Update app.py

app.py

@@ -72,11 +72,95 @@ optimize_pipeline_(pipe,
 )
 print("All models loaded and optimized. Gradio app is ready.")
 
-HF_MODEL = os.environ.get("HF_UPLOAD_REPO", "rahul7star/VideoExplain")
+
 from huggingface_hub import HfApi, upload_file
 import os
 import uuid
 
+import subprocess
+import tempfile
+import logging
+import shutil
+import os
+from huggingface_hub import HfApi, upload_file
+from datetime import datetime
+import uuid
+
+HF_MODEL = os.environ.get("HF_UPLOAD_REPO", "rahul7star/VideoExplain")
+
+def upscale_and_upload_4k(input_video_path: str, summary_text: str) -> str:
+    """
+    Upscale a video to 4K and upload it to Hugging Face Hub without replacing the original file.
+
+    Args:
+        input_video_path (str): Path to the original video.
+        summary_text (str): Text summary to upload alongside the video.
+
+    Returns:
+        str: Hugging Face folder path where the video and summary were uploaded.
+    """
+    logging.info(f"Upscaling video to 4K for upload: {input_video_path}")
+
+    # Create a temporary file for the upscaled video
+    with tempfile.NamedTemporaryFile(suffix=".mp4", delete=False) as tmp_upscaled:
+        upscaled_path = tmp_upscaled.name
+
+    # FFmpeg upscale command
+    cmd = [
+        "ffmpeg",
+        "-i", input_video_path,
+        "-vf", "scale=3840:2160:flags=lanczos",
+        "-c:v", "libx264",
+        "-crf", "18",
+        "-preset", "slow",
+        "-y",
+        upscaled_path,
+    ]
+    try:
+        subprocess.run(cmd, check=True, capture_output=True)
+        logging.info(f"✅ Upscaled video created at: {upscaled_path}")
+    except subprocess.CalledProcessError as e:
+        logging.error(f"FFmpeg failed:\n{e.stderr.decode()}")
+        raise
+
+    # Create a date-based folder on HF
+    today_str = datetime.now().strftime("%Y-%m-%d")
+    unique_subfolder = f"Upload-4K-{uuid.uuid4().hex[:8]}"
+    hf_folder = f"{today_str}/{unique_subfolder}"
+
+    # Upload video
+    video_filename = os.path.basename(input_video_path)
+    video_hf_path = f"{hf_folder}/{video_filename}"
+    upload_file(
+        path_or_fileobj=upscaled_path,
+        path_in_repo=video_hf_path,
+        repo_id=HF_MODEL,
+        repo_type="model",
+        token=os.environ.get("HUGGINGFACE_HUB_TOKEN"),
+    )
+    logging.info(f"✅ Uploaded 4K video to HF: {video_hf_path}")
+
+    # Upload summary.txt
+    summary_file = tempfile.NamedTemporaryFile(delete=False, suffix=".txt").name
+    with open(summary_file, "w", encoding="utf-8") as f:
+        f.write(summary_text)
+
+    summary_hf_path = f"{hf_folder}/summary.txt"
+    upload_file(
+        path_or_fileobj=summary_file,
+        path_in_repo=summary_hf_path,
+        repo_id=HF_MODEL,
+        repo_type="model",
+        token=os.environ.get("HUGGINGFACE_HUB_TOKEN"),
+    )
+    logging.info(f"✅ Uploaded summary to HF: {summary_hf_path}")
+
+    # Cleanup temporary files
+    os.remove(upscaled_path)
+    os.remove(summary_file)
+
+    return hf_folder
+
 def upload_to_hf(video_path, summary_text):
     api = HfApi()
     unique_folder = f"WANI2V-FFLF-Vvideo_{uuid.uuid4().hex[:8]}"
@@ -228,7 +312,7 @@ def generate_video(
     export_to_video(output_frames_list, video_path, fps=FIXED_FPS)
 
     progress(1.0, desc="Done!")
-    hf_folder = upload_to_hf(video_path, prompt)
+    hf_folder = upscale_and_upload_4k(video_path,prompt)
     return video_path, current_seed