feat: Add PR comment statistics behind HIDE_PR_STATISTICS configuration flag

README.md

@@ -156,6 +156,7 @@ This action can be configured to authenticate with GitHub App Installation or Pe
 | `HIDE_TIME_TO_FIRST_RESPONSE` | False    | False                                      | If set to `true`, the time to first response will not be displayed in the generated Markdown file.                                                                                                                                                                                                         |
 | `HIDE_STATUS`                 | False    | True                                       | If set to `true`, the status column will not be shown                                                                                                                                                                                                                                                      |
 | `HIDE_CREATED_AT`             | False    | True                                       | If set to `true`, the creation timestamp will not be displayed in the generated Markdown file.                                                                                                                                                                                                             |
+| `HIDE_PR_STATISTICS`          | False    | True                                       | If set to `true`, PR comment statistics (mean, median, 90th percentile, and individual PR comment counts) will not be displayed in the generated Markdown file.                                                                                                                                           |
 | `DRAFT_PR_TRACKING`           | False    | False                                      | If set to `true`, draft PRs will be included in the metrics as a new column and in the summary stats.                                                                                                                                                                                                      |
 | `IGNORE_USERS`                | False    | False                                      | A comma separated list of users to ignore when calculating metrics. (ie. `IGNORE_USERS: 'user1,user2'`). To ignore bots, append `[bot]` to the user (ie. `IGNORE_USERS: 'github-actions[bot]'`) Users in this list will also have their authored issues and pull requests removed from the Markdown table. |
 | `ENABLE_MENTOR_COUNT`         | False    | False                                      | If set to 'TRUE' count number of comments users left on discussions, issues and PRs and display number of active mentors                                                                                                                                                                                   |

classes.py

@@ -25,6 +25,7 @@ class IssueWithMetrics:
         mentor_activity (dict, optional): A dictionary containing active mentors
         created_at (datetime, optional): The time the issue was created.
         status (str, optional): The status of the issue, e.g., "open", "closed as completed",
+        pr_comment_count (int, optional): The number of comments on the PR (excluding bots).
     """
 
     # pylint: disable=too-many-instance-attributes
@@ -44,6 +45,7 @@ def __init__(
         assignee=None,
         assignees=None,
         status=None,
+        pr_comment_count=None,
     ):
         self.title = title
         self.html_url = html_url
@@ -58,3 +60,4 @@ def __init__(
         self.mentor_activity = mentor_activity
         self.created_at = created_at
         self.status = status
+        self.pr_comment_count = pr_comment_count

config.py

@@ -57,6 +57,7 @@ class EnvVars:
         rate_limit_bypass (bool): If set to TRUE, bypass the rate limit for the GitHub API
         draft_pr_tracking (bool): If set to TRUE, track PR time in draft state
             in addition to other metrics
+        hide_pr_statistics (bool): If set to TRUE, hide PR comment statistics in the output
     """
 
     def __init__(
@@ -88,6 +89,7 @@ def __init__(
         output_file: str,
         rate_limit_bypass: bool = False,
         draft_pr_tracking: bool = False,
+        hide_pr_statistics: bool = True,
     ):
         self.gh_app_id = gh_app_id
         self.gh_app_installation_id = gh_app_installation_id
@@ -116,6 +118,7 @@ def __init__(
         self.output_file = output_file
         self.rate_limit_bypass = rate_limit_bypass
         self.draft_pr_tracking = draft_pr_tracking
+        self.hide_pr_statistics = hide_pr_statistics
 
     def __repr__(self):
         return (
@@ -147,6 +150,7 @@ def __repr__(self):
             f"{self.output_file}"
             f"{self.rate_limit_bypass}"
             f"{self.draft_pr_tracking}"
+            f"{self.hide_pr_statistics}"
         )
 
 
@@ -244,6 +248,7 @@ def get_env_vars(test: bool = False) -> EnvVars:
     hide_time_to_first_response = get_bool_env_var("HIDE_TIME_TO_FIRST_RESPONSE", False)
     hide_created_at = get_bool_env_var("HIDE_CREATED_AT", True)
     hide_status = get_bool_env_var("HIDE_STATUS", True)
+    hide_pr_statistics = get_bool_env_var("HIDE_PR_STATISTICS", True)
     enable_mentor_count = get_bool_env_var("ENABLE_MENTOR_COUNT", False)
     min_mentor_comments = os.getenv("MIN_MENTOR_COMMENTS", "10")
     max_comments_eval = os.getenv("MAX_COMMENTS_EVAL", "20")
@@ -278,4 +283,5 @@ def get_env_vars(test: bool = False) -> EnvVars:
         output_file,
         rate_limit_bypass,
         draft_pr_tracking,
+        hide_pr_statistics,
     )

issue_metrics.py

@@ -30,6 +30,7 @@
 from markdown_helpers import markdown_too_large_for_issue_body, split_markdown_file
 from markdown_writer import write_to_markdown
 from most_active_mentors import count_comments_per_user, get_mentor_count
+from pr_comments import count_pr_comments, get_stats_pr_comments
 from search import get_owners_and_repositories, search_issues
 from time_in_draft import get_stats_time_in_draft, measure_time_in_draft
 from time_to_answer import get_stats_time_to_answer, measure_time_to_answer
@@ -153,6 +154,12 @@ def get_per_issue_metrics(
                         f"An error occurred processing review comments. Perhaps the review contains a ghost user. {e}"
                     )
 
+            # Count PR comments if this is a pull request and statistics are not hidden
+            if pull_request and not env_vars.hide_pr_statistics:
+                issue_with_metrics.pr_comment_count = count_pr_comments(
+                    issue, pull_request, ignore_users
+                )
+
             if env_vars.hide_time_to_first_response is False:
                 issue_with_metrics.time_to_first_response = (
                     measure_time_to_first_response(
@@ -302,6 +309,7 @@ def main():  # pragma: no cover
                 average_time_to_answer=None,
                 average_time_in_draft=None,
                 average_time_in_labels=None,
+                stats_pr_comments=None,
                 num_issues_opened=None,
                 num_issues_closed=None,
                 num_mentor_count=None,
@@ -329,6 +337,7 @@ def main():  # pragma: no cover
                 average_time_to_answer=None,
                 average_time_in_draft=None,
                 average_time_in_labels=None,
+                stats_pr_comments=None,
                 num_issues_opened=None,
                 num_issues_closed=None,
                 num_mentor_count=None,
@@ -362,6 +371,7 @@ def main():  # pragma: no cover
 
     stats_time_to_answer = get_stats_time_to_answer(issues_with_metrics)
     stats_time_in_draft = get_stats_time_in_draft(issues_with_metrics)
+    stats_pr_comments = get_stats_pr_comments(issues_with_metrics)
 
     num_mentor_count = 0
     if enable_mentor_count:
@@ -379,6 +389,7 @@ def main():  # pragma: no cover
         stats_time_to_answer=stats_time_to_answer,
         stats_time_in_draft=stats_time_in_draft,
         stats_time_in_labels=stats_time_in_labels,
+        stats_pr_comments=stats_pr_comments,
         num_issues_opened=num_issues_open,
         num_issues_closed=num_issues_closed,
         num_mentor_count=num_mentor_count,
@@ -393,6 +404,7 @@ def main():  # pragma: no cover
         average_time_to_answer=stats_time_to_answer,
         average_time_in_draft=stats_time_in_draft,
         average_time_in_labels=stats_time_in_labels,
+        stats_pr_comments=stats_pr_comments,
         num_issues_opened=num_issues_open,
         num_issues_closed=num_issues_closed,
         num_mentor_count=num_mentor_count,

json_writer.py

@@ -21,7 +21,7 @@
 import json
 import os
 from datetime import timedelta
-from typing import Any, List, Union
+from typing import Any, Dict, List, Union
 
 from classes import IssueWithMetrics
 
@@ -33,6 +33,7 @@ def write_to_json(
     stats_time_to_answer: Union[dict[str, timedelta], None],
     stats_time_in_draft: Union[dict[str, timedelta], None],
     stats_time_in_labels: Union[dict[str, dict[str, timedelta]], None],
+    stats_pr_comments: Union[Dict[str, float], None],
     num_issues_opened: Union[int, None],
     num_issues_closed: Union[int, None],
     num_mentor_count: Union[int, None],
@@ -142,6 +143,15 @@ def write_to_json(
         for label, time in stats_time_in_labels["90p"].items():
             p90_time_in_labels[label] = str(time)
 
+    # PR comments statistics
+    average_pr_comments = None
+    med_pr_comments = None
+    p90_pr_comments = None
+    if stats_pr_comments is not None:
+        average_pr_comments = stats_pr_comments["avg"]
+        med_pr_comments = stats_pr_comments["med"]
+        p90_pr_comments = stats_pr_comments["90p"]
+
     # Create a dictionary with the metrics
     metrics: dict[str, Any] = {
         "average_time_to_first_response": str(average_time_to_first_response),
@@ -159,6 +169,9 @@ def write_to_json(
         "90_percentile_time_to_answer": str(p90_time_to_answer),
         "90_percentile_time_in_draft": str(p90_time_in_draft),
         "90_percentile_time_in_labels": p90_time_in_labels,
+        "average_pr_comments": average_pr_comments,
+        "median_pr_comments": med_pr_comments,
+        "90_percentile_pr_comments": p90_pr_comments,
         "num_items_opened": num_issues_opened,
         "num_items_closed": num_issues_closed,
         "num_mentor_count": num_mentor_count,
@@ -184,6 +197,7 @@ def write_to_json(
                 "time_to_answer": str(issue.time_to_answer),
                 "time_in_draft": str(issue.time_in_draft),
                 "label_metrics": formatted_label_metrics,
+                "pr_comment_count": issue.pr_comment_count,
                 "created_at": str(issue.created_at),
             }
         )

markdown_writer.py

@@ -91,6 +91,10 @@ def get_non_hidden_columns(labels) -> List[str]:
     if not hide_status:
         columns.append("Status")
 
+    hide_pr_statistics = env_vars.hide_pr_statistics
+    if not hide_pr_statistics:
+        columns.append("PR Comments")
+
     return columns
 
 
@@ -101,6 +105,7 @@ def write_to_markdown(
     average_time_to_answer: Union[dict[str, timedelta], None],
     average_time_in_draft: Union[dict[str, timedelta], None],
     average_time_in_labels: Union[dict, None],
+    stats_pr_comments: Union[dict[str, float], None],
     num_issues_opened: Union[int, None],
     num_issues_closed: Union[int, None],
     num_mentor_count: Union[int, None],
@@ -146,6 +151,7 @@ def write_to_markdown(
 
     """
     columns = get_non_hidden_columns(labels)
+    env_vars = get_env_vars()
     output_file_name = output_file if output_file else "issue_metrics.md"
     with open(output_file_name, "w", encoding="utf-8") as file:
         file.write(f"# {report_title}\n\n")
@@ -169,6 +175,7 @@ def write_to_markdown(
             average_time_to_answer,
             average_time_in_draft,
             average_time_in_labels,
+            stats_pr_comments,
             num_issues_opened,
             num_issues_closed,
             num_mentor_count,
@@ -178,6 +185,7 @@ def write_to_markdown(
             hide_label_metrics,
             hide_items_closed_count,
             enable_mentor_count,
+            env_vars.hide_pr_statistics,
         )
 
         # Write second table with individual issue/pr/discussion metrics
@@ -238,6 +246,8 @@ def write_to_markdown(
                 file.write(f" {issue.created_at} |")
             if "Status" in columns:
                 file.write(f" {issue.status} |")
+            if "PR Comments" in columns:
+                file.write(f" {issue.pr_comment_count or 'N/A'} |")
             file.write("\n")
         file.write(
             "\n_This report was generated with the \
@@ -256,6 +266,7 @@ def write_overall_metrics_tables(
     stats_time_to_answer,
     average_time_in_draft,
     stats_time_in_labels,
+    stats_pr_comments,
     num_issues_opened,
     num_issues_closed,
     num_mentor_count,
@@ -265,17 +276,23 @@ def write_overall_metrics_tables(
     hide_label_metrics,
     hide_items_closed_count=False,
     enable_mentor_count=False,
+    hide_pr_statistics=True,
 ):
     """Write the overall metrics tables to the markdown file."""
-    if any(
-        column in columns
-        for column in [
-            "Time to first response",
-            "Time to close",
-            "Time to answer",
-            "Time in draft",
-        ]
-    ) or (hide_label_metrics is False and len(labels) > 0):
+
+    if (
+        any(
+            column in columns
+            for column in [
+                "Time to first response",
+                "Time to close",
+                "Time to answer",
+                "Time in draft",
+            ]
+        )
+        or (hide_label_metrics is False and len(labels) > 0)
+        or (not hide_pr_statistics and stats_pr_comments is not None)
+    ):
         file.write("| Metric | Average | Median | 90th percentile |\n")
         file.write("| --- | --- | --- | ---: |\n")
         if "Time to first response" in columns:
@@ -330,6 +347,16 @@ def write_overall_metrics_tables(
                         f"| {stats_time_in_labels['med'][label]} "
                         f"| {stats_time_in_labels['90p'][label]} |\n"
                     )
+
+        # Add PR comment statistics if not hidden
+        if not hide_pr_statistics and stats_pr_comments is not None:
+            file.write(
+                f"| Number of comments per PR "
+                f"| {stats_pr_comments['avg']} "
+                f"| {stats_pr_comments['med']} "
+                f"| {stats_pr_comments['90p']} |\n"
+            )
+
         if "Status" in columns:  # Add logic for the 'status' column
             file.write("| Status | | | |\n")