search

whyhow-ai / knowledge-table

Knowledge Table is an open-source package designed to simplify extracting and exploring structured data from unstructured documents.
MIT License
196 stars 25 forks source link

change milvus config format #18

Closed zc277584121 closed 16 hours ago

zc277584121 commented 1 week ago

hello there, milvus recommends a config format like URI/TOKEN, because in this format, it can flexibly support both milvus-lite/milvus docker/zilliz cloud

here is an docuement example https://milvus.io/docs/build-rag-with-milvus.md#Load-data-into-Milvus

image

tomsmoker commented 1 week ago

Thanks @zc277584121! I'm just working on a refactor to extract dependencies (see PR here: https://github.com/whyhow-ai/knowledge-table/pull/17) so I'll bring it in once that's done.

zc277584121 commented 1 week ago

that's great, thank you

tomsmoker commented 5 days ago

@zc277584121 merged in the refactor, can you update?

zc277584121 commented 4 days ago

@tomsmoker I update this PR by merging origin main branch. The uri/token pattern looks good to me. But maybe the default name of vector_db_provider can be milvus, which represent for all of milvus-lite/milvus docker serive/zilliz cloud. I think maybe in the code side, we need not to differentiate them. Users can pass in any kind of milvus settings by themselves. So I updated them in this PR. This is just a small suggestion. If you have any inconsistent opinions from your business perspective, you can also discard this PR or get further discussion.