[tasks/github] Batch processing for PR reviews collection

[shlokgilda]

Description

• Adds batched processing to collect_pull_request_reviews to reduce memory usage
• Processes reviews in batches of 1000 instead of accumulating all in memory before insertion
• Combines contributor and review extraction into a single pass (was two separate loops)
• Removes redundant repo_id database lookup
• Fixes incorrect argument order in extract_needed_pr_review_data call (was storing wrong metadata values)

Notes for Reviewers

• The original code had a subtle bug: it passed tool_source, tool_version where tool_version, data_source was expected. This didn't crash because tool_source is hardcoded inside the function, but it stored "pull_request_reviews_reviews" as tool_version and "2.0" as data_source instead of the correct values.
• Memory impact: For a repo with 5000 PRs averaging 3 reviews each, old code held ~15,000 review objects in memory. New code caps at ~1000.
• Follows the same batching pattern already used in collect_pull_requests.

Testing

• Testing this code with large repos with around 20-30K PRs each. I can see the logs and lower memory usage for workers working on collect_pull_request_reviews. Will post testing related updates in comments if any.

Signed commits

• Yes, I signed my commits.

AI Disclosure: I used Claude Code to generate docstrings and write this PR draft.

[MoralCode]

Looks good! Im a little hesitant about introducing new functions for inserting things into the DB, but I can see that this was added to deduplicate the code since you need to repeat the call after the batches are processed to get any remaining items that don't fill a full batch.

[shlokgilda]

Testing update: Did not encounter any errors. Table values correctly updated. Tested with https://github.com/tensorflow/tensorflow (large repo; ~75K PRs).

[sgoggins]

@shlokgilda : I merged a PR that fixed a seperate issue on PR reviews from @MoralCode just now ... which created a conflict with your code. I think it looks pretty straightforward as both changes add safety. While I am nearly certain I know how to resolve the conflict while incorporating both changes, I would like @shlokgilda to change it in their fork/branch and update the PR ... So I don't mess something up editing online.

[sgoggins]

Just needs the conflict I introduced by merging another PR fixed.

[shlokgilda]

@sgoggins Fixed and pushed.

[sgoggins]

LGTM

[sgoggins]

@shlokgilda : We will merge this as soon as the conflict is resolved! Thanks!

[shlokgilda]

@sgoggins Merge conflict resolved.

[MoralCode]

My comments were addressed. I trust that this has been tested by the author and the code looks reasonable and well documented

[sgoggins]

LGTM