146,7 @@ Once you've embedded the query, you can find its nearest neighbors using the fol
 
 === "Python"
       ```python
-      tbl.search([100, 100]).limit(2).to_df()
+      tbl.search([100, 100]).limit(2).to_pandas()
       ```
 
       This returns a pandas DataFrame with the results.

docs/src/cli_config.md

@@ -0,0 +1,37 @@
+
+## LanceDB CLI
+Once lanceDB is installed, you can access the CLI using `lancedb` command on the console
+```
+lancedb
+```
+This lists out all the various command-line options available. You can get the usage or help for a particular command
+```
+lancedb {command} --help
+```
+
+## LanceDB config
+LanceDB uses a global config file to store certain settings. These settings are configurable using the lanceDB cli.
+To view your config settings, you can use:
+```
+lancedb config
+```
+These config parameters can be tuned using the cli.
+```
+lancedb {config_name} --{argument}
+```
+
+## LanceDB Opt-in Diagnostics
+When enabled, LanceDB will send anonymous events to help us improve LanceDB. These diagnostics are used only for error reporting and no data is collected. Error & stats allow us to automate certain aspects of bug reporting, prioritization of fixes and feature requests.
+These diagnostics are opt-in and can be enabled or disabled using the `lancedb diagnostics` command. These are enabled by default.
+Get usage help.
+```
+lancedb diagnostics --help
+```
+Disable diagnostics
+```
+lancedb diagnostics --disabled
+```
+Enable diagnostics
+```
+lancedb diagnostics --enabled
+```

docs/src/embedding.md

@@ -118,7 +118,7 @@ belong in the same latent space and your results will be nonsensical.
      ```python
      query = "What's the best pizza topping?"
      query_vector = embed_func([query])[0]
-     tbl.search(query_vector).limit(10).to_df()
+     tbl.search(query_vector).limit(10).to_pandas()
      ```
 
      The above snippet returns a pandas DataFrame with the 10 closest vectors to the query.

docs/src/examples/serverless_lancedb_with_s3_and_lambda.md

@@ -80,14 +80,14 @@ def handler(event, context):
     # Shape of SIFT is (128,1M), d=float32
     query_vector = np.array(event['query_vector'], dtype=np.float32)
 
-    rs = table.search(query_vector).limit(2).to_df()
+    rs = table.search(query_vector).limit(2).to_list()
 
     return {
         "statusCode": status_code,
         "headers": {
             "Content-Type": "application/json"
         },
-        "body": rs.to_json()
+        "body": json.dumps(rs)
     }
 ``` 
 

docs/src/fts.md

@@ -43,7 +43,13 @@ table.create_fts_index("text")
 To search:
 
 ```python
-df = table.search("puppy").limit(10).select(["text"]).to_df()
+table.search("puppy").limit(10).select(["text"]).to_list()
+```
+
+Which returns a list of dictionaries:
+
+```python
+[{'text': 'Frodo was a happy puppy', 'score': 0.6931471824645996}]
 ```
 
 LanceDB automatically looks for an FTS index if the input is str.

docs/src/guides/tables.md

@@ -364,6 +364,48 @@ Use the `delete()` method on tables to delete rows from a table. To choose which
     await tbl.countRows() // Returns 1
     ```
 
+### Updating a Table [Experimental]
+EXPERIMENTAL: Update rows in the table (not threadsafe).
+
+This can be used to update zero to all rows depending on how many rows match the where clause.
+
+| Parameter | Type | Description |
+|---|---|---|
+| `where` | `str` | The SQL where clause to use when updating rows. For example, `'x = 2'` or `'x IN (1, 2, 3)'`. The filter must not be empty, or it will error. |
+| `values` | `dict` | The values to update. The keys are the column names and the values are the values to set. |
+
+
+=== "Python"
+
+    ```python
+    import lancedb
+    import pandas as pd
+
+    # Create a lancedb connection
+    db = lancedb.connect("./.lancedb")
+
+    # Create a table from a pandas DataFrame
+    data = pd.DataFrame({"x": [1, 2, 3], "vector": [[1, 2], [3, 4], [5, 6]]})
+    table = db.create_table("my_table", data)
+
+    # Update the table where x = 2
+    table.update(where="x = 2", values={"vector": [10, 10]})
+
+    # Get the updated table as a pandas DataFrame
+    df = table.to_pandas()
+
+    # Print the DataFrame
+    print(df)
+    ```
+
+    Output
+    ```shell
+        x  vector
+    0  1  [1.0, 2.0]
+    1  3  [5.0, 6.0]
+    2  2  [10.0, 10.0]
+    ```
+
 ## What's Next?
 
 Learn how to Query your tables and create indices

docs/src/index.md

@@ -36,7 +36,7 @@ LanceDB's core is written in Rust 🦀 and is built using <a href="https://githu
       table = db.create_table("my_table",
                               data=[{"vector": [3.1, 4.1], "item": "foo", "price": 10.0},
                                     {"vector": [5.9, 26.5], "item": "bar", "price": 20.0}])
-      result = table.search([100, 100]).limit(2).to_df()
+      result = table.search([100, 100]).limit(2).to_list()
       ```
 
 === "Javascript"

docs/src/notebooks/multimodal_search.ipynb

@@ -19,11 +19,11 @@
      "output_type": "stream",
      "text": [
       "\n",
-      "\u001b[1m[\u001b[0m\u001b[34;49mnotice\u001b[0m\u001b[1;39;49m]\u001b[0m\u001b[39;49m A new release of pip available: \u001b[0m\u001b[31;49m22.3.1\u001b[0m\u001b[39;49m -> \u001b[0m\u001b[32;49m23.1.2\u001b[0m\n",
+      "\u001B[1m[\u001B[0m\u001B[34;49mnotice\u001B[0m\u001B[1;39;49m]\u001B[0m\u001B[39;49m A new release of pip available: \u001B[0m\u001B[31;49m22.3.1\u001B[0m\u001B[39;49m -> \u001B[0m\u001B[32;49m23.1.2\u001B[0m\n",
-      "\u001b[1m[\u001b[0m\u001b[34;49mnotice\u001b[0m\u001b[1;39;49m]\u001b[0m\u001b[39;49m To update, run: \u001b[0m\u001b[32;49mpip install --upgrade pip\u001b[0m\n",
+      "\u001B[1m[\u001B[0m\u001B[34;49mnotice\u001B[0m\u001B[1;39;49m]\u001B[0m\u001B[39;49m To update, run: \u001B[0m\u001B[32;49mpip install --upgrade pip\u001B[0m\n",
       "\n",
-      "\u001b[1m[\u001b[0m\u001b[34;49mnotice\u001b[0m\u001b[1;39;49m]\u001b[0m\u001b[39;49m A new release of pip available: \u001b[0m\u001b[31;49m22.3.1\u001b[0m\u001b[39;49m -> \u001b[0m\u001b[32;49m23.1.2\u001b[0m\n",
+      "\u001B[1m[\u001B[0m\u001B[34;49mnotice\u001B[0m\u001B[1;39;49m]\u001B[0m\u001B[39;49m A new release of pip available: \u001B[0m\u001B[31;49m22.3.1\u001B[0m\u001B[39;49m -> \u001B[0m\u001B[32;49m23.1.2\u001B[0m\n",
-      "\u001b[1m[\u001b[0m\u001b[34;49mnotice\u001b[0m\u001b[1;39;49m]\u001b[0m\u001b[39;49m To update, run: \u001b[0m\u001b[32;49mpip install --upgrade pip\u001b[0m\n"
+      "\u001B[1m[\u001B[0m\u001B[34;49mnotice\u001B[0m\u001B[1;39;49m]\u001B[0m\u001B[39;49m To update, run: \u001B[0m\u001B[32;49mpip install --upgrade pip\u001B[0m\n"
      ]
     }
    ],
@@ -39,6 +39,7 @@
    "outputs": [],
    "source": [
     "import io\n",
+    "\n",
     "import PIL\n",
     "import duckdb\n",
     "import lancedb"
@@ -158,18 +159,18 @@
     "        \"db = lancedb.connect('~/datasets/demo')\\n\"\n",
     "        \"tbl = db.open_table('diffusiondb')\\n\\n\"\n",
     "        f\"embedding = embed_func('{query}')\\n\"\n",
-    "        \"tbl.search(embedding).limit(9).to_df()\"\n",
+    "        \"tbl.search(embedding).limit(9).to_pandas()\"\n",
     "    )\n",
-    "    return (_extract(tbl.search(emb).limit(9).to_df()), code)\n",
+    "    return (_extract(tbl.search(emb).limit(9).to_pandas()), code)\n",
     "\n",
     "def find_image_keywords(query):\n",
     "    code = (\n",
     "        \"import lancedb\\n\"\n",
     "        \"db = lancedb.connect('~/datasets/demo')\\n\"\n",
     "        \"tbl = db.open_table('diffusiondb')\\n\\n\"\n",
-    "        f\"tbl.search('{query}').limit(9).to_df()\"\n",
+    "        f\"tbl.search('{query}').limit(9).to_pandas()\"\n",
     "    )\n",
-    "    return (_extract(tbl.search(query).limit(9).to_df()), code)\n",
+    "    return (_extract(tbl.search(query).limit(9).to_pandas()), code)\n",
     "\n",
     "def find_image_sql(query):\n",
     "    code = (\n",

docs/src/notebooks/youtube_transcript_search.ipynb

@@ -27,11 +27,11 @@
      "output_type": "stream",
      "text": [
       "\n",
-      "\u001b[1m[\u001b[0m\u001b[34;49mnotice\u001b[0m\u001b[1;39;49m]\u001b[0m\u001b[39;49m A new release of pip is available: \u001b[0m\u001b[31;49m23.0\u001b[0m\u001b[39;49m -> \u001b[0m\u001b[32;49m23.1.1\u001b[0m\n",
+      "\u001B[1m[\u001B[0m\u001B[34;49mnotice\u001B[0m\u001B[1;39;49m]\u001B[0m\u001B[39;49m A new release of pip is available: \u001B[0m\u001B[31;49m23.0\u001B[0m\u001B[39;49m -> \u001B[0m\u001B[32;49m23.1.1\u001B[0m\n",
-      "\u001b[1m[\u001b[0m\u001b[34;49mnotice\u001b[0m\u001b[1;39;49m]\u001b[0m\u001b[39;49m To update, run: \u001b[0m\u001b[32;49mpip install --upgrade pip\u001b[0m\n",
+      "\u001B[1m[\u001B[0m\u001B[34;49mnotice\u001B[0m\u001B[1;39;49m]\u001B[0m\u001B[39;49m To update, run: \u001B[0m\u001B[32;49mpip install --upgrade pip\u001B[0m\n",
       "\n",
-      "\u001b[1m[\u001b[0m\u001b[34;49mnotice\u001b[0m\u001b[1;39;49m]\u001b[0m\u001b[39;49m A new release of pip is available: \u001b[0m\u001b[31;49m23.0\u001b[0m\u001b[39;49m -> \u001b[0m\u001b[32;49m23.1.1\u001b[0m\n",
+      "\u001B[1m[\u001B[0m\u001B[34;49mnotice\u001B[0m\u001B[1;39;49m]\u001B[0m\u001B[39;49m A new release of pip is available: \u001B[0m\u001B[31;49m23.0\u001B[0m\u001B[39;49m -> \u001B[0m\u001B[32;49m23.1.1\u001B[0m\n",
-      "\u001b[1m[\u001b[0m\u001b[34;49mnotice\u001b[0m\u001b[1;39;49m]\u001b[0m\u001b[39;49m To update, run: \u001b[0m\u001b[32;49mpip install --upgrade pip\u001b[0m\n"
+      "\u001B[1m[\u001B[0m\u001B[34;49mnotice\u001B[0m\u001B[1;39;49m]\u001B[0m\u001B[39;49m To update, run: \u001B[0m\u001B[32;49mpip install --upgrade pip\u001B[0m\n"
      ]
     }
    ],
@@ -184,7 +184,7 @@
     "df = (contextualize(data.to_pandas())\n",
     "      .groupby(\"title\").text_col(\"text\")\n",
     "      .window(20).stride(4)\n",
-    "      .to_df())\n",
+    "      .to_pandas())\n",
     "df.head(1)"
    ]
   },
@@ -603,7 +603,7 @@
    "outputs": [],
    "source": [
     "# Use LanceDB to get top 3 most relevant context\n",
-    "context = tbl.search(emb).limit(3).to_df()"
+    "context = tbl.search(emb).limit(3).to_pandas()"
    ]
   },
   {

docs/src/python/arrow.md

@@ -74,7 +74,7 @@ table = db.open_table("pd_table")
 
 query_vector = [100, 100]
 # Pandas DataFrame
-df = table.search(query_vector).limit(1).to_df()
+df = table.search(query_vector).limit(1).to_pandas()
 print(df)
 ```
 
@@ -89,12 +89,12 @@ If you have more complex criteria, you can always apply the filter to the result
 ```python
 
 # Apply the filter via LanceDB
-results = table.search([100, 100]).where("price < 15").to_df()
+results = table.search([100, 100]).where("price < 15").to_pandas()
 assert len(results) == 1
 assert results["item"].iloc[0] == "foo"
 
 # Apply the filter via Pandas
-df = results = table.search([100, 100]).to_df()
+df = results = table.search([100, 100]).to_pandas()
 results = df[df.price < 15]
 assert len(results) == 1
 assert results["item"].iloc[0] == "foo"

docs/src/search.md

@@ -67,7 +67,7 @@ await db_setup.createTable('my_vectors', data)
 
     df = tbl.search(np.random.random((1536))) \
         .limit(10) \
-        .to_df()
+        .to_list()
     ```
 
 === "JavaScript"
@@ -92,7 +92,7 @@ as well.
     df = tbl.search(np.random.random((1536))) \
         .metric("cosine") \
         .limit(10) \
-        .to_df()
+        .to_list()
     ```
 
 

node/src/index.ts

@@ -23,7 +23,7 @@ import { Query } from './query'
 import { isEmbeddingFunction } from './embedding/embedding_function'
 
 // eslint-disable-next-line @typescript-eslint/no-var-requires
-const { databaseNew, databaseTableNames, databaseOpenTable, databaseDropTable, tableCreate, tableAdd, tableCreateVectorIndex, tableCountRows, tableDelete } = require('../native.js')
+const { databaseNew, databaseTableNames, databaseOpenTable, databaseDropTable, tableCreate, tableAdd, tableCreateVectorIndex, tableCountRows, tableDelete, tableCleanupOldVersions, tableCompactFiles } = require('../native.js')
 
 export { Query }
 export type { EmbeddingFunction }
@@ -459,6 +459,111 @@ export class LocalTable<T = number[]> implements Table<T> {
   async delete (filter: string): Promise<void> {
     return tableDelete.call(this._tbl, filter).then((newTable: any) => { this._tbl = newTable })
   }
+
+  /**
+   * Clean up old versions of the table, freeing disk space.
+   *
+   * @param olderThan The minimum age in minutes of the versions to delete. If not
+   *                  provided, defaults to two weeks.
+   * @param deleteUnverified Because they may be part of an in-progress
+   *                  transaction, uncommitted files newer than 7 days old are
+   *                  not deleted by default. This means that failed transactions
+   *                  can leave around data that takes up disk space for up to
+   *                  7 days. You can override this safety mechanism by setting
+   *                 this option to `true`, only if you promise there are no
+   *                 in progress writes while you run this operation. Failure to
+   *                 uphold this promise can lead to corrupted tables.
+   * @returns
+   */
+  async cleanupOldVersions (olderThan?: number, deleteUnverified?: boolean): Promise<CleanupStats> {
+    return tableCleanupOldVersions.call(this._tbl, olderThan, deleteUnverified)
+      .then((res: { newTable: any, metrics: CleanupStats }) => {
+        this._tbl = res.newTable
+        return res.metrics
+      })
+  }
+
+  /**
+   * Run the compaction process on the table.
+   *
+   * This can be run after making several small appends to optimize the table
+   * for faster reads.
+   *
+   * @param options Advanced options configuring compaction. In most cases, you
+   *               can omit this arguments, as the default options are sensible
+   *               for most tables.
+   * @returns Metrics about the compaction operation.
+   */
+  async compactFiles (options?: CompactionOptions): Promise<CompactionMetrics> {
+    const optionsArg = options ?? {}
+    return tableCompactFiles.call(this._tbl, optionsArg)
+      .then((res: { newTable: any, metrics: CompactionMetrics }) => {
+        this._tbl = res.newTable
+        return res.metrics
+      })
+  }
+}
+
+export interface CleanupStats {
+  /**
+   * The number of bytes removed from disk.
+   */
+  bytesRemoved: number
+  /**
+   * The number of old table versions removed.
+   */
+  oldVersions: number
+}
+
+export interface CompactionOptions {
+  /**
+   * The number of rows per fragment to target. Fragments that have fewer rows
+   * will be compacted into adjacent fragments to produce larger fragments.
+   * Defaults to 1024 * 1024.
+   */
+  targetRowsPerFragment?: number
+  /**
+   * The maximum number of rows per group. Defaults to 1024.
+   */
+  maxRowsPerGroup?: number
+  /**
+   * If true, fragments that have rows that are deleted may be compacted to
+   * remove the deleted rows. This can improve the performance of queries.
+   * Default is true.
+   */
+  materializeDeletions?: boolean
+  /**
+   * A number between 0 and 1, representing the proportion of rows that must be
+   * marked deleted before a fragment is a candidate for compaction to remove
+   * the deleted rows. Default is 10%.
+   */
+  materializeDeletionsThreshold?: number
+  /**
+   * The number of threads to use for compaction. If not provided, defaults to
+   * the number of cores on the machine.
+   */
+  numThreads?: number
+}
+
+export interface CompactionMetrics {
+  /**
+   * The number of fragments that were removed.
+   */
+  fragmentsRemoved: number
+  /**
+   * The number of new fragments that were created.
+   */
+  fragmentsAdded: number
+  /**
+   * The number of files that were removed. Each fragment may have more than one
+   * file.
+   */
+  filesRemoved: number
+  /**
+   * The number of files added. This is typically equal to the number of
+   * fragments added.
+   */
+  filesAdded: number
 }
 
 /// Config to build IVF_PQ index.

node/src/test/test.ts

@@ -18,7 +18,7 @@ import * as chai from 'chai'
 import * as chaiAsPromised from 'chai-as-promised'
 
 import * as lancedb from '../index'
-import { type AwsCredentials, type EmbeddingFunction, MetricType, Query, WriteMode, DefaultWriteOptions, isWriteOptions } from '../index'
+import { type AwsCredentials, type EmbeddingFunction, MetricType, Query, WriteMode, DefaultWriteOptions, isWriteOptions, type LocalTable } from '../index'
 import { FixedSizeList, Field, Int32, makeVector, Schema, Utf8, Table as ArrowTable, vectorFromArray, Float32 } from 'apache-arrow'
 
 const expect = chai.expect
@@ -446,3 +446,45 @@ describe('WriteOptions', function () {
     })
   })
 })
+
+describe('Compact and cleanup', function () {
+  it('can cleanup after compaction', async function () {
+    const dir = await track().mkdir('lancejs')
+    const con = await lancedb.connect(dir)
+
+    const data = [
+      { price: 10, name: 'foo', vector: [1, 2, 3] },
+      { price: 50, name: 'bar', vector: [4, 5, 6] }
+    ]
+    const table = await con.createTable('t1', data) as LocalTable
+
+    const newData = [
+      { price: 30, name: 'baz', vector: [7, 8, 9] }
+    ]
+    await table.add(newData)
+
+    const compactionMetrics = await table.compactFiles({
+      numThreads: 2
+    })
+    assert.equal(compactionMetrics.fragmentsRemoved, 2)
+    assert.equal(compactionMetrics.fragmentsAdded, 1)
+    assert.equal(await table.countRows(), 3)
+
+    await table.cleanupOldVersions()
+    assert.equal(await table.countRows(), 3)
+
+    // should have no effect, but this validates the arguments are parsed.
+    await table.compactFiles({
+      targetRowsPerFragment: 1024 * 10,
+      maxRowsPerGroup: 1024,
+      materializeDeletions: true,
+      materializeDeletionsThreshold: 0.5,
+      numThreads: 2
+    })
+
+    const cleanupMetrics = await table.cleanupOldVersions(0, true)
+    assert.isAtLeast(cleanupMetrics.bytesRemoved, 1)
+    assert.isAtLeast(cleanupMetrics.oldVersions, 1)
+    assert.equal(await table.countRows(), 3)
+  })
+})

python/.bumpversion.cfg

@@ -1,5 +1,5 @@
 [bumpversion]
-current_version = 0.3.0
+current_version = 0.3.1
 commit = True
 message = [python] Bump version: {current_version} → {new_version}
 tag = True

python/README.md

@@ -16,7 +16,7 @@ pip install lancedb
 import lancedb
 db = lancedb.connect('<PATH_TO_LANCEDB_DATASET>')
 table = db.open_table('my_table')
-results = table.search([0.1, 0.3]).limit(20).to_df()
+results = table.search([0.1, 0.3]).limit(20).to_list()
 print(results)
 ```
 

python/lancedb/__init__.py

@@ -14,11 +14,12 @@
 import importlib.metadata
 from typing import Optional
 
+__version__ = importlib.metadata.version("lancedb")
+
 from .db import URI, DBConnection, LanceDBConnection
 from .remote.db import RemoteDBConnection
 from .schema import vector
-
+from .utils import sentry_log
-__version__ = importlib.metadata.version("lancedb")
 
 
 def connect(

python/lancedb/cli/__init__.py

@@ -0,0 +1,12 @@
+#  Copyright 2023 LanceDB Developers
+#
+#  Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+#  You may obtain a copy of the License at
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+#  Unless required by applicable law or agreed to in writing, software
+#  distributed under the License is distributed on an "AS IS" BASIS,
+#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#  See the License for the specific language governing permissions and
+#  limitations under the License.

python/lancedb/cli/cli.py

@@ -0,0 +1,46 @@
+#  Copyright 2023 LanceDB Developers
+#
+#  Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+#  You may obtain a copy of the License at
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+#  Unless required by applicable law or agreed to in writing, software
+#  distributed under the License is distributed on an "AS IS" BASIS,
+#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#  See the License for the specific language governing permissions and
+#  limitations under the License.
+
+import click
+
+from lancedb.utils import CONFIG
+
+
+@click.group()
+@click.version_option(help="LanceDB command line interface entry point")
+def cli():
+    "LanceDB command line interface"
+
+
+diagnostics_help = """
+Enable or disable LanceDB diagnostics. When enabled, LanceDB will send anonymous events to help us improve LanceDB.
+These diagnostics are used only for error reporting and no data is collected. You can find more about diagnosis on
+our docs: https://lancedb.github.io/lancedb/cli_config/
+"""
+
+
+@cli.command(help=diagnostics_help)
+@click.option("--enabled/--disabled", default=True)
+def diagnostics(enabled):
+    CONFIG.update({"diagnostics": True if enabled else False})
+    click.echo("LanceDB diagnostics is %s" % ("enabled" if enabled else "disabled"))
+
+
+@cli.command(help="Show current LanceDB configuration")
+def config():
+    # TODO: pretty print as table with colors and formatting
+    click.echo("Current LanceDB configuration:")
+    cfg = CONFIG.copy()
+    cfg.pop("uuid")  # Don't show uuid as it is not configurable
+    for item, amount in cfg.items():
+        click.echo("{} ({})".format(item, amount))

python/lancedb/context.py

@@ -12,6 +12,9 @@
 #  limitations under the License.
 from __future__ import annotations
 
+import deprecation
+
+from . import __version__
 from .exceptions import MissingColumnError, MissingValueError
 from .util import safe_import_pandas
 
@@ -43,7 +46,7 @@ def contextualize(raw_df: "pd.DataFrame") -> Contextualizer:
     this how many tokens, but depending on the input data, it could be sentences,
     paragraphs, messages, etc.
 
-    >>> contextualize(data).window(3).stride(1).text_col('token').to_df()
+    >>> contextualize(data).window(3).stride(1).text_col('token').to_pandas()
                     token  document_id
     0     The quick brown            1
     1     quick brown fox            1
@@ -56,7 +59,7 @@ def contextualize(raw_df: "pd.DataFrame") -> Contextualizer:
     8          dog I love            1
     9   I love sandwiches            2
     10    love sandwiches            2
-    >>> contextualize(data).window(7).stride(1).min_window_size(7).text_col('token').to_df()
+    >>> contextualize(data).window(7).stride(1).min_window_size(7).text_col('token').to_pandas()
                                       token  document_id
     0   The quick brown fox jumped over the            1
     1  quick brown fox jumped over the lazy            1
@@ -68,7 +71,7 @@ def contextualize(raw_df: "pd.DataFrame") -> Contextualizer:
     ``stride`` determines how many rows to skip between each window start. This can
     be used to reduce the total number of windows generated.
 
-    >>> contextualize(data).window(4).stride(2).text_col('token').to_df()
+    >>> contextualize(data).window(4).stride(2).text_col('token').to_pandas()
                         token  document_id
     0     The quick brown fox            1
     2   brown fox jumped over            1
@@ -81,7 +84,7 @@ def contextualize(raw_df: "pd.DataFrame") -> Contextualizer:
     context windows that don't cross document boundaries. In this case, we can
     pass ``document_id`` as the group by.
 
-    >>> contextualize(data).window(4).stride(2).text_col('token').groupby('document_id').to_df()
+    >>> contextualize(data).window(4).stride(2).text_col('token').groupby('document_id').to_pandas()
                        token  document_id
     0    The quick brown fox            1
     2  brown fox jumped over            1
@@ -93,14 +96,14 @@ def contextualize(raw_df: "pd.DataFrame") -> Contextualizer:
     This can be used to trim the last few context windows which have size less than
     ``min_window_size``. By default context windows of size 1 are skipped.
 
-    >>> contextualize(data).window(6).stride(3).text_col('token').groupby('document_id').to_df()
+    >>> contextualize(data).window(6).stride(3).text_col('token').groupby('document_id').to_pandas()
                                  token  document_id
     0  The quick brown fox jumped over            1
     3     fox jumped over the lazy dog            1
     6                     the lazy dog            1
     9                I love sandwiches            2
 
-    >>> contextualize(data).window(6).stride(3).min_window_size(4).text_col('token').groupby('document_id').to_df()
+    >>> contextualize(data).window(6).stride(3).min_window_size(4).text_col('token').groupby('document_id').to_pandas()
                                  token  document_id
     0  The quick brown fox jumped over            1
     3     fox jumped over the lazy dog            1
@@ -176,7 +179,16 @@ class Contextualizer:
         self._min_window_size = min_window_size
         return self
 
+    @deprecation.deprecated(
+        deprecated_in="0.3.1",
+        removed_in="0.4.0",
+        current_version=__version__,
+        details="Use the bar function instead",
+    )
     def to_df(self) -> "pd.DataFrame":
+        return self.to_pandas()
+
+    def to_pandas(self) -> "pd.DataFrame":
         """Create the context windows and return a DataFrame."""
         if pd is None:
             raise ImportError(

python/lancedb/embeddings/__init__.py

@@ -12,6 +12,7 @@
 #  limitations under the License.
 
 
+from .cohere import CohereEmbeddingFunction
 from .functions import (
     EmbeddingFunction,
     EmbeddingFunctionConfig,

python/lancedb/embeddings/cohere.py

@@ -0,0 +1,86 @@
+#  Copyright (c) 2023. LanceDB Developers
+#
+#  Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+#  You may obtain a copy of the License at
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+#  Unless required by applicable law or agreed to in writing, software
+#  distributed under the License is distributed on an "AS IS" BASIS,
+#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#  See the License for the specific language governing permissions and
+#  limitations under the License.
+
+import os
+from typing import ClassVar, List, Union
+
+import numpy as np
+
+from .functions import TextEmbeddingFunction, register
+from .utils import api_key_not_found_help
+
+
+@register("cohere")
+class CohereEmbeddingFunction(TextEmbeddingFunction):
+    """
+    An embedding function that uses the Cohere API
+
+    https://docs.cohere.com/docs/multilingual-language-models
+
+    Parameters
+    ----------
+    name: str, default "embed-multilingual-v2.0"
+        The name of the model to use. See the Cohere documentation for a list of available models.
+
+    Examples
+    --------
+    import lancedb
+    from lancedb.pydantic import LanceModel, Vector
+    from lancedb.embeddings import EmbeddingFunctionRegistry
+
+    cohere = EmbeddingFunctionRegistry.get_instance().get("cohere").create(name="embed-multilingual-v2.0")
+
+    class TextModel(LanceModel):
+        text: str = cohere.SourceField()
+        vector: Vector(cohere.ndims()) =  cohere.VectorField()
+
+    data = [ { "text": "hello world" },
+            { "text": "goodbye world" }]
+
+    db = lancedb.connect("~/.lancedb")
+    tbl = db.create_table("test", schema=TextModel, mode="overwrite")
+
+    tbl.add(data)
+
+    """
+
+    name: str = "embed-multilingual-v2.0"
+    client: ClassVar = None
+
+    def ndims(self):
+        # TODO: fix hardcoding
+        return 768
+
+    def generate_embeddings(
+        self, texts: Union[List[str], np.ndarray]
+    ) -> List[np.array]:
+        """
+        Get the embeddings for the given texts
+
+        Parameters
+        ----------
+        texts: list[str] or np.ndarray (of str)
+            The texts to embed
+        """
+        # TODO retry, rate limit, token limit
+        self._init_client()
+        rs = CohereEmbeddingFunction.client.embed(texts=texts, model=self.name)
+
+        return [emb for emb in rs.embeddings]
+
+    def _init_client(self):
+        cohere = self.safe_import("cohere")
+        if CohereEmbeddingFunction.client is None:
+            if os.environ.get("COHERE_API_KEY") is None:
+                api_key_not_found_help("cohere")
+            CohereEmbeddingFunction.client = cohere.Client(os.environ["COHERE_API_KEY"])

python/lancedb/embeddings/utils.py

@@ -21,6 +21,7 @@ from lance.vector import vec_to_table
 from retry import retry
 
 from ..util import safe_import_pandas
+from ..utils.general import LOGGER
 
 pd = safe_import_pandas()
 DATA = Union[pa.Table, "pd.DataFrame"]
@@ -152,3 +153,8 @@ class FunctionWrapper:
             yield from tqdm(_chunker(arr), total=math.ceil(length / self._batch_size))
         else:
             yield from _chunker(arr)
+
+
+def api_key_not_found_help(provider):
+    LOGGER.error(f"Could not find API key for {provider}.")
+    raise ValueError(f"Please set the {provider.upper()}_API_KEY environment variable.")

python/lancedb/query.py

@@ -16,10 +16,12 @@ from __future__ import annotations
 from abc import ABC, abstractmethod
 from typing import List, Literal, Optional, Type, Union
 
+import deprecation
 import numpy as np
 import pyarrow as pa
 import pydantic
 
+from . import __version__
 from .common import VECTOR_COLUMN_NAME
 from .pydantic import LanceModel
 from .util import safe_import_pandas
@@ -127,7 +129,24 @@ class LanceQueryBuilder(ABC):
         self._columns = None
         self._where = None
 
+    @deprecation.deprecated(
+        deprecated_in="0.3.1",
+        removed_in="0.4.0",
+        current_version=__version__,
+        details="Use the bar function instead",
+    )
     def to_df(self) -> "pd.DataFrame":
+        """
+        Deprecated alias for `to_pandas()`. Please use `to_pandas()` instead.
+
+        Execute the query and return the results as a pandas DataFrame.
+        In addition to the selected columns, LanceDB also returns a vector
+        and also the "_distance" column which is the distance between the query
+        vector and the returned vector.
+        """
+        return self.to_pandas()
+
+    def to_pandas(self) -> "pd.DataFrame":
         """
         Execute the query and return the results as a pandas DataFrame.
         In addition to the selected columns, LanceDB also returns a vector
@@ -148,6 +167,16 @@ class LanceQueryBuilder(ABC):
         """
         raise NotImplementedError
 
+    def to_list(self) -> List[dict]:
+        """
+        Execute the query and return the results as a list of dictionaries.
+
+        Each list entry is a dictionary with the selected column names as keys,
+        or all table columns if `select` is not called. The vector and the "_distance"
+        fields are returned whether or not they're explicitly selected.
+        """
+        return self.to_arrow().to_pylist()
+
     def to_pydantic(self, model: Type[LanceModel]) -> List[LanceModel]:
         """Return the table as a list of pydantic models.
 
@@ -232,7 +261,7 @@ class LanceVectorQueryBuilder(LanceQueryBuilder):
     ...       .where("b < 10")
     ...       .select(["b"])
     ...       .limit(2)
-    ...       .to_df())
+    ...       .to_pandas())
        b      vector  _distance
     0  6  [0.4, 0.4]        0.0
     """

python/lancedb/table.py

@@ -16,6 +16,7 @@ from __future__ import annotations
 import inspect
 import os
 from abc import ABC, abstractmethod
+from datetime import timedelta
 from functools import cached_property
 from typing import Any, Iterable, List, Optional, Union
 
@@ -24,7 +25,7 @@ import numpy as np
 import pyarrow as pa
 import pyarrow.compute as pc
 from lance import LanceDataset
-from lance.dataset import ReaderLike
+from lance.dataset import CleanupStats, ReaderLike
 from lance.vector import vec_to_table
 
 from .common import DATA, VEC, VECTOR_COLUMN_NAME
@@ -33,6 +34,7 @@ from .embeddings.functions import EmbeddingFunctionConfig
 from .pydantic import LanceModel
 from .query import LanceQueryBuilder, Query
 from .util import fs_from_uri, safe_import_pandas
+from .utils.events import register_event
 
 pd = safe_import_pandas()
 
@@ -136,7 +138,7 @@ class Table(ABC):
 
     Can query the table with [Table.search][lancedb.table.Table.search].
 
-    >>> table.search([0.4, 0.4]).select(["b"]).to_df()
+    >>> table.search([0.4, 0.4]).select(["b"]).to_pandas()
        b      vector  _distance
     0  4  [0.5, 1.3]       0.82
     1  2  [1.1, 1.2]       1.13
@@ -394,6 +396,17 @@ class LanceTable(Table):
             raise ValueError(f"Invalid version {version}")
         self._reset_dataset(version=version)
 
+        try:
+            # Accessing the property updates the cached value
+            _ = self._dataset
+        except Exception as e:
+            if "not found" in str(e):
+                raise ValueError(
+                    f"Version {version} no longer exists. Was it cleaned up?"
+                )
+            else:
+                raise e
+
     def restore(self, version: int = None):
         """Restore a version of the table. This is an in-place operation.
 
@@ -496,6 +509,7 @@ class LanceTable(Table):
             accelerator=accelerator,
         )
         self._reset_dataset()
+        register_event("create_index")
 
     def create_fts_index(self, field_names: Union[str, List[str]]):
         """Create a full-text search index on the table.
@@ -514,6 +528,7 @@ class LanceTable(Table):
             field_names = [field_names]
         index = create_index(self._get_fts_index_path(), field_names)
         populate_index(index, self, field_names)
+        register_event("create_fts_index")
 
     def _get_fts_index_path(self):
         return os.path.join(self._dataset_uri, "_indices", "tantivy")
@@ -566,6 +581,7 @@ class LanceTable(Table):
         )
         lance.write_dataset(data, self._dataset_uri, schema=self.schema, mode=mode)
         self._reset_dataset()
+        register_event("add")
 
     def merge(
         self,
@@ -629,6 +645,7 @@ class LanceTable(Table):
             other_table, left_on=left_on, right_on=right_on, schema=schema
         )
         self._reset_dataset()
+        register_event("merge")
 
     @cached_property
     def embedding_functions(self) -> dict:
@@ -679,6 +696,7 @@ class LanceTable(Table):
             and also the "_distance" column which is the distance between the query
             vector and the returned vector.
         """
+        register_event("search")
         return LanceQueryBuilder.create(
             self, query, query_type, vector_column_name=vector_column_name
         )
@@ -782,6 +800,7 @@ class LanceTable(Table):
         if data is not None:
             table.add(data)
 
+        register_event("create_table")
         return table
 
     @classmethod
@@ -847,6 +866,7 @@ class LanceTable(Table):
         self.delete(where)
         self.add(orig_data, mode="append")
         self._reset_dataset()
+        register_event("update")
 
     def _execute_query(self, query: Query) -> pa.Table:
         ds = self.to_lance()
@@ -870,6 +890,48 @@ class LanceTable(Table):
             },
         )
 
+    def cleanup_old_versions(
+        self,
+        older_than: Optional[timedelta] = None,
+        *,
+        delete_unverified: bool = False,
+    ) -> CleanupStats:
+        """
+        Clean up old versions of the table, freeing disk space.
+
+        Parameters
+        ----------
+        older_than: timedelta, default None
+            The minimum age of the version to delete. If None, then this defaults
+            to two weeks.
+        delete_unverified: bool, default False
+            Because they may be part of an in-progress transaction, files newer
+            than 7 days old are not deleted by default. If you are sure that
+            there are no in-progress transactions, then you can set this to True
+            to delete all files older than `older_than`.
+
+        Returns
+        -------
+        CleanupStats
+            The stats of the cleanup operation, including how many bytes were
+            freed.
+        """
+        return self.to_lance().cleanup_old_versions(
+            older_than, delete_unverified=delete_unverified
+        )
+
+    def compact_files(self, *args, **kwargs):
+        """
+        Run the compaction process on the table.
+
+        This can be run after making several small appends to optimize the table
+        for faster reads.
+
+        Arguments are passed onto :meth:`lance.dataset.DatasetOptimizer.compact_files`.
+        For most cases, the default should be fine.
+        """
+        return self.to_lance().optimize.compact_files(*args, **kwargs)
+
 
 def _sanitize_schema(
     data: pa.Table,

python/lancedb/utils/__init__.py

@@ -0,0 +1,15 @@
+#  Copyright 2023 LanceDB Developers
+#
+#  Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+#  You may obtain a copy of the License at
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+#  Unless required by applicable law or agreed to in writing, software
+#  distributed under the License is distributed on an "AS IS" BASIS,
+#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#  See the License for the specific language governing permissions and
+#  limitations under the License.
+from .config import Config
+
+CONFIG = Config()

python/lancedb/utils/config.py

@@ -0,0 +1,116 @@
+#  Copyright 2023 LanceDB Developers
+#
+#  Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+#  You may obtain a copy of the License at
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+#  Unless required by applicable law or agreed to in writing, software
+#  distributed under the License is distributed on an "AS IS" BASIS,
+#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#  See the License for the specific language governing permissions and
+#  limitations under the License.
+
+import copy
+import hashlib
+import os
+import platform
+import uuid
+from pathlib import Path
+
+from .general import LOGGER, is_dir_writeable, yaml_load, yaml_save
+
+
+def get_user_config_dir(sub_dir="lancedb"):
+    """
+    Get the user config directory.
+
+    Args:
+        sub_dir (str): The name of the subdirectory to create.
+
+    Returns:
+        (Path): The path to the user config directory.
+    """
+    # Return the appropriate config directory for each operating system
+    if platform.system() == "Windows":
+        path = Path.home() / "AppData" / "Roaming" / sub_dir
+    elif platform.system() == "Darwin":
+        path = Path.home() / "Library" / "Application Support" / sub_dir
+    elif platform.system() == "Linux":
+        path = Path.home() / ".config" / sub_dir
+    else:
+        raise ValueError(f"Unsupported operating system: {platform.system()}")
+
+    # GCP and AWS lambda fix, only /tmp is writeable
+    if not is_dir_writeable(path.parent):
+        LOGGER.warning(
+            f"WARNING ⚠️ user config directory '{path}' is not writeable, defaulting to '/tmp' or CWD."
+            "Alternatively you can define a LANCEDB_CONFIG_DIR environment variable for this path."
+        )
+        path = (
+            Path("/tmp") / sub_dir
+            if is_dir_writeable("/tmp")
+            else Path().cwd() / sub_dir
+        )
+
+    # Create the subdirectory if it does not exist
+    path.mkdir(parents=True, exist_ok=True)
+
+    return path
+
+
+USER_CONFIG_DIR = Path(os.getenv("LANCEDB_CONFIG_DIR") or get_user_config_dir())
+CONFIG_FILE = USER_CONFIG_DIR / "config.yaml"
+
+
+class Config(dict):
+    """
+    Manages lancedb config stored in a YAML file.
+
+    Args:
+        file (str | Path): Path to the lancedb config YAML file. Default is USER_CONFIG_DIR / 'config.yaml'.
+    """
+
+    def __init__(self, file=CONFIG_FILE):
+        self.file = Path(file)
+        self.defaults = {  # Default global config values
+            "diagnostics": True,
+            "uuid": hashlib.sha256(str(uuid.getnode()).encode()).hexdigest(),
+        }
+
+        super().__init__(copy.deepcopy(self.defaults))
+
+        if not self.file.exists():
+            self.save()
+
+        self.load()
+        correct_keys = self.keys() == self.defaults.keys()
+        correct_types = all(
+            type(a) is type(b) for a, b in zip(self.values(), self.defaults.values())
+        )
+        if not (correct_keys and correct_types):
+            LOGGER.warning(
+                "WARNING ⚠️ LanceDB settings reset to default values. This may be due to a possible problem "
+                "with your settings or a recent package update. "
+                f"\nView settings & usage with 'lancedb settings' or at '{self.file}'"
+            )
+            self.reset()
+
+    def load(self):
+        """Loads settings from the YAML file."""
+        super().update(yaml_load(self.file))
+
+    def save(self):
+        """Saves the current settings to the YAML file."""
+        yaml_save(self.file, dict(self))
+
+    def update(self, *args, **kwargs):
+        """Updates a setting value in the current settings."""
+        super().update(*args, **kwargs)
+        self.save()
+
+    def reset(self):
+        """Resets the settings to default and saves them."""
+        self.clear()
+        self.update(self.defaults)
+        self.save()

python/lancedb/utils/events.py

@@ -0,0 +1,161 @@
+#  Copyright 2023 LanceDB Developers
+#
+#  Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+#  You may obtain a copy of the License at
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+#  Unless required by applicable law or agreed to in writing, software
+#  distributed under the License is distributed on an "AS IS" BASIS,
+#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#  See the License for the specific language governing permissions and
+#  limitations under the License.
+
+import datetime
+import importlib.metadata
+import platform
+import random
+import sys
+import time
+
+from lancedb.utils import CONFIG
+from lancedb.utils.general import TryExcept
+
+from .general import (
+    PLATFORMS,
+    get_git_origin_url,
+    is_git_dir,
+    is_github_actions_ci,
+    is_online,
+    is_pip_package,
+    is_pytest_running,
+    threaded_request,
+)
+
+
+class _Events:
+    """
+    A class for collecting anonymous event analytics. Event analytics are enabled when ``diagnostics=True`` in config and
+    disabled when ``diagnostics=False``.
+
+    You can enable or disable diagnostics by running ``lancedb diagnostics --enabled`` or ``lancedb diagnostics --disabled``.
+
+    Attributes
+    ----------
+    url : str
+        The URL to send anonymous events.
+    rate_limit : float
+        The rate limit in seconds for sending events.
+    metadata : dict
+        A dictionary containing metadata about the environment.
+    enabled : bool
+        A flag to enable or disable Events based on certain conditions.
+    """
+
+    _instance = None
+
+    url = "https://app.posthog.com/capture/"
+    headers = {"Content-Type": "application/json"}
+    api_key = "phc_oENDjGgHtmIDrV6puUiFem2RB4JA8gGWulfdulmMdZP"
+    # This api-key is write only and is safe to expose in the codebase.
+
+    def __init__(self):
+        """
+        Initializes the Events object with default values for events, rate_limit, and metadata.
+        """
+        self.events = []  # events list
+        self.max_events = 25  # max events to store in memory
+        self.rate_limit = 60.0  # rate limit (seconds)
+        self.time = 0.0
+
+        if is_git_dir():
+            install = "git"
+        elif is_pip_package():
+            install = "pip"
+        else:
+            install = "other"
+        self.metadata = {
+            "cli": sys.argv[0],
+            "install": install,
+            "python": ".".join(platform.python_version_tuple()[:2]),
+            "version": importlib.metadata.version("lancedb"),
+            "platforms": PLATFORMS,
+            "session_id": round(random.random() * 1e15),
+            # 'engagement_time_msec': 1000 # TODO: In future we might be interested in this metric
+        }
+
+        TESTS_RUNNING = is_pytest_running() or is_github_actions_ci()
+        ONLINE = is_online()
+        self.enabled = (
+            CONFIG["diagnostics"]
+            and not TESTS_RUNNING
+            and ONLINE
+            and (
+                is_pip_package()
+                or get_git_origin_url() == "https://github.com/lancedb/lancedb.git"
+            )
+        )
+
+    def __call__(self, event_name, params={}):
+        """
+        Attempts to add a new event to the events list and send events if the rate limit is reached.
+
+        Args
+        ----
+        event_name : str
+            The name of the event to be logged.
+        params : dict, optional
+            A dictionary of additional parameters to be logged with the event.
+        """
+        ### NOTE: We might need a way to tag a session with a label to check usage from a source. Setting label should be exposed to the user.
+        if not self.enabled:
+            return
+        if (
+            len(self.events) < self.max_events
+        ):  # Events list limited to 25 events (drop any events past this)
+            params.update(self.metadata)
+            self.events.append(
+                {
+                    "event": event_name,
+                    "properties": params,
+                    "timestamp": datetime.datetime.now(
+                        tz=datetime.timezone.utc
+                    ).isoformat(),
+                    "distinct_id": CONFIG["uuid"],
+                }
+            )
+
+        # Check rate limit
+        t = time.time()
+        if (t - self.time) < self.rate_limit:
+            return
+        # Time is over rate limiter, send now
+        data = {
+            "api_key": self.api_key,
+            "distinct_id": CONFIG["uuid"],  # posthog needs this to accepts the event
+            "batch": self.events,
+        }
+
+        # POST equivalent to requests.post(self.url, json=data).
+        # threaded request is used to avoid blocking, retries are disabled, and verbose is disabled
+        # to avoid any possible disruption in the console.
+        threaded_request(
+            method="post",
+            url=self.url,
+            headers=self.headers,
+            json=data,
+            retry=0,
+            verbose=False,
+        )
+
+        # Flush & Reset
+        self.events = []
+        self.time = t
+
+
+@TryExcept(verbose=False)
+def register_event(name: str, **kwargs):
+    if _Events._instance is None:
+        _Events._instance = _Events()
+
+    _Events._instance(name, **kwargs)

python/lancedb/utils/general.py

@@ -0,0 +1,445 @@
+#  Copyright 2023 LanceDB Developers
+#
+#  Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+#  You may obtain a copy of the License at
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+#  Unless required by applicable law or agreed to in writing, software
+#  distributed under the License is distributed on an "AS IS" BASIS,
+#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#  See the License for the specific language governing permissions and
+#  limitations under the License.
+
+import contextlib
+import importlib
+import logging.config
+import os
+import platform
+import subprocess
+import sys
+import threading
+import time
+from pathlib import Path
+from typing import Union
+
+import requests
+import yaml
+
+LOGGING_NAME = "lancedb"
+VERBOSE = (
+    str(os.getenv("LANCEDB_VERBOSE", True)).lower() == "true"
+)  # global verbose mode
+
+
+def set_logging(name=LOGGING_NAME, verbose=True):
+    """Sets up logging for the given name.
+
+    Parameters
+    ----------
+    name : str, optional
+        The name of the logger. Default is 'lancedb'.
+    verbose : bool, optional
+        Whether to enable verbose logging. Default is True.
+    """
+
+    rank = int(os.getenv("RANK", -1))  # rank in world for Multi-GPU trainings
+    level = logging.INFO if verbose and rank in {-1, 0} else logging.ERROR
+    logging.config.dictConfig(
+        {
+            "version": 1,
+            "disable_existing_loggers": False,
+            "formatters": {name: {"format": "%(message)s"}},
+            "handlers": {
+                name: {
+                    "class": "logging.StreamHandler",
+                    "formatter": name,
+                    "level": level,
+                }
+            },
+            "loggers": {name: {"level": level, "handlers": [name], "propagate": False}},
+        }
+    )
+
+
+set_logging(LOGGING_NAME, verbose=VERBOSE)
+LOGGER = logging.getLogger(LOGGING_NAME)
+
+
+def is_pip_package(filepath: str = __name__) -> bool:
+    """Determines if the file at the given filepath is part of a pip package.
+
+    Parameters
+    ----------
+    filepath : str, optional
+        The filepath to check. Default is the current file.
+
+    Returns
+    -------
+    bool
+        True if the file is part of a pip package, False otherwise.
+    """
+    # Get the spec for the module
+    spec = importlib.util.find_spec(filepath)
+
+    # Return whether the spec is not None and the origin is not None (indicating it is a package)
+    return spec is not None and spec.origin is not None
+
+
+def is_pytest_running():
+    """Determines whether pytest is currently running or not.
+
+    Returns
+    -------
+    bool
+        True if pytest is running, False otherwise.
+    """
+    return (
+        ("PYTEST_CURRENT_TEST" in os.environ)
+        or ("pytest" in sys.modules)
+        or ("pytest" in Path(sys.argv[0]).stem)
+    )
+
+
+def is_github_actions_ci() -> bool:
+    """
+    Determine if the current environment is a GitHub Actions CI Python runner.
+
+    Returns
+    -------
+    bool
+        True if the current environment is a GitHub Actions CI Python runner, False otherwise.
+    """
+
+    return (
+        "GITHUB_ACTIONS" in os.environ
+        and "RUNNER_OS" in os.environ
+        and "RUNNER_TOOL_CACHE" in os.environ
+    )
+
+
+def is_git_dir():
+    """
+    Determines whether the current file is part of a git repository.
+    If the current file is not part of a git repository, returns None.
+
+    Returns
+    -------
+    bool
+        True if current file is part of a git repository.
+    """
+    return get_git_dir() is not None
+
+
+def is_online() -> bool:
+    """
+    Check internet connectivity by attempting to connect to a known online host.
+
+    Returns
+    -------
+    bool
+        True if connection is successful, False otherwise.
+    """
+    import socket
+
+    for host in "1.1.1.1", "8.8.8.8", "223.5.5.5":  # Cloudflare, Google, AliDNS:
+        try:
+            test_connection = socket.create_connection(address=(host, 53), timeout=2)
+        except (socket.timeout, socket.gaierror, OSError):
+            continue
+        else:
+            # If the connection was successful, close it to avoid a ResourceWarning
+            test_connection.close()
+            return True
+    return False
+
+
+def is_dir_writeable(dir_path: Union[str, Path]) -> bool:
+    """Check if a directory is writeable.
+
+    Parameters
+    ----------
+    dir_path : Union[str, Path]
+        The path to the directory.
+
+    Returns
+    -------
+    bool
+        True if the directory is writeable, False otherwise.
+    """
+    return os.access(str(dir_path), os.W_OK)
+
+
+def is_colab():
+    """Check if the current script is running inside a Google Colab notebook.
+
+    Returns
+    -------
+    bool
+        True if running inside a Colab notebook, False otherwise.
+    """
+    return "COLAB_RELEASE_TAG" in os.environ or "COLAB_BACKEND_VERSION" in os.environ
+
+
+def is_kaggle():
+    """Check if the current script is running inside a Kaggle kernel.
+
+    Returns
+    -------
+    bool
+        True if running inside a Kaggle kernel, False otherwise.
+    """
+    return (
+        os.environ.get("PWD") == "/kaggle/working"
+        and os.environ.get("KAGGLE_URL_BASE") == "https://www.kaggle.com"
+    )
+
+
+def is_jupyter():
+    """Check if the current script is running inside a Jupyter Notebook.
+
+    Returns
+    -------
+    bool
+        True if running inside a Jupyter Notebook, False otherwise.
+    """
+    with contextlib.suppress(Exception):
+        from IPython import get_ipython
+
+        return get_ipython() is not None
+    return False
+
+
+def is_docker() -> bool:
+    """Determine if the script is running inside a Docker container.
+
+    Returns
+    -------
+    bool
+        True if the script is running inside a Docker container, False otherwise.
+    """
+    file = Path("/proc/self/cgroup")
+    if file.exists():
+        with open(file) as f:
+            return "docker" in f.read()
+    else:
+        return False
+
+
+def get_git_dir():
+    """Determine whether the current file is part of a git repository and if so, returns the repository root directory.
+    If the current file is not part of a git repository, returns None.
+
+    Returns
+    -------
+    Path | None
+        Git root directory if found or None if not found.
+    """
+    for d in Path(__file__).parents:
+        if (d / ".git").is_dir():
+            return d
+
+
+def get_git_origin_url():
+    """Retrieve the origin URL of a git repository.
+
+    Returns
+    -------
+    str | None
+        The origin URL of the git repository or None if not git directory.
+    """
+    if is_git_dir():
+        with contextlib.suppress(subprocess.CalledProcessError):
+            origin = subprocess.check_output(
+                ["git", "config", "--get", "remote.origin.url"]
+            )
+            return origin.decode().strip()
+
+
+def yaml_save(file="data.yaml", data=None, header=""):
+    """Save YAML data to a file.
+
+    Parameters
+    ----------
+    file : str, optional
+        File name, by default 'data.yaml'.
+    data : dict, optional
+        Data to save in YAML format, by default None.
+    header : str, optional
+        YAML header to add, by default "".
+    """
+    if data is None:
+        data = {}
+    file = Path(file)
+    if not file.parent.exists():
+        # Create parent directories if they don't exist
+        file.parent.mkdir(parents=True, exist_ok=True)
+
+    # Convert Path objects to strings
+    for k, v in data.items():
+        if isinstance(v, Path):
+            data[k] = str(v)
+
+    # Dump data to file in YAML format
+    with open(file, "w", errors="ignore", encoding="utf-8") as f:
+        if header:
+            f.write(header)
+        yaml.safe_dump(data, f, sort_keys=False, allow_unicode=True)
+
+
+def yaml_load(file="data.yaml", append_filename=False):
+    """
+    Load YAML data from a file.
+
+    Parameters
+    ----------
+    file : str, optional
+        File name. Default is 'data.yaml'.
+    append_filename : bool, optional
+        Add the YAML filename to the YAML dictionary. Default is False.
+
+    Returns
+    -------
+    dict
+        YAML data and file name.
+    """
+    assert Path(file).suffix in (
+        ".yaml",
+        ".yml",
+    ), f"Attempting to load non-YAML file {file} with yaml_load()"
+    with open(file, errors="ignore", encoding="utf-8") as f:
+        s = f.read()  # string
+
+        # Add YAML filename to dict and return
+        data = (
+            yaml.safe_load(s) or {}
+        )  # always return a dict (yaml.safe_load() may return None for empty files)
+        if append_filename:
+            data["yaml_file"] = str(file)
+        return data
+
+
+def yaml_print(yaml_file: Union[str, Path, dict]) -> None:
+    """
+    Pretty prints a YAML file or a YAML-formatted dictionary.
+
+    Parameters
+    ----------
+    yaml_file : Union[str, Path, dict]
+        The file path of the YAML file or a YAML-formatted dictionary.
+
+    Returns
+    -------
+    None
+    """
+    yaml_dict = (
+        yaml_load(yaml_file) if isinstance(yaml_file, (str, Path)) else yaml_file
+    )
+    dump = yaml.dump(yaml_dict, sort_keys=False, allow_unicode=True)
+    LOGGER.info(f"Printing '{yaml_file}'\n\n{dump}")
+
+
+PLATFORMS = [platform.system()]
+if is_colab():
+    PLATFORMS.append("Colab")
+if is_kaggle():
+    PLATFORMS.append("Kaggle")
+if is_jupyter():
+    PLATFORMS.append("Jupyter")
+if is_docker():
+    PLATFORMS.append("Docker")
+
+PLATFORMS = "|".join(PLATFORMS)
+
+
+class TryExcept(contextlib.ContextDecorator):
+    """
+    TryExcept context manager.
+    Usage: @TryExcept() decorator or 'with TryExcept():' context manager.
+    """
+
+    def __init__(self, msg="", verbose=True):
+        """
+        Parameters
+        ----------
+        msg : str, optional
+            Custom message to display in case of exception, by default "".
+        verbose : bool, optional
+            Whether to display the message, by default True.
+        """
+        self.msg = msg
+        self.verbose = verbose
+
+    def __enter__(self):
+        pass
+
+    def __exit__(self, exc_type, value, traceback):
+        if self.verbose and value:
+            LOGGER.info(f"{self.msg}{': ' if self.msg else ''}{value}")
+        return True
+
+
+def threaded_request(
+    method, url, retry=3, timeout=30, thread=True, code=-1, verbose=True, **kwargs
+):
+    """
+    Makes an HTTP request using the 'requests' library, with exponential backoff retries up to a specified timeout.
+
+    Parameters
+    ----------
+    method : str
+        The HTTP method to use for the request. Choices are 'post' and 'get'.
+    url : str
+        The URL to make the request to.
+    retry : int, optional
+        Number of retries to attempt before giving up, by default 3.
+    timeout : int, optional
+        Timeout in seconds after which the function will give up retrying, by default 30.
+    thread : bool, optional
+        Whether to execute the request in a separate daemon thread, by default True.
+    code : int, optional
+        An identifier for the request, used for logging purposes, by default -1.
+    verbose : bool, optional
+        A flag to determine whether to print out to console or not, by default True.
+
+    Returns
+    -------
+    requests.Response
+        The HTTP response object. If the request is executed in a separate thread, returns the thread itself.
+    """
+    retry_codes = ()  # retry only these codes TODO: add codes if needed in future (500, 408)
+
+    @TryExcept(verbose=verbose)
+    def func(method, url, **kwargs):
+        """Make HTTP requests with retries and timeouts, with optional progress tracking."""
+        response = None
+        t0 = time.time()
+        for i in range(retry + 1):
+            if (time.time() - t0) > timeout:
+                break
+            response = requests.request(method, url, **kwargs)
+            if response.status_code < 300:  # good return codes in the 2xx range
+                break
+            try:
+                m = response.json().get("message", "No JSON message.")
+            except AttributeError:
+                m = "Unable to read JSON."
+            if i == 0:
+                if response.status_code in retry_codes:
+                    m += f" Retrying {retry}x for {timeout}s." if retry else ""
+                elif response.status_code == 429:  # rate limit
+                    m = f"Rate limit reached"
+                if verbose:
+                    LOGGER.warning(f"{response.status_code} #{code}")
+                if response.status_code not in retry_codes:
+                    return response
+            time.sleep(2**i)  # exponential standoff
+        return response
+
+    args = method, url
+    if thread:
+        return threading.Thread(
+            target=func, args=args, kwargs=kwargs, daemon=True
+        ).start()
+    else:
+        return func(*args, **kwargs)

python/lancedb/utils/sentry_log.py

@@ -0,0 +1,112 @@
+#  Copyright 2023 LanceDB Developers
+#
+#  Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+#  You may obtain a copy of the License at
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+#  Unless required by applicable law or agreed to in writing, software
+#  distributed under the License is distributed on an "AS IS" BASIS,
+#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#  See the License for the specific language governing permissions and
+#  limitations under the License.
+
+import bdb
+import importlib.metadata
+import logging
+import sys
+from pathlib import Path
+
+from lancedb.utils import CONFIG
+
+from .general import (
+    PLATFORMS,
+    TryExcept,
+    is_git_dir,
+    is_github_actions_ci,
+    is_online,
+    is_pip_package,
+    is_pytest_running,
+)
+
+
+@TryExcept(verbose=False)
+def set_sentry():
+    """
+    Initialize the Sentry SDK for error tracking and reporting. Only used if sentry_sdk package is installed and
+    sync=True in settings. Run 'lancedb settings' to see and update settings YAML file.
+
+    Conditions required to send errors (ALL conditions must be met or no errors will be reported):
+        - sentry_sdk package is installed
+        - sync=True in  settings
+        - pytest is not running
+        - running in a pip package installation
+        - running in a non-git directory
+        - online environment
+
+    The function also configures Sentry SDK to ignore KeyboardInterrupt and FileNotFoundError
+    exceptions for now.
+
+    Additionally, the function sets custom tags and user information for Sentry events.
+    """
+
+    def before_send(event, hint):
+        """
+        Modify the event before sending it to Sentry based on specific exception types and messages.
+
+        Args:
+            event (dict): The event dictionary containing information about the error.
+            hint (dict): A dictionary containing additional information about the error.
+
+        Returns:
+            dict: The modified event or None if the event should not be sent to Sentry.
+        """
+        if "exc_info" in hint:
+            exc_type, exc_value, tb = hint["exc_info"]
+            if "out of memory" in str(exc_value).lower():
+                return None
+
+        if is_git_dir():
+            install = "git"
+        elif is_pip_package():
+            install = "pip"
+        else:
+            install = "other"
+
+        event["tags"] = {
+            "sys_argv": sys.argv[0],
+            "sys_argv_name": Path(sys.argv[0]).name,
+            "install": install,
+            "platforms": PLATFORMS,
+            "version": importlib.metadata.version("lancedb"),
+        }
+        return event
+
+    TESTS_RUNNING = is_pytest_running() or is_github_actions_ci()
+    ONLINE = is_online()
+    if CONFIG["diagnostics"] and not TESTS_RUNNING and ONLINE and is_pip_package():
+        # and not is_git_dir(): # not running inside a git dir. Maybe too restrictive?
+
+        # If sentry_sdk package is not installed then return and do not use Sentry
+        try:
+            import sentry_sdk  # noqa
+        except ImportError:
+            return
+
+        sentry_sdk.init(
+            dsn="https://c63ef8c64e05d1aa1a96513361f3ca2f@o4505950840946688.ingest.sentry.io/4505950933614592",
+            debug=False,
+            include_local_variables=False,
+            traces_sample_rate=1.0,
+            environment="production",  # 'dev' or 'production'
+            before_send=before_send,
+            ignore_errors=[KeyboardInterrupt, FileNotFoundError, bdb.BdbQuit],
+        )
+        sentry_sdk.set_user({"id": CONFIG["uuid"]})  # SHA-256 anonymized UUID hash
+
+        # Disable all sentry logging
+        for logger in "sentry_sdk", "sentry_sdk.errors":
+            logging.getLogger(logger).setLevel(logging.CRITICAL)
+
+
+set_sentry()

python/pyproject.toml

@@ -1,7 +1,8 @@
 [project]
 name = "lancedb"
-version = "0.3.0"
+version = "0.3.1"
 dependencies = [
+    "deprecation",
     "pylance==0.8.3",
     "ratelimiter~=1.0",
     "retry>=0.9.2",
@@ -10,7 +11,10 @@ dependencies = [
     "pydantic>=1.10",
     "attrs>=21.3.0",
     "semver>=3.0",
-    "cachetools"
+    "cachetools",
+    "pyyaml>=6.0",
+    "click>=8.1.7",
+    "requests>=2.31.0"
 ]
 description = "lancedb"
 authors = [{ name = "LanceDB Devs", email = "dev@lancedb.com" }]
@@ -48,7 +52,10 @@ tests = ["pandas>=1.4", "pytest", "pytest-mock", "pytest-asyncio", "requests"]
 dev = ["ruff", "pre-commit", "black"]
 docs = ["mkdocs", "mkdocs-jupyter", "mkdocs-material", "mkdocstrings[python]"]
 clip = ["torch", "pillow", "open-clip"]
-embeddings = ["openai", "sentence-transformers", "torch", "pillow", "open-clip"]
+embeddings = ["openai", "sentence-transformers", "torch", "pillow", "open-clip", "cohere"]
+
+[project.scripts]
+lancedb = "lancedb.cli.cli:cli"
 
 [build-system]
 requires = ["setuptools", "wheel"]

python/tests/test_cli.py

@@ -0,0 +1,35 @@
+from click.testing import CliRunner
+
+from lancedb.cli.cli import cli
+from lancedb.utils import CONFIG
+
+
+def test_entry():
+    runner = CliRunner()
+    result = runner.invoke(cli)
+    assert result.exit_code == 0  # Main check
+    assert "lancedb" in result.output.lower()  # lazy check
+
+
+def test_diagnostics():
+    runner = CliRunner()
+    result = runner.invoke(cli, ["diagnostics", "--disabled"])
+    assert result.exit_code == 0  # Main check
+    assert CONFIG["diagnostics"] == False
+
+    result = runner.invoke(cli, ["diagnostics", "--enabled"])
+    assert result.exit_code == 0  # Main check
+    assert CONFIG["diagnostics"] == True
+
+
+def test_config():
+    runner = CliRunner()
+    result = runner.invoke(cli, ["config"])
+    assert result.exit_code == 0  # Main check
+    cfg = CONFIG.copy()
+    cfg.pop("uuid")
+    for (
+        item,
+        _,
+    ) in cfg.items():  # check for keys only as formatting is subject to change
+        assert item in result.output

python/tests/test_context.py

@@ -47,7 +47,7 @@ def test_contextualizer(raw_df: pd.DataFrame):
         .stride(3)
         .text_col("token")
         .groupby("document_id")
-        .to_df()["token"]
+        .to_pandas()["token"]
         .to_list()
     )
 
@@ -67,7 +67,7 @@ def test_contextualizer_with_threshold(raw_df: pd.DataFrame):
         .text_col("token")
         .groupby("document_id")
         .min_window_size(4)
-        .to_df()["token"]
+        .to_pandas()["token"]
         .to_list()
     )
 

python/tests/test_db.py

@@ -33,11 +33,11 @@ def test_basic(tmp_path):
             {"vector": [5.9, 26.5], "item": "bar", "price": 20.0},
         ],
     )
-    rs = table.search([100, 100]).limit(1).to_df()
+    rs = table.search([100, 100]).limit(1).to_pandas()
     assert len(rs) == 1
     assert rs["item"].iloc[0] == "bar"
 
-    rs = table.search([100, 100]).where("price < 15").limit(2).to_df()
+    rs = table.search([100, 100]).where("price < 15").limit(2).to_pandas()
     assert len(rs) == 1
     assert rs["item"].iloc[0] == "foo"
 
@@ -62,11 +62,11 @@ def test_ingest_pd(tmp_path):
         }
     )
     table = db.create_table("test", data=data)
-    rs = table.search([100, 100]).limit(1).to_df()
+    rs = table.search([100, 100]).limit(1).to_pandas()
     assert len(rs) == 1
     assert rs["item"].iloc[0] == "bar"
 
-    rs = table.search([100, 100]).where("price < 15").limit(2).to_df()
+    rs = table.search([100, 100]).where("price < 15").limit(2).to_pandas()
     assert len(rs) == 1
     assert rs["item"].iloc[0] == "foo"
 
@@ -137,8 +137,8 @@ def test_ingest_iterator(tmp_path):
         db = lancedb.connect(tmp_path)
         tbl = db.create_table("table2", make_batches(), schema=schema, mode="overwrite")
         tbl.to_pandas()
-        assert tbl.search([3.1, 4.1]).limit(1).to_df()["_distance"][0] == 0.0
+        assert tbl.search([3.1, 4.1]).limit(1).to_pandas()["_distance"][0] == 0.0
-        assert tbl.search([5.9, 26.5]).limit(1).to_df()["_distance"][0] == 0.0
+        assert tbl.search([5.9, 26.5]).limit(1).to_pandas()["_distance"][0] == 0.0
         tbl_len = len(tbl)
         tbl.add(make_batches())
         assert tbl_len == 50

python/tests/test_e2e_remote_db.py

@@ -23,5 +23,5 @@ from lancedb import LanceDBConnection
 def test_against_local_server():
     conn = LanceDBConnection("lancedb+http://localhost:10024")
     table = conn.open_table("sift1m_ivf1024_pq16")
-    df = table.search(np.random.rand(128)).to_df()
+    df = table.search(np.random.rand(128)).to_pandas()
     assert len(df) == 10

python/tests/test_embeddings_slow.py

@@ -11,6 +11,7 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 import io
+import os
 
 import numpy as np
 import pandas as pd
@@ -123,3 +124,26 @@ def test_openclip(tmp_path):
         arrow_table["vector"].combine_chunks().values.to_numpy(),
         arrow_table["vec_from_bytes"].combine_chunks().values.to_numpy(),
     )
+
+
+@pytest.mark.slow
+@pytest.mark.skipif(
+    os.environ.get("COHERE_API_KEY") is None, reason="COHERE_API_KEY not set"
+)  # also skip if cohere not installed
+def test_cohere_embedding_function():
+    cohere = (
+        EmbeddingFunctionRegistry.get_instance()
+        .get("cohere")
+        .create(name="embed-multilingual-v2.0")
+    )
+
+    class TextModel(LanceModel):
+        text: str = cohere.SourceField()
+        vector: Vector(cohere.ndims()) = cohere.VectorField()
+
+    df = pd.DataFrame({"text": ["hello world", "goodbye world"]})
+    db = lancedb.connect("~/lancedb")
+    tbl = db.create_table("test", schema=TextModel, mode="overwrite")
+
+    tbl.add(df)
+    assert len(tbl.to_pandas()["vector"][0]) == cohere.ndims()

python/tests/test_fts.py

@@ -71,14 +71,14 @@ def test_search_index(tmp_path, table):
 
 def test_create_index_from_table(tmp_path, table):
     table.create_fts_index("text")
-    df = table.search("puppy").limit(10).select(["text"]).to_df()
+    df = table.search("puppy").limit(10).select(["text"]).to_pandas()
     assert len(df) == 10
     assert "text" in df.columns
 
 
 def test_create_index_multiple_columns(tmp_path, table):
     table.create_fts_index(["text", "text2"])
-    df = table.search("puppy").limit(10).to_df()
+    df = table.search("puppy").limit(10).to_pandas()
     assert len(df) == 10
     assert "text" in df.columns
     assert "text2" in df.columns
@@ -87,5 +87,5 @@ def test_create_index_multiple_columns(tmp_path, table):
 def test_empty_rs(tmp_path, table, mocker):
     table.create_fts_index(["text", "text2"])
     mocker.patch("lancedb.fts.search_index", return_value=([], []))
-    df = table.search("puppy").limit(10).to_df()
+    df = table.search("puppy").limit(10).to_pandas()
     assert len(df) == 0

python/tests/test_io.py

@@ -36,11 +36,11 @@ def test_s3_io():
             {"vector": [5.9, 26.5], "item": "bar", "price": 20.0},
         ],
     )
-    rs = table.search([100, 100]).limit(1).to_df()
+    rs = table.search([100, 100]).limit(1).to_pandas()
     assert len(rs) == 1
     assert rs["item"].iloc[0] == "bar"
 
-    rs = table.search([100, 100]).where("price < 15").limit(2).to_df()
+    rs = table.search([100, 100]).where("price < 15").limit(2).to_pandas()
     assert len(rs) == 1
     assert rs["item"].iloc[0] == "foo"
 

python/tests/test_query.py

@@ -85,17 +85,20 @@ def test_cast(table):
 
 
 def test_query_builder(table):
-    df = (
+    rs = (
-        LanceVectorQueryBuilder(table, [0, 0], "vector").limit(1).select(["id"]).to_df()
+        LanceVectorQueryBuilder(table, [0, 0], "vector")
+        .limit(1)
+        .select(["id"])
+        .to_list()
     )
-    assert df["id"].values[0] == 1
+    assert rs[0]["id"] == 1
-    assert all(df["vector"].values[0] == [1, 2])
+    assert all(np.array(rs[0]["vector"]) == [1, 2])
 
 
 def test_query_builder_with_filter(table):
-    df = LanceVectorQueryBuilder(table, [0, 0], "vector").where("id = 2").to_df()
+    rs = LanceVectorQueryBuilder(table, [0, 0], "vector").where("id = 2").to_list()
-    assert df["id"].values[0] == 2
+    assert rs[0]["id"] == 2
-    assert all(df["vector"].values[0] == [3, 4])
+    assert all(np.array(rs[0]["vector"]) == [3, 4])
 
 
 def test_query_builder_with_prefilter(table):
@@ -103,7 +106,7 @@ def test_query_builder_with_prefilter(table):
         LanceVectorQueryBuilder(table, [0, 0], "vector")
         .where("id = 2")
         .limit(1)
-        .to_df()
+        .to_pandas()
     )
     assert len(df) == 0
 
@@ -111,7 +114,7 @@ def test_query_builder_with_prefilter(table):
         LanceVectorQueryBuilder(table, [0, 0], "vector")
         .where("id = 2", prefilter=True)
         .limit(1)
-        .to_df()
+        .to_pandas()
     )
     assert df["id"].values[0] == 2
     assert all(df["vector"].values[0] == [3, 4])
@@ -120,9 +123,11 @@ def test_query_builder_with_prefilter(table):
 def test_query_builder_with_metric(table):
     query = [4, 8]
     vector_column_name = "vector"
-    df_default = LanceVectorQueryBuilder(table, query, vector_column_name).to_df()
+    df_default = LanceVectorQueryBuilder(table, query, vector_column_name).to_pandas()
     df_l2 = (
-        LanceVectorQueryBuilder(table, query, vector_column_name).metric("L2").to_df()
+        LanceVectorQueryBuilder(table, query, vector_column_name)
+        .metric("L2")
+        .to_pandas()
     )
     tm.assert_frame_equal(df_default, df_l2)
 
@@ -130,7 +135,7 @@ def test_query_builder_with_metric(table):
         LanceVectorQueryBuilder(table, query, vector_column_name)
         .metric("cosine")
         .limit(1)
-        .to_df()
+        .to_pandas()
     )
     assert df_cosine._distance[0] == pytest.approx(
         cosine_distance(query, df_cosine.vector[0]),

python/tests/test_remote_client.py

@@ -86,7 +86,7 @@ async def test_e2e_with_mock_server():
                     columns=["id", "vector"],
                 ),
             )
-        ).to_df()
+        ).to_pandas()
 
         assert "vector" in df.columns
         assert "id" in df.columns

python/tests/test_remote_db.py

@@ -32,4 +32,4 @@ def test_remote_db():
     setattr(conn, "_client", FakeLanceDBClient())
 
     table = conn["test"]
-    table.search([1.0, 2.0]).to_df()
+    table.search([1.0, 2.0]).to_pandas()

python/tests/test_table.py

@@ -12,6 +12,7 @@
 #  limitations under the License.
 
 import functools
+from datetime import timedelta
 from pathlib import Path
 from typing import List
 from unittest.mock import PropertyMock, patch
@@ -427,8 +428,8 @@ def test_multiple_vector_columns(db):
     table.add(df)
 
     q = np.random.randn(10)
-    result1 = table.search(q, vector_column_name="vector1").limit(1).to_df()
+    result1 = table.search(q, vector_column_name="vector1").limit(1).to_pandas()
-    result2 = table.search(q, vector_column_name="vector2").limit(1).to_df()
+    result2 = table.search(q, vector_column_name="vector2").limit(1).to_pandas()
 
     assert result1["text"].iloc[0] != result2["text"].iloc[0]
 
@@ -439,6 +440,34 @@ def test_empty_query(db):
         "my_table",
         data=[{"text": "foo", "id": 0}, {"text": "bar", "id": 1}],
     )
-    df = table.search().select(["id"]).where("text='bar'").limit(1).to_df()
+    df = table.search().select(["id"]).where("text='bar'").limit(1).to_pandas()
     val = df.id.iloc[0]
     assert val == 1
+
+
+def test_compact_cleanup(db):
+    table = LanceTable.create(
+        db,
+        "my_table",
+        data=[{"text": "foo", "id": 0}, {"text": "bar", "id": 1}],
+    )
+
+    table.add([{"text": "baz", "id": 2}])
+    assert len(table) == 3
+    assert table.version == 3
+
+    stats = table.compact_files()
+    assert len(table) == 3
+    assert table.version == 4
+    assert stats.fragments_removed > 0
+    assert stats.fragments_added == 1
+
+    stats = table.cleanup_old_versions()
+    assert stats.bytes_removed == 0
+
+    stats = table.cleanup_old_versions(older_than=timedelta(0), delete_unverified=True)
+    assert stats.bytes_removed > 0
+    assert table.version == 4
+
+    with pytest.raises(Exception, match="Version 3 no longer exists"):
+        table.checkout(3)

python/tests/test_telemetry.py

@@ -0,0 +1,60 @@
+import json
+
+import pytest
+
+import lancedb
+from lancedb.utils.events import _Events
+
+
+@pytest.fixture(autouse=True)
+def request_log_path(tmp_path):
+    return tmp_path / "request.json"
+
+
+def mock_register_event(name: str, **kwargs):
+    if _Events._instance is None:
+        _Events._instance = _Events()
+
+    _Events._instance.enabled = True
+    _Events._instance.rate_limit = 0
+    _Events._instance(name, **kwargs)
+
+
+def test_event_reporting(monkeypatch, request_log_path, tmp_path) -> None:
+    def mock_request(**kwargs):
+        json_data = kwargs.get("json", {})
+        with open(request_log_path, "w") as f:
+            json.dump(json_data, f)
+
+    monkeypatch.setattr(
+        lancedb.table, "register_event", mock_register_event
+    )  # Force enable registering events and strip exception handling
+    monkeypatch.setattr(lancedb.utils.events, "threaded_request", mock_request)
+
+    db = lancedb.connect(tmp_path)
+    db.create_table(
+        "test",
+        data=[
+            {"vector": [3.1, 4.1], "item": "foo", "price": 10.0},
+            {"vector": [5.9, 26.5], "item": "bar", "price": 20.0},
+        ],
+        mode="overwrite",
+    )
+
+    assert request_log_path.exists()  # test if event was registered
+
+    with open(request_log_path, "r") as f:
+        json_data = json.load(f)
+
+    # TODO: don't hardcode these here. Instead create a module level json scehma in lancedb.utils.events for better evolvability
+    batch_keys = ["api_key", "distinct_id", "batch"]
+    event_keys = ["event", "properties", "timestamp", "distinct_id"]
+    property_keys = ["cli", "install", "platforms", "version", "session_id"]
+
+    assert all([key in json_data for key in batch_keys])
+    assert all([key in json_data["batch"][0] for key in event_keys])
+    assert all([key in json_data["batch"][0]["properties"] for key in property_keys])
+
+    # cleanup & reset
+    monkeypatch.undo()
+    _Events._instance = None

rust/ffi/node/Cargo.toml

@@ -13,6 +13,7 @@ crate-type = ["cdylib"]
 arrow-array = { workspace = true }
 arrow-ipc = { workspace = true }
 arrow-schema = { workspace = true }
+chrono = { workspace = true }
 conv = "0.3.3"
 once_cell = "1"
 futures = "0.3"

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

@@ -78,9 +78,11 @@ fn get_index_params_builder(
 
             num_partitions.map(|np| {
                 let max_iters = max_iters.unwrap_or(50);
-                let mut ivf_params = IvfBuildParams::default();
+                let ivf_params = IvfBuildParams {
-                ivf_params.num_partitions = np;
+                    num_partitions: np,
-                ivf_params.max_iters = max_iters;
+                    max_iters,
+                    ..Default::default()
+                };
                 index_builder.ivf_params(ivf_params)
             });
 

rust/ffi/node/src/lib.rs

@@ -237,6 +237,8 @@ fn main(mut cx: ModuleContext) -> NeonResult<()> {
     cx.export_function("tableAdd", JsTable::js_add)?;
     cx.export_function("tableCountRows", JsTable::js_count_rows)?;
     cx.export_function("tableDelete", JsTable::js_delete)?;
+    cx.export_function("tableCleanupOldVersions", JsTable::js_cleanup)?;
+    cx.export_function("tableCompactFiles", JsTable::js_compact)?;
     cx.export_function(
         "tableCreateVectorIndex",
         index::vector::table_create_vector_index,

rust/ffi/node/src/table.rs

@@ -13,6 +13,7 @@
 // limitations under the License.
 
 use arrow_array::RecordBatchIterator;
+use lance::dataset::optimize::CompactionOptions;
 use lance::dataset::{WriteMode, WriteParams};
 use lance::io::object_store::ObjectStoreParams;
 
@@ -163,4 +164,116 @@ impl JsTable {
         });
         Ok(promise)
     }
+
+    pub(crate) fn js_cleanup(mut cx: FunctionContext) -> JsResult<JsPromise> {
+        let js_table = cx.this().downcast_or_throw::<JsBox<JsTable>, _>(&mut cx)?;
+        let rt = runtime(&mut cx)?;
+        let (deferred, promise) = cx.promise();
+        let table = js_table.table.clone();
+        let channel = cx.channel();
+
+        let older_than: i64 = cx
+            .argument_opt(0)
+            .and_then(|val| val.downcast::<JsNumber, _>(&mut cx).ok())
+            .map(|val| val.value(&mut cx) as i64)
+            .unwrap_or_else(|| 2 * 7 * 24 * 60); // 2 weeks
+        let older_than = chrono::Duration::minutes(older_than);
+        let delete_unverified: bool = cx
+            .argument_opt(1)
+            .and_then(|val| val.downcast::<JsBoolean, _>(&mut cx).ok())
+            .map(|val| val.value(&mut cx))
+            .unwrap_or_default();
+
+        rt.spawn(async move {
+            let stats = table
+                .cleanup_old_versions(older_than, Some(delete_unverified))
+                .await;
+
+            deferred.settle_with(&channel, move |mut cx| {
+                let stats = stats.or_throw(&mut cx)?;
+
+                let output_metrics = JsObject::new(&mut cx);
+                let bytes_removed = cx.number(stats.bytes_removed as f64);
+                output_metrics.set(&mut cx, "bytesRemoved", bytes_removed)?;
+
+                let old_versions = cx.number(stats.old_versions as f64);
+                output_metrics.set(&mut cx, "oldVersions", old_versions)?;
+
+                let output_table = cx.boxed(JsTable::from(table));
+
+                let output = JsObject::new(&mut cx);
+                output.set(&mut cx, "metrics", output_metrics)?;
+                output.set(&mut cx, "newTable", output_table)?;
+
+                Ok(output)
+            })
+        });
+        Ok(promise)
+    }
+
+    pub(crate) fn js_compact(mut cx: FunctionContext) -> JsResult<JsPromise> {
+        let js_table = cx.this().downcast_or_throw::<JsBox<JsTable>, _>(&mut cx)?;
+        let rt = runtime(&mut cx)?;
+        let (deferred, promise) = cx.promise();
+        let mut table = js_table.table.clone();
+        let channel = cx.channel();
+
+        let js_options = cx.argument::<JsObject>(0)?;
+        let mut options = CompactionOptions::default();
+
+        if let Some(target_rows) =
+            js_options.get_opt::<JsNumber, _, _>(&mut cx, "targetRowsPerFragment")?
+        {
+            options.target_rows_per_fragment = target_rows.value(&mut cx) as usize;
+        }
+        if let Some(max_per_group) =
+            js_options.get_opt::<JsNumber, _, _>(&mut cx, "maxRowsPerGroup")?
+        {
+            options.max_rows_per_group = max_per_group.value(&mut cx) as usize;
+        }
+        if let Some(materialize_deletions) =
+            js_options.get_opt::<JsBoolean, _, _>(&mut cx, "materializeDeletions")?
+        {
+            options.materialize_deletions = materialize_deletions.value(&mut cx);
+        }
+        if let Some(materialize_deletions_threshold) =
+            js_options.get_opt::<JsNumber, _, _>(&mut cx, "materializeDeletionsThreshold")?
+        {
+            options.materialize_deletions_threshold =
+                materialize_deletions_threshold.value(&mut cx) as f32;
+        }
+        if let Some(num_threads) = js_options.get_opt::<JsNumber, _, _>(&mut cx, "numThreads")? {
+            options.num_threads = num_threads.value(&mut cx) as usize;
+        }
+
+        rt.spawn(async move {
+            let stats = table.compact_files(options).await;
+
+            deferred.settle_with(&channel, move |mut cx| {
+                let stats = stats.or_throw(&mut cx)?;
+
+                let output_metrics = JsObject::new(&mut cx);
+                let fragments_removed = cx.number(stats.fragments_removed as f64);
+                output_metrics.set(&mut cx, "fragmentsRemoved", fragments_removed)?;
+
+                let fragments_added = cx.number(stats.fragments_added as f64);
+                output_metrics.set(&mut cx, "fragmentsAdded", fragments_added)?;
+
+                let files_removed = cx.number(stats.files_removed as f64);
+                output_metrics.set(&mut cx, "filesRemoved", files_removed)?;
+
+                let files_added = cx.number(stats.files_added as f64);
+                output_metrics.set(&mut cx, "filesAdded", files_added)?;
+
+                let output_table = cx.boxed(JsTable::from(table));
+
+                let output = JsObject::new(&mut cx);
+                output.set(&mut cx, "metrics", output_metrics)?;
+                output.set(&mut cx, "newTable", output_table)?;
+
+                Ok(output)
+            })
+        });
+        Ok(promise)
+    }
 }

rust/vectordb/Cargo.toml

@@ -16,6 +16,7 @@ arrow-data = { workspace = true }
 arrow-schema = { workspace = true }
 arrow-ord = { workspace = true }
 arrow-cast = { workspace = true }
+chrono = { workspace = true }
 object_store = { workspace = true }
 snafu = { workspace = true }
 half = { workspace = true }

rust/vectordb/src/io/object_store.rs

@@ -14,7 +14,6 @@
 
 //! A mirroring object store that mirror writes to a secondary object store
 
-
 use std::{
     fmt::Formatter,
     pin::Pin,

rust/vectordb/src/table.rs

@@ -12,10 +12,13 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+use chrono::Duration;
 use std::sync::Arc;
 
 use arrow_array::{Float32Array, RecordBatchReader};
 use arrow_schema::SchemaRef;
+use lance::dataset::cleanup::RemovalStats;
+use lance::dataset::optimize::{compact_files, CompactionMetrics, CompactionOptions};
 use lance::dataset::{Dataset, WriteParams};
 use lance::index::IndexType;
 use lance::io::object_store::WrappingObjectStore;
@@ -305,6 +308,41 @@ impl Table {
         self.dataset = Arc::new(dataset);
         Ok(())
     }
+
+    /// Remove old versions of the dataset from disk.
+    ///
+    /// # Arguments
+    /// * `older_than` - The duration of time to keep versions of the dataset.
+    /// * `delete_unverified` - Because they may be part of an in-progress
+    ///   transaction, files newer than 7 days old are not deleted by default.
+    ///   If you are sure that there are no in-progress transactions, then you
+    ///   can set this to True to delete all files older than `older_than`.
+    ///
+    /// This calls into [lance::dataset::Dataset::cleanup_old_versions] and
+    /// returns the result.
+    pub async fn cleanup_old_versions(
+        &self,
+        older_than: Duration,
+        delete_unverified: Option<bool>,
+    ) -> Result<RemovalStats> {
+        Ok(self
+            .dataset
+            .cleanup_old_versions(older_than, delete_unverified)
+            .await?)
+    }
+
+    /// Compact files in the dataset.
+    ///
+    /// This can be run after making several small appends to optimize the table
+    /// for faster reads.
+    ///
+    /// This calls into [lance::dataset::optimize::compact_files].
+    pub async fn compact_files(&mut self, options: CompactionOptions) -> Result<CompactionMetrics> {
+        let mut dataset = self.dataset.as_ref().clone();
+        let metrics = compact_files(&mut dataset, options).await?;
+        self.dataset = Arc::new(dataset);
+        Ok(metrics)
+    }
 }
 
 #[cfg(test)]