From d74befc1169af04c001783f22bf028b1a8a5c9e2 Mon Sep 17 00:00:00 2001
From: Holly Gong <gongh@google.com>
Date: Thu, 15 Jun 2023 11:49:19 +1000
Subject: [PATCH] Fix pr helper

---
 .github/workflows/pr_helper.yml | 12 -----------
 infra/pr_helper.py              | 37 ++++++++++++++++++---------------
 2 files changed, 20 insertions(+), 29 deletions(-)

diff --git a/.github/workflows/pr_helper.yml b/.github/workflows/pr_helper.yml
index 65740d916d87..30eba719f776 100644
--- a/.github/workflows/pr_helper.yml
+++ b/.github/workflows/pr_helper.yml
@@ -56,15 +56,3 @@ jobs:
               repo: context.repo.repo,
               body: '${{env.MESSAGE}}'
             })
-
-      - name: Add labels for valid PR
-        if: env.IS_READY_FOR_MERGE == 'True'
-        uses: actions/github-script@v6
-        with:
-          script: |
-            github.rest.issues.addLabels({
-              issue_number: context.issue.number,
-              owner: context.repo.owner,
-              repo: context.repo.repo,
-              labels: ['Ready to merge']
-            })
\ No newline at end of file
diff --git a/infra/pr_helper.py b/infra/pr_helper.py
index 45abc749da86..37c378ab78f1 100644
--- a/infra/pr_helper.py
+++ b/infra/pr_helper.py
@@ -26,7 +26,7 @@
 OWNER = 'google'
 REPO = 'oss-fuzz'
 GITHUB_URL = 'https://github.com/'
-GITHUB_NOREF_URL = 'https://www.github.com/'  # Github URL that doesn't send emails on linked issues.
+GITHUB_NONREF_URL = f'https://www.github.com/{OWNER}/{REPO}'  # Github URL that doesn't send emails on linked issues.
 API_URL = 'https://api.github.com'
 BASE_URL = f'{API_URL}/repos/{OWNER}/{REPO}'
 BRANCH = 'master'
@@ -75,7 +75,7 @@ def main():
   pr_author = github.get_pr_author()
   # Gets all modified projects path.
   projects_path = github.get_projects_path()
-  verified, email = github.get_author_email()
+  email = github.get_author_email()
 
   for project_path in projects_path:
     project_url = f'{GITHUB_URL}/{OWNER}/{REPO}/tree/{BRANCH}/{project_path}'
@@ -103,6 +103,7 @@ def main():
         message += (
             f'{pr_author} is either the primary contact or is in the CCs list '
             f'of [{project_path}]({project_url}).<br/>')
+        continue
 
     # Checks the previous commits.
     commit_sha = github.has_author_modified_project(project_path)
@@ -115,16 +116,16 @@ def main():
       continue
 
     # If the previous commit is not associated with a pull request.
-    pr_message = (
-        f'{pr_author} has previously contributed to '
-        f'[{project_path}]({project_url}). The previous commit was '
-        f'{GITHUB_NONREF_URL}/{OWNER}/{REPO}/commit/{commit_sha}<br/>')
+    pr_message = (f'{pr_author} has previously contributed to '
+                  f'[{project_path}]({project_url}). The previous commit was '
+                  f'{GITHUB_NONREF_URL}/commit/{commit_sha}<br/>')
 
-    pr_url = github.get_pull_request_url(commit_sha)
-    if pr_url is not None:
+    previous_pr_number = github.get_pull_request_number(commit_sha)
+    if previous_pr_number is not None:
       pr_message = (f'{pr_author} has previously contributed to '
                     f'[{project_path}]({project_url}). '
-                    f'The previous PR was {pr_url}<br/>')
+                    f'The previous PR was [#{previous_pr_number}]'
+                    f'({GITHUB_NONREF_URL}/pull/{previous_pr_number})<br/>')
     message += pr_message
 
   save_env(message, is_ready_for_merge, False)
@@ -151,14 +152,16 @@ def get_projects_path(self):
     """Gets the current project path."""
     response = requests.get(f'{BASE_URL}/pulls/{self._pr_number}/files',
                             headers=self._headers)
+    if not response.ok:
+      return []
 
     projects_path = set()
     for file in response.json():
       file_path = file['filename']
-      dir_path = os.path.dirname(file_path)
-      if dir_path is not None and dir_path.split(os.sep)[0] == 'projects':
-        projects_path.add(dir_path)
-    return list(set(projects_path))
+      dir_path = file_path.split(os.sep)
+      if len(dir_path) > 1 and dir_path[0] == 'projects':
+        projects_path.add(os.sep.join(dir_path[0:2]))
+    return list(projects_path)
 
   def get_author_email(self):
     """Retrieves the author's email address for a pull request,
@@ -200,15 +203,15 @@ def get_integrated_project_info(self):
       if 'project.yaml' in file_path:
         return self.get_yaml_file_content(file['contents_url'])
 
-    return None
+    return {}
 
-  def get_pull_request_url(self, commit):
-    """Gets the pull request url."""
+  def get_pull_request_number(self, commit):
+    """Gets the pull request number."""
     pr_response = requests.get(f'{BASE_URL}/commits/{commit}/pulls',
                                headers=self._headers)
     if not pr_response.ok:
       return None
-    return pr_response.json()[0]['html_url']
+    return pr_response.json()[0]['number']
 
   def is_author_internal_member(self):
     """Returns if the author is an internal member."""