OpenSPG
[Bug] [kag] invalid vectorizer config: 'No configuration setting found for key path'
Search before asking
- [X] I had searched in the issues and found no similar issues.
Operating system information
Linux
What happened
when i save the global configuration ,the errors occur， how can i fex it
How to reproduce
download https://raw.githubusercontent.com/OpenSPG/openspg/refs/heads/master/dev/release/docker-compose-west.yml run docker-compose -f docker-compose.yml up -d login and go to this page, full the form and submit, the error will occur
Are you willing to submit PR?
- [ ] Yes I am willing to submit a PR!
Search before asking
- [x] I had searched in the issues and found no similar issues.
Operating system information
Linux
What happened
when i save the global configuration ,the errors occur， how can i fex it
How to reproduce
download https://raw.githubusercontent.com/OpenSPG/openspg/refs/heads/master/dev/release/docker-compose-west.yml run docker-compose -f docker-compose.yml up -d login and go to this page, full the form and submit, the error will occur
Are you willing to submit PR?
- [ ] Yes I am willing to submit a PR!
It looks like your vectorizer configuration is NULL, and should be filled with representation model service conf.
Search before asking
- [x] I had searched in the issues and found no similar issues.
Operating system information
Linux
What happened
when i save the global configuration ,the errors occur， how can i fex it
How to reproduce
download https://raw.githubusercontent.com/OpenSPG/openspg/refs/heads/master/dev/release/docker-compose-west.yml run docker-compose -f docker-compose.yml up -d login and go to this page, full the form and submit, the error will occur
Are you willing to submit PR?
- [ ] Yes I am willing to submit a PR!
It looks like your vectorizer configuration is NULL, and should be filled with representation model service conf.
my configuration, i use the maas,not local embedding model.
i tried use command knext to create knowadge, it works.
but use the web page, it not works, it has the errors.
Search before asking
- [x] I had searched in the issues and found no similar issues.
Operating system information
Linux
What happened
when i save the global configuration ,the errors occur， how can i fex it
How to reproduce
download https://raw.githubusercontent.com/OpenSPG/openspg/refs/heads/master/dev/release/docker-compose-west.yml run docker-compose -f docker-compose.yml up -d login and go to this page, full the form and submit, the error will occur
Are you willing to submit PR?
- [ ] Yes I am willing to submit a PR!
It looks like your vectorizer configuration is NULL, and should be filled with representation model service conf.
my configuration, i use the maas,not local embedding model. i tried use command knext to create knowadge, it works. but use the web page, it not works, it has the errors.
api_key filed of vector configuration should be filled with real api_key value, which can be acquired from commercial model service provider such as openai or siliconflow.
Same here, i tried using OLLAMA but it doesn't work.
It looks like your local model loaded by OLLAMA is not accessible by openspg-server container. You can refer to User Guide for Model Service Configuration to test accessibility.
maybe it is not this reason,I filled the real api_key .There are two situations
the first is: common configuration,and the errors:
the second is : model configuration,and the errors:
And I searched the database，the value is empty: select * from kg_config where config_id = 'KAG_CONFIG'
And I decompiled the source code , because of the json key graph_store is not exists ,so the errors occur.
follow the same idea, I decompiled the source code for the first error, the error occured wher check the json key "vectorizer". and the python executed errors.
maybe it is not this reason,I filled the real api_key .There are two situations the first is: common configuration,and the errors:
the second is : model configuration,and the errors:
And I searched the database，the value is empty: select * from kg_config where config_id = 'KAG_CONFIG'
And I decompiled the source code , because of the json key graph_store is not exists ,so the errors occur.
follow the same idea, I decompiled the source code for the first error, the error occured wher check the json key "vectorizer". and the python executed errors.
It looks like you have not settled common configuration yet，in which graph store should be filled.
You can refer to User Guide for Startup to get detail information of Create a knowledge base.
https://github.com/OpenSPG/openspg/pull/105/files#diff-831cd845e2cd8e24b72cce0c80930f14a6bda5310cc7b579791b35fff5051fbf
I tried update the database using :update kg_config set config = '{"graph_store":{"database":"neo4j","password":"neo4j@openspg","uri":"neo4j://release-openspg-neo4j:7687","user":"neo4j"},"vectorizer":{"type":"bge_m3","model":"BAAI/bge-m3","base_url":"https://api.siliconflow.cn/v1","api_key":"sk-gqdmllzbuybpqblewpxxxxxxxxjdkfxqcvgfyrg"},"prompt":{"biz_scene":"default","language":"zh"},"llm":{"type":"maas","creator":"openspg","createTime":"2025-01-10 09:31:49","model":"glm-4-flash","base_url":"https://open.bigmodel.cn/api/paas/v4","desc":"glm","api_key":"8253c27fxxxxxx5a798ffc809bcc0b.xxxxxx","temperature":0.7,"stream":"False","default":true}}' where config_id = 'KAG_CONFIG' and version = '1';
after that ,the web can show the configuration, but it cannnot changed ,it will occur the same error. Every time it will check "vectorizer"
1: change the common configuration:
2:add new model configuration:
3:create the knowledge:
type: bge_m3is incorrect, it should betype: openai.
thanks ,it works