[diffusion] CI: change ground truth upload path and improve publish script (#24120)

scripts/ci/utils/diffusion/publish_diffusion_gt.py

@@ -4,9 +4,12 @@ via the GitHub API (same pattern as publish_traces.py).
 """
 
 import argparse
+import hashlib
+import json
 import os
 import sys
 from pathlib import Path
+from urllib.error import HTTPError
 
 # Reuse GitHub API helpers from publish_traces.
 # Support both direct script execution and package-style imports.
@@ -19,6 +22,7 @@ if __package__:
         get_tree_sha,
         is_permission_error,
         is_rate_limit_error,
+        make_github_request,
         update_branch_ref,
         verify_token_permissions,
     )
@@ -32,6 +36,7 @@ else:
         get_tree_sha,
         is_permission_error,
         is_rate_limit_error,
+        make_github_request,
         update_branch_ref,
         verify_token_permissions,
     )
@@ -39,7 +44,7 @@ else:
 REPO_OWNER = "sglang-bot"
 REPO_NAME = "sglang-ci-data"
 BRANCH = "main"
-DEFAULT_TARGET_DIR = "diffusion-ci/consistency_gt"
+DEFAULT_TARGET_DIR = "diffusion-ci/consistency_gt/sglang_generated"
 
 IMAGE_EXTENSIONS = {".png", ".jpg", ".jpeg", ".webp"}
 
@@ -61,6 +66,38 @@ def collect_images(source_dir, target_dir):
     return files
 
 
+def git_blob_sha(content):
+    header = f"blob {len(content)}\0".encode()
+    return hashlib.sha1(header + content).hexdigest()
+
+
+def get_remote_blob_shas(repo_owner, repo_name, target_dir, token):
+    url = (
+        f"https://api.github.com/repos/{repo_owner}/{repo_name}/contents/"
+        f"{target_dir}?ref={BRANCH}"
+    )
+    try:
+        response = make_github_request(url, token)
+    except HTTPError as e:
+        if e.code == 404:
+            return {}
+        raise
+    entries = json.loads(response)
+    return {
+        item["path"]: item["sha"]
+        for item in entries
+        if item.get("type") == "file" and "sha" in item
+    }
+
+
+def filter_changed_files(files, remote_blob_shas):
+    return [
+        (path, content)
+        for path, content in files
+        if remote_blob_shas.get(path) != git_blob_sha(content)
+    ]
+
+
 def publish(source_dir, target_dir=None):
     target_dir = target_dir or DEFAULT_TARGET_DIR
     token = os.getenv("GITHUB_TOKEN")
@@ -86,37 +123,48 @@ def publish(source_dir, target_dir=None):
         print("Token permission verification failed.")
         sys.exit(1)
 
-    # Create blobs
-    try:
-        tree_items = create_blobs(REPO_OWNER, REPO_NAME, files_to_upload, token)
-    except Exception as e:
-        if is_rate_limit_error(e):
-            print("Rate-limited during blob creation, skipping.")
-            return
-        if is_permission_error(e):
-            print(
-                f"ERROR: Token lacks write permission to {REPO_OWNER}/{REPO_NAME}. "
-                "Update GH_PAT_FOR_NIGHTLY_CI_DATA with a token that has contents:write."
-            )
-            sys.exit(1)
-        raise
-
     # Commit with retry (handle concurrent pushes)
     max_retries = 5
     for attempt in range(max_retries):
         try:
             branch_sha = get_branch_sha(REPO_OWNER, REPO_NAME, BRANCH, token)
             tree_sha = get_tree_sha(REPO_OWNER, REPO_NAME, branch_sha, token)
+            remote_blob_shas = get_remote_blob_shas(
+                REPO_OWNER, REPO_NAME, target_dir, token
+            )
+            changed_files = filter_changed_files(files_to_upload, remote_blob_shas)
+            if not changed_files:
+                print("No image changes to publish.")
+                return
+
+            try:
+                tree_items = create_blobs(REPO_OWNER, REPO_NAME, changed_files, token)
+            except Exception as e:
+                if is_rate_limit_error(e):
+                    print("Rate-limited during blob creation, skipping.")
+                    return
+                if is_permission_error(e):
+                    print(
+                        f"ERROR: Token lacks write permission to {REPO_OWNER}/{REPO_NAME}. "
+                        "Update GH_PAT_FOR_NIGHTLY_CI_DATA with a token that has contents:write."
+                    )
+                    sys.exit(1)
+                raise
+
             new_tree_sha = create_tree(
                 REPO_OWNER, REPO_NAME, tree_sha, tree_items, token
             )
-            commit_msg = f"diffusion-ci: update consistency_gt images ({len(files_to_upload)} files) [automated]"
+            if new_tree_sha == tree_sha:
+                print("No tree changes to publish.")
+                return
+
+            commit_msg = f"diffusion-ci: update images in {target_dir} ({len(changed_files)} files) [automated]"
             commit_sha = create_commit(
                 REPO_OWNER, REPO_NAME, new_tree_sha, branch_sha, commit_msg, token
             )
             update_branch_ref(REPO_OWNER, REPO_NAME, BRANCH, commit_sha, token)
             print(
-                f"Successfully pushed {len(files_to_upload)} images (commit {commit_sha[:10]})"
+                f"Successfully pushed {len(changed_files)} changed images (commit {commit_sha[:10]})"
             )
             return
         except Exception as e:
@@ -134,8 +182,6 @@ def publish(source_dir, target_dir=None):
                 elif "Object does not exist" in e.error_body:
                     retryable = True
 
-            from urllib.error import HTTPError
-
             if isinstance(e, HTTPError) and e.code in [422, 500, 502, 503, 504]:
                 retryable = True