Upload 16 files

app.py

@@ -434,6 +434,14 @@ def run_v3_retry():
     except Exception as e:
         return f"Error: {e}"
 
+def run_verify_repo(repo_id):
+    cmd = ["python", "check_repo_db.py", "--repo-id", repo_id]
+    try:
+        res = subprocess.run(cmd, capture_output=True, text=True, cwd=os.path.dirname(os.path.abspath(__file__)))
+        return f"--- REPO VERIFICATION OUTPUT ---\n{res.stdout}\n{res.stderr}"
+    except Exception as e:
+        return f"Error: {e}"
+
 with gr.Blocks() as app:
     gr.Markdown("# YouTube Video Orchestrator Host")
     gr.Markdown("Deploy this to Hugging Face Spaces or run locally. It fetches pending links from your database and delegates them to FastAPI workers or Kaggle.")
@@ -452,12 +460,20 @@ with gr.Blocks() as app:
                 v3_status_btn = gr.Button("Check Status")
                 v3_retry_btn = gr.Button("Retry Failed Videos")
                 
+            gr.Markdown("---")
+            gr.Markdown("### Verify Database against any HF Repository")
+            gr.Markdown("List files in a target HF repo and mark any matching pending videos as 'done' in your tracker database.")
+            with gr.Row():
+                verify_repo_id = gr.Textbox(label="Hugging Face Repo ID", value="AdhyanshVerma/YT")
+                verify_btn = gr.Button("Verify and Update DB", variant="primary")
+                
             v3_output_log = gr.Textbox(label="Terminal Output", lines=15)
             
             v3_launch_btn.click(fn=run_v3_launch, inputs=[v3_max_workers, v3_videos], outputs=v3_output_log)
             v3_sync_btn.click(fn=run_v3_sync, inputs=[], outputs=v3_output_log)
             v3_status_btn.click(fn=run_v3_status, inputs=[], outputs=v3_output_log)
             v3_retry_btn.click(fn=run_v3_retry, inputs=[], outputs=v3_output_log)
+            verify_btn.click(fn=run_verify_repo, inputs=[verify_repo_id], outputs=v3_output_log)
 
         with gr.Tab("Database Sync"):
             gr.Markdown("Sync (Add) new API keys and YouTube Links into your database.")

check_repo_db.py

@@ -0,0 +1,114 @@
+import os
+import sqlite3
+import argparse
+from pathlib import Path
+from huggingface_hub import HfApi, login
+
+def get_db_path():
+    # Priority list of where the DB might be
+    paths = [
+        "/data/tracker.db",
+        "/home/adhyansh/datasets/yt/hf_space_host/db/tracker.db",
+        "/home/adhyansh/datasets/yt/scripts/db/tracker.db"
+    ]
+    for p in paths:
+        if os.path.exists(p):
+            return p
+    return None
+
+def load_hf_token():
+    token = os.environ.get("HF_TOKEN", "")
+    if not token:
+        p = Path.home() / ".cache" / "huggingface" / "token"
+        if p.exists():
+            token = p.read_text().strip()
+    return token
+
+def main():
+    parser = argparse.ArgumentParser(description="List files in an HF Repo and mark matching videos as 'done' in the local database.")
+    parser.add_argument("--repo-id", default="AdhyanshVerma/YT", help="Hugging Face Repository ID")
+    parser.add_argument("--db-path", default=None, help="Explicit path to the tracker.db SQLite database")
+    args = parser.parse_args()
+
+    # 1. Connect to Database
+    db_path = args.db_path or get_db_path()
+    if not db_path:
+        print("❌ Error: Could not find tracker.db in default locations. Please specify with --db-path")
+        return
+        
+    print(f"✅ Using Database: {db_path}")
+    conn = sqlite3.connect(db_path)
+    cursor = conn.cursor()
+
+    # Fetch all video_ids that are NOT done
+    try:
+        cursor.execute("SELECT video_id FROM videos WHERE status != 'done'")
+        pending_videos = [row[0] for row in cursor.fetchall()]
+        print(f"📊 Found {len(pending_videos)} videos in DB that are not marked as 'done'.")
+    except Exception as e:
+        print(f"❌ Database error: {e}")
+        return
+
+    if not pending_videos:
+        print("✅ All videos are already marked as done. Exiting.")
+        return
+
+    # 2. Authenticate with Hugging Face
+    token = load_hf_token()
+    if token:
+        login(token=token)
+        print("✅ Authenticated with Hugging Face token.")
+    else:
+        print("⚠️ Warning: No HF_TOKEN found. If the repo is private, this might fail.")
+
+    api = HfApi()
+    
+    # 3. List all files from the target repository
+    repo_types_to_try = ["dataset", "space", "model"]
+    files = []
+    
+    print(f"🔍 Searching for files in repository '{args.repo_id}'...")
+    for repo_type in repo_types_to_try:
+        try:
+            # We use recursive=True to list ALL files without downloading them
+            files = api.list_repo_files(repo_id=args.repo_id, repo_type=repo_type)
+            print(f"✅ Successfully listed {len(files)} files as a '{repo_type}' repo.")
+            break
+        except Exception as e:
+            # Continue checking other repo types
+            pass
+            
+    if not files:
+        print(f"❌ Error: Could not find or list files in repository '{args.repo_id}'. Make sure it exists and you have access.")
+        return
+        
+    # 4. Check files against database and modify DB
+    print("\n🔄 Analyzing files against the database...")
+    
+    marked_done_count = 0
+
+    for vid in pending_videos:
+        match_found = False
+        
+        # Search for the video ID anywhere in the full file paths (this catches directory names like staging/.../Jb9Jx0KkJl8/...)
+        for f in files:
+            if vid in f:
+                match_found = True
+                break
+                
+        if match_found:
+            print(f"  🟢 Match found for video: {vid} - Marking as 'done'!")
+            cursor.execute("UPDATE videos SET status='done' WHERE video_id=?", (vid,))
+            marked_done_count += 1
+            
+    # 5. Save changes
+    if marked_done_count > 0:
+        conn.commit()
+        print(f"\n🎉 Success! Updated {marked_done_count} videos to 'done' in the database.")
+    else:
+        print("\n⚠️ No new matching videos found in the repository. No database changes made.")
+        
+    conn.close()
+
+if __name__ == "__main__":
+    main()