Better handling of PR updates

app.py

@@ -11,8 +11,11 @@ from huggingface_hub import (
     create_commit,
     create_repo,
     delete_repo,
+    get_repo_discussions,
     snapshot_download,
+    space_info,
 )
+from huggingface_hub.repocard import RepoCard
 from pydantic import BaseModel
 from requests import HTTPError
 
@@ -29,6 +32,7 @@ class WebhookPayloadRepo(BaseModel):
     type: Literal["dataset", "model", "space"]
     name: str
     private: bool
+    headSha: str
 
 
 class WebhookPayloadDiscussion(BaseModel):
@@ -65,32 +69,61 @@ async def post_webhook(
     if payload.repo.type != "space":
         raise HTTPException(400, f"Must be a Space, not {payload.repo.type}")
 
-    if not payload.event.scope.startswith("discussion"):
-        return "Not a discussion"
+    space_id = payload.repo.name
 
-    if payload.discussion is None:
-        return "Couldn't parse 'payload.discussion'"
-
-    if not payload.discussion.isPullRequest:
-        return "Not a Pull Request"
-
-    if payload.event.action == "create" or payload.event.action == "update":
+    if (
+        payload.event.scope.startswith("discussion")
+        and payload.discussion is not None
+        and payload.discussion.isPullRequest
+        and payload.event.action == "create"
+    ):
+        # New PR!
         task_queue.add_task(
             sync_ci_space,
-            space_id=payload.repo.name,
+            space_id=space_id,
             pr_num=payload.discussion.num,
             private=payload.repo.private,
         )
-    elif payload.event.action == "delete":
+    elif (
+        payload.event.scope.startswith("repo.content")
+        and payload.event.action == "update"
+    ):
+        # New repo change. Is it a commit on a PR?
+        # => loop through all PRs and check if new changes happened
+        for discussion in get_repo_discussions(
+            repo_id=space_id, repo_type="space", token=HF_TOKEN
+        ):
+            if discussion.is_pull_request and discussion.status == "open":
+                if not is_pr_synced(space_id=space_id, pr_num=discussion.num):
+                    task_queue.add_task(
+                        sync_ci_space,
+                        space_id=space_id,
+                        pr_num=discussion.num,
+                        private=payload.repo.private,
+                    )
+    elif False:
+        # PR merged!
         task_queue.add_task(
             delete_ci_space,
             space_id=payload.repo.name,
             pr_num=payload.discussion.num,
         )
-    else:
-        return f"Couldn't handle action {payload.event.action}"
 
-    return "Processed"
+    return "Processed."
+
+
+def is_pr_synced(space_id: str, pr_num: int) -> bool:
+    # What is the last synced commit for this PR?
+    ci_space_id = _get_ci_space_id(space_id=space_id, pr_num=pr_num)
+    card = RepoCard.load(repo_id_or_path=ci_space_id, repo_type="space", token=HF_TOKEN)
+    last_synced_sha = getattr(card.data, "synced_sha", None)
+
+    # What is the last commit id for this PR?
+    info = space_info(repo_id=space_id, revision=f"refs/pr/{pr_num}")
+    last_pr_sha = info.sha
+
+    # Is it up to date ?
+    return last_synced_sha == last_pr_sha
 
 
 def sync_ci_space(space_id: str, pr_num: int, private: bool) -> None:
@@ -113,25 +146,37 @@ def sync_ci_space(space_id: str, pr_num: int, private: bool) -> None:
             raise
 
     # Download space codebase from PR revision
-    snapshot_path = snapshot_download(
-        repo_id=space_id,
-        revision=f"refs/pr/{pr_num}",
-        repo_type="space",
-        token=HF_TOKEN,
+    snapshot_path = Path(
+        snapshot_download(
+            repo_id=space_id,
+            revision=f"refs/pr/{pr_num}",
+            repo_type="space",
+            token=HF_TOKEN,
+        )
     )
 
     # Sync space codebase with PR revision
     operations = [  # little aggressive but works
         CommitOperationDelete(".", is_folder=True)
     ]
-    operations += [
-        CommitOperationAdd(
-            path_in_repo=str(filepath.relative_to(snapshot_path)),
-            path_or_fileobj=filepath,
-        )
-        for filepath in Path(snapshot_path).glob("**/*")
-        if filepath.is_file()
-    ]
+    for filepath in snapshot_path.glob("**/*"):
+        if filepath.is_file():
+            path_in_repo = str(filepath.relative_to(snapshot_path))
+
+            # Upload all files without changes except for the README file
+            if path_in_repo == "README.md":
+                card = RepoCard.load(filepath)
+                setattr(card.data, "synced_sha", snapshot_path.name)  # latest sha
+                path_or_fileobj = str(card).encode()
+            else:
+                path_or_fileobj = filepath
+
+            operations.append(
+                CommitOperationAdd(
+                    path_in_repo=path_in_repo, path_or_fileobj=path_or_fileobj
+                )
+            )
+
     create_commit(
         repo_id=ci_space_id,
         repo_type="space",