From 770d098f05bd5d63d077586e78e58fefc67ca720 Mon Sep 17 00:00:00 2001 From: Wenjie Zhang Date: Wed, 23 Jul 2025 19:21:45 +0800 Subject: [PATCH] =?UTF-8?q?refactor(database):=20=E6=96=87=E4=BB=B6?= =?UTF-8?q?=E7=BB=93=E6=9E=84=E4=BC=98=E5=8C=96=EF=BC=8C=E9=A1=BA=E4=BE=BF?= =?UTF-8?q?=E4=BF=AE=E5=A4=8D=E4=BA=86=20chroma=20=E7=9A=84=20embedding=20?= =?UTF-8?q?=E9=94=99=E8=AF=AF?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- server/routers/chat_router.py | 2 +- src/__init__.py | 12 +- src/{core => knowledge}/__init__.py | 0 src/{core => knowledge}/chroma_kb.py | 193 ++++------------------ src/{core => knowledge}/graphbase.py | 0 src/{core => knowledge}/history.py | 0 src/{core => knowledge}/indexing.py | 0 src/{core => knowledge}/kb_factory.py | 2 +- src/{core => knowledge}/kb_manager.py | 4 +- src/knowledge/kb_utils.py | 137 +++++++++++++++ src/{core => knowledge}/knowledge_base.py | 2 +- src/{core => knowledge}/lightrag_kb.py | 54 ++---- src/{core => knowledge}/milvus_kb.py | 95 ++--------- web/src/views/DataBaseInfoView.vue | 6 +- web/src/views/GraphView.vue | 11 +- 15 files changed, 218 insertions(+), 300 deletions(-) rename src/{core => knowledge}/__init__.py (100%) rename src/{core => knowledge}/chroma_kb.py (66%) rename src/{core => knowledge}/graphbase.py (100%) rename src/{core => knowledge}/history.py (100%) rename src/{core => knowledge}/indexing.py (100%) rename src/{core => knowledge}/kb_factory.py (97%) rename src/{core => knowledge}/kb_manager.py (98%) create mode 100644 src/knowledge/kb_utils.py rename src/{core => knowledge}/knowledge_base.py (99%) rename src/{core => knowledge}/lightrag_kb.py (83%) rename src/{core => knowledge}/milvus_kb.py (83%) diff --git a/server/routers/chat_router.py b/server/routers/chat_router.py index 344907fb..c7bfa90e 100644 --- a/server/routers/chat_router.py +++ b/server/routers/chat_router.py @@ -11,7 +11,7 @@ from sqlalchemy.orm import Session from pydantic import BaseModel from src import executor, config -from src.core import HistoryManager +from src.knowledge import HistoryManager from src.agents import agent_manager from src.models import select_model from src.utils.logging_config import logger diff --git a/src/__init__.py b/src/__init__.py index 97c00169..c95f6ad2 100644 --- a/src/__init__.py +++ b/src/__init__.py @@ -10,11 +10,11 @@ from src.config import Config # noqa: E402 config = Config() # 导入知识库相关模块 -from src.core.kb_factory import KnowledgeBaseFactory # noqa: E402 -from src.core.kb_manager import KnowledgeBaseManager # noqa: E402 -from src.core.lightrag_kb import LightRagKB # noqa: E402 -from src.core.chroma_kb import ChromaKB # noqa: E402 -from src.core.milvus_kb import MilvusKB # noqa: E402 +from src.knowledge.kb_factory import KnowledgeBaseFactory # noqa: E402 +from src.knowledge.kb_manager import KnowledgeBaseManager # noqa: E402 +from src.knowledge.lightrag_kb import LightRagKB # noqa: E402 +from src.knowledge.chroma_kb import ChromaKB # noqa: E402 +from src.knowledge.milvus_kb import MilvusKB # noqa: E402 # 注册知识库类型 KnowledgeBaseFactory.register("lightrag", LightRagKB, { @@ -37,5 +37,5 @@ KnowledgeBaseFactory.register("milvus", MilvusKB, { work_dir = os.path.join(config.save_dir, "knowledge_base_data") knowledge_base = KnowledgeBaseManager(work_dir) -from src.core import GraphDatabase # noqa: E402 +from src.knowledge import GraphDatabase # noqa: E402 graph_base = GraphDatabase() diff --git a/src/core/__init__.py b/src/knowledge/__init__.py similarity index 100% rename from src/core/__init__.py rename to src/knowledge/__init__.py diff --git a/src/core/chroma_kb.py b/src/knowledge/chroma_kb.py similarity index 66% rename from src/core/chroma_kb.py rename to src/knowledge/chroma_kb.py index 99c7ebd9..77c75400 100644 --- a/src/core/chroma_kb.py +++ b/src/knowledge/chroma_kb.py @@ -6,81 +6,19 @@ from pathlib import Path from typing import Optional, Dict, List, Any from datetime import datetime -try: - import chromadb - from chromadb.config import Settings - from chromadb.api.types import EmbeddingFunction, Documents, Embeddings -except ImportError: - chromadb = None - EmbeddingFunction = None - Documents = None - Embeddings = None +import chromadb +from chromadb.config import Settings +from chromadb.api.types import EmbeddingFunction, Documents, Embeddings +from chromadb.utils.embedding_functions import OpenAIEmbeddingFunction -from src.core.knowledge_base import KnowledgeBase + + +from src.knowledge.knowledge_base import KnowledgeBase +from src.knowledge.kb_utils import split_text_into_chunks, prepare_item_metadata, get_embedding_config from src.utils import logger, hashstr from src import config -if EmbeddingFunction is not None: - class OpenAIEmbeddingFunction(EmbeddingFunction): - """ - 符合 ChromaDB 0.4.16+ 接口的 OpenAI 兼容嵌入函数 - """ - - def __init__(self, model: str, api_key: str, base_url: str): - self.model = model - self.api_key = api_key - self.base_url = base_url.replace("/embeddings", "") - - def __call__(self, input: Documents) -> Embeddings: - """ - 生成文档嵌入向量 - - Args: - input: 文档列表（字符串列表） - - Returns: - Embeddings: 嵌入向量列表 - """ - import asyncio - import concurrent.futures - from lightrag.llm.openai import openai_embed - - # 确保输入是列表格式 - if isinstance(input, str): - texts = [input] - else: - texts = list(input) - - # 在新线程中运行异步函数，避免事件循环冲突 - def run_embedding(): - loop = asyncio.new_event_loop() - asyncio.set_event_loop(loop) - try: - return loop.run_until_complete( - openai_embed( - texts=texts, - model=self.model, - api_key=self.api_key, - base_url=self.base_url, - ) - ) - finally: - loop.close() - - # 使用线程池执行异步函数 - with concurrent.futures.ThreadPoolExecutor() as executor: - future = executor.submit(run_embedding) - embeddings = future.result() - - return embeddings -else: - # 如果 ChromaDB 没有安装，提供一个空的类 - class OpenAIEmbeddingFunction: - def __init__(self, *args, **kwargs): - pass - - class ChromaKB(KnowledgeBase): """基于 ChromaDB 的向量知识库实现""" @@ -136,7 +74,10 @@ class ChromaKB(KnowledgeBase): try: # 尝试获取现有集合 - collection = self.chroma_client.get_collection(name=collection_name) + collection = self.chroma_client.get_collection( + name=collection_name, + embedding_function=embedding_function + ) logger.info(f"Retrieved existing collection: {collection_name}") # 检查现有集合的配置是否匹配当前的 embed_info @@ -144,6 +85,7 @@ class ChromaKB(KnowledgeBase): collection_metadata = collection.metadata or {} current_model = collection_metadata.get("embedding_model", "unknown") + logger.debug(f"Collection {collection_name} uses model '{current_model}', but expected '{expected_model}'.") # 如果模型不匹配，删除现有集合并重新创建 if current_model != expected_model: logger.warning(f"Collection {collection_name} uses model '{current_model}', but expected '{expected_model}'. Recreating collection.") @@ -173,22 +115,12 @@ class ChromaKB(KnowledgeBase): def _get_embedding_function(self, embed_info: Dict): """获取 embedding 函数""" - if embed_info: - model = embed_info["name"] - api_key = os.getenv(embed_info["api_key"], embed_info["api_key"]) - base_url = embed_info["base_url"] - else: - from src.models import select_embedding_model - default_model = select_embedding_model(config.embed_model) - model = default_model.model - api_key = default_model.api_key - base_url = default_model.base_url + config_dict = get_embedding_config(embed_info) - # 返回符合 ChromaDB 0.4.16+ 接口的 EmbeddingFunction 实例 return OpenAIEmbeddingFunction( - model=model, - api_key=api_key, - base_url=base_url + model_name=config_dict["model"], + api_key=config_dict["api_key"], + api_base=config_dict["base_url"].replace('/embeddings', '') ) async def _get_chroma_collection(self, db_id: str): @@ -214,60 +146,15 @@ class ChromaKB(KnowledgeBase): def _split_text_into_chunks(self, text: str, file_id: str, filename: str) -> List[Dict]: """将文本分割成块""" - chunks = [] + chunks = split_text_into_chunks(text, file_id, filename, self.chunk_size, self.chunk_overlap) - # 简单的分块策略：按段落和长度分割 - paragraphs = text.split('\n\n') - - current_chunk = "" - chunk_index = 0 - - for paragraph in paragraphs: - paragraph = paragraph.strip() - if not paragraph: - continue - - # 如果当前块加上新段落会超过限制，保存当前块 - if len(current_chunk) + len(paragraph) > self.chunk_size and current_chunk: - chunks.append({ - "id": f"{file_id}_chunk_{chunk_index}", - "content": current_chunk.strip(), - "file_id": file_id, - "filename": filename, - "chunk_index": chunk_index, - "metadata": { - "source": filename, - "chunk_id": f"{file_id}_chunk_{chunk_index}", - "full_doc_id": file_id - } - }) - - # 开始新块，包含重叠内容 - if len(current_chunk) > self.chunk_overlap: - current_chunk = current_chunk[-self.chunk_overlap:] + "\n\n" + paragraph - else: - current_chunk = paragraph - chunk_index += 1 - else: - if current_chunk: - current_chunk += "\n\n" + paragraph - else: - current_chunk = paragraph - - # 添加最后一块 - if current_chunk.strip(): - chunks.append({ - "id": f"{file_id}_chunk_{chunk_index}", - "content": current_chunk.strip(), - "file_id": file_id, - "filename": filename, - "chunk_index": chunk_index, - "metadata": { - "source": filename, - "chunk_id": f"{file_id}_chunk_{chunk_index}", - "full_doc_id": file_id - } - }) + # 为 ChromaDB 添加特定的 metadata 格式 + for chunk in chunks: + chunk["metadata"] = { + "source": chunk["source"], + "chunk_id": chunk["chunk_id"], + "full_doc_id": file_id + } return chunks @@ -285,35 +172,17 @@ class ChromaKB(KnowledgeBase): processed_items_info = [] for item in items: - # 根据内容类型生成不同的ID和文件名 - if content_type == "file": - file_path = Path(item) - file_id = f"file_{hashstr(str(file_path) + str(time.time()), 6)}" - file_type = file_path.suffix.lower().replace(".", "") - filename = file_path.name - item_path = str(file_path) - else: # URL - file_id = f"url_{hashstr(item + str(time.time()), 6)}" - file_type = "url" - filename = f"webpage_{hashstr(item, 6)}.md" - item_path = item + # 准备文件元数据 + metadata = prepare_item_metadata(item, content_type, db_id) + file_id = metadata["file_id"] + filename = metadata["filename"] + item_path = metadata["path"] # 添加文件记录 - file_record = { - "database_id": db_id, - "filename": filename, - "path": item_path, - "file_type": file_type, - "status": "processing", - "created_at": time.time() - } + file_record = metadata.copy() self.files_meta[file_id] = file_record self._save_metadata() - # 添加 file_id 到返回数据 - file_record = file_record.copy() - file_record["file_id"] = file_id - try: # 根据内容类型处理内容 if content_type == "file": diff --git a/src/core/graphbase.py b/src/knowledge/graphbase.py similarity index 100% rename from src/core/graphbase.py rename to src/knowledge/graphbase.py diff --git a/src/core/history.py b/src/knowledge/history.py similarity index 100% rename from src/core/history.py rename to src/knowledge/history.py diff --git a/src/core/indexing.py b/src/knowledge/indexing.py similarity index 100% rename from src/core/indexing.py rename to src/knowledge/indexing.py diff --git a/src/core/kb_factory.py b/src/knowledge/kb_factory.py similarity index 97% rename from src/core/kb_factory.py rename to src/knowledge/kb_factory.py index ee783bff..529966e0 100644 --- a/src/core/kb_factory.py +++ b/src/knowledge/kb_factory.py @@ -1,5 +1,5 @@ from typing import Dict, Type, Any -from src.core.knowledge_base import KnowledgeBase, KBNotFoundError +from src.knowledge.knowledge_base import KnowledgeBase, KBNotFoundError from src.utils import logger diff --git a/src/core/kb_manager.py b/src/knowledge/kb_manager.py similarity index 98% rename from src/core/kb_manager.py rename to src/knowledge/kb_manager.py index bd2b461b..59e0bd58 100644 --- a/src/core/kb_manager.py +++ b/src/knowledge/kb_manager.py @@ -4,8 +4,8 @@ import time from typing import Dict, Optional, List, Any from datetime import datetime -from src.core.knowledge_base import KnowledgeBase, KBNotFoundError, KBOperationError -from src.core.kb_factory import KnowledgeBaseFactory +from src.knowledge.knowledge_base import KnowledgeBase, KBNotFoundError, KBOperationError +from src.knowledge.kb_factory import KnowledgeBaseFactory from src.utils import logger diff --git a/src/knowledge/kb_utils.py b/src/knowledge/kb_utils.py new file mode 100644 index 00000000..a0b5cdec --- /dev/null +++ b/src/knowledge/kb_utils.py @@ -0,0 +1,137 @@ +import os +import time +from pathlib import Path +from typing import Dict, List, Any +from src.utils import hashstr, get_docker_safe_url, logger +from src import config + + +def split_text_into_chunks(text: str, file_id: str, filename: str, + chunk_size: int = 1000, chunk_overlap: int = 200) -> List[Dict]: + """ + 将文本分割成块 + + Args: + text: 要分割的文本 + file_id: 文件ID + filename: 文件名 + chunk_size: 块大小 + chunk_overlap: 块重叠大小 + + Returns: + List[Dict]: 分割后的文本块列表 + """ + chunks = [] + + # 简单的分块策略：按段落和长度分割 + paragraphs = text.split('\n\n') + + current_chunk = "" + chunk_index = 0 + + for paragraph in paragraphs: + paragraph = paragraph.strip() + if not paragraph: + continue + + # 如果当前块加上新段落会超过限制，保存当前块 + if len(current_chunk) + len(paragraph) > chunk_size and current_chunk: + chunks.append({ + "id": f"{file_id}_chunk_{chunk_index}", + "content": current_chunk.strip(), + "file_id": file_id, + "filename": filename, + "chunk_index": chunk_index, + "source": filename, + "chunk_id": f"{file_id}_chunk_{chunk_index}" + }) + + # 开始新块，包含重叠内容 + if len(current_chunk) > chunk_overlap: + current_chunk = current_chunk[-chunk_overlap:] + "\n\n" + paragraph + else: + current_chunk = paragraph + chunk_index += 1 + else: + if current_chunk: + current_chunk += "\n\n" + paragraph + else: + current_chunk = paragraph + + # 添加最后一块 + if current_chunk.strip(): + chunks.append({ + "id": f"{file_id}_chunk_{chunk_index}", + "content": current_chunk.strip(), + "file_id": file_id, + "filename": filename, + "chunk_index": chunk_index, + "source": filename, + "chunk_id": f"{file_id}_chunk_{chunk_index}" + }) + + return chunks + + +def prepare_item_metadata(item: str, content_type: str, db_id: str) -> Dict: + """ + 准备文件或URL的元数据 + + Args: + item: 文件路径或URL + content_type: 内容类型 ('file' 或 'url') + db_id: 数据库ID + + Returns: + Dict: 包含元数据的字典 + """ + if content_type == "file": + file_path = Path(item) + file_id = f"file_{hashstr(str(file_path) + str(time.time()), 6)}" + file_type = file_path.suffix.lower().replace(".", "") + filename = file_path.name + item_path = str(file_path) + else: # URL + file_id = f"url_{hashstr(item + str(time.time()), 6)}" + file_type = "url" + filename = f"webpage_{hashstr(item, 6)}.md" + item_path = item + + return { + "database_id": db_id, + "filename": filename, + "path": item_path, + "file_type": file_type, + "status": "processing", + "created_at": time.time(), + "file_id": file_id + } + + +def get_embedding_config(embed_info: Dict) -> Dict: + """ + 获取嵌入模型配置 + + Args: + embed_info: 嵌入信息字典 + + Returns: + Dict: 标准化的嵌入配置 + """ + config_dict = {} + + if embed_info: + config_dict['model'] = embed_info["name"] + config_dict['api_key'] = os.getenv(embed_info["api_key"], embed_info["api_key"]) + config_dict['base_url'] = embed_info["base_url"] + config_dict['dimension'] = embed_info.get("dimension", 1024) + else: + from src.models import select_embedding_model + default_model = select_embedding_model(config.embed_model) + config_dict['model'] = default_model.model + config_dict['api_key'] = default_model.api_key + config_dict['base_url'] = default_model.base_url + config_dict['dimension'] = getattr(default_model, 'dimension', 1024) + + logger.debug(f"Embedding config: {config_dict}") + return config_dict \ No newline at end of file diff --git a/src/core/knowledge_base.py b/src/knowledge/knowledge_base.py similarity index 99% rename from src/core/knowledge_base.py rename to src/knowledge/knowledge_base.py index bff4020a..f854502f 100644 --- a/src/core/knowledge_base.py +++ b/src/knowledge/knowledge_base.py @@ -388,7 +388,7 @@ class KnowledgeBase(ABC): if file_ext == '.pdf': # 使用 OCR 处理 PDF - from src.core.indexing import parse_pdf_async + from src.knowledge.indexing import parse_pdf_async text = await parse_pdf_async(str(file_path_obj), params=params) return f"# {file_path_obj.name}\n\n{text}" diff --git a/src/core/lightrag_kb.py b/src/knowledge/lightrag_kb.py similarity index 83% rename from src/core/lightrag_kb.py rename to src/knowledge/lightrag_kb.py index e8efcdfd..c128dc49 100644 --- a/src/core/lightrag_kb.py +++ b/src/knowledge/lightrag_kb.py @@ -10,7 +10,8 @@ from lightrag.llm.openai import openai_complete_if_cache, openai_embed from lightrag.utils import EmbeddingFunc, setup_logger from lightrag.kg.shared_storage import initialize_pipeline_status -from src.core.knowledge_base import KnowledgeBase +from src.knowledge.knowledge_base import KnowledgeBase +from src.knowledge.kb_utils import split_text_into_chunks, prepare_item_metadata, get_embedding_config from src import config from src.utils import logger, hashstr, get_docker_safe_url @@ -126,27 +127,16 @@ class LightRagKB(KnowledgeBase): def _get_embedding_func(self, embed_info: Dict): """获取 embedding 函数""" - if embed_info: - model = embed_info["name"] - api_key = os.getenv(embed_info["api_key"], embed_info["api_key"]) - base_url = get_docker_safe_url(embed_info["base_url"]) - dimension = embed_info["dimension"] - else: - from src.models import select_embedding_model - default_model = select_embedding_model(config.embed_model) - model = default_model.model - api_key = default_model.api_key - base_url = default_model.base_url - dimension = default_model.dimension + config_dict = get_embedding_config(embed_info) return EmbeddingFunc( - embedding_dim=dimension, + embedding_dim=config_dict["dimension"], max_token_size=4096, func=lambda texts: openai_embed( texts=texts, - model=model, - api_key=api_key, - base_url=base_url.replace("/embeddings", ""), + model=config_dict["model"], + api_key=config_dict["api_key"], + base_url=config_dict["base_url"].replace("/embeddings", ""), ), ) @@ -164,35 +154,17 @@ class LightRagKB(KnowledgeBase): processed_items_info = [] for item in items: - # 根据内容类型生成不同的ID和文件名 - if content_type == "file": - file_path = Path(item) - file_id = f"file_{hashstr(str(file_path) + str(time.time()), 6)}" - file_type = file_path.suffix.lower().replace(".", "") - filename = file_path.name - item_path = str(file_path) - else: # URL - file_id = f"url_{hashstr(item + str(time.time()), 6)}" - file_type = "url" - filename = f"webpage_{hashstr(item, 6)}.md" - item_path = item + # 准备文件元数据 + metadata = prepare_item_metadata(item, content_type, db_id) + file_id = metadata["file_id"] + filename = metadata["filename"] + item_path = metadata["path"] # 添加文件记录 - file_record = { - "database_id": db_id, - "filename": filename, - "path": item_path, - "file_type": file_type, - "status": "processing", - "created_at": time.time() - } + file_record = metadata.copy() self.files_meta[file_id] = file_record self._save_metadata() - # 添加 file_id 到返回数据 - file_record = file_record.copy() - file_record["file_id"] = file_id - try: # 根据内容类型处理内容 if content_type == "file": diff --git a/src/core/milvus_kb.py b/src/knowledge/milvus_kb.py similarity index 83% rename from src/core/milvus_kb.py rename to src/knowledge/milvus_kb.py index bee31287..46f12631 100644 --- a/src/core/milvus_kb.py +++ b/src/knowledge/milvus_kb.py @@ -18,7 +18,8 @@ except ImportError: utility = None Collection = None -from src.core.knowledge_base import KnowledgeBase +from src.knowledge.knowledge_base import KnowledgeBase +from src.knowledge.kb_utils import split_text_into_chunks, prepare_item_metadata, get_embedding_config from src.utils import logger, hashstr from src import config @@ -171,16 +172,10 @@ class MilvusKB(KnowledgeBase): def _get_embedding_function(self, embed_info: Dict): """获取 embedding 函数""" - if embed_info: - model = embed_info["name"] - api_key = os.getenv(embed_info["api_key"], embed_info["api_key"]) - base_url = embed_info["base_url"] - else: - from src.models import select_embedding_model - default_model = select_embedding_model(config.embed_model) - model = default_model.model - api_key = default_model.api_key - base_url = default_model.base_url + config_dict = get_embedding_config(embed_info) + model = config_dict["model"] + api_key = config_dict["api_key"] + base_url = config_dict["base_url"] # 返回同步的嵌入函数 def embedding_function(texts): @@ -232,56 +227,7 @@ class MilvusKB(KnowledgeBase): def _split_text_into_chunks(self, text: str, file_id: str, filename: str) -> List[Dict]: """将文本分割成块""" - chunks = [] - - # 简单的分块策略：按段落和长度分割 - paragraphs = text.split('\n\n') - - current_chunk = "" - chunk_index = 0 - - for paragraph in paragraphs: - paragraph = paragraph.strip() - if not paragraph: - continue - - # 如果当前块加上新段落会超过限制，保存当前块 - if len(current_chunk) + len(paragraph) > self.chunk_size and current_chunk: - chunks.append({ - "id": f"{file_id}_chunk_{chunk_index}", - "content": current_chunk.strip(), - "file_id": file_id, - "filename": filename, - "chunk_index": chunk_index, - "source": filename, - "chunk_id": f"{file_id}_chunk_{chunk_index}" - }) - - # 开始新块，包含重叠内容 - if len(current_chunk) > self.chunk_overlap: - current_chunk = current_chunk[-self.chunk_overlap:] + "\n\n" + paragraph - else: - current_chunk = paragraph - chunk_index += 1 - else: - if current_chunk: - current_chunk += "\n\n" + paragraph - else: - current_chunk = paragraph - - # 添加最后一块 - if current_chunk.strip(): - chunks.append({ - "id": f"{file_id}_chunk_{chunk_index}", - "content": current_chunk.strip(), - "file_id": file_id, - "filename": filename, - "chunk_index": chunk_index, - "source": filename, - "chunk_id": f"{file_id}_chunk_{chunk_index}" - }) - - return chunks + return split_text_into_chunks(text, file_id, filename, self.chunk_size, self.chunk_overlap) async def add_content(self, db_id: str, items: List[str], params: Optional[Dict] = None) -> List[Dict]: @@ -300,28 +246,15 @@ class MilvusKB(KnowledgeBase): processed_items_info = [] for item in items: - # 根据内容类型生成不同的ID和文件名 - if content_type == "file": - file_path = Path(item) - file_id = f"file_{hashstr(str(file_path) + str(time.time()), 6)}" - file_type = file_path.suffix.lower().replace(".", "") - filename = file_path.name - item_path = str(file_path) - else: # URL - file_id = f"url_{hashstr(item + str(time.time()), 6)}" - file_type = "url" - filename = f"webpage_{hashstr(item, 6)}.md" - item_path = item + # 准备文件元数据 + metadata = prepare_item_metadata(item, content_type, db_id) + file_id = metadata["file_id"] + filename = metadata["filename"] + item_path = metadata["path"] # 添加文件记录 - file_record = { - "database_id": db_id, - "filename": filename, - "path": item_path, - "file_type": file_type, - "status": "processing", - "created_at": time.time() - } + file_record = metadata.copy() + del file_record["file_id"] # 从记录中移除file_id，因为它是key self.files_meta[file_id] = file_record self._save_metadata() diff --git a/web/src/views/DataBaseInfoView.vue b/web/src/views/DataBaseInfoView.vue index 4909dae2..0f942b91 100644 --- a/web/src/views/DataBaseInfoView.vue +++ b/web/src/views/DataBaseInfoView.vue @@ -89,7 +89,7 @@ :loading="state.chunkLoading" :disabled="(uploadMode === 'file' && fileList.length === 0) || (uploadMode === 'url' && !urlList.trim())" > - 生成分块 + 添加到知识库
@@ -219,7 +219,7 @@
@@ -1550,7 +1550,7 @@ const getKbTypeColor = (type) => { } .header-container { - padding: 8px; + padding: 8px 16px; height: 54px; } diff --git a/web/src/views/GraphView.vue b/web/src/views/GraphView.vue index b550d41a..7009c17c 100644 --- a/web/src/views/GraphView.vue +++ b/web/src/views/GraphView.vue @@ -17,6 +17,9 @@
+ + Neo4j 浏览器 + 上传文件 为{{ unindexedCount }}个节点添加索引 @@ -89,10 +92,10 @@