fix(rust): infer vector dimensions from list data

python/python/lancedb/db.py

@@ -707,9 +707,6 @@ class LanceDBConnection(DBConnection):
         self._namespace_client_properties = namespace_client_properties
         if _inner is not None:
             self._conn = _inner
-            # Native-derived wrappers resolve this in their async reconstruction
-            # path so construction never synchronously re-enters LOOP.
-            self._read_consistency_interval = read_consistency_interval
             self._cached_namespace_client = None
             return
 
@@ -759,14 +756,11 @@ class LanceDBConnection(DBConnection):
         # storage_options.  Also, this class really shouldn't be holding any state
         # beyond _conn.
         self._conn = AsyncConnection(LOOP.run(do_connect()))
-        # Keep property access synchronous so debugger introspection cannot wait on
-        # the background loop while that thread is suspended at a breakpoint.
-        self._read_consistency_interval = read_consistency_interval
         self._cached_namespace_client: Optional[LanceNamespace] = None
 
     @property
     def read_consistency_interval(self) -> Optional[timedelta]:
-        return self._read_consistency_interval
+        return LOOP.run(self._conn.get_read_consistency_interval())
 
     @property
     def session(self) -> Optional[Session]:
@@ -777,16 +771,8 @@ class LanceDBConnection(DBConnection):
         return self._conn.uri
 
     @classmethod
-    def from_inner(
-        cls,
-        inner: LanceDbConnection,
-        read_consistency_interval: Optional[timedelta],
-    ):
-        return cls(
-            None,
-            read_consistency_interval=read_consistency_interval,
-            _inner=inner,
-        )
+    def from_inner(cls, inner: LanceDbConnection):
+        return cls(None, _inner=inner)
 
     def __repr__(self) -> str:
         return f"{self.__class__.__name__}(uri={self._conn.uri!r})"

python/python/lancedb/embeddings/instructor.py

@@ -3,7 +3,6 @@
 
 
 from typing import List
-from urllib.parse import unquote, urlparse
 
 import numpy as np
 
@@ -126,20 +125,9 @@ class InstructorEmbeddingFunction(TextEmbeddingFunction):
 
     @weak_lru(maxsize=1)
     def get_model(self):
-        huggingface_hub = attempt_import_or_raise("huggingface_hub", "huggingface-hub")
-        missing = object()
-        original_cached_download = getattr(huggingface_hub, "cached_download", missing)
-        if original_cached_download is missing:
-            huggingface_hub.cached_download = _cached_download(huggingface_hub)
-
-        try:
-            instructor_embedding = attempt_import_or_raise(
-                "InstructorEmbedding", "InstructorEmbedding"
-            )
-        finally:
-            if original_cached_download is missing:
-                del huggingface_hub.cached_download
-
+        instructor_embedding = attempt_import_or_raise(
+            "InstructorEmbedding", "InstructorEmbedding"
+        )
         torch = attempt_import_or_raise("torch", "torch")
 
         model = instructor_embedding.INSTRUCTOR(self.name)
@@ -152,44 +140,3 @@ class InstructorEmbeddingFunction(TextEmbeddingFunction):
                 model, {torch.nn.Linear}, dtype=torch.qint8
             )
         return model
-
-
-def _cached_download(huggingface_hub):
-    """Provide the legacy download API used by sentence-transformers 2.2.x."""
-
-    def cached_download(
-        *,
-        url,
-        cache_dir=None,
-        force_filename=None,
-        library_name=None,
-        library_version=None,
-        user_agent=None,
-        use_auth_token=None,
-        **_,
-    ):
-        path = urlparse(url).path.lstrip("/")
-        try:
-            repo_id, resolved_path = path.split("/resolve/", maxsplit=1)
-            revision, filename = resolved_path.split("/", maxsplit=1)
-        except ValueError as err:
-            raise ValueError(f"Unsupported Hugging Face Hub URL: {url}") from err
-
-        repo_id = unquote(repo_id)
-        revision = unquote(revision)
-        filename = unquote(filename)
-        # sentence-transformers derives force_filename from this Hub path with
-        # os.path.join. Using the URL path beneath local_dir produces the same
-        # local destination without sending Windows separators to the Hub.
-        return huggingface_hub.hf_hub_download(
-            repo_id=repo_id,
-            filename=filename,
-            revision=revision,
-            local_dir=cache_dir,
-            library_name=library_name,
-            library_version=library_version,
-            user_agent=user_agent,
-            token=use_auth_token,
-        )
-
-    return cached_download

python/python/lancedb/permutation.py

@@ -226,7 +226,7 @@ class PermutationBuilder:
 
         async def do_execute():
             inner_tbl = await self._async.execute()
-            return await LanceTable.from_inner(inner_tbl)
+            return LanceTable.from_inner(inner_tbl)
 
         return LOOP.run(do_execute())
 

python/python/lancedb/table.py

@@ -2182,15 +2182,11 @@ class LanceTable(Table):
         return self.name
 
     @classmethod
-    async def from_inner(cls, tbl: LanceDBTable):
-        from .db import AsyncConnection, LanceDBConnection
+    def from_inner(cls, tbl: LanceDBTable):
+        from .db import LanceDBConnection
 
         async_tbl = AsyncTable(tbl)
-        inner_conn = tbl.database()
-        read_consistency_interval = await AsyncConnection(
-            inner_conn
-        ).get_read_consistency_interval()
-        conn = LanceDBConnection.from_inner(inner_conn, read_consistency_interval)
+        conn = LanceDBConnection.from_inner(tbl.database())
         return cls(
             conn,
             async_tbl.name,

python/python/tests/test_db.py

@@ -77,23 +77,6 @@ def test_sync_repr_does_not_use_background_loop(tmp_path, monkeypatch):
     assert repr(table) == f"LanceTable(name='test', _conn={db!r})"
 
 
-def test_read_consistency_interval_does_not_use_background_loop(tmp_path, monkeypatch):
-    from lancedb.background_loop import LOOP
-    from lancedb.db import LanceDBConnection
-
-    consistency_interval = timedelta(seconds=5)
-    db = lancedb.connect(tmp_path, read_consistency_interval=consistency_interval)
-    db_from_inner = LanceDBConnection.from_inner(db._inner, consistency_interval)
-
-    def fail_run(*args, **kwargs):
-        raise AssertionError("properties should not use the Python background loop")
-
-    monkeypatch.setattr(LOOP, "run", fail_run)
-
-    assert db.read_consistency_interval == consistency_interval
-    assert db_from_inner.read_consistency_interval == consistency_interval
-
-
 def test_ingest_pd(tmp_path):
     db = lancedb.connect(tmp_path)
 

python/python/tests/test_embeddings.py

@@ -1,11 +1,8 @@
 # SPDX-License-Identifier: Apache-2.0
 # SPDX-FileCopyrightText: Copyright The LanceDB Authors
 
-import ntpath
 import os
 import pickle
-import sys
-from types import ModuleType
 from typing import List, Optional, Union
 from unittest.mock import MagicMock, patch
 
@@ -525,59 +522,6 @@ def test_embedding_function_safe_model_dump(embedding_type):
             )
 
 
-def test_instructor_embedding_supports_huggingface_hub_without_cached_download(
-    tmp_path, monkeypatch
-):
-    from lancedb.embeddings.instructor import InstructorEmbeddingFunction
-
-    hub_download = MagicMock(return_value="/cache/1_Pooling/config.json")
-    huggingface_hub = ModuleType("huggingface_hub")
-    huggingface_hub.hf_hub_download = hub_download
-    torch = ModuleType("torch")
-    monkeypatch.setitem(sys.modules, "huggingface_hub", huggingface_hub)
-    monkeypatch.setitem(sys.modules, "torch", torch)
-    monkeypatch.delitem(sys.modules, "InstructorEmbedding", raising=False)
-    monkeypatch.syspath_prepend(str(tmp_path))
-
-    (tmp_path / "InstructorEmbedding.py").write_text(
-        "from huggingface_hub import cached_download\n\n"
-        "class INSTRUCTOR:\n"
-        "    def __init__(self, name):\n"
-        "        self.name = name\n"
-    )
-
-    embedding = InstructorEmbeddingFunction.create(show_progress_bar=False)
-    instructor_model = embedding.get_model()
-
-    assert instructor_model.name == "hkunlp/instructor-base"
-    assert not hasattr(huggingface_hub, "cached_download")
-
-    instructor_embedding = sys.modules["InstructorEmbedding"]
-    path = instructor_embedding.cached_download(
-        url=(
-            "https://huggingface.co/hkunlp/instructor-base/resolve/abc123/"
-            "1_Pooling/config.json"
-        ),
-        cache_dir="/cache",
-        force_filename=ntpath.join("1_Pooling", "config.json"),
-        library_name="sentence-transformers",
-        library_version="2.2.2",
-        use_auth_token="token",
-    )
-
-    assert path == "/cache/1_Pooling/config.json"
-    hub_download.assert_called_once_with(
-        repo_id="hkunlp/instructor-base",
-        filename="1_Pooling/config.json",
-        revision="abc123",
-        local_dir="/cache",
-        library_name="sentence-transformers",
-        library_version="2.2.2",
-        user_agent=None,
-        token="token",
-    )
-
-
 @patch("time.sleep")
 def test_retry(mock_sleep):
     test_function = MagicMock(side_effect=[Exception] * 9 + ["result"])

python/python/tests/test_permutation.py

@@ -6,7 +6,6 @@ import math
 import pytest
 
 from lancedb import DBConnection, Table, connect
-from lancedb.background_loop import LOOP
 from lancedb.permutation import Permutation, Permutations, permutation_builder
 
 
@@ -32,25 +31,6 @@ def test_split_random_ratios(mem_db):
     assert 65 <= split_1_count <= 75  # ~70% ± tolerance
 
 
-def test_execute_does_not_reenter_background_loop(tmp_path, monkeypatch):
-    import threading
-
-    db = connect(tmp_path)
-    tbl = db.create_table("test_table", pa.table({"x": range(10)}))
-    original_run = LOOP.run
-
-    def fail_on_reentry(future):
-        assert threading.current_thread() is not LOOP.thread
-        return original_run(future)
-
-    monkeypatch.setattr(LOOP, "run", fail_on_reentry)
-
-    permutation_tbl = permutation_builder(tbl).execute()
-
-    assert permutation_tbl.count_rows() == 10
-    assert permutation_tbl._conn.read_consistency_interval is None
-
-
 def test_split_random_counts(mem_db):
     """Test random splitting with absolute counts."""
     tbl = mem_db.create_table(

python/python/tests/test_table.py

@@ -6,7 +6,6 @@ import os
 import sys
 import threading
 import warnings
-from concurrent.futures import ThreadPoolExecutor
 from datetime import date, datetime, timedelta
 from time import sleep
 from typing import List
@@ -2125,27 +2124,6 @@ def test_delete(mem_db: DBConnection):
     assert table.to_arrow()["id"].to_pylist() == [1]
 
 
-def test_concurrent_deletes_are_thread_safe(mem_db: DBConnection):
-    num_workers = 8
-    table = mem_db.create_table(
-        "my_table", data=[{"id": row_id} for row_id in range(num_workers)]
-    )
-    barrier = threading.Barrier(num_workers)
-
-    def delete(row_id: int):
-        barrier.wait()
-        return table.delete(f"id = {row_id}")
-
-    with ThreadPoolExecutor(max_workers=num_workers) as pool:
-        results = list(pool.map(delete, range(num_workers)))
-
-    assert all(result.num_deleted_rows == 1 for result in results)
-    assert sorted(result.version for result in results) == list(
-        range(2, num_workers + 2)
-    )
-    assert table.count_rows() == 0
-
-
 def test_delete_expr(mem_db: DBConnection):
     table = mem_db.create_table(
         "my_table",

python/src/table.rs

@@ -745,9 +745,6 @@ impl Table {
 
     #[allow(private_interfaces)]
     pub fn delete(self_: PyRef<'_, Self>, condition: PredicateArg) -> PyResult<Bound<'_, PyAny>> {
-        // Do not hold the Python borrow across the await.  The cloned Rust table
-        // handle is thread-safe and allows deletes on the same Python table to
-        // run concurrently without PyO3 reporting "Already borrowed".
         let inner = self_.inner_ref()?.clone();
         future_into_py(self_.py(), async move {
             let result = match &condition {

rust/lancedb/src/connection.rs

@@ -420,6 +420,11 @@ impl Connection {
     ///
     /// * `name` - The name of the table
     /// * `initial_data` - The initial data to write to the table
+    ///
+    /// Floating-point `List` columns named `vec`, or with `vector` or `embedding`
+    /// in their name, are inferred as vector columns when the first batch has a
+    /// uniform, non-zero list length. The inferred dimension is validated for all
+    /// subsequent batches and stored as a `FixedSizeList`.
     pub fn create_table<T: Scannable + 'static>(
         &self,
         name: impl Into<String>,

rust/lancedb/src/connection/create_table.rs

@@ -8,7 +8,7 @@ use lance_io::object_store::StorageOptionsProvider;
 use crate::{
     Error, Result, Table,
     connection::{merge_storage_options, set_storage_options_provider},
-    data::scannable::{Scannable, WithEmbeddingsScannable},
+    data::scannable::{Scannable, WithEmbeddingsScannable, maybe_infer_vector_schema},
     database::{CreateTableMode, CreateTableRequest, Database},
     embeddings::{EmbeddingDefinition, EmbeddingFunction, EmbeddingRegistry},
     table::WriteOptions,
@@ -147,6 +147,8 @@ impl CreateTableBuilder {
         let embedding_registry = self.embedding_registry.clone();
         let parent = self.parent.clone();
 
+        self.request.data = maybe_infer_vector_schema(self.request.data).await?;
+
         // If embeddings were configured via add_embedding(), wrap the data
         if !self.embeddings.is_empty() {
             let wrapped_data: Box<dyn Scannable> = Box::new(WithEmbeddingsScannable::try_new(

rust/lancedb/src/data/scannable.rs

@@ -18,13 +18,16 @@ use crate::embeddings::{
 };
 use crate::table::{ColumnDefinition, ColumnKind, TableDefinition};
 use crate::{Error, Result};
-use arrow_array::{ArrayRef, RecordBatch, RecordBatchIterator, RecordBatchReader};
-use arrow_schema::{ArrowError, SchemaRef};
+use arrow_array::{ArrayRef, RecordBatch, RecordBatchIterator, RecordBatchReader, cast::AsArray};
+use arrow_cast::{CastOptions, cast_with_options};
+use arrow_schema::{ArrowError, DataType, Schema, SchemaRef};
 use async_trait::async_trait;
 use futures::StreamExt;
 use futures::stream::once;
 use lance_datafusion::utils::StreamingWriteSource;
 
+use super::inspect::infer_dimension;
+
 pub trait Scannable: Send {
     /// Returns the schema of the data.
     fn schema(&self) -> SchemaRef;
@@ -497,6 +500,146 @@ impl Scannable for PeekedScannable {
     }
 }
 
+fn name_suggests_vector_column(name: &str) -> bool {
+    let name = name.to_ascii_lowercase();
+    name == "vec" || name.contains("vector") || name.contains("embedding")
+}
+
+/// Infer fixed dimensions for vector-like floating-point list columns.
+///
+/// Lance vector search requires `FixedSizeList` columns, but Arrow data assembled
+/// from runtime embedding models is often represented as `List`. For vector-like
+/// column names, inspect the first batch and convert uniform, non-empty lists to a
+/// fixed-size schema. Every subsequent batch is cast with strict length checking.
+pub(crate) async fn maybe_infer_vector_schema(
+    data: Box<dyn Scannable>,
+) -> Result<Box<dyn Scannable>> {
+    let input_schema = data.schema();
+    let candidates = input_schema
+        .fields()
+        .iter()
+        .enumerate()
+        .filter(|(_, field)| {
+            name_suggests_vector_column(field.name())
+                && matches!(
+                    field.data_type(),
+                    DataType::List(item) | DataType::LargeList(item)
+                        if item.data_type().is_floating()
+                )
+        })
+        .map(|(index, _)| index)
+        .collect::<Vec<_>>();
+    if candidates.is_empty() {
+        return Ok(data);
+    }
+
+    let mut peeked = PeekedScannable::new(data);
+    let Some(first_batch) = peeked.peek().await else {
+        return Ok(Box::new(peeked));
+    };
+
+    let mut fields = input_schema.fields().iter().cloned().collect::<Vec<_>>();
+    let mut changed = false;
+    for index in candidates {
+        let array = first_batch.column(index);
+        let dimension = match array.data_type() {
+            DataType::List(_) => {
+                infer_dimension::<arrow_array::types::Int32Type>(array.as_list::<i32>())?
+                    .map(i64::from)
+            }
+            DataType::LargeList(_) => {
+                infer_dimension::<arrow_array::types::Int64Type>(array.as_list::<i64>())?
+            }
+            _ => unreachable!(),
+        };
+        let Some(dimension) = dimension.filter(|dimension| *dimension > 0) else {
+            continue;
+        };
+        let dimension = i32::try_from(dimension).map_err(|_| Error::InvalidInput {
+            message: format!(
+                "Vector column '{}' has a dimension larger than i32::MAX",
+                fields[index].name()
+            ),
+        })?;
+        let item = match fields[index].data_type() {
+            DataType::List(item) | DataType::LargeList(item) => item.clone(),
+            _ => unreachable!(),
+        };
+        fields[index] = Arc::new(
+            fields[index]
+                .as_ref()
+                .clone()
+                .with_data_type(DataType::FixedSizeList(item, dimension)),
+        );
+        changed = true;
+    }
+
+    if !changed {
+        return Ok(Box::new(peeked));
+    }
+
+    let output_schema = Arc::new(Schema::new_with_metadata(
+        fields,
+        input_schema.metadata().clone(),
+    ));
+    Ok(Box::new(InferredVectorScannable {
+        inner: peeked,
+        output_schema,
+    }))
+}
+
+struct InferredVectorScannable {
+    inner: PeekedScannable,
+    output_schema: SchemaRef,
+}
+
+impl Scannable for InferredVectorScannable {
+    fn schema(&self) -> SchemaRef {
+        self.output_schema.clone()
+    }
+
+    fn scan_as_stream(&mut self) -> SendableRecordBatchStream {
+        let output_schema = self.output_schema.clone();
+        let stream_schema = output_schema.clone();
+        let stream = self.inner.scan_as_stream().map(move |batch| {
+            let batch = batch?;
+            let columns = batch
+                .columns()
+                .iter()
+                .zip(output_schema.fields())
+                .map(|(array, field)| {
+                    if array.data_type() == field.data_type() {
+                        Ok(array.clone())
+                    } else {
+                        cast_with_options(
+                            array,
+                            field.data_type(),
+                            &CastOptions {
+                                safe: false,
+                                ..Default::default()
+                            },
+                        )
+                        .map_err(Error::from)
+                    }
+                })
+                .collect::<Result<Vec<_>>>()?;
+            Ok(RecordBatch::try_new(output_schema.clone(), columns)?)
+        });
+        Box::pin(SimpleRecordBatchStream {
+            schema: stream_schema,
+            stream,
+        })
+    }
+
+    fn num_rows(&self) -> Option<usize> {
+        self.inner.num_rows()
+    }
+
+    fn rescannable(&self) -> bool {
+        self.inner.rescannable()
+    }
+}
+
 /// Compute the number of write partitions based on data size estimates.
 ///
 /// `sample_bytes` and `sample_rows` come from a representative batch and are

rust/lancedb/src/lib.rs

@@ -72,7 +72,9 @@
 //!
 //! LanceDB uses [arrow-rs](https://github.com/apache/arrow-rs) to define schema, data types and array itself.
 //! It treats [`FixedSizeList<Float16/Float32>`](https://docs.rs/arrow/latest/arrow/array/struct.FixedSizeListArray.html)
-//! columns as vector columns.
+//! columns as vector columns. When creating a table with a floating-point `List`
+//! column named `vec`, or with `vector` or `embedding` in its name, LanceDB infers
+//! a uniform dimension from the first batch and stores it as a `FixedSizeList`.
 //!
 //! For more details, please refer to the [LanceDB documentation](https://docs.lancedb.com).
 //!
@@ -82,6 +84,8 @@
 //! schema of the `RecordBatch` determines the schema of the table.
 //!
 //! Vector columns should be represented as `FixedSizeList<Float16/Float32>` data type.
+//! A vector-like `List<Float16/Float32>` input is also accepted when every vector
+//! has the same runtime dimension.
 //!
 //! ```rust
 //! # use std::sync::Arc;

rust/lancedb/src/query.rs

@@ -1648,8 +1648,8 @@ mod tests {
     use super::*;
     use arrow::{array::downcast_array, compute::concat_batches, datatypes::Int32Type};
     use arrow_array::{
-        FixedSizeListArray, Float32Array, Int32Array, RecordBatch, StringArray, cast::AsArray,
-        types::Float32Type,
+        FixedSizeListArray, Float32Array, Int32Array, ListArray, RecordBatch, StringArray,
+        cast::AsArray, types::Float32Type,
     };
     use arrow_schema::{DataType, Field as ArrowField, Schema as ArrowSchema};
     use futures::{StreamExt, TryStreamExt};
@@ -2282,6 +2282,91 @@ mod tests {
         );
     }
 
+    #[tokio::test]
+    async fn vector_search_infers_dimension_from_list_array() {
+        let tmp_dir = tempdir().unwrap();
+        let schema = Arc::new(ArrowSchema::new(vec![
+            ArrowField::new("id", DataType::Int32, false),
+            ArrowField::new(
+                "vec",
+                DataType::List(Arc::new(ArrowField::new("item", DataType::Float32, true))),
+                true,
+            ),
+        ]));
+        let vectors = ListArray::from_iter_primitive::<Float32Type, _, _>([
+            Some([Some(0.0), Some(0.0)]),
+            Some([Some(1.0), Some(1.0)]),
+        ]);
+        let batch = RecordBatch::try_new(
+            schema,
+            vec![Arc::new(Int32Array::from(vec![0, 1])), Arc::new(vectors)],
+        )
+        .unwrap();
+        let table = connect(tmp_dir.path().to_str().unwrap())
+            .execute()
+            .await
+            .unwrap()
+            .create_table("vectors", batch)
+            .execute()
+            .await
+            .unwrap();
+        assert!(matches!(
+            table.schema().await.unwrap().field(1).data_type(),
+            DataType::FixedSizeList(_, 2)
+        ));
+
+        let results = table
+            .vector_search(&[0.0, 0.0])
+            .unwrap()
+            .limit(1)
+            .execute()
+            .await
+            .unwrap()
+            .try_collect::<Vec<_>>()
+            .await
+            .unwrap();
+
+        assert_eq!(results[0]["id"].as_primitive::<Int32Type>().value(0), 0);
+    }
+
+    #[tokio::test]
+    async fn inferred_vector_dimension_is_validated_across_batches() {
+        let tmp_dir = tempdir().unwrap();
+        let schema = Arc::new(ArrowSchema::new(vec![ArrowField::new(
+            "vec",
+            DataType::List(Arc::new(ArrowField::new("item", DataType::Float32, true))),
+            true,
+        )]));
+        let first = RecordBatch::try_new(
+            schema.clone(),
+            vec![Arc::new(
+                ListArray::from_iter_primitive::<Float32Type, _, _>([Some([Some(0.0), Some(0.0)])]),
+            )],
+        )
+        .unwrap();
+        let wrong_dimension = RecordBatch::try_new(
+            schema,
+            vec![Arc::new(
+                ListArray::from_iter_primitive::<Float32Type, _, _>([Some([
+                    Some(1.0),
+                    Some(1.0),
+                    Some(1.0),
+                ])]),
+            )],
+        )
+        .unwrap();
+
+        let result = connect(tmp_dir.path().to_str().unwrap())
+            .execute()
+            .await
+            .unwrap()
+            .create_table("vectors", vec![first, wrong_dimension])
+            .execute()
+            .await;
+
+        assert!(result.is_err());
+    }
+
     #[tokio::test]
     async fn test_fast_search_plan() {
         let tmp_dir = tempdir().unwrap();