Skip to content

fix(qdrant): add pandas dep, fix cross-repo ID collisions, fix comment indexing #2323

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
gvago wants to merge 2 commits into
base: main
Choose a base branch
from
+17 −4
Open

fix(qdrant): add pandas dep, fix cross-repo ID collisions, fix comment indexing #2323

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
f114d4c
fix(qdrant): add pandas dep, fix cross-repo ID collisions, fix commen…
gvago Apr 14, 2026
d1b7a85
fix: address review feedback — break long PointStruct line, clarify p…
gvago Apr 15, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 4 additions & 4 deletions pr_agent/tools/pr_similar_issue.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -445,7 +445,7 @@ def _update_index_with_issues(self, issues_list, repo_name_for_index, upsert=Fal
if len(comment_body) < 8000 or \
self.token_handler.count_tokens(comment_body) < MAX_TOKENS[MODEL]:
comment_record = Record(
id=issue_key + ".comment_" + str(j + 1),
id=issue_key + ".comment_" + str(j),
text=comment_body,
metadata=Metadata(repo=repo_name_for_index,
username=username, # use issue username for all comments
Expand Down Expand Up @@ -541,7 +541,7 @@ def _update_table_with_issues(self, issues_list, repo_name_for_index, ingest=Fal
if len(comment_body) < 8000 or \
self.token_handler.count_tokens(comment_body) < MAX_TOKENS[MODEL]:
comment_record = Record(
id=issue_key + ".comment_" + str(j + 1),
id=issue_key + ".comment_" + str(j),
text=comment_body,
metadata=Metadata(repo=repo_name_for_index,
username=username, # use issue username for all comments
Expand Down Expand Up @@ -639,7 +639,7 @@ def _update_qdrant_with_issues(self, issues_list, repo_name_for_index, ingest=Fa
if len(comment_body) < 8000 or \
self.token_handler.count_tokens(comment_body) < MAX_TOKENS[MODEL]:
comment_record = Record(
id=issue_key + ".comment_" + str(j + 1),
id=issue_key + ".comment_" + str(j),
text=comment_body,
metadata=Metadata(repo=repo_name_for_index,
username=username,
Expand Down Expand Up @@ -673,7 +673,7 @@ def _update_qdrant_with_issues(self, issues_list, repo_name_for_index, ingest=Fa
points = []
for row in df.to_dict(orient="records"):
points.append(
PointStruct(id=uuid.uuid5(uuid.NAMESPACE_DNS, row["id"]).hex, vector=row["vector"], payload={"id": row["id"], "text": row["text"], "metadata": row["metadata"]})
PointStruct(id=uuid.uuid5(uuid.NAMESPACE_DNS, f"{repo_name_for_index}:{row['id']}").hex, vector=row["vector"], payload={"id": row["id"], "text": row["text"], "metadata": row["metadata"]})
Comment thread
qodo-free-for-open-source-projects[bot] marked this conversation as resolved.
Outdated
Copy link
Copy Markdown
Contributor

@qodo-free-for-open-source-projects qodo-free-for-open-source-projects Bot Apr 14, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Action required

3. Qdrant upsert creates duplicates 🐞 Bug ≡ Correctness

Changing Qdrant PointStruct.id generation to include repo_name_for_index alters point IDs for
existing repos, so re-ingesting into an existing collection will add a second copy of each point
instead of overwriting the old ones. This can reduce result diversity (duplicates consume top_k)
because querying/parsing uses payload["id"], not the Qdrant point ID.
Agent Prompt 
### Issue description
Qdrant point IDs changed (UUID seed now includes `repo_name_for_index`), so re-ingesting an already-indexed repo into an existing collection will create duplicate points rather than overwrite the existing ones.

### Issue Context
- `_update_qdrant_with_issues()` always calls `qdrant.upsert(...)` with newly generated point IDs.
- When re-indexing (e.g., `force_update_dataset`), the code does not delete existing points for that repo first.
- Querying later reads `payload['id']`, so duplicates remain queryable and can consume the top_k.

### Fix Focus Areas
- pr_agent/tools/pr_similar_issue.py[212-257]
- pr_agent/tools/pr_similar_issue.py[587-679]

### What to change
1. Before calling `_update_qdrant_with_issues(..., ingest=True)` for an already-existing collection, delete existing points for that repo using a filter on `metadata.repo == repo_name_for_index` (and optionally `level in {issue,comment}` if needed).
2. Alternatively (or additionally), set `PointStruct.id` to a stable, human-readable, repo-scoped string like `f"{repo_name_for_index}:{row['id']}"` (Qdrant supports string point IDs) to avoid needing UUIDs and make overwrite semantics explicit.
3. Add a small log/metric indicating how many points were deleted before reindexing so operators can validate migration behavior.

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools

)
self.qdrant.upsert(collection_name=self.index_name, points=points)
get_logger().info('Done')
Expand Down
1 change: 1 addition & 0 deletions requirements.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -39,6 +39,7 @@ giteapy==1.0.8
# pinecone-datasets @ git+https://github.com/mrT23/pinecone-datasets.git@main
# lancedb==0.5.1
# qdrant-client==1.15.1
# pandas # required by qdrant indexing path
Copy link
Copy Markdown
Contributor

@qodo-free-for-open-source-projects qodo-free-for-open-source-projects Bot Apr 14, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Action required

1. requirements.txt adds commented pandas 📘 Rule violation ⚙ Maintainability

The PR introduces a new commented-out dependency line for pandas, which is inactive code and
violates the no-commented-out-code requirement. This can also cause runtime failures for users who
enable Qdrant without actually installing pandas.
Agent Prompt 
## Issue description
A new commented-out dependency line (`# pandas ...`) was added to `requirements.txt`, violating the requirement to avoid commented-out code and leaving the dependency inactive.

## Issue Context
The code path `_update_qdrant_with_issues` imports/uses `pandas`, so leaving it commented relies on manual user action and can still lead to runtime `ImportError`.

## Fix Focus Areas
- requirements.txt[42-42]

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools

# uncomment this to support language LangChainOpenAIHandler
# langchain==0.2.0
# langchain-core==0.2.28
Expand Down
Loading