Refresh user photos from github profiles (boostorg#1853) (boostorg#1855)

Author: daveoconnor

config/celery.py

@@ -63,3 +63,9 @@ def setup_periodic_tasks(sender, **kwargs):
         crontab(day_of_week="sat", hour=20, minute=3),
         app.signature("libraries.tasks.release_tasks", generate_report=True),
     )
+
+    # Update users' profile photos from GitHub. Executes daily at 3:30 AM.
+    sender.add_periodic_task(
+        crontab(hour=3, minute=30),
+        app.signature("users.tasks.refresh_users_github_photos"),
+    )

core/githubhelper.py

@@ -510,7 +510,11 @@ def get_tree(self, repo_slug: str = None, tree_sha: str = None) -> dict:
 
     def get_user_by_username(self, username: str) -> dict:
         """Return the response from GitHub's /users/{username}/"""
-        return self.api.users.get_by_username(username=username)
+        try:
+            return self.api.users.get_by_username(username=username)
+        except HTTP404NotFoundError:
+            logger.warning(f"{username=} not found on GitHub.")
+        return None
 
     def get_artifacts(self, owner="", repo_slug="", name=None):
         """Return a list of artifacts from the GH api.

users/tasks.py

@@ -1,3 +1,5 @@
+from time import sleep
+
 import structlog
 
 from django.contrib.auth import get_user_model
@@ -34,11 +36,35 @@ def update_user_github_photo(user_pk):
 
     client = GithubAPIClient()
     response = client.get_user_by_username(user.github_username)
+    # response can be None if the user does not exist on GitHub
+    if not response:
+        return
     avatar_url = response["avatar_url"]
-    user.save_image_from_github(avatar_url)
+    user.save_image_from_provider(avatar_url)
     logger.info("users_tasks_update_gh_photo_finished", user_pk=user_pk)
 
 
+@app.task
+def refresh_users_github_photos():
+    """
+    Refreshes the GitHub photos for all users who have a GitHub username.
+    This is intended to be run periodically to ensure user photos are up-to-date.
+    """
+    users = User.objects.exclude(github_username="")
+    for user in users:
+        try:
+            logger.info(f"updating {user.pk=}")
+            update_user_github_photo.delay(user.pk)
+            # not strictly necessary, but helps to avoid hammering the GitHub API
+            sleep(0.5)
+        except UserMissingGithubUsername:
+            logger.warning(
+                "users_tasks_refresh_gh_photos_no_github_username",
+                user_pk=user.pk,
+                github_username=user.github_username,
+            )
+
+
 # OAuth2 Tasks