Merge pull request #78 from Okabe-Junya/junya/feat/add-ignore-users

feat: Filter Out CI User Responses from Time to First Response

.env-example

@@ -1,3 +1,4 @@
 GH_TOKEN = " "
 SEARCH_QUERY = "repo:owner/repo is:open is:issue"
 LABELS_TO_MEASURE = "waiting-for-review,waiting-for-manager"
+IGNORE_USERS = "user1,user2"

README.md

@@ -52,6 +52,7 @@ Below are the allowed configuration options:
 | `HIDE_TIME_TO_CLOSE` | False |         | If set to any value, the time to close will not be displayed in the generated markdown file. |
 | `HIDE_TIME_TO_ANSWER` | False |         | If set to any value, the time to answer a discussion will not be displayed in the generated markdown file. |
 | `HIDE_LABEL_METRICS` | False |         | If set to any value, the time in label metrics will not be displayed in the generated markdown file. |
+| `IGNORE_USERS` | False |         | A comma separated list of users to ignore when calculating metrics. (ie. `IGNORE_USERS: 'user1,user2'`) |
 
 ### Example workflows
 
@@ -425,4 +426,4 @@ jobs:
 
 ## License
 
-[MIT](LICENSE)
+[MIT](LICENSE)

issue_metrics.py

@@ -13,7 +13,7 @@
         Searches for issues in a GitHub repository that match the given search query.
     auth_to_github() -> github3.GitHub: Connect to GitHub API with token authentication.
     get_per_issue_metrics(issues: Union[List[dict], List[github3.issues.Issue]],
-        discussions: bool = False) -> tuple[List, int, int]:
+        discussions: bool = False), labels: Union[List[str], None] = None, ignore_users: List[str] = [] -> tuple[List, int, int]:
         Calculate the metrics for each issue in a list of GitHub issues.
     get_owner(search_query: str) -> Union[str, None]]:
         Get the owner from the search query.
@@ -41,13 +41,14 @@
 )
 
 
-def get_env_vars() -> tuple[str, str]:
+def get_env_vars() -> tuple[str, str, List[str]]:
     """
     Get the environment variables for use in the script.
 
     Returns:
         str: the search query used to filter issues, prs, and discussions
         str: the github token used to authenticate to github.com
+        List[str]: a list of users to ignore when calculating metrics
     """
     search_query = os.getenv("SEARCH_QUERY")
     if not search_query:
@@ -57,7 +58,13 @@ def get_env_vars() -> tuple[str, str]:
     if not token:
         raise ValueError("GITHUB_TOKEN environment variable not set")
 
-    return search_query, token
+    ignore_users = os.getenv("IGNORE_USERS")
+    if ignore_users:
+        ignore_users = ignore_users.split(",")
+    else:
+        ignore_users = []
+
+    return search_query, token, ignore_users
 
 
 def search_issues(
@@ -125,6 +132,7 @@ def get_per_issue_metrics(
     issues: Union[List[dict], List[github3.search.IssueSearchResult]],  # type: ignore
     discussions: bool = False,
     labels: Union[List[str], None] = None,
+    ignore_users: List[str] = [],
 ) -> tuple[List, int, int]:
     """
     Calculate the metrics for each issue/pr/discussion in a list provided.
@@ -135,6 +143,7 @@ def get_per_issue_metrics(
         discussions (bool, optional): Whether the issues are discussions or not.
             Defaults to False.
         labels (List[str]): A list of labels to measure time spent in. Defaults to empty list.
+        ignore_users (List[str]): A list of users to ignore when calculating metrics.
 
     Returns:
         tuple[List[IssueWithMetrics], int, int]: A tuple containing a
@@ -157,7 +166,7 @@ def get_per_issue_metrics(
                 None,
             )
             issue_with_metrics.time_to_first_response = measure_time_to_first_response(
-                None, issue
+                None, issue, ignore_users
             )
             issue_with_metrics.time_to_answer = measure_time_to_answer(issue)
             if issue["closedAt"]:
@@ -175,7 +184,7 @@ def get_per_issue_metrics(
                 None,
             )
             issue_with_metrics.time_to_first_response = measure_time_to_first_response(
-                issue, None
+                issue, None, ignore_users
             )
             if labels:
                 issue_with_metrics.label_metrics = get_label_metrics(issue, labels)
@@ -238,6 +247,7 @@ def main():
     env_vars = get_env_vars()
     search_query = env_vars[0]
     token = env_vars[1]
+    ignore_users = env_vars[2]
 
     # Get the repository owner and name from the search query
     owner = get_owner(search_query)
@@ -280,6 +290,7 @@ def main():
         issues,
         discussions="type:discussions" in search_query,
         labels=labels,
+        ignore_users=ignore_users,
     )
 
     average_time_to_first_response = get_average_time_to_first_response(

test_time_to_first_response.py

@@ -61,6 +61,56 @@ def test_measure_time_to_first_response_no_comments(self):
         # Check the results
         self.assertEqual(result, expected_result)
 
+    def test_measure_time_to_first_response_ignore_users(self):
+        """Test that measure_time_to_first_response ignores comments from ignored users."""
+        # Set up the mock GitHub issues
+        mock_issue1 = MagicMock()
+        mock_issue1.comments = 1
+        mock_issue1.created_at = "2023-01-01T00:00:00Z"
+
+        # Set up the mock GitHub comments (one ignored, one not ignored)
+        mock_comment1 = MagicMock()
+        mock_comment1.user.login = "ignored_user"
+        mock_comment1.created_at = datetime.fromisoformat("2023-01-02T00:00:00Z")
+
+        mock_comment2 = MagicMock()
+        mock_comment2.user.login = "not_ignored_user"
+        mock_comment2.created_at = datetime.fromisoformat("2023-01-03T00:00:00Z")
+
+        mock_issue1.issue.comments.return_value = [mock_comment1, mock_comment2]
+
+        # Call the function
+        result = measure_time_to_first_response(mock_issue1, None, ["ignored_user"])
+        expected_result = timedelta(days=2)
+
+        # Check the results
+        self.assertEqual(result, expected_result)
+
+    def test_measure_time_to_first_response_only_ignored_users(self):
+        """Test that measure_time_to_first_response returns empty for an issue with only ignored users."""
+        # Set up the mock GitHub issues
+        mock_issue1 = MagicMock()
+        mock_issue1.comments = 1
+        mock_issue1.created_at = "2023-01-01T00:00:00Z"
+
+        # Set up the mock GitHub comments (all ignored)
+        mock_comment1 = MagicMock()
+        mock_comment1.user.login = "ignored_user"
+        mock_comment1.created_at = datetime.fromisoformat("2023-01-02T00:00:00Z")
+
+        mock_comment2 = MagicMock()
+        mock_comment2.user.login = "ignored_user2"
+        mock_comment2.created_at = datetime.fromisoformat("2023-01-03T00:00:00Z")
+
+        mock_issue1.issue.comments.return_value = [mock_comment1, mock_comment2]
+
+        # Call the function
+        result = measure_time_to_first_response(mock_issue1, None, ["ignored_user", "ignored_user2"])
+        expected_result = None
+
+        # Check the results
+        self.assertEqual(result, expected_result)
+
 
 class TestGetAverageTimeToFirstResponse(unittest.TestCase):
     """Test the get_average_time_to_first_response function."""

time_to_first_response.py

@@ -27,12 +27,14 @@
 def measure_time_to_first_response(
     issue: Union[github3.issues.Issue, None],  # type: ignore
     discussion: Union[dict, None],
+    ignore_users: List[str] = [],
 ) -> Union[timedelta, None]:
     """Measure the time to first response for a single issue or a discussion.
 
     Args:
         issue (Union[github3.issues.Issue, None]): A GitHub issue.
         discussion (Union[dict, None]): A GitHub discussion.
+        ignore_users (List[str]): A list of GitHub usernames to ignore.
 
     Returns:
         Union[timedelta, None]: The time to first response for the issue/discussion.
@@ -46,17 +48,21 @@ def measure_time_to_first_response(
     # Get the first comment time
     if issue:
         comments = issue.issue.comments(
-            number=1, sort="created", direction="asc"
+            number=20, sort="created", direction="asc"
         )  # type: ignore
         for comment in comments:
+            if comment.user.login in ignore_users:
+                continue
             first_comment_time = comment.created_at
 
         # Check if the issue is actually a pull request
         # so we may also get the first review comment time
         if issue.issue.pull_request_urls:
             pull_request = issue.issue.pull_request()
-            review_comments = pull_request.reviews(number=1)  # type: ignore
+            review_comments = pull_request.reviews(number=50)  # type: ignore
             for review_comment in review_comments:
+                if review_comment.user.login in ignore_users:
+                    continue
                 first_review_comment_time = review_comment.submitted_at
 
         # Figure out the earliest response timestamp