rework: switch to ChromaDB (vector_store.py)

vector_store.py

@@ -1,17 +1,16 @@
-"""Векторное хранилище Qdrant + эмбеддинги Ollama."""
+"""Векторное хранилище ChromaDB + эмбеддинги Ollama."""
 from __future__ import annotations
 
 import os
+from pathlib import Path
 
-from langchain_qdrant import QdrantVectorStore
+from langchain_chroma import Chroma
 from langchain_core.documents import Document
 from langchain_ollama import OllamaEmbeddings
 from langchain_text_splitters import RecursiveCharacterTextSplitter
-from qdrant_client import QdrantClient
-from qdrant_client.http.models import Distance, VectorParams
 
 COLLECTION_NAME = "knowledge_base"
-QDRANT_PATH = os.getenv("QDRANT_PATH", "./qdrant_data")
+CHROMA_PATH = os.getenv("CHROMA_PATH", "./chroma_data")
 OLLAMA_BASE_URL = os.getenv("OLLAMA_BASE_URL", "http://localhost:11434")
 EMBED_MODEL = os.getenv("OLLAMA_EMBED_MODEL", "nomic-embed-text")
 
@@ -23,18 +22,12 @@ def get_embeddings() -> OllamaEmbeddings:
     )
 
 
-def get_vector_store() -> QdrantVectorStore:
-    client = QdrantClient(path=QDRANT_PATH)
-    collections = client.get_collections().collections
-    if not any(col.name == COLLECTION_NAME for col in collections):
-        client.create_collection(
+def get_vector_store() -> Chroma:
+    Path(CHROMA_PATH).mkdir(parents=True, exist_ok=True)
+    return Chroma(
         collection_name=COLLECTION_NAME,
-            vectors_config=VectorParams(size=768, distance=Distance.COSINE),
-        )
-    return QdrantVectorStore(
-        client=client,
-        collection_name=COLLECTION_NAME,
-        embedding=get_embeddings(),
+        embedding_function=get_embeddings(),
+        persist_directory=CHROMA_PATH,
     )