AnswerDotAI / AnswerDotAI/RAGatouille

search not working after add_to_index

Open
#266 1 comment 0 reactions 0 assignees View on GitHub
Dominant language
Python
Stars
4k
Forks
276
PR merge metrics
No merged PRs in 30d

Description

I was able to use add_to_index successfully. The new file contents appear in the collection.json and also the updated pid_docid_map.json (in the index directory). When I perform a search, the model seems to return the right index (no 55 in my case). However, it is blowing up as below:

File /opt/conda/lib/python3.11/site-packages/ragatouille/models/colbert.py:411, in ColBERT.search(self, query, index_name, k, force_fast, zero_index_ranks, doc_ids)
409 result_for_query = []
410 for id_, rank, score in zip(*result):
--> 411 document_id = self.pid_docid_map[id_]
412 result_dict = {
413 "content": self.collection[id_],
414 "score": score,
(...)
417 "passage_id": id_,
418 }
420 if self.docid_metadata_map is not None:

KeyError: 55

Is pid_docid_map getting properly read during search? Please note that the search works with the other indexes from the first document.

Contributor guide

No contributing guide indexed for this repository

Research direction

Start in models/colbert.py at the search path around line 411, then inspect how collection.json and pid_docid_map.json are loaded and updated by add_to_index. Reproduce by adding a file and searching the updated index; done means the new content can be returned without a KeyError for its passage ID.

Written by the indexing model from the issue text.

Assessment

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

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.