Added: Generate image to video

Dockerfile

@@ -39,4 +39,5 @@ USER flaskuser
 EXPOSE 7860
 
 # Run the Flask app using gunicorn with an infinite request timeout
-CMD ["gunicorn", "--bind", "0.0.0.0:7860", "--timeout", "0", "app:app"]
+# CMD ["gunicorn", "--bind", "0.0.0.0:7860", "--timeout", "0", "app:app"]
+CMD ["gunicorn", "--bind", "0.0.0.0:7860", "--timeout", "0", "--access-logfile", "-", "--error-logfile", "-", "app:app"]

app.py

@@ -261,8 +261,8 @@ def get_video_data(filename, subfolder, token):
                 ##################################################
 
 # Route: Image to Video
-@app.route('/image_to_video', methods=['POST'])
-def image_to_video():
+@app.route('/v1/image_to_video', methods=['POST'])
+def v1_image_to_video():
     data = request.json
 
     # Extract and validate token
@@ -273,9 +273,21 @@ def image_to_video():
 
     # Validate text prompt
     text_prompt = data.get('text_prompt')
-    if not text_prompt:
+    frame_rate = data.get('frame_rate')
+    if not text_prompt or not text_prompt.strip():
         return jsonify({'error': 'Text prompt is required'}), 400
 
+    # Check if frame_rate is missing or invalid
+    if not frame_rate:  # If frame_rate is None, empty, or 0
+        frame_rate = 24  # Default to 24 fps
+    else:
+        try:
+            frame_rate = int(frame_rate)
+            if frame_rate not in [8, 12, 24]:  # Ensure it's one of the allowed values
+                return jsonify({'error': 'Frame rate must be a valid number (8, 12, or 24).'}), 400
+        except ValueError:
+            return jsonify({'error': 'Frame rate must be a valid number (8, 12, or 24).'}), 400
+
     # Handle uploaded image or base64 image
     image_file = request.files.get('image')
     base64_image = data.get('base64_image')
@@ -322,8 +334,9 @@ def image_to_video():
 
         # Modify workflow with inputs
         workflow["30"]["inputs"]["prompt"] = text_prompt
+        workflow["44"]["inputs"]["frame_rate"] = frame_rate
         workflow["73"]["inputs"]["url"] = image_url
-        workflow["31"]["inputs"]["prompt"] = "Low quality, watermark, strange motion"
+        workflow["31"]["inputs"]["prompt"] = "Low quality, watermark, strange motion, blur"
 
         # WebSocket connection to queue the prompt
         ws = websocket.WebSocket()
@@ -345,9 +358,9 @@ def image_to_video():
         #     os.remove(image_path)
         #     print(f"Deleted temporary image: {image_path}", flush=True)
 
-# Get video route
-@app.route('/get_video/<prompt_id>', methods=['GET'])
-def get_video(prompt_id):
+# Get video_tasks route
+@app.route('/v1/video_tasks/<prompt_id>', methods=['GET'])
+def video_tasks(prompt_id):
     token = request.headers.get('Authorization')
     if not token or not token.startswith("Bearer "):
         return jsonify({'error': 'Valid Bearer token required'}), 400
@@ -408,8 +421,8 @@ def get_video(prompt_id):
         return jsonify({'error': str(e)}), 500
 
 # Route: Image to Video old
-@app.route('/v1/image_to_video', methods=['POST'])
-def v1_image_to_video():
+@app.route('/image_to_video', methods=['POST'])
+def image_to_video():
     data = request.json
 
     # Extract and validate token
@@ -420,9 +433,21 @@ def v1_image_to_video():
 
     # Validate text prompt
     text_prompt = data.get('text_prompt')
-    if not text_prompt:
+    frame_rate = data.get('frame_rate')
+    if not text_prompt or not text_prompt.strip():
         return jsonify({'error': 'Text prompt is required'}), 400
 
+    # Check if frame_rate is missing or invalid
+    if not frame_rate:  # If frame_rate is None, empty, or 0
+        frame_rate = 24  # Default to 24 fps
+    else:
+        try:
+            frame_rate = int(frame_rate)
+            if frame_rate not in [8, 12, 24]:
+                return jsonify({'error': 'Frame rate must be a valid number (8, 12, or 24).'}), 400
+        except ValueError:
+            return jsonify({'error': 'Frame rate must be a valid number (8, 12, or 24).'}), 400
+
     # Handle uploaded image or base64 image
     image_file = request.files.get('image')
     base64_image = data.get('base64_image')
@@ -469,6 +494,7 @@ def v1_image_to_video():
 
         # Modify workflow with inputs
         workflow["30"]["inputs"]["prompt"] = text_prompt
+        workflow["44"]["inputs"]["frame_rate"] = frame_rate
         workflow["73"]["inputs"]["url"] = image_url
         workflow["31"]["inputs"]["prompt"] = "Low quality, watermark, strange motion"
 
@@ -518,8 +544,10 @@ def v1_image_to_video():
         local_video_path = f"static/generated_image_to_video_{prompt_id}.mp4"
         with open(local_video_path, 'wb') as f:
             f.write(video_data)
+
         # Construct the public URL for the video
-        video_url = f"https://gosign-de-comfyui-api.hf.space/{local_video_path}"
+
+        # video_url = f"https://gosign-de-comfyui-api.hf.space/{local_video_path}"
 
         # Prepare the response with the video URL
         # response_data = {

workflows/cogvideox_image_to_video_workflow_api.json

@@ -17,7 +17,6 @@
       "clip_name": "t5\\google_t5-v1_1-xxl_encoderonly-fp8_e4m3fn.safetensors",
       "type": "sd3"
     },
-    "class_type": "CLIPLoader",
     "_meta": {
       "title": "Load CLIP"
     }
@@ -72,7 +71,7 @@
   },
   "44": {
     "inputs": {
-      "frame_rate": 12,
+      "frame_rate": 24,
       "loop_count": 0,
       "filename_prefix": "CogVideoX-I2V",
       "format": "video/h264-mp4",