[BUG] 简洁阐述问题 / langchain-chatchat 部署后再导入知识库时候，向量库为打×

[jiexiuqi1566]

2025-05-28 21:25:23.864 | WARNING | chatchat.server.utils:detect_xf_models:107 - cannot connect to xinference host: http://127.0.0.1:9997, please check your configuration.
2025-05-28 21:25:23.888 | INFO | chatchat.startup:run_webui:80 - Webui MODEL_PLATFORMS: [PlatformConfig(platform_name='xinference', platform_type='xinference', api_base_url='http://127.0.0.1:9997/v1', api_key='EMPTY', api_proxy='', api_concurrencies=5, auto_detect_model=True, llm_models=[], embed_models=[], text2image_models=[], image2text_models=[], rerank_models=[], speech2text_models=[], text2speech_models=[]), PlatformConfig(platform_name='ollama', platform_type='ollama', api_base_url='http://127.0.0.1:11434/v1', api_key='EMPTY', api_proxy='', api_concurrencies=5, auto_detect_model=False, llm_models=['qwen:7b', 'qwen2:7b', 'deepseek-r1'], embed_models=['quentinz/bge-large-zh-v1.5', 'bge-m3'], text2image_models=[], image2text_models=[], rerank_models=[], speech2text_models=[], text2speech_models=[]), PlatformConfig(platform_name='oneapi', platform_type='oneapi', api_base_url='http://127.0.0.1:3000/v1', api_key='sk-', api_proxy='', api_concurrencies=5, auto_detect_model=False, llm_models=['chatglm_pro', 'chatglm_turbo', 'chatglm_std', 'chatglm_lite', 'qwen-turbo', 'qwen-plus', 'qwen-max', 'qwen-max-longcontext', 'ERNIE-Bot', 'ERNIE-Bot-turbo', 'ERNIE-Bot-4', 'SparkDesk'], embed_models=['text-embedding-v1', 'Embedding-V1'], text2image_models=[], image2text_models=[], rerank_models=[], speech2text_models=[], text2speech_models=[]), PlatformConfig(platform_name='openai', platform_type='openai', api_base_url='https://api.openai.com/v1', api_key='sk-proj-', api_proxy='', api_concurrencies=5, auto_detect_model=False, llm_models=['gpt-4o', 'gpt-3.5-turbo'], embed_models=['text-embedding-3-small', 'text-embedding-3-large'], text2image_models=[], image2text_models=[], rerank_models=[], speech2text_models=[], text2speech_models=[])]
2025-05-28 21:25:24.148 Port 8501 is already in use
INFO: 127.0.0.1:6497 - "GET /knowledge_base/list_knowledge_bases HTTP/1.1" 200 OK
2025-05-28 21:26:20.332 | INFO | chatchat.server.knowledge_base.utils:file2docs:336 - UnstructuredFileLoader used for C:\Users\Administrator\data\knowledge_base\samples\content\参数信息.txt

**2025-05-28 21:26:23,624 pikepdf._core 19140 INFO pikepdf C++ to Python logger bridge initialized

2025-05-28 21:26:23.655 | ERROR | chatchat.server.knowledge_base.utils:files2docs_in_thread_file2docs:419 - ImportError: 从文件 samples/参数信息.txt 加载文档时出错：cannot import name 'PSSyntaxError' from 'pdfminer.pdfparser' (D:\Anaconda\envs\langchain_new\lib\site-packages\pdfminer\pdfparser.py)**
2025-05-28 21:26:23.655 | INFO | chatchat.server.knowledge_base.kb_cache.faiss_cache:load_vector_store:109 - loading vector store in 'samples/vector_store/bge-m3' from disk.
2025-05-28 21:26:25.812 | WARNING | chatchat.server.utils:detect_xf_models:107 - cannot connect to xinference host: http://127.0.0.1:9997, please check your configuration.
2025-05-28 21:26:25,902 faiss.loader 19140 INFO Loading faiss with AVX2 support.
2025-05-28 21:26:25,918 faiss.loader 19140 INFO Could not load library with AVX2 support due to:
ModuleNotFoundError("No module named 'faiss.swigfaiss_avx2'")
2025-05-28 21:26:25,918 faiss.loader 19140 INFO Loading faiss.
2025-05-28 21:26:25,949 faiss.loader 19140 INFO Successfully loaded faiss.
2025-05-28 21:26:25.965 | INFO | chatchat.server.knowledge_base.kb_cache.faiss_cache:save:40 - 已将向量库 ('samples', 'bge-m3') 保存到磁盘
INFO: 127.0.0.1:6524 - "POST /knowledge_base/upload_docs HTTP/1.1" 200 OK

[dosubot]

Hi @jiexiuqi1566! I'm Dosu and I’m helping the Langchain-Chatchat team.

针对你遇到的三个主要问题，整理如下解决思路：

1. xinference host 无法连接

   • 日志显示无法连接到 http://127.0.0.1:9997。请确认 xinference 服务已启动，并监听在正确端口。你可以参考项目的 startup.sh 脚本 和 docker-compose.yaml 配置，确保 xinference 服务已正确运行。如果用 Docker，注意 network_mode: "host" 可能影响端口暴露，必要时可手动指定端口映射。更多模型注册和启动方法可参考 README_xinference.md。

2. pdfminer ImportError: cannot import name 'PSSyntaxError'

   • 项目依赖的是 pdfminer-six，但没有找到关于 PSSyntaxError 的相关说明。pdfminer-six 的新版本中可能已经移除了 PSSyntaxError，导致无法导入。建议检查你的 pdfminer-six 版本，或尝试升级/降级到兼容的版本。如果依赖代码确实需要 PSSyntaxError，可以考虑用其他 PDF 解析库（如 PyMuPDF/fitz）替代，或修改相关代码以适配当前 pdfminer 版本。

3. faiss.swigfaiss_avx2 ModuleNotFoundError

   • 项目依赖的是 faiss-cpu==1.7.4，这是纯 CPU 版本，不包含 AVX2 优化模块（faiss.swigfaiss_avx2）。日志显示虽然加载 AVX2 失败，但已自动回退到普通 CPU 版本并成功加载，无需强制解决。如果你确实需要 AVX2 支持，需要自行编译带 AVX2 的 faiss 或安装官方带 AVX2 的二进制包，但这通常不是必须的，除非你有极高的性能需求。requirements.txt 已明确指定使用 CPU 版本。

如需进一步定位 pdfminer 的兼容性问题，建议提供 pdfminer-six 的具体版本号。其余问题如 xinference 端口占用、API key 配置等，也建议逐一排查服务是否正常启动、端口未被占用、配置项填写正确。

To reply, just mention my name @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}   

[azhe1234]

2025-05-28 21:25:23.864 | WARNING | chatchat.server.utils:detect_xf_models:107 - cannot connect to xinference host: http://127.0.0.1:9997, please check your configuration. 2025-05-28 21:25:23.888 | INFO | chatchat.startup:run_webui:80 - Webui MODEL_PLATFORMS: [PlatformConfig(platform_name='xinference', platform_type='xinference', api_base_url='http://127.0.0.1:9997/v1', api_key='EMPTY', api_proxy='', api_concurrencies=5, auto_detect_model=True, llm_models=[], embed_models=[], text2image_models=[], image2text_models=[], rerank_models=[], speech2text_models=[], text2speech_models=[]), PlatformConfig(platform_name='ollama', platform_type='ollama', api_base_url='http://127.0.0.1:11434/v1', api_key='EMPTY', api_proxy='', api_concurrencies=5, auto_detect_model=False, llm_models=['qwen:7b', 'qwen2:7b', 'deepseek-r1'], embed_models=['quentinz/bge-large-zh-v1.5', 'bge-m3'], text2image_models=[], image2text_models=[], rerank_models=[], speech2text_models=[], text2speech_models=[]), PlatformConfig(platform_name='oneapi', platform_type='oneapi', api_base_url='http://127.0.0.1:3000/v1', api_key='sk-', api_proxy='', api_concurrencies=5, auto_detect_model=False, llm_models=['chatglm_pro', 'chatglm_turbo', 'chatglm_std', 'chatglm_lite', 'qwen-turbo', 'qwen-plus', 'qwen-max', 'qwen-max-longcontext', 'ERNIE-Bot', 'ERNIE-Bot-turbo', 'ERNIE-Bot-4', 'SparkDesk'], embed_models=['text-embedding-v1', 'Embedding-V1'], text2image_models=[], image2text_models=[], rerank_models=[], speech2text_models=[], text2speech_models=[]), PlatformConfig(platform_name='openai', platform_type='openai', api_base_url='https://api.openai.com/v1', api_key='sk-proj-', api_proxy='', api_concurrencies=5, auto_detect_model=False, llm_models=['gpt-4o', 'gpt-3.5-turbo'], embed_models=['text-embedding-3-small', 'text-embedding-3-large'], text2image_models=[], image2text_models=[], rerank_models=[], speech2text_models=[], text2speech_models=[])] 2025-05-28 21:25:24.148 Port 8501 is already in use INFO: 127.0.0.1:6497 - "GET /knowledge_base/list_knowledge_bases HTTP/1.1" 200 OK 2025-05-28 21:26:20.332 | INFO | chatchat.server.knowledge_base.utils:file2docs:336 - UnstructuredFileLoader used for C:\Users\Administrator\data\knowledge_base\samples\content\参数信息.txt

**2025-05-28 21:26:23,624 pikepdf._core 19140 INFO pikepdf C++ to Python logger bridge initialized

2025-05-28 21:26:23.655 | ERROR | chatchat.server.knowledge_base.utils:files2docs_in_thread_file2docs:419 - ImportError: 从文件 samples/参数信息.txt 加载文档时出错：cannot import name 'PSSyntaxError' from 'pdfminer.pdfparser' (D:\Anaconda\envs\langchain_new\lib\site-packages\pdfminer\pdfparser.py)** 2025-05-28 21:26:23.655 | INFO | chatchat.server.knowledge_base.kb_cache.faiss_cache:load_vector_store:109 - loading vector store in 'samples/vector_store/bge-m3' from disk. 2025-05-28 21:26:25.812 | WARNING | chatchat.server.utils:detect_xf_models:107 - cannot connect to xinference host: http://127.0.0.1:9997, please check your configuration. 2025-05-28 21:26:25,902 faiss.loader 19140 INFO Loading faiss with AVX2 support. 2025-05-28 21:26:25,918 faiss.loader 19140 INFO Could not load library with AVX2 support due to: ModuleNotFoundError("No module named 'faiss.swigfaiss_avx2'") 2025-05-28 21:26:25,918 faiss.loader 19140 INFO Loading faiss. 2025-05-28 21:26:25,949 faiss.loader 19140 INFO Successfully loaded faiss. 2025-05-28 21:26:25.965 | INFO | chatchat.server.knowledge_base.kb_cache.faiss_cache:save:40 - 已将向量库 ('samples', 'bge-m3') 保存到磁盘 INFO: 127.0.0.1:6524 - "POST /knowledge_base/upload_docs HTTP/1.1" 200 OK

看起来像PSSyntaxError库没有导入，你看看你的向量模型是否在xinference界面上成功启动了

[jiexiuqi1566]

2025-05-28 21:25:23.864 | WARNING | chatchat.server.utils:detect_xf_models:107 - cannot connect to xinference host: http://127.0.0.1:9997, please check your configuration. 2025-05-28 21:25:23.888 | INFO | chatchat.startup:run_webui:80 - Webui MODEL_PLATFORMS: [PlatformConfig(platform_name='xinference', platform_type='xinference', api_base_url='http://127.0.0.1:9997/v1', api_key='EMPTY', api_proxy='', api_concurrencies=5, auto_detect_model=True, llm_models=[], embed_models=[], text2image_models=[], image2text_models=[], rerank_models=[], speech2text_models=[], text2speech_models=[]), PlatformConfig(platform_name='ollama', platform_type='ollama', api_base_url='http://127.0.0.1:11434/v1', api_key='EMPTY', api_proxy='', api_concurrencies=5, auto_detect_model=False, llm_models=['qwen:7b', 'qwen2:7b', 'deepseek-r1'], embed_models=['quentinz/bge-large-zh-v1.5', 'bge-m3'], text2image_models=[], image2text_models=[], rerank_models=[], speech2text_models=[], text2speech_models=[]), PlatformConfig(platform_name='oneapi', platform_type='oneapi', api_base_url='http://127.0.0.1:3000/v1', api_key='sk-', api_proxy='', api_concurrencies=5, auto_detect_model=False, llm_models=['chatglm_pro', 'chatglm_turbo', 'chatglm_std', 'chatglm_lite', 'qwen-turbo', 'qwen-plus', 'qwen-max', 'qwen-max-longcontext', 'ERNIE-Bot', 'ERNIE-Bot-turbo', 'ERNIE-Bot-4', 'SparkDesk'], embed_models=['text-embedding-v1', 'Embedding-V1'], text2image_models=[], image2text_models=[], rerank_models=[], speech2text_models=[], text2speech_models=[]), PlatformConfig(platform_name='openai', platform_type='openai', api_base_url='https://api.openai.com/v1', api_key='sk-proj-', api_proxy='', api_concurrencies=5, auto_detect_model=False, llm_models=['gpt-4o', 'gpt-3.5-turbo'], embed_models=['text-embedding-3-small', 'text-embedding-3-large'], text2image_models=[], image2text_models=[], rerank_models=[], speech2text_models=[], text2speech_models=[])] 2025-05-28 21:25:24.148 Port 8501 is already in use INFO: 127.0.0.1:6497 - "GET /knowledge_base/list_knowledge_bases HTTP/1.1" 200 OK 2025-05-28 21:26:20.332 | INFO | chatchat.server.knowledge_base.utils:file2docs:336 - UnstructuredFileLoader used for C:\Users\Administrator\data\knowledge_base\samples\content\参数信息.txt

**2025-05-28 21:26:23,624 pikepdf._core 19140 INFO pikepdf C++ to Python logger bridge initialized

2025-05-28 21:26:23.655 | ERROR | chatchat.server.knowledge_base.utils:files2docs_in_thread_file2docs:419 - ImportError: 从文件 samples/参数信息.txt 加载文档时出错：cannot import name 'PSSyntaxError' from 'pdfminer.pdfparser' (D:\Anaconda\envs\langchain_new\lib\site-packages\pdfminer\pdfparser.py)** 2025-05-28 21:26:23.655 | INFO | chatchat.server.knowledge_base.kb_cache.faiss_cache:load_vector_store:109 - loading vector store in 'samples/vector_store/bge-m3' from disk. 2025-05-28 21:26:25.812 | WARNING | chatchat.server.utils:detect_xf_models:107 - cannot connect to xinference host: http://127.0.0.1:9997, please check your configuration. 2025-05-28 21:26:25,902 faiss.loader 19140 INFO Loading faiss with AVX2 support. 2025-05-28 21:26:25,918 faiss.loader 19140 INFO Could not load library with AVX2 support due to: ModuleNotFoundError("No module named 'faiss.swigfaiss_avx2'") 2025-05-28 21:26:25,918 faiss.loader 19140 INFO Loading faiss. 2025-05-28 21:26:25,949 faiss.loader 19140 INFO Successfully loaded faiss. 2025-05-28 21:26:25.965 | INFO | chatchat.server.knowledge_base.kb_cache.faiss_cache:save:40 - 已将向量库 ('samples', 'bge-m3') 保存到磁盘 INFO: 127.0.0.1:6524 - "POST /knowledge_base/upload_docs HTTP/1.1" 200 OK

看起来像PSSyntaxError库没有导入，你看看你的向量模型是否在xinference界面上成功启动了

怎么看呢

[azhe1234]

2025-05-28 21:25:23.864 | WARNING | chatchat.server.utils:detect_xf_models:107 - cannot connect to xinference host: http://127.0.0.1:9997, please check your configuration. 2025-05-28 21:25:23.888 | INFO | chatchat.startup:run_webui:80 - Webui MODEL_PLATFORMS: [PlatformConfig(platform_name='xinference', platform_type='xinference', api_base_url='http://127.0.0.1:9997/v1', api_key='EMPTY', api_proxy='', api_concurrencies=5, auto_detect_model=True, llm_models=[], embed_models=[], text2image_models=[], image2text_models=[], rerank_models=[], speech2text_models=[], text2speech_models=[]), PlatformConfig(platform_name='ollama', platform_type='ollama', api_base_url='http://127.0.0.1:11434/v1', api_key='EMPTY', api_proxy='', api_concurrencies=5, auto_detect_model=False, llm_models=['qwen:7b', 'qwen2:7b', 'deepseek-r1'], embed_models=['quentinz/bge-large-zh-v1.5', 'bge-m3'], text2image_models=[], image2text_models=[], rerank_models=[], speech2text_models=[], text2speech_models=[]), PlatformConfig(platform_name='oneapi', platform_type='oneapi', api_base_url='http://127.0.0.1:3000/v1', api_key='sk-', api_proxy='', api_concurrencies=5, auto_detect_model=False, llm_models=['chatglm_pro', 'chatglm_turbo', 'chatglm_std', 'chatglm_lite', 'qwen-turbo', 'qwen-plus', 'qwen-max', 'qwen-max-longcontext', 'ERNIE-Bot', 'ERNIE-Bot-turbo', 'ERNIE-Bot-4', 'SparkDesk'], embed_models=['text-embedding-v1', 'Embedding-V1'], text2image_models=[], image2text_models=[], rerank_models=[], speech2text_models=[], text2speech_models=[]), PlatformConfig(platform_name='openai', platform_type='openai', api_base_url='https://api.openai.com/v1', api_key='sk-proj-', api_proxy='', api_concurrencies=5, auto_detect_model=False, llm_models=['gpt-4o', 'gpt-3.5-turbo'], embed_models=['text-embedding-3-small', 'text-embedding-3-large'], text2image_models=[], image2text_models=[], rerank_models=[], speech2text_models=[], text2speech_models=[])] 2025-05-28 21:25:24.148 Port 8501 is already in use INFO: 127.0.0.1:6497 - "GET /knowledge_base/list_knowledge_bases HTTP/1.1" 200 OK 2025-05-28 21:26:20.332 | INFO | chatchat.server.knowledge_base.utils:file2docs:336 - UnstructuredFileLoader used for C:\Users\Administrator\data\knowledge_base\samples\content\参数信息.txt

**2025-05-28 21:26:23,624 pikepdf._core 19140 INFO pikepdf C++ to Python logger bridge initialized

2025-05-28 21:26:23.655 | ERROR | chatchat.server.knowledge_base.utils:files2docs_in_thread_file2docs:419 - ImportError: 从文件 samples/参数信息.txt 加载文档时出错：cannot import name 'PSSyntaxError' from 'pdfminer.pdfparser' (D:\Anaconda\envs\langchain_new\lib\site-packages\pdfminer\pdfparser.py)** 2025-05-28 21:26:23.655 | INFO | chatchat.server.knowledge_base.kb_cache.faiss_cache:load_vector_store:109 - loading vector store in 'samples/vector_store/bge-m3' from disk. 2025-05-28 21:26:25.812 | WARNING | chatchat.server.utils:detect_xf_models:107 - cannot connect to xinference host: http://127.0.0.1:9997, please check your configuration. 2025-05-28 21:26:25,902 faiss.loader 19140 INFO Loading faiss with AVX2 support. 2025-05-28 21:26:25,918 faiss.loader 19140 INFO Could not load library with AVX2 support due to: ModuleNotFoundError("No module named 'faiss.swigfaiss_avx2'") 2025-05-28 21:26:25,918 faiss.loader 19140 INFO Loading faiss. 2025-05-28 21:26:25,949 faiss.loader 19140 INFO Successfully loaded faiss. 2025-05-28 21:26:25.965 | INFO | chatchat.server.knowledge_base.kb_cache.faiss_cache:save:40 - 已将向量库 ('samples', 'bge-m3') 保存到磁盘 INFO: 127.0.0.1:6524 - "POST /knowledge_base/upload_docs HTTP/1.1" 200 OK

看起来像PSSyntaxError库没有导入，你看看你的向量模型是否在xinference界面上成功启动了

怎么看呢

localhost:9997,查看里面running models的ui界面是否启动了模型