Feat/qdrant vdb (#1649)

* feat: Qdrant vector search support

Signed-off-by: Anush008 <anushshetty90@gmail.com>

* fix: modify env

* fix: fix the old version problem

* fix: For older versions

* perf: minor perf

---------

Signed-off-by: Anush008 <anushshetty90@gmail.com>
Co-authored-by: Anush008 <anushshetty90@gmail.com>
Co-authored-by: Junyan Qin <rockchinq@gmail.com>

pkg/rag/knowledge/services/retriever.py

@@ -24,23 +24,23 @@ class Retriever(base_service.BaseService):
             extra_args={},  # TODO: add extra args
         )
 
-        chroma_results = await self.ap.vector_db_mgr.vector_db.search(kb_id, query_embedding[0], k)
+        vector_results = await self.ap.vector_db_mgr.vector_db.search(kb_id, query_embedding[0], k)
 
-        # 'ids' is always returned by ChromaDB, even if not explicitly in 'include'
-        matched_chroma_ids = chroma_results.get('ids', [[]])[0]
-        distances = chroma_results.get('distances', [[]])[0]
-        chroma_metadatas = chroma_results.get('metadatas', [[]])[0]
+        # 'ids' shape mirrors the Chroma-style response contract for compatibility
+        matched_vector_ids = vector_results.get('ids', [[]])[0]
+        distances = vector_results.get('distances', [[]])[0]
+        vector_metadatas = vector_results.get('metadatas', [[]])[0]
 
-        if not matched_chroma_ids:
-            self.ap.logger.info('No relevant chunks found in Chroma.')
+        if not matched_vector_ids:
+            self.ap.logger.info('No relevant chunks found in vector database.')
             return []
 
         result: list[retriever_entities.RetrieveResultEntry] = []
 
-        for i, id in enumerate(matched_chroma_ids):
+        for i, id in enumerate(matched_vector_ids):
             entry = retriever_entities.RetrieveResultEntry(
                 id=id,
-                metadata=chroma_metadatas[i],
+                metadata=vector_metadatas[i],
                 distance=distances[i],
             )
             result.append(entry)

pkg/vector/mgr.py

@@ -3,6 +3,7 @@ from __future__ import annotations
 from ..core import app
 from .vdb import VectorDatabase
 from .vdbs.chroma import ChromaVectorDatabase
+from .vdbs.qdrant import QdrantVectorDatabase
 
 
 class VectorDBManager:
@@ -13,6 +14,17 @@ class VectorDBManager:
         self.ap = ap
 
     async def initialize(self):
-        # 初始化 Chroma 向量数据库（可扩展为多种实现）
-        if self.vector_db is None:
+        kb_config = self.ap.instance_config.data.get('vdb')
+        if kb_config:
+            if kb_config.get('use') == 'chroma':
+                self.vector_db = ChromaVectorDatabase(self.ap)
+                self.ap.logger.info('Initialized Chroma vector database backend.')
+            elif kb_config.get('use') == 'qdrant':
+                self.vector_db = QdrantVectorDatabase(self.ap)
+                self.ap.logger.info('Initialized Qdrant vector database backend.')
+            else:
+                self.vector_db = ChromaVectorDatabase(self.ap)
+                self.ap.logger.warning('No valid vector database backend configured, defaulting to Chroma.')
+        else:
             self.vector_db = ChromaVectorDatabase(self.ap)
+            self.ap.logger.warning('No vector database backend configured, defaulting to Chroma.')

pkg/vector/vdb.py

@@ -14,24 +14,25 @@ class VectorDatabase(abc.ABC):
         metadatas: list[dict[str, Any]],
         documents: list[str],
     ) -> None:
-        """向指定 collection 添加向量数据。"""
+        """Add vector data to the specified collection."""
         pass
 
     @abc.abstractmethod
     async def search(self, collection: str, query_embedding: np.ndarray, k: int = 5) -> Dict[str, Any]:
-        """在指定 collection 中检索最相似的向量。"""
+        """Search for the most similar vectors in the specified collection."""
         pass
 
     @abc.abstractmethod
     async def delete_by_file_id(self, collection: str, file_id: str) -> None:
-        """根据 file_id 删除指定 collection 中的向量。"""
+        """Delete vectors from the specified collection by file_id."""
         pass
 
     @abc.abstractmethod
     async def get_or_create_collection(self, collection: str):
-        """获取或创建 collection。"""
+        """Get or create collection."""
         pass
 
     @abc.abstractmethod
     async def delete_collection(self, collection: str):
+        """Delete collection."""
         pass

pkg/vector/vdbs/qdrant.py

@@ -0,0 +1,104 @@
+from __future__ import annotations
+
+from typing import Any, Dict, List
+
+from qdrant_client import AsyncQdrantClient, models
+from pkg.core import app
+from pkg.vector.vdb import VectorDatabase
+
+
+class QdrantVectorDatabase(VectorDatabase):
+    def __init__(self, ap: app.Application):
+        self.ap = ap
+        url = self.ap.instance_config.data['vdb']['qdrant']['url']
+        host = self.ap.instance_config.data['vdb']['qdrant']['host']
+        port = self.ap.instance_config.data['vdb']['qdrant']['port']
+        api_key = self.ap.instance_config.data['vdb']['qdrant']['api_key']
+
+        if url:
+            self.client = AsyncQdrantClient(url=url, api_key=api_key)
+        else:
+            self.client = AsyncQdrantClient(host=host, port=int(port), api_key=api_key)
+
+        self._collections: set[str] = set()
+
+    async def _ensure_collection(self, collection: str, vector_size: int) -> None:
+        if collection in self._collections:
+            return
+
+        exists = await self.client.collection_exists(collection)
+        if exists:
+            self._collections.add(collection)
+            return
+
+        await self.client.create_collection(
+            collection_name=collection,
+            vectors_config=models.VectorParams(size=vector_size, distance=models.Distance.COSINE),
+        )
+        self._collections.add(collection)
+        self.ap.logger.info(f"Qdrant collection '{collection}' created with dim={vector_size}.")
+
+    async def get_or_create_collection(self, collection: str):
+        # Qdrant requires vector size to create a collection; no-op here.
+        pass
+
+    async def add_embeddings(
+        self,
+        collection: str,
+        ids: List[str],
+        embeddings_list: List[List[float]],
+        metadatas: List[Dict[str, Any]],
+    ) -> None:
+        if not embeddings_list:
+            return
+
+        await self._ensure_collection(collection, len(embeddings_list[0]))
+
+        points = [
+            models.PointStruct(id=ids[i], vector=embeddings_list[i], payload=metadatas[i]) for i in range(len(ids))
+        ]
+        await self.client.upsert(collection_name=collection, points=points)
+        self.ap.logger.info(f"Added {len(ids)} embeddings to Qdrant collection '{collection}'.")
+
+    async def search(self, collection: str, query_embedding: list[float], k: int = 5) -> dict[str, Any]:
+        exists = await self.client.collection_exists(collection)
+        if not exists:
+            return {'ids': [[]], 'metadatas': [[]], 'distances': [[]]}
+
+        hits = (
+            await self.client.query_points(
+                collection_name=collection,
+                query=query_embedding,
+                limit=k,
+                with_payload=True,
+            )
+        ).points
+        ids = [str(hit.id) for hit in hits]
+        metadatas = [hit.payload or {} for hit in hits]
+        # Qdrant's score is similarity; convert to a pseudo-distance for consistency
+        distances = [1 - float(hit.score) if hit.score is not None else 1.0 for hit in hits]
+        results = {'ids': [ids], 'metadatas': [metadatas], 'distances': [distances]}
+
+        self.ap.logger.info(f"Qdrant search in '{collection}' returned {len(results.get('ids', [[]])[0])} results.")
+        return results
+
+    async def delete_by_file_id(self, collection: str, file_id: str) -> None:
+        exists = await self.client.collection_exists(collection)
+        if not exists:
+            return
+
+        await self.client.delete(
+            collection_name=collection,
+            points_selector=models.Filter(
+                must=[models.FieldCondition(key='file_id', match=models.MatchValue(value=file_id))]
+            ),
+        )
+        self.ap.logger.info(f"Deleted embeddings from Qdrant collection '{collection}' with file_id: {file_id}")
+
+    async def delete_collection(self, collection: str):
+        try:
+            await self.client.delete_collection(collection)
+            self._collections.discard(collection)
+            self.ap.logger.info(f"Qdrant collection '{collection}' deleted.")
+        except Exception:
+            self.ap.logger.warning(f"Qdrant collection '{collection}' not found.")

pyproject.toml

@@ -1,9 +1,9 @@
 [project]
 name = "langbot"
 version = "4.2.2"
-description = "高稳定、支持扩展、多模态 - 大模型原生即时通信机器人平台"
+description = "Easy-to-use global IM bot platform designed for LLM era"
 readme = "README.md"
-requires-python = ">=3.10.1"
+requires-python = ">=3.10.1,<4.0"
 dependencies = [
     "aiocqhttp>=1.4.4",
     "aiofiles>=24.1.0",
@@ -60,6 +60,7 @@ dependencies = [
     "html2text>=2024.2.26",
     "langchain>=0.2.0",
     "chromadb>=0.4.24",
+    "qdrant-client (>=1.15.1,<2.0.0)",
 ]
 keywords = [
     "bot",

templates/config.yaml

@@ -20,3 +20,10 @@ system:
     jwt:
         expire: 604800
         secret: ''
+vdb:
+    use: chroma
+    qdrant:
+        url: ''
+        host: localhost
+        port: 6333
+        api_key: ''