Update release notes labeling (#2249)

Summary:
- fix retrieve PR script to handle commits with unrecognized/invalid PR numbers, such as in 7b6b2d00
- add modifications similar to pytorch's [#71917](https://github.com/pytorch/pytorch/pull/71917), [#72085](https://github.com/pytorch/pytorch/pull/72085)

Pull Request resolved: https://github.com/pytorch/audio/pull/2249

Reviewed By: nateanl, mthrok

Differential Revision: D34304210

Pulled By: carolineechen

fbshipit-source-id: 245784219317e355b5cece4a139dee71d65bfdd1

.github/process_commit.py

@@ -5,6 +5,8 @@ Note: we only ping the person who pulls the pr, not the reviewers, as the review
 to torchaudio with no labeling responsibility, so we don't want to bother them.
 """
 
+import json
+import os
 import sys
 from typing import Any, Optional, Set, Tuple
 
@@ -36,15 +38,18 @@ SECONDARY_LABELS = {
     "perf",
     "other",
 }
+GITHUB_TOKEN = os.environ.get("GITHUB_TOKEN")
+REQUEST_HEADERS = {"Accept": "application/vnd.github.v3+json", "Authorization": f"token {GITHUB_TOKEN}"}
+TORCHAUDIO_REPO = "https://api.github.com/repos/pytorch/audio"
 
 
-def query_torchaudio(cmd: str, *, accept) -> Any:
-    response = requests.get(f"https://api.github.com/repos/pytorch/audio/{cmd}", headers=dict(Accept=accept))
+def query_torchaudio(cmd: str) -> Any:
+    response = requests.get(f"{TORCHAUDIO_REPO}/{cmd}", headers=REQUEST_HEADERS)
     return response.json()
 
 
 def get_pr_merger_and_number(commit_hash: str) -> Optional[str]:
-    data = query_torchaudio(f"commits/{commit_hash}", accept="application/vnd.github.v3+json")
+    data = query_torchaudio(f"commits/{commit_hash}")
     commit_message = data["commit"]["message"]
 
     pulled_by = commit_message.split("Pulled By: ")
@@ -57,11 +62,25 @@ def get_pr_merger_and_number(commit_hash: str) -> Optional[str]:
 
 
 def get_labels(pr_number: int) -> Set[str]:
-    data = query_torchaudio(f"pulls/{pr_number}", accept="application/vnd.github.v3+json")
+    data = query_torchaudio(f"pulls/{pr_number}")
     labels = {label["name"] for label in data["labels"]}
     return labels
 
 
+def post_github_comment(pr_number: int, merger: str) -> Any:
+    message = {
+        "body": f"Hey @{merger}."
+        + """
+You merged this PR, but labels were not properly added. Please add a primary and secondary label \
+(See https://github.com/pytorch/audio/blob/main/.github/process_commit.py)"""
+    }
+
+    response = requests.post(
+        f"{TORCHAUDIO_REPO}/issues/{pr_number}/comments", json.dumps(message), headers=REQUEST_HEADERS
+    )
+    return response.json()
+
+
 if __name__ == "__main__":
     commit_hash = sys.argv[1]
 
@@ -71,4 +90,4 @@ if __name__ == "__main__":
         is_properly_labeled = bool(PRIMARY_LABELS.intersection(labels) and SECONDARY_LABELS.intersection(labels))
 
         if not is_properly_labeled:
-            print(f"@{merger}")
+            post_github_comment(pr_number, merger)

.github/workflows/pr-labels.yml

 name: pr-labels
 
 on:
-  pull_request:
-    types:
-      - closed
+  push:
+    branches:
+      - main
 
 jobs:
   is-properly-labeled:
@@ -21,15 +21,11 @@ jobs:
 
       - name: Process commit and find merger responsible for labeling
         id: commit
-        run: echo "::set-output name=merger::$(python .github/process_commit.py ${{ github.sha }})"
-
-      - name: Ping merger responsible for labeling if necessary
-        if: ${{ steps.commit.outputs.merger != '' }}
-        uses: mshick/add-pr-comment@v1
         env:
+          SHA1: ${{ github.sha }}
           GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-        with:
-          message: |
-            Hey ${{ steps.commit.outputs.merger }}!
-            You merged this PR, but labels were not properly added. Please add a primary and secondary label
-            (See https://github.com/pytorch/audio/blob/main/.github/process_commit.py)
+        run: python .github/process_commit.py "${SHA1}"
+
+concurrency:
+  group: pr-labels-${{ github.event.pull_request.number || github.sha }}-${{ github.event_name == 'workflow_dispatch' }}
+  cancel-in-progress: true

tools/release_notes/retrieve_prs.py

@@ -81,7 +81,11 @@ def gh_labels(pr_number):
     }}
     """
     query = run_query(query)
-    edges = query["data"]["repository"]["pullRequest"]["labels"]["edges"]
+    pr = query["data"]["repository"]["pullRequest"]
+    if not pr:
+        # to account for unrecognized PR numbers from commits originating from fb internal
+        return []
+    edges = pr["labels"]["edges"]
     return [edge["node"]["name"] for edge in edges]