fix: improve review quality multiplier observability + harden is_valid_issue

[dataCenter430]

Summary

Two fixes merged in this PR:

1. Review quality multiplier observability

• PR number missing from penalty log — calculate_review_quality_multiplier logged penalties without the PR number, making log lines untraceable during a validator run. Added optional pr_number param (same pattern as PR fix: add PR number context to file-content fetch failure warning #711) so the call site in calculate_pr_multipliers can pass pr.number.

2. Harden is_valid_issue close-window check

The original guard was if issue.closed_at and pr.merged_at: — silently skipping issues with a missing closed_at instead of flagging them. Two changes:

• Explicit closed_at guard — issues that are CLOSED but have no closed_at timestamp now log a warning and are rejected instead of being silently accepted.
• abs() on days_diff — the sign check (days_diff < 0) was replaced with abs() so the window is symmetric and the rejection log no longer shows a signed delta.

Note: the redundant inner state_reason check that was briefly added has been removed — upstream's _is_completed_when_closed helper already enforces this earlier in the function.

Test plan

• Validator logs show (PR #N) in review quality penalty lines
• Issues with a missing closed_at timestamp are rejected with a warning
• Close-window rejection fires symmetrically (issue closed before or after merge)
• Existing review quality multiplier and is_valid_issue tests pass

Closes: #742

[anderdc]

remove the database changes

[dataCenter430]

Thanks for your review, @anderdc
Okay, I will got it right away.

[dataCenter430]

Just done and ready to review!

[dataCenter430]

Hi, @anderdc
I already updated the PR as you mentioned.
Plesae review the PR, thanks.

[anderdc]

Prior DB-changes complaint is resolved. Two new asks before merge:

1. The body claims abs() replaced the < 0 sign check on days_diff, but the actual diff at scoring.py:482 keeps if days_diff > MAX_ISSUE_CLOSE_WINDOW_DAYS or days_diff < 0: unchanged. Fix the description or actually make the change.

2. Please scope down to #742. Keep: the pr_number parameter on calculate_review_quality_multiplier and its call site in calculate_pr_multipliers. Drop: the is_valid_issue close-window changes — that's unrelated to #742 and is its own scoring-eligibility decision, file separately. If you want to also pick up #742 ask 3 (logging in calculate_issue_review_quality_multiplier), do it here; otherwise note it as a follow-up.

Replacing with cleaner scope-down ask.

[anderdc]

Prior DB-changes ask is resolved. One ask before merge: please scope this down to #742.

Keep: the pr_number parameter on calculate_review_quality_multiplier and its call site in calculate_pr_multipliers.

Drop: the is_valid_issue close-window changes — unrelated to #742 and a separate scoring-eligibility decision.

If you also want to pick up #742 ask 3 (logging in calculate_issue_review_quality_multiplier), include it here; otherwise note it as a follow-up.

[dataCenter430]

Okay, thanks for your review.
I will got it right away.