AnswerDotAI / AnswerDotAI/RAGatouille

Delete from index problem

Open
#101 2 comments 1 reaction 0 assignees View on GitHub
bug CRUD (experimental) help wanted
Dominant language
Python
Stars
4k
Forks
276
PR merge metrics
No merged PRs in 30d

Description

I am having problems when deleting docs from an index, when its files are not in the standard '{index_root}/colbert/indexes/{index_name}'.

My index files are in the path '{index_root}/{index_name}', and when loading the Searcher in the [delete_from_index](https://github.com/bclavie/RAGatouille/blob/864dfb7f95c0160bd0efda00db05c50fd2f7aef9/ragatouille/models/colbert.py#L239) you don't pass the 'index_root' attribute to the Searcher, which then picks the default index_root ('indexes/colbert/indexes/') as you can see [in the searcher code](https://github.com/stanford-futuredata/ColBERT/blob/8fb3abbeead17c506a323de7108603d559c061b1/colbert/searcher.py#L31)

I suppose this is a bug because in the [add_to_index](https://github.com/bclavie/RAGatouille/blob/864dfb7f95c0160bd0efda00db05c50fd2f7aef9/ragatouille/models/colbert.py#L151) you indeed pass the index_root param to the Searcher.

Thanks

Contributor guide

No contributing guide indexed for this repository

Research direction

Read ragatouille/models/colbert.py at delete_from_index and compare its Searcher construction with add_to_index around line 151. Verify how the custom index path is passed, then confirm that deleting documents from an index under {index_root}/{index_name} uses the intended files rather than the default root.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
backend, search
Issue type
Bug
Difficulty
2/5
Estimated time
1-3 hours
Activity status
Stale
Clarity
Clearly specified
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.