PR Comment Filtering Tool

Problem/Feature Description

A DevOps team at a mid-size company uses GitHub PRs extensively. Their engineering workflow involves an automated agent that periodically processes unresolved review comments. However, the agent currently re-processes every comment on every run, including ones it already addressed — leading to duplicate replies and redundant commits. The team needs a robust filtering module that determines which comments actually need attention on any given run.

You have a JSON payload of PR review comments (provided below) from the GitHub API. Build a Python module called comment_filter.py that takes a list of PR review comment objects and returns only the comments that still need to be addressed. The module should also include a shell script fetch_comments.sh that demonstrates how to retrieve and filter comments from the GitHub API using gh CLI commands.

Output Specification

Produce the following files:

1. comment_filter.py — A Python module with a function filter_unaddressed_comments(comments: list, my_login: str, last_agent_commit_timestamp: str | None) -> list that filters the input list and returns only unaddressed comments.

2. fetch_comments.sh — A bash script that shows the full sequence of gh CLI commands for: (a) getting the authenticated user's login, (b) finding the last agent commit timestamp, (c) fetching review comments with the appropriate API endpoint and since parameter, and (d) calling the filter. The script should be well-commented explaining each step.

3. test_filter.py — Unit tests for the filtering function covering edge cases: comments already replied to, old comments from prior runs, first run with no prior commits, and comments that still need attention.

Input Files

The following files are provided as inputs. Extract them before beginning.

=============== FILE: inputs/sample_comments.json =============== [ { "id": 1001, "body": "This function should validate the input before processing", "user": {"login": "reviewer-alice"}, "in_reply_to_id": null, "created_at": "2024-11-01T10:00:00Z", "updated_at": "2024-11-01T10:00:00Z" }, { "id": 1002, "body": "Fixed in abc123 — added input validation.", "user": {"login": "agent-bot"}, "in_reply_to_id": 1001, "created_at": "2024-11-01T11:00:00Z", "updated_at": "2024-11-01T11:00:00Z" }, { "id": 1003, "body": "The error message here is misleading, it says 'not found' but the actual issue is a permission error", "user": {"login": "reviewer-bob"}, "in_reply_to_id": null, "created_at": "2024-11-01T12:00:00Z", "updated_at": "2024-11-01T12:00:00Z" }, { "id": 1004, "body": "Consider using a constant for this magic number", "user": {"login": "reviewer-alice"}, "in_reply_to_id": null, "created_at": "2024-11-02T09:00:00Z", "updated_at": "2024-11-02T09:00:00Z" }, { "id": 1005, "body": "This reply is from another user, not the agent", "user": {"login": "reviewer-carol"}, "in_reply_to_id": 1003, "created_at": "2024-11-02T10:00:00Z", "updated_at": "2024-11-02T10:00:00Z" }, { "id": 1006, "body": "The return type annotation is wrong here", "user": {"login": "reviewer-bob"}, "in_reply_to_id": null, "created_at": "2024-10-25T08:00:00Z", "updated_at": "2024-10-25T08:00:00Z" } ]