adjust backend docker deps and scipy version req

README.md

@@ -47,6 +47,7 @@ A simple GUI app to synchronize recorded audio with video lip movements using th
 - In the root project directory, run
 ```bash
 docker-compose build
+docker-compose up
 ```
 
 

backend/docker/Dockerfile.backend

@@ -5,9 +5,10 @@ FROM pytorch/pytorch:1.8.1-cuda11.1-cudnn8-runtime
 WORKDIR /backend
 
 # Install system dependencies
-RUN apt-get update && apt-get install -y git wget && \
+RUN apt-get update && apt-get install -y git wget libgl1-mesa-glx && \
     rm -rf /var/lib/apt/lists/*
 
+
 # Install backend dependencies
 RUN pip install poetry
 

backend/docker/download_models.py

@@ -1,5 +1,8 @@
 import gdown
 
+base_directory = "/backend/wav2lip-hq/"
+checkpoints_directory = base_directory + "checkpoints/"
+
 urls = {
     "wav2lip_gan.pth": "10Iu05Modfti3pDbxCFPnofmfVlbkvrCm",
     "face_segmentation.pth": "154JgKpzCPW82qINcVieuPH3fZ2e0P812",
@@ -8,6 +11,6 @@
 
 for name, id in urls.items():
     url = f"https://drive.google.com/uc?id={id}"
-    output = f"checkpoints/{name}"
+    output = checkpoints_directory + name
     gdown.download(url, output, quiet=False)
     print(f"Loaded {name}")

backend/pyproject.toml

@@ -10,6 +10,8 @@ python = ">=3.8, <=3.13"
 flask = "^2.3.3"
 flask-cors = "^4.0.0"
 ffmpeg = "^1.4"
+numpy = ">=1.19.5, <1.27.0"
+scipy = "^1.7"
 
 [build-system]
 requires = ["poetry-core>=1.0.0"]

backend/services/merge_service.py

@@ -35,6 +35,6 @@ def merge_audio_video(audio_file, video_file):
         )
     except Exception as e:
         clean_files(audio_path, video_path)
-        raise Exception("Output video not found after processing. Error: " + str(e))
+        raise Exception("Error: " + str(e))
 
     return output_path

backend/services/wav2lip_inference.py

@@ -12,12 +12,14 @@ def run_command(cmd, workdir=None):
         cmd, text=True, cwd=workdir, stdout=subprocess.PIPE, stderr=subprocess.PIPE
     )
     if result.returncode != 0:
-        raise Exception(
-            f"Command {cmd[0]} failed with return code {result.returncode}."
-        )
+        error_details = f"Command '{' '.join(cmd)}' failed with return code {result.returncode}. "
+        error_details += f"STDOUT: {result.stdout} "
+        error_details += f"STDERR: {result.stderr}"
+        raise Exception(error_details)
     return result
 
 
+
 def convert_webm_to_wav(input_path, output_path):
     cmd = ["ffmpeg", "-y", "-i", input_path, output_path]
     run_command(cmd)
@@ -71,6 +73,6 @@ def run_wav2lip_inference(face_path, audio_path, outfile_path):
 
     if not os.path.exists(outfile_path):
         error_msg = result.stderr if result.stderr else result.stdout
-        raise Exception(f"Output video not found after processing. Error: {error_msg}")
+        raise Exception(f"Error: {error_msg}")
 
     return outfile_path

backend/tests/services/test_merge_service.py

@@ -63,7 +63,7 @@ def test_merge_audio_video_failure(self, mock_run, mock_remove, mock_exists):
         mock_run.return_value = MagicMock(returncode=1, stderr="Error in Wav2Lip")
 
         with self.assertRaisesRegex(
-            Exception, "Output video not found after processing"
+            Exception, "Error:"
         ):
             merge_service.merge_audio_video(
                 self.sample_audio_file, self.sample_video_file

backend/tests/services/test_wav2lip_inference.py

@@ -88,10 +88,11 @@ def test_wav2lip_inference_processing_error(self, mock_run, mock_exists):
 
         with self.assertRaises(Exception) as context:
             run_wav2lip_inference(self.face, self.audio, self.outfile)
+
+        self.assertTrue(str(context.exception).startswith(f"Command 'ffmpeg -y -i test_audio.webm test_audio.wav' failed with return code 1."))
+        self.assertTrue("Error in Wav2Lip" in str(context.exception))
+
 
-        self.assertTrue(
-            "Command ffmpeg failed with return code 1." in str(context.exception)
-        )
 
     @patch("os.path.exists", return_value=True)
     @patch(

backend/tests/test_routes.py

@@ -79,7 +79,7 @@ def test_merge_disallowed_file_extensions(self):
         self.assertEqual(response.json["error"], "File type not allowed")
 
     def test_merge_error_during_process(self):
-        error_msg_prefix = "Output video not found after processing."
+        error_msg_prefix = "Error:"
 
         with patch(
             "services.merge_service.merge_audio_video",

frontend/.env.example

@@ -0,0 +1 @@
+BACKEND_URL=http://localhost:5000

frontend/nginx/nginx.conf

@@ -7,7 +7,7 @@ server {
     }
 
     location /api/ {
-        proxy_pass http://lipsync-backend:5000;
+        proxy_pass http://backend:5000;
         proxy_set_header Host $host;
         proxy_set_header X-Real-IP $remote_addr;
         proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;