From 4be8d50dbe1f3ee866697017737a37db102169af Mon Sep 17 00:00:00 2001
From: Tom Dugovic <tom.dugovic@gmail.com>
Date: Tue, 25 Jun 2024 12:13:32 -0500
Subject: [PATCH] Extract CLA check into standalone Python script

---
 .github/scripts/check-cla.py    | 72 +++++++++++++++++++++++++++++++++
 .github/workflows/check-cla.yml | 42 ++++++++-----------
 2 files changed, 89 insertions(+), 25 deletions(-)
 create mode 100644 .github/scripts/check-cla.py

diff --git a/.github/scripts/check-cla.py b/.github/scripts/check-cla.py
new file mode 100644
index 000000000000..391a4218151e
--- /dev/null
+++ b/.github/scripts/check-cla.py
@@ -0,0 +1,72 @@
+import json
+import sys
+import urllib.request
+
+JSON_HEADER = {'Content-Type': 'application/json'}
+TEXT_ENCODING = 'utf-8'
+
+URL_GITHUB = 'https://api.github.com/repos'
+URL_JUCE_CLA_CHECK = 'https://cla.juce.com/check'
+
+ALLOWED_AUTHORS = set(
+    'web-flow'
+)
+
+
+class NoAuthorsException(Exception):
+    __MESSAGE = "No author or committer user IDs contained within commit information."
+    
+    def __init__(self, commits: list):
+        message = f'{self.__MESSAGE}\n\n{commits}'
+        super().__init__(message)
+
+
+class UnsignedAuthorsException(Exception):
+    __MESSAGE_0 = "The following GitHub users need to sign the JUCE CLA:",
+    __MESSAGE_1 = "Please go to https://cla.juce.com to sign the JUCE Contributor Licence Agreement."
+    
+    def __init__(self, authors: set[str]):
+        authors_str = ', '.join(authors)
+        message = f'{self.__MESSAGE_0} {authors_str}\n\n{self.__MESSAGE_1}'
+        super().__init__(message)
+
+
+def _json_request(url: str, data: dict | None = None) -> dict:
+    if data is not None:
+        data = json.dumps(data).encode(TEXT_ENCODING)
+    request = urllib.request.Request(url, data, JSON_HEADER, JSON_HEADER)
+    with urllib.request.urlopen(request) as response:
+        raw_response = response.read().decode(TEXT_ENCODING)
+    return json.loads(raw_response)
+
+
+def _get_authors(github_repository: str, github_event_number: int) -> set[str]:
+    pr_commit_list = _json_request(f'{URL_GITHUB}/{github_repository}/pulls/{github_event_number}/commits')
+    authors = set()
+    for commit in pr_commit_list:
+        for author_type in ['author', 'committer']:
+            if author_type in commit:
+                authors.add(commit[author_type]['login'])
+    return authors - ALLOWED_AUTHORS
+
+
+def _verify_authors(authors: set[str]) -> None:
+    if not authors:
+        raise NoAuthorsException()
+    verification = _json_request(URL_JUCE_CLA_CHECK, {'logins': list(authors)})
+    if verification['unsigned']:
+        raise UnsignedAuthorsException(verification['unsigned'])
+
+
+def main(github_repository: str, github_event_number: int):
+    authors = _get_authors(github_repository, github_event_number)
+    try:
+        _verify_authors(authors)
+    except Exception as e:
+        print(e.message)
+
+
+if __name__ == '__main__':
+    github_repository = sys.argv[1]
+    github_event_number = int(sys.argv[2])
+    main(github_repository, github_event_number)
\ No newline at end of file
diff --git a/.github/workflows/check-cla.yml b/.github/workflows/check-cla.yml
index 70d082413d39..dcb4b89e8eaa 100644
--- a/.github/workflows/check-cla.yml
+++ b/.github/workflows/check-cla.yml
@@ -1,32 +1,24 @@
-name: check-CLA
-on: [pull_request_target]
+name: Check CLA
+
+on:
+  pull_request:
+
 jobs:
   check-cla:
     runs-on: ubuntu-latest
+    continue-on-error: ${{ github.repository_owner != 'juce' }}
     env:
       PR_NUMBER: ${{ github.event.number }}
     steps:
+      - name: Checkout Scripts
+        uses: actions/checkout@v4
+        with:
+          sparse-checkout: |
+            .github/scripts
+          sparse-checkout-cone-mode: false
+
+      - name: LS
+        run: ls -la
+
       - name: check-CLA
-        run: |
-          import urllib.request
-          import json
-          import sys
-          def jsonRequest(url, data={}):
-              req = urllib.request.Request(url,
-                                           headers={'Content-Type': 'application/json'},
-                                           data=json.dumps(data).encode('utf-8') if data else None)
-              with urllib.request.urlopen(req) as response:
-                  return json.loads(response.read().decode('utf-8'))
-          prCommits = jsonRequest('https://api.github.com/repos/juce-framework/JUCE/pulls/${{ github.event.number }}/commits')
-          allAuthors = [commit[authorType]['login'] for authorType in ['author', 'committer'] for commit in prCommits if commit[authorType]]
-          uniqueAuthors = [name for name in list(set(allAuthors)) if name != 'web-flow']
-          if (len(uniqueAuthors) == 0):
-              print(f'\nNo author or committer user IDs contained within commit information\n\n{prCommits}\n')
-              sys.exit(1)
-          print(f'Authors: {uniqueAuthors}')
-          claResult = jsonRequest('https://cla.juce.com/check', {'logins': uniqueAuthors})
-          unsignedLogins = claResult['unsigned']
-          if (len(unsignedLogins) != 0):
-              print(f'\nThe following GitHub users need to sign the JUCE CLA: {", ".join(unsignedLogins)}\n\nPlease go to https://cla.juce.com to sign the JUCE Contributor Licence Agreement\n')
-              sys.exit(1)
-        shell: python
+        run: python .github/scripts/check-cla.py ${{github.repository}} ${{ github.event.number }}