[Bug]: too many requests time out when extracting knowledge graph

[zznova]

Is there an existing issue for the same bug?

• I have checked the existing issues.

RAGFlow workspace code commit ID

448fa1c

RAGFlow image version

v0.16.0 slim

Other environment information

Actual behavior

对知识库中的一个数据集中的一个word文档进行解析时，在进度查看界面看到的大量超时请求，但也有成功的请求。模型使用的是LMstudio部署的qwen2.5-14b-instruct-1m

Expected behavior

No response

Steps to reproduce

对知识库中的一个数据集中的一个word文档进行解析时，在进度查看界面看到的大量超时请求，但也有成功的请求。模型使用的是LMstudio部署的qwen2.5-14b-instruct-1m

Additional information

开始于:
Wed, 26 Feb 2025 08:41:31 GMT
持续时间:
4329.05 s
进度:
08:41:31 Page(0~100000000): Reused previous task's chunks.
09:25:05 Knowledge graph extraction error:ERROR: Request timed out.
09:25:05 Knowledge graph extraction error:ERROR: Request timed out.
09:33:57 Entities extraction progress ... 85/142 (9848 tokens)
09:34:59 Entities extraction progress ... 86/142 (11750 tokens)
09:34:59 Knowledge graph extraction error:ERROR: Request timed out.
09:34:59 Knowledge graph extraction error:ERROR: Request timed out.
09:34:59 Knowledge graph extraction error:ERROR: Request timed out.
09:35:00 Knowledge graph extraction error:ERROR: Request timed out.

[zznova]

10:02:34 Entities extraction progress ... 139/142 (6241 tokens)
10:02:34 Entities extraction progress ... 140/142 (5281 tokens)
10:02:34 Entities extraction progress ... 141/142 (8021 tokens)
10:02:34 Entities extraction progress ... 142/142 (6220 tokens)
10:05:05 [ERROR]Fail to bind LLM used by Knowledge Graph: Error tokenizing 'ledge_graph_kwd IN ('entity') AND entity_kwd='筛选列''
10:05:05 [ERROR][Exception]: Error tokenizing 'ledge_graph_kwd IN ('entity') AND entity_kwd='筛选列''

最终该文档解析失败

[KevinHuSh]

What about setting GRAPH_EXTRACTOR_MAX_WORKERS=1 as an environment variable.