[python] Bump version: 0.4.2 → 0.4.3

chore: bump pylance to 0.9.2 (#754 )
docs: fix link (#752 )
2025-12-23 05:19:58 +00:00 · 2023-12-30 00:52:54 +00:00 · 2023-12-29 16:39:45 -08:00 · 2023-12-29 15:33:24 -08:00 · 2023-12-29 15:33:03 -08:00 · 2023-12-29 00:19:21 +00:00
16 changed files with 269 additions and 70 deletions
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -5,10 +5,10 @@ exclude = ["python"]
 resolver = "2"

 [workspace.dependencies]
-lance = { "version" = "=0.9.1", "features" = ["dynamodb"] }
-lance-index = { "version" = "=0.9.1" }
-lance-linalg = { "version" = "=0.9.1" }
-lance-testing = { "version" = "=0.9.1" }
+lance = { "version" = "=0.9.2", "features" = ["dynamodb"] }
+lance-index = { "version" = "=0.9.2" }
+lance-linalg = { "version" = "=0.9.2" }
+lance-testing = { "version" = "=0.9.2" }
 # Note that this one does not include pyarrow
 arrow = { version = "49.0.0", optional = false }
 arrow-array = "49.0"
--- a/docs/src/fts.md
+++ b/docs/src/fts.md
@@ -29,8 +29,9 @@ uri = "data/sample-lancedb"
 db = lancedb.connect(uri)

 table = db.create_table("my_table",
-            data=[{"vector": [3.1, 4.1], "text": "Frodo was a happy puppy"},
-                  {"vector": [5.9, 26.5], "text": "There are several kittens playing"}])
+            data=[{"vector": [3.1, 4.1], "text": "Frodo was a happy puppy", "meta": "foo"},
+                  {"vector": [5.9, 26.5], "text": "Sam was a loyal puppy", "meta": "bar"},
+                  {"vector": [15.9, 6.5], "text": "There are several kittens playing"}])

 ```

@@ -64,10 +65,23 @@ table.create_fts_index(["text1", "text2"])

 Note that the search API call does not change - you can search over all indexed columns at once.

+## Filtering
+
+Currently the LanceDB full text search feature supports *post-filtering*, meaning filters are
+applied on top of the full text search results. This can be invoked via the familiar
+`where` syntax:
+
+```python
+table.search("puppy").limit(10).where("meta='foo'").to_list()
+```
+
 ## Current limitations

 1. Currently we do not yet support incremental writes.
-If you add data after fts index creation, it won't be reflected
-in search results until you do a full reindex.
+   If you add data after fts index creation, it won't be reflected
+   in search results until you do a full reindex.
+
+2. We currently only support local filesystem paths for the fts index. 
+   This is a tantivy limitation. We've implemented an object store plugin
+   but there's no way in tantivy-py to specify to use it.

-2. We currently only support local filesystem paths for the fts index.
--- a/docs/src/guides/tables.md
+++ b/docs/src/guides/tables.md
@@ -118,6 +118,84 @@ This guide will show how to create tables, insert data into them, and update the
    table = db.create_table(table_name, schema=Content)
    ```

+    #### Nested schemas
+
+    Sometimes your data model may contain nested objects.
+    For example, you may want to store the document string
+    and the document soure name as a nested Document object:
+
+    ```python
+    class Document(BaseModel):
+        content: str
+        source: str
+    ```
+
+    This can be used as the type of a LanceDB table column:
+
+    ```python
+    class NestedSchema(LanceModel):
+        id: str
+        vector: Vector(1536)
+        document: Document
+
+    tbl = db.create_table("nested_table", schema=NestedSchema, mode="overwrite")
+    ```
+
+    This creates a struct column called "document" that has two subfields 
+    called "content" and "source":
+
+    ```
+    In [28]: tbl.schema
+    Out[28]:
+    id: string not null
+    vector: fixed_size_list<item: float>[1536] not null
+        child 0, item: float
+    document: struct<content: string not null, source: string not null> not null
+        child 0, content: string not null
+        child 1, source: string not null    
+    ```
+
+    #### Validators
+
+    Note that neither pydantic nor pyarrow automatically validates that input data
+    is of the *correct* timezone, but this is easy to add as a custom field validator:
+
+    ```python
+    from datetime import datetime
+    from zoneinfo import ZoneInfo
+
+    from lancedb.pydantic import LanceModel
+    from pydantic import Field, field_validator, ValidationError, ValidationInfo
+
+    tzname = "America/New_York"
+    tz = ZoneInfo(tzname)
+
+    class TestModel(LanceModel):
+        dt_with_tz: datetime = Field(json_schema_extra={"tz": tzname})
+
+        @field_validator('dt_with_tz')
+        @classmethod
+        def tz_must_match(cls, dt: datetime) -> datetime:
+            assert dt.tzinfo == tz
+            return dt        
+
+    ok = TestModel(dt_with_tz=datetime.now(tz))
+
+    try:
+        TestModel(dt_with_tz=datetime.now(ZoneInfo("Asia/Shanghai")))
+        assert 0 == 1, "this should raise ValidationError"
+    except ValidationError:
+        print("A ValidationError was raised.")
+        pass
+    ```
+
+    When you run this code it should print "A ValidationError was raised."
+
+    #### Pydantic custom types
+
+    LanceDB does NOT yet support converting pydantic custom types. If this is something you need,
+    please file a feature request on the [LanceDB Github repo](https://github.com/lancedb/lancedb/issues/new).
+
    ### Using Iterators / Writing Large Datasets

    It is recommended to use itertators to add large datasets in batches when creating your table in one go. This does not create multiple versions of your dataset unlike manually adding batches using `table.add()`
@@ -153,7 +231,7 @@ This guide will show how to create tables, insert data into them, and update the
    You can also use iterators of other types like Pandas dataframe or Pylists directly in the above example.

    ## Creating Empty Table
-    You can also create empty tables in python. Initialize it with schema and later ingest data into it.
+    You can create empty tables in python. Initialize it with schema and later ingest data into it.

    ```python
    import lancedb
--- a/docs/src/python/pydantic.md
+++ b/docs/src/python/pydantic.md
@@ -7,7 +7,7 @@ LanceDB integrates with Pydantic for schema inference, data ingestion, and query

 LanceDB supports to create Apache Arrow Schema from a
 [Pydantic BaseModel](https://docs.pydantic.dev/latest/api/main/#pydantic.main.BaseModel)
-via [pydantic_to_schema()](python.md##lancedb.pydantic.pydantic_to_schema) method.
+via [pydantic_to_schema()](python.md#lancedb.pydantic.pydantic_to_schema) method.

 ::: lancedb.pydantic.pydantic_to_schema

--- a/node/package-lock.json
+++ b/node/package-lock.json
@@ -1,12 +1,12 @@
 {
  "name": "vectordb",
-  "version": "0.4.0",
+  "version": "0.4.1",
  "lockfileVersion": 2,
  "requires": true,
  "packages": {
    "": {
      "name": "vectordb",
-      "version": "0.4.0",
+      "version": "0.4.1",
      "cpu": [
        "x64",
        "arm64"
@@ -53,11 +53,11 @@
        "uuid": "^9.0.0"
      },
      "optionalDependencies": {
-        "@lancedb/vectordb-darwin-arm64": "0.4.0",
-        "@lancedb/vectordb-darwin-x64": "0.4.0",
-        "@lancedb/vectordb-linux-arm64-gnu": "0.4.0",
-        "@lancedb/vectordb-linux-x64-gnu": "0.4.0",
-        "@lancedb/vectordb-win32-x64-msvc": "0.4.0"
+        "@lancedb/vectordb-darwin-arm64": "0.4.1",
+        "@lancedb/vectordb-darwin-x64": "0.4.1",
+        "@lancedb/vectordb-linux-arm64-gnu": "0.4.1",
+        "@lancedb/vectordb-linux-x64-gnu": "0.4.1",
+        "@lancedb/vectordb-win32-x64-msvc": "0.4.1"
      }
    },
    "node_modules/@apache-arrow/ts": {
@@ -317,9 +317,9 @@
      }
    },
    "node_modules/@lancedb/vectordb-darwin-arm64": {
-      "version": "0.4.0",
-      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-darwin-arm64/-/vectordb-darwin-arm64-0.4.0.tgz",
-      "integrity": "sha512-cP6zGtBWXEcJHCI4uLNIP5ILtRvexvwmL8Uri1dnHG8dT8g12Ykug3BHO6Wt6wp/xASd2jJRIF/VAJsN9IeP1A==",
+      "version": "0.4.1",
+      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-darwin-arm64/-/vectordb-darwin-arm64-0.4.1.tgz",
+      "integrity": "sha512-ul/Hvv5RX2RThpKSuiUjJRVrmXuBPvpU+HrLjcBmu4dzpuWN4+IeHIUM6xe79gLxOKlwkscVweTOuZnmMfsZeg==",
      "cpu": [
        "arm64"
      ],
@@ -329,9 +329,9 @@
      ]
    },
    "node_modules/@lancedb/vectordb-darwin-x64": {
-      "version": "0.4.0",
-      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-darwin-x64/-/vectordb-darwin-x64-0.4.0.tgz",
-      "integrity": "sha512-ig0gV5ol1sFe2lb1HOatK0rizyj9I91WbnH79i7OdUl3nAQIcWm70CnxrPLtx0DS2NTGh2kFJbYCWcaUlu6YfA==",
+      "version": "0.4.1",
+      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-darwin-x64/-/vectordb-darwin-x64-0.4.1.tgz",
+      "integrity": "sha512-sJtF2Cv6T9RhUpdeHNkryiJwPuW9QPQ3aMs5fID1hMCJA2U3BX27t/WlkiPT2+kTLeUcwF1JvAOgsfvZkfvI8w==",
      "cpu": [
        "x64"
      ],
@@ -341,9 +341,9 @@
      ]
    },
    "node_modules/@lancedb/vectordb-linux-arm64-gnu": {
-      "version": "0.4.0",
-      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-linux-arm64-gnu/-/vectordb-linux-arm64-gnu-0.4.0.tgz",
-      "integrity": "sha512-gMXIDT2kriAPDwWIRKXdaTCNdOeFGEok1S9Y30AOruHXddW1vCIo4JNJIYbBqHnwAeI4wI3ae6GRCFaf1UxO3g==",
+      "version": "0.4.1",
+      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-linux-arm64-gnu/-/vectordb-linux-arm64-gnu-0.4.1.tgz",
+      "integrity": "sha512-tNnziT0BRjPsznKI4GgWROFdCOsCGx0inFu0z+WV1UomwXKcMWGslpWBqKE8IUiCq14duPVx/ie7Wwcf51IeJQ==",
      "cpu": [
        "arm64"
      ],
@@ -353,9 +353,9 @@
      ]
    },
    "node_modules/@lancedb/vectordb-linux-x64-gnu": {
-      "version": "0.4.0",
-      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-linux-x64-gnu/-/vectordb-linux-x64-gnu-0.4.0.tgz",
-      "integrity": "sha512-ZQ3lDrDSz1IKdx/mS9Lz08agFO+OD5oSFrrcFNCoT1+H93eS1mCLdmCoEARu3jKbx0tMs38l5J9yXZ2QmJye3w==",
+      "version": "0.4.1",
+      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-linux-x64-gnu/-/vectordb-linux-x64-gnu-0.4.1.tgz",
+      "integrity": "sha512-PAcF2p1FUsC0AD+qkLfgE5+ZlQwlHe9eTP9dSsX43V/NGPDQ9+gBzaBTEDbvyHj1wl2Wft2NwOqB1HAFhilSDg==",
      "cpu": [
        "x64"
      ],
@@ -365,9 +365,9 @@
      ]
    },
    "node_modules/@lancedb/vectordb-win32-x64-msvc": {
-      "version": "0.4.0",
-      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-win32-x64-msvc/-/vectordb-win32-x64-msvc-0.4.0.tgz",
-      "integrity": "sha512-toNcNwBRE1sdsSf5hr7W8QiqZ33csc/knVEek4CyvYkZHJGh4Z6WI+DJUIASo5wzUez4TX7qUPpRPL9HuaPMCg==",
+      "version": "0.4.1",
+      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-win32-x64-msvc/-/vectordb-win32-x64-msvc-0.4.1.tgz",
+      "integrity": "sha512-8mvThCppI/AfSPby6Y3t6xpCfbo8IY6JH5exO8fDGTwBFHOqgwR4Izb2K7FgXxkwUYcN4EfGSsk/6B1GpwMudg==",
      "cpu": [
        "x64"
      ],
@@ -4869,33 +4869,33 @@
      }
    },
    "@lancedb/vectordb-darwin-arm64": {
-      "version": "0.4.0",
-      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-darwin-arm64/-/vectordb-darwin-arm64-0.4.0.tgz",
-      "integrity": "sha512-cP6zGtBWXEcJHCI4uLNIP5ILtRvexvwmL8Uri1dnHG8dT8g12Ykug3BHO6Wt6wp/xASd2jJRIF/VAJsN9IeP1A==",
+      "version": "0.4.1",
+      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-darwin-arm64/-/vectordb-darwin-arm64-0.4.1.tgz",
+      "integrity": "sha512-ul/Hvv5RX2RThpKSuiUjJRVrmXuBPvpU+HrLjcBmu4dzpuWN4+IeHIUM6xe79gLxOKlwkscVweTOuZnmMfsZeg==",
      "optional": true
    },
    "@lancedb/vectordb-darwin-x64": {
-      "version": "0.4.0",
-      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-darwin-x64/-/vectordb-darwin-x64-0.4.0.tgz",
-      "integrity": "sha512-ig0gV5ol1sFe2lb1HOatK0rizyj9I91WbnH79i7OdUl3nAQIcWm70CnxrPLtx0DS2NTGh2kFJbYCWcaUlu6YfA==",
+      "version": "0.4.1",
+      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-darwin-x64/-/vectordb-darwin-x64-0.4.1.tgz",
+      "integrity": "sha512-sJtF2Cv6T9RhUpdeHNkryiJwPuW9QPQ3aMs5fID1hMCJA2U3BX27t/WlkiPT2+kTLeUcwF1JvAOgsfvZkfvI8w==",
      "optional": true
    },
    "@lancedb/vectordb-linux-arm64-gnu": {
-      "version": "0.4.0",
-      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-linux-arm64-gnu/-/vectordb-linux-arm64-gnu-0.4.0.tgz",
-      "integrity": "sha512-gMXIDT2kriAPDwWIRKXdaTCNdOeFGEok1S9Y30AOruHXddW1vCIo4JNJIYbBqHnwAeI4wI3ae6GRCFaf1UxO3g==",
+      "version": "0.4.1",
+      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-linux-arm64-gnu/-/vectordb-linux-arm64-gnu-0.4.1.tgz",
+      "integrity": "sha512-tNnziT0BRjPsznKI4GgWROFdCOsCGx0inFu0z+WV1UomwXKcMWGslpWBqKE8IUiCq14duPVx/ie7Wwcf51IeJQ==",
      "optional": true
    },
    "@lancedb/vectordb-linux-x64-gnu": {
-      "version": "0.4.0",
-      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-linux-x64-gnu/-/vectordb-linux-x64-gnu-0.4.0.tgz",
-      "integrity": "sha512-ZQ3lDrDSz1IKdx/mS9Lz08agFO+OD5oSFrrcFNCoT1+H93eS1mCLdmCoEARu3jKbx0tMs38l5J9yXZ2QmJye3w==",
+      "version": "0.4.1",
+      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-linux-x64-gnu/-/vectordb-linux-x64-gnu-0.4.1.tgz",
+      "integrity": "sha512-PAcF2p1FUsC0AD+qkLfgE5+ZlQwlHe9eTP9dSsX43V/NGPDQ9+gBzaBTEDbvyHj1wl2Wft2NwOqB1HAFhilSDg==",
      "optional": true
    },
    "@lancedb/vectordb-win32-x64-msvc": {
-      "version": "0.4.0",
-      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-win32-x64-msvc/-/vectordb-win32-x64-msvc-0.4.0.tgz",
-      "integrity": "sha512-toNcNwBRE1sdsSf5hr7W8QiqZ33csc/knVEek4CyvYkZHJGh4Z6WI+DJUIASo5wzUez4TX7qUPpRPL9HuaPMCg==",
+      "version": "0.4.1",
+      "resolved": "https://registry.npmjs.org/@lancedb/vectordb-win32-x64-msvc/-/vectordb-win32-x64-msvc-0.4.1.tgz",
+      "integrity": "sha512-8mvThCppI/AfSPby6Y3t6xpCfbo8IY6JH5exO8fDGTwBFHOqgwR4Izb2K7FgXxkwUYcN4EfGSsk/6B1GpwMudg==",
      "optional": true
    },
    "@neon-rs/cli": {
--- a/node/src/remote/index.ts
+++ b/node/src/remote/index.ts
@@ -267,7 +267,7 @@ export class RemoteTable<T = number[]> implements Table<T> {
    const column = indexParams.column ?? 'vector'
    const indexType = 'vector' // only vector index is supported for remote connections
    const metricType = indexParams.metric_type ?? 'L2'
-    const indexCacheSize = indexParams ?? null
+    const indexCacheSize = indexParams.index_cache_size ?? null

    const data = {
      column,
--- a/python/.bumpversion.cfg
+++ b/python/.bumpversion.cfg
@@ -1,5 +1,5 @@
 [bumpversion]
-current_version = 0.4.1
+current_version = 0.4.3
 commit = True
 message = [python] Bump version: {current_version} → {new_version}
 tag = True
--- a/python/lancedb/embeddings/openai.py
+++ b/python/lancedb/embeddings/openai.py
@@ -10,6 +10,7 @@
 #  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 functools import cached_property
 from typing import List, Union

 import numpy as np
@@ -44,6 +45,10 @@ class OpenAIEmbeddings(TextEmbeddingFunction):
            The texts to embed
        """
        # TODO retry, rate limit, token limit
+        rs = self._openai_client.embeddings.create(input=texts, model=self.name)
+        return [v.embedding for v in rs.data]
+
+    @cached_property
+    def _openai_client(self):
        openai = self.safe_import("openai")
-        rs = openai.Embedding.create(input=texts, model=self.name)["data"]
-        return [v["embedding"] for v in rs]
+        return openai.OpenAI()
--- a/python/lancedb/embeddings/utils.py
+++ b/python/lancedb/embeddings/utils.py
@@ -249,7 +249,7 @@ def retry_with_exponential_backoff(

                if num_retries > max_retries:
                    raise Exception(
-                        f"Maximum number of retries ({max_retries}) exceeded."
+                        f"Maximum number of retries ({max_retries}) exceeded.", e
                    )

                delay *= exponential_base * (1 + jitter * random.random())
--- a/python/lancedb/pydantic.py
+++ b/python/lancedb/pydantic.py
@@ -26,6 +26,7 @@ import numpy as np
 import pyarrow as pa
 import pydantic
 import semver
+from pydantic.fields import FieldInfo

 from .embeddings import EmbeddingFunctionRegistry

@@ -142,8 +143,8 @@ def Vector(
    return FixedSizeList


-def _py_type_to_arrow_type(py_type: Type[Any]) -> pa.DataType:
-    """Convert Python Type to Arrow DataType.
+def _py_type_to_arrow_type(py_type: Type[Any], field: FieldInfo) -> pa.DataType:
+    """Convert a field with native Python type to Arrow data type.

    Raises
    ------
@@ -163,9 +164,13 @@ def _py_type_to_arrow_type(py_type: Type[Any]) -> pa.DataType:
    elif py_type == date:
        return pa.date32()
    elif py_type == datetime:
-        return pa.timestamp("us")
+        tz = get_extras(field, "tz")
+        return pa.timestamp("us", tz=tz)
+    elif getattr(py_type, "__origin__", None) in (list, tuple):
+        child = py_type.__args__[0]
+        return pa.list_(_py_type_to_arrow_type(child, field))
    raise TypeError(
-        f"Converting Pydantic type to Arrow Type: unsupported type {py_type}"
+        f"Converting Pydantic type to Arrow Type: unsupported type {py_type}."
    )


@@ -194,10 +199,10 @@ def _pydantic_to_arrow_type(field: pydantic.fields.FieldInfo) -> pa.DataType:
        args = field.annotation.__args__
        if origin == list:
            child = args[0]
-            return pa.list_(_py_type_to_arrow_type(child))
+            return pa.list_(_py_type_to_arrow_type(child, field))
        elif origin == Union:
            if len(args) == 2 and args[1] == type(None):
-                return _py_type_to_arrow_type(args[0])
+                return _py_type_to_arrow_type(args[0], field)
    elif inspect.isclass(field.annotation):
        if issubclass(field.annotation, pydantic.BaseModel):
            # Struct
@@ -205,7 +210,7 @@ def _pydantic_to_arrow_type(field: pydantic.fields.FieldInfo) -> pa.DataType:
            return pa.struct(fields)
        elif issubclass(field.annotation, FixedSizeListMixin):
            return pa.list_(field.annotation.value_arrow_type(), field.annotation.dim())
-    return _py_type_to_arrow_type(field.annotation)
+    return _py_type_to_arrow_type(field.annotation, field)


 def is_nullable(field: pydantic.fields.FieldInfo) -> bool:
--- a/python/lancedb/query.py
+++ b/python/lancedb/query.py
@@ -70,7 +70,7 @@ class Query(pydantic.BaseModel):
    vector_column: str = VECTOR_COLUMN_NAME

    # vector to search for
-    vector: List[float]
+    vector: Union[List[float], List[List[float]]]

    # sql filter to refine the query with
    filter: Optional[str] = None
@@ -421,6 +421,8 @@ class LanceVectorQueryBuilder(LanceQueryBuilder):
        vector and the returned vectors.
        """
        vector = self._query if isinstance(self._query, list) else self._query.tolist()
+        if isinstance(vector[0], np.ndarray):
+            vector = [v.tolist() for v in vector]
        query = Query(
            vector=vector,
            filter=self._where,
@@ -488,6 +490,27 @@ class LanceFtsQueryBuilder(LanceQueryBuilder):
        scores = pa.array(scores)
        output_tbl = self._table.to_lance().take(row_ids, columns=self._columns)
        output_tbl = output_tbl.append_column("score", scores)
+
+        if self._where is not None:
+            try:
+                # TODO would be great to have Substrait generate pyarrow compute expressions
+                # or conversely have pyarrow support SQL expressions using Substrait
+                import duckdb
+
+                output_tbl = (
+                    duckdb.sql(f"SELECT * FROM output_tbl")
+                    .filter(self._where)
+                    .to_arrow_table()
+                )
+            except ImportError:
+                import lance
+                import tempfile
+
+                # TODO Use "memory://" instead once that's supported
+                with tempfile.TemporaryDirectory() as tmp:
+                    ds = lance.write_dataset(output_tbl, tmp)
+                    output_tbl = ds.to_table(filter=self._where)
+
        return output_tbl


--- a/python/lancedb/remote/table.py
+++ b/python/lancedb/remote/table.py
@@ -11,6 +11,7 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.

+import asyncio
 import uuid
 from functools import cached_property
 from typing import Dict, Optional, Union
@@ -227,8 +228,24 @@ class RemoteTable(Table):
        return LanceVectorQueryBuilder(self, query, vector_column_name)

    def _execute_query(self, query: Query) -> pa.Table:
-        result = self._conn._client.query(self._name, query)
-        return self._conn._loop.run_until_complete(result).to_arrow()
+        if (
+            query.vector is not None
+            and len(query.vector) > 0
+            and not isinstance(query.vector[0], float)
+        ):
+            futures = []
+            for v in query.vector:
+                v = list(v)
+                q = query.copy()
+                q.vector = v
+                futures.append(self._conn._client.query(self._name, q))
+            result = self._conn._loop.run_until_complete(asyncio.gather(*futures))
+            return pa.concat_tables(
+                [add_index(r.to_arrow(), i) for i, r in enumerate(result)]
+            )
+        else:
+            result = self._conn._client.query(self._name, query)
+            return self._conn._loop.run_until_complete(result).to_arrow()

    def delete(self, predicate: str):
        """Delete rows from the table.
@@ -342,3 +359,11 @@ class RemoteTable(Table):
        self._conn._loop.run_until_complete(
            self._conn._client.post(f"/v1/table/{self._name}/update/", data=payload)
        )
+
+
+def add_index(tbl: pa.Table, i: int) -> pa.Table:
+    return tbl.add_column(
+        0,
+        pa.field("query_index", pa.uint32()),
+        pa.array([i] * len(tbl), pa.uint32()),
+    )
--- a/python/pyproject.toml
+++ b/python/pyproject.toml
@@ -1,9 +1,9 @@
 [project]
 name = "lancedb"
-version = "0.4.1"
+version = "0.4.3"
 dependencies = [
    "deprecation",
-    "pylance==0.9.1",
+    "pylance==0.9.2",
    "ratelimiter~=1.0",
    "retry>=0.9.2",
    "tqdm>=4.27.0",
@@ -49,11 +49,11 @@ classifiers = [
 repository = "https://github.com/lancedb/lancedb"

 [project.optional-dependencies]
-tests = ["pandas>=1.4", "pytest", "pytest-mock", "pytest-asyncio", "requests"]
+tests = ["pandas>=1.4", "pytest", "pytest-mock", "pytest-asyncio", "requests", "duckdb", "pytz"]
 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-torch", "cohere", "InstructorEmbedding"]
+embeddings = ["openai>=1.6.1", "sentence-transformers", "torch", "pillow", "open-clip-torch", "cohere", "InstructorEmbedding"]

 [project.scripts]
 lancedb = "lancedb.cli.cli:cli"
--- a/python/tests/test_embeddings_slow.py
+++ b/python/tests/test_embeddings_slow.py
@@ -29,7 +29,7 @@ from lancedb.pydantic import LanceModel, Vector

@pytest.mark.slow
@pytest.mark.parametrize("alias", ["sentence-transformers", "openai"])
-def test_sentence_transformer(alias, tmp_path):
+def test_basic_text_embeddings(alias, tmp_path):
    db = lancedb.connect(tmp_path)
    registry = get_registry()
    func = registry.get(alias).create(max_retries=0)
--- a/python/tests/test_fts.py
+++ b/python/tests/test_fts.py
@@ -12,6 +12,7 @@
 #  limitations under the License.
 import os
 import random
+from unittest import mock

 import numpy as np
 import pandas as pd
@@ -47,6 +48,7 @@ def table(tmp_path) -> ldb.table.LanceTable:
        data=pd.DataFrame(
            {
                "vector": vectors,
+                "id": [i % 2 for i in range(100)],
                "text": text,
                "text2": text,
                "nested": [{"text": t} for t in text],
@@ -88,6 +90,7 @@ def test_create_index_from_table(tmp_path, table):
        [
            {
                "vector": np.random.randn(128),
+                "id": 101,
                "text": "gorilla",
                "text2": "gorilla",
                "nested": {"text": "gorilla"},
@@ -121,3 +124,26 @@ def test_nested_schema(tmp_path, table):
    table.create_fts_index("nested.text")
    rs = table.search("puppy").limit(10).to_list()
    assert len(rs) == 10
+
+
+def test_search_index_with_filter(table):
+    table.create_fts_index("text")
+    orig_import = __import__
+
+    def import_mock(name, *args):
+        if name == "duckdb":
+            raise ImportError
+        return orig_import(name, *args)
+
+    # no duckdb
+    with mock.patch("builtins.__import__", side_effect=import_mock):
+        rs = table.search("puppy").where("id=1").limit(10).to_list()
+        for r in rs:
+            assert r["id"] == 1
+
+    # yes duckdb
+    rs2 = table.search("puppy").where("id=1").limit(10).to_list()
+    for r in rs2:
+        assert r["id"] == 1
+
+    assert rs == rs2
--- a/python/tests/test_pydantic.py
+++ b/python/tests/test_pydantic.py
@@ -13,9 +13,10 @@


 import json
+import pytz
 import sys
 from datetime import date, datetime
-from typing import List, Optional
+from typing import List, Optional, Tuple

 import pyarrow as pa
 import pydantic
@@ -38,11 +39,14 @@ def test_pydantic_to_arrow():
        id: int
        s: str
        vec: list[float]
-        li: List[int]
+        li: list[int]
+        lili: list[list[float]]
+        litu: list[tuple[float, float]]
        opt: Optional[str] = None
        st: StructModel
        dt: date
        dtt: datetime
+        dt_with_tz: datetime = Field(json_schema_extra={"tz": "Asia/Shanghai"})
        # d: dict

    m = TestModel(
@@ -50,9 +54,12 @@ def test_pydantic_to_arrow():
        s="hello",
        vec=[1.0, 2.0, 3.0],
        li=[2, 3, 4],
+        lili=[[2.5, 1.5], [3.5, 4.5], [5.5, 6.5]],
+        litu=[(2.5, 1.5), (3.5, 4.5), (5.5, 6.5)],
        st=StructModel(a="a", b=1.0),
        dt=date.today(),
        dtt=datetime.now(),
+        dt_with_tz=datetime.now(pytz.timezone("Asia/Shanghai")),
    )

    schema = pydantic_to_schema(TestModel)
@@ -63,6 +70,8 @@ def test_pydantic_to_arrow():
            pa.field("s", pa.utf8(), False),
            pa.field("vec", pa.list_(pa.float64()), False),
            pa.field("li", pa.list_(pa.int64()), False),
+            pa.field("lili", pa.list_(pa.list_(pa.float64())), False),
+            pa.field("litu", pa.list_(pa.list_(pa.float64())), False),
            pa.field("opt", pa.utf8(), True),
            pa.field(
                "st",
@@ -73,11 +82,16 @@ def test_pydantic_to_arrow():
            ),
            pa.field("dt", pa.date32(), False),
            pa.field("dtt", pa.timestamp("us"), False),
+            pa.field("dt_with_tz", pa.timestamp("us", tz="Asia/Shanghai"), False),
        ]
    )
    assert schema == expect_schema


+@pytest.mark.skipif(
+    sys.version_info > (3, 8),
+    reason="using native type alias requires python3.9 or higher",
+)
 def test_pydantic_to_arrow_py38():
    class StructModel(pydantic.BaseModel):
        a: str
@@ -88,10 +102,13 @@ def test_pydantic_to_arrow_py38():
        s: str
        vec: List[float]
        li: List[int]
+        lili: List[List[float]]
+        litu: List[Tuple[float, float]]
        opt: Optional[str] = None
        st: StructModel
        dt: date
        dtt: datetime
+        dt_with_tz: datetime = Field(json_schema_extra={"tz": "Asia/Shanghai"})
        # d: dict

    m = TestModel(
@@ -99,9 +116,12 @@ def test_pydantic_to_arrow_py38():
        s="hello",
        vec=[1.0, 2.0, 3.0],
        li=[2, 3, 4],
+        lili=[[2.5, 1.5], [3.5, 4.5], [5.5, 6.5]],
+        litu=[(2.5, 1.5), (3.5, 4.5), (5.5, 6.5)],
        st=StructModel(a="a", b=1.0),
        dt=date.today(),
        dtt=datetime.now(),
+        dt_with_tz=datetime.now(pytz.timezone("Asia/Shanghai")),
    )

    schema = pydantic_to_schema(TestModel)
@@ -112,6 +132,8 @@ def test_pydantic_to_arrow_py38():
            pa.field("s", pa.utf8(), False),
            pa.field("vec", pa.list_(pa.float64()), False),
            pa.field("li", pa.list_(pa.int64()), False),
+            pa.field("lili", pa.list_(pa.list_(pa.float64())), False),
+            pa.field("litu", pa.list_(pa.list_(pa.float64())), False),
            pa.field("opt", pa.utf8(), True),
            pa.field(
                "st",
@@ -122,6 +144,7 @@ def test_pydantic_to_arrow_py38():
            ),
            pa.field("dt", pa.date32(), False),
            pa.field("dtt", pa.timestamp("us"), False),
+            pa.field("dt_with_tz", pa.timestamp("us", tz="Asia/Shanghai"), False),
        ]
    )
    assert schema == expect_schema
Author	SHA1	Message	Date
Lance Release	c3059dc689	[python] Bump version: 0.4.2 → 0.4.3	2023-12-30 00:52:54 +00:00
Lei Xu	a9caa5f2d4	chore: bump pylance to 0.9.2 (#754 )	2023-12-29 16:39:45 -08:00
Xin Hao	8411c36b96	docs: fix link (#752 )	2023-12-29 15:33:24 -08:00
Chang She	7773bda7ee	feat(python): first cut batch queries for remote api (#753 ) issue separate requests under the hood and concatenate results	2023-12-29 15:33:03 -08:00
Lance Release	392777952f	[python] Bump version: 0.4.1 → 0.4.2	2023-12-29 00:19:21 +00:00
Chang She	7e75e50d3a	chore(python): update embedding API to use openai 1.6.1 (#751 ) API has changed significantly, namely `openai.Embedding.create` no longer exists. https://github.com/openai/openai-python/discussions/742 Update the OpenAI embedding function and put a minimum on the openai sdk version.	2023-12-28 15:05:57 -08:00
Chang She	4b8af261a3	feat: add timezone handling for datetime in pydantic (#578 ) If you add timezone information in the Field annotation for a datetime then that will now be passed to the pyarrow data type. I'm not sure how pyarrow enforces timezones, right now, it silently coerces to the timezone given in the column regardless of whether the input had the matching timezone or not. This is probably not the right behavior. Though we could just make it so the user has to make the pydantic model do the validation instead of doing that at the pyarrow conversion layer.	2023-12-28 11:02:56 -08:00
Chang She	c8728d4ca1	feat(python): add post filtering for full text search (#739 ) Closes #721 fts will return results as a pyarrow table. Pyarrow tables has a `filter` method but it does not take sql filter strings (only pyarrow compute expressions). Instead, we do one of two things to support `tbl.search("keywords").where("foo=5").limit(10).to_arrow()`: Default path: If duckdb is available then use duckdb to execute the sql filter string on the pyarrow table. Backup path: Otherwise, write the pyarrow table to a lance dataset and then do `to_table(filter=<filter>)` Neither is ideal. Default path has two issues: 1. requires installing an extra library (duckdb) 2. duckdb mangles some fields (like fixed size list => list) Backup path incurs a latency penalty (~20ms on ssd) to write the resultset to disk. In the short term, once #676 is addressed, we can write the dataset to "memory://" instead of disk, this makes the post filter evaluate much quicker (ETA next week). In the longer term, we'd like to be able to evaluate the filter string on the pyarrow Table directly, one possibility being that we use Substrait to generate pyarrow compute expressions from sql string. Or if there's enough progress on pyarrow, it could support Substrait expressions directly (no ETA) --------- Co-authored-by: Will Jones <willjones127@gmail.com>	2023-12-27 09:31:04 -08:00
Aidan	446f837335	fix: createIndex index cache size (#741 )	2023-12-27 09:25:13 -08:00
Chang She	8f9ad978f5	feat(python): support list of list fields from pydantic schema (#747 ) For object detection, each row may correspond to an image and each image can have multiple bounding boxes of x-y coordinates. This means that a `bbox` field is potentially "list of list of float". This adds support in our pydantic-pyarrow conversion for nested lists.	2023-12-27 09:10:09 -08:00
Lance Release	0df38341d5	Updating package-lock.json	2023-12-26 17:21:51 +00:00