feat(store): add vector store implementation with LanceDB integration

- Implemented VectorStoreTable class for managing vector storage operations
- Added StoreDocument and SearchedDocument data structures for document handling
- Integrated LanceDB for vector database operations with proper schema management
- Implemented add_documents method for batch document insertion with UUID generation
- Added search_document method for similarity search using vector embeddings
- Created helper methods for record batch processing and column extraction
- Added metadata handling with JSON serialization support
- Implemented error handling for missing columns and type validation
- Updated package version from 0.2.3 to 0.2.4

Author: Whth

packages/fabricatio-rag/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "fabricatio-rag"
-version = "0.2.3"
+version = "0.2.4"
 description = "A Python library for Retrieval-Augmented Generation (RAG) capabilities in LLM applications."
 readme = "README.md"
 license = { file = "LICENSE" }

packages/fabricatio-rag/src/store.rs

@@ -6,7 +6,7 @@ use arrow_array::array::*;
 use arrow_array::cast::AsArray;
 use arrow_array::types::*;
 
-use arrow_array::{RecordBatch, RecordBatchIterator};
+use arrow_array::RecordBatch;
 use error_mapping::AsPyErr;
 use futures_util::TryStreamExt;
 use lancedb::Table;
@@ -49,7 +49,7 @@ type DataContainers = (
     Vec<Option<JsonString>>,
 );
 
-#[gen_stub_pyclass]
+#[cfg_attr(feature = "stubgen", gen_stub_pyclass)]
 #[pyclass(set_all, get_all, from_py_object)]
 #[derive(Clone, Debug)]
 pub struct StoreDocument {
@@ -63,7 +63,7 @@ pub struct StoreDocument {
 /// This structure contains the core information of a searched document including
 /// its unique identifier, content, timestamp of creation/modification, and any
 /// associated metadata stored as JSON string.
-#[gen_stub_pyclass]
+#[cfg_attr(feature = "stubgen", gen_stub_pyclass)]
 #[pyclass(get_all, skip_from_py_object)]
 #[derive(Clone, Debug)]
 pub struct SearchedDocument {