feat: deduplicate issues across review runs to avoid repeated feedback

[Sarthak-commits]

Implements the feature request from #231 — Gito now deduplicates issues across review runs.

Problem: When Gito reviews a PR, the user addresses some issues, and then Gito re-reviews, it reports all the same issues again — including ones already fixed.

Solution: Before posting a new review comment, Gito now:

1. Fetches previous review comments on the PR (identified by the HTML_CR_COMMENT_MARKER)
2. Extracts issue titles from the previous review (parsing ## #N <title> sections)
3. Compares against the new review and filters out any issues whose titles match
4. If all issues are duplicates, posts a "✅ No new issues found" message instead

Changes to gito/commands/gh_post_review_comment.py:

• _extract_issue_titles() — parses issue titles from markdown review body
• _filter_duplicate_issues() — orchestrates fetching previous reviews and filtering
• _remove_duplicate_sections() — removes matched issue sections from the new review

All logic is wrapped in try/except so failure to fetch previous reviews is non-fatal (falls through to posting the full review).

Closes #231

[github-actions]

I've Reviewed the Code

The logging statement incorrectly reports the total number of previously seen issue titles instead of the actual number of duplicates removed from the current review body.

⚠️ 1 issue found across 6 files

#1 Incorrect count logged for filtered duplicate issues

gito/commands/gh_post_review_comment.py L117-L123

The log message reports len(previous_titles) as the number of filtered duplicates, but this is the total number of titles found in all previous reviews, not the number of issues actually removed from the current body. For example, if previous reviews contained 10 unique titles but only 2 matched the current review, the log would incorrectly say "Filtered 10 duplicate issue(s)".
Tags: bug
Affected code:

117:         filtered_body = _remove_duplicate_sections(body, previous_titles)
118: 
119:         if filtered_body != body:
120:             logging.info(
121:                 "Filtered %d duplicate issue(s) from previous review",
122:                 len(previous_titles),
123:             )

Proposed change:

        filtered_body = _remove_duplicate_sections(body, previous_titles)

        if filtered_body != body:
            current_titles = _extract_issue_titles(body)
            filtered_titles = _extract_issue_titles(filtered_body)
            removed_count = len(current_titles) - len(filtered_titles)
            logging.info(
                "Filtered %d duplicate issue(s) from previous review",
                removed_count,
            )