mirror of
https://github.com/RYDE-WORK/Langchain-Chatchat.git
synced 2026-01-19 13:23:16 +08:00
65592a45c3
新功能：
- 支持在线 Embeddings：zhipu-api, qwen-api, minimax-api, qianfan-api
- API 增加 /other/embed_texts 接口
- init_database.py 增加 --embed-model 参数，可以指定使用的嵌入模型（本地或在线均可）
- 对于 FAISS 知识库，支持多向量库，默认位置：{KB_PATH}/vector_store/{embed_model}
- Lite 模式支持所有知识库相关功能。此模式下最主要的限制是：
- 不能使用本地 LLM 和 Embeddings 模型
- 知识库不支持 PDF 文件
- init_database.py 重建知识库时不再默认情况数据库表，增加 clear-tables 参数手动控制。
- API 和 WEBUI 中 score_threshold 参数范围改为 [0, 2]，以更好的适应在线嵌入模型
问题修复：
- API 中 list_config_models 会删除 ONLINE_LLM_MODEL 中的敏感信息，导致第二轮API请求错误
开发者：
- 统一向量库的识别：以（kb_name,embed_model）为判断向量库唯一性的依据，避免 FAISS 知识库缓存加载逻辑错误
- KBServiceFactory.get_service_by_name 中添加 default_embed_model 参数，用于在构建新知识库时设置 embed_model
- 优化 kb_service 中 Embeddings 操作：
- 统一加载接口： server.utils.load_embeddings，利用全局缓存避免各处 Embeddings 传参
- 统一文本嵌入接口：server.knowledge_base.kb_service.base.[embed_texts, embed_documents]
- 重写 normalize 函数，去除对 scikit-learn/scipy 的依赖
91 lines
3.6 KiB
Python
91 lines
3.6 KiB
Python
import json
|
|
from typing import List, Dict, Optional
|
|
|
|
from langchain.schema import Document
|
|
from langchain.vectorstores.pgvector import PGVector, DistanceStrategy
|
|
from sqlalchemy import text
|
|
|
|
from configs import kbs_config
|
|
|
|
from server.knowledge_base.kb_service.base import SupportedVSType, KBService, EmbeddingsFunAdapter, \
|
|
score_threshold_process
|
|
from server.knowledge_base.utils import KnowledgeFile
|
|
|
|
|
|
class PGKBService(KBService):
|
|
pg_vector: PGVector
|
|
|
|
def _load_pg_vector(self):
|
|
self.pg_vector = PGVector(embedding_function=EmbeddingsFunAdapter(self.embed_model),
|
|
collection_name=self.kb_name,
|
|
distance_strategy=DistanceStrategy.EUCLIDEAN,
|
|
connection_string=kbs_config.get("pg").get("connection_uri"))
|
|
|
|
def get_doc_by_id(self, id: str) -> Optional[Document]:
|
|
with self.pg_vector.connect() as connect:
|
|
stmt = text("SELECT document, cmetadata FROM langchain_pg_embedding WHERE collection_id=:id")
|
|
results = [Document(page_content=row[0], metadata=row[1]) for row in
|
|
connect.execute(stmt, parameters={'id': id}).fetchall()]
|
|
if len(results) > 0:
|
|
return results[0]
|
|
|
|
def do_init(self):
|
|
self._load_pg_vector()
|
|
|
|
def do_create_kb(self):
|
|
pass
|
|
|
|
def vs_type(self) -> str:
|
|
return SupportedVSType.PG
|
|
|
|
def do_drop_kb(self):
|
|
with self.pg_vector.connect() as connect:
|
|
connect.execute(text(f'''
|
|
-- 删除 langchain_pg_embedding 表中关联到 langchain_pg_collection 表中 的记录
|
|
DELETE FROM langchain_pg_embedding
|
|
WHERE collection_id IN (
|
|
SELECT uuid FROM langchain_pg_collection WHERE name = '{self.kb_name}'
|
|
);
|
|
-- 删除 langchain_pg_collection 表中 记录
|
|
DELETE FROM langchain_pg_collection WHERE name = '{self.kb_name}';
|
|
'''))
|
|
connect.commit()
|
|
|
|
def do_search(self, query: str, top_k: int, score_threshold: float):
|
|
self._load_pg_vector()
|
|
embed_func = EmbeddingsFunAdapter(self.embed_model)
|
|
embeddings = embed_func.embed_query(query)
|
|
docs = self.pg_vector.similarity_search_with_score_by_vector(embeddings, top_k)
|
|
return score_threshold_process(score_threshold, top_k, docs)
|
|
|
|
def do_add_doc(self, docs: List[Document], **kwargs) -> List[Dict]:
|
|
ids = self.pg_vector.add_documents(docs)
|
|
doc_infos = [{"id": id, "metadata": doc.metadata} for id, doc in zip(ids, docs)]
|
|
return doc_infos
|
|
|
|
def do_delete_doc(self, kb_file: KnowledgeFile, **kwargs):
|
|
with self.pg_vector.connect() as connect:
|
|
filepath = kb_file.filepath.replace('\\', '\\\\')
|
|
connect.execute(
|
|
text(
|
|
''' DELETE FROM langchain_pg_embedding WHERE cmetadata::jsonb @> '{"source": "filepath"}'::jsonb;'''.replace(
|
|
"filepath", filepath)))
|
|
connect.commit()
|
|
|
|
def do_clear_vs(self):
|
|
self.pg_vector.delete_collection()
|
|
self.pg_vector.create_collection()
|
|
|
|
|
|
if __name__ == '__main__':
|
|
from server.db.base import Base, engine
|
|
|
|
# Base.metadata.create_all(bind=engine)
|
|
pGKBService = PGKBService("test")
|
|
# pGKBService.create_kb()
|
|
# pGKBService.add_doc(KnowledgeFile("README.md", "test"))
|
|
# pGKBService.delete_doc(KnowledgeFile("README.md", "test"))
|
|
# pGKBService.drop_kb()
|
|
print(pGKBService.get_doc_by_id("f1e51390-3029-4a19-90dc-7118aaa25772"))
|
|
# print(pGKBService.search_docs("如何启动api服务"))
|