Update app.py

app.py

@@ -16,7 +16,6 @@ B_KEY = os.getenv("B_KEY")
 
 # URLs
 API_URL = os.getenv("API_URL")
-UPLOAD_URL = os.getenv("UPLOAD_URL")
 
 def get_voices():
     # OpenAI TTS voices
@@ -29,7 +28,7 @@ def get_voices():
         ("shimmer", "shimmer")
     ]
 
-def text_to_speech(voice, text, session_id):
+def text_to_speech(voice, text):
     url = "https://api.openai.com/v1/audio/speech"
     
     headers = {
@@ -47,30 +46,20 @@ def text_to_speech(voice, text, session_id):
     if response.status_code != 200:
         return None
     
-    # Save temporary audio file with session ID
-    audio_file_path = f'temp_voice_{session_id}.mp3'
-    with open(audio_file_path, 'wb') as audio_file:
-        audio_file.write(response.content)
-    return audio_file_path
-
-def upload_file(file_path):
-    with open(file_path, 'rb') as file:
-        files = {'fileToUpload': (os.path.basename(file_path), file)}
-        data = {'reqtype': 'fileupload'}
-        response = requests.post(UPLOAD_URL, files=files, data=data)
-    
-    if response.status_code == 200:
-        return response.text.strip()
-    return None
+    return response.content
 
-def lipsync_api_call(video_url, audio_url):
+def lipsync_api_call(video_url, audio_content):
     headers = {
         "Content-Type": "application/json",
         "x-api-key": B_KEY
     }
     
+    # Create a multipart form-data request
+    files = {
+        'audio': ('audio.mp3', audio_content, 'audio/mpeg')
+    }
+    
     data = {
-        "audioUrl": audio_url,
         "videoUrl": video_url,
         "maxCredits": 1000,
         "model": "sync-1.7.1-beta",
@@ -79,7 +68,7 @@ def lipsync_api_call(video_url, audio_url):
         "synergizerStrength": 1
     }
     
-    response = requests.post(API_URL, headers=headers, data=json.dumps(data))
+    response = requests.post(API_URL, headers=headers, data=data, files=files)
     return response.json()
 
 def check_job_status(job_id):
@@ -104,15 +93,20 @@ def get_media_duration(file_path):
     result = subprocess.run(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
     return float(result.stdout.strip())
 
-def combine_audio_video(video_path, audio_path, output_path):
+def combine_audio_video(video_path, audio_content, output_path):
+    # Save audio content to a temporary file
+    temp_audio_path = f'temp_audio_{uuid.uuid4()}.mp3'
+    with open(temp_audio_path, 'wb') as audio_file:
+        audio_file.write(audio_content)
+
     # Get durations of both video and audio
     video_duration = get_media_duration(video_path)
-    audio_duration = get_media_duration(audio_path)
+    audio_duration = get_media_duration(temp_audio_path)
 
     if video_duration > audio_duration:
         # Trim video to match the audio length
         cmd = [
-            'ffmpeg', '-i', video_path, '-i', audio_path,
+            'ffmpeg', '-i', video_path, '-i', temp_audio_path,
             '-t', str(audio_duration),  # Trim video to audio duration
             '-map', '0:v', '-map', '1:a',
             '-c:v', 'copy', '-c:a', 'aac',
@@ -122,7 +116,7 @@ def combine_audio_video(video_path, audio_path, output_path):
         # Loop video if it's shorter than audio
         loop_count = int(audio_duration // video_duration) + 1  # Calculate how many times to loop
         cmd = [
-            'ffmpeg', '-stream_loop', str(loop_count), '-i', video_path, '-i', audio_path,
+            'ffmpeg', '-stream_loop', str(loop_count), '-i', video_path, '-i', temp_audio_path,
             '-t', str(audio_duration),  # Match the duration of the final video with the audio
             '-map', '0:v', '-map', '1:a',
             '-c:v', 'copy', '-c:a', 'aac',
@@ -131,24 +125,21 @@ def combine_audio_video(video_path, audio_path, output_path):
 
     subprocess.run(cmd, check=True)
 
+    # Clean up temporary audio file
+    os.remove(temp_audio_path)
+
 def process_video(voice, video_url, text, progress=gr.Progress()):
     session_id = str(uuid.uuid4())  # Generate a unique session ID
     progress(0, desc="Generating speech...")
-    audio_path = text_to_speech(voice, text, session_id)
-    if not audio_path:
+    audio_content = text_to_speech(voice, text)
+    if not audio_content:
         return None, "Failed to generate speech audio."
     
     progress(0.2, desc="Processing video...")
     
     try:
-        progress(0.3, desc="Uploading audio...")
-        audio_url = upload_file(audio_path)
-        
-        if not audio_url:
-            raise Exception("Failed to upload audio file")
-        
         progress(0.4, desc="Initiating lipsync...")
-        job_data = lipsync_api_call(video_url, audio_url)
+        job_data = lipsync_api_call(video_url, audio_content)
         
         if "error" in job_data or "message" in job_data:
             raise Exception(job_data.get("error", job_data.get("message", "Unknown error")))
@@ -179,15 +170,13 @@ def process_video(voice, video_url, text, progress=gr.Progress()):
                 f.write(video_response.content)
             
             output_path = f"output_{session_id}.mp4"
-            combine_audio_video(video_path, audio_path, output_path)
+            combine_audio_video(video_path, audio_content, output_path)
             progress(1.0, desc="Complete!")
             return output_path, f"Used fallback method. Original error: {str(e)}"
         except Exception as fallback_error:
             return None, f"All methods failed. Error: {str(fallback_error)}"
     finally:
         # Cleanup
-        if os.path.exists(audio_path):
-            os.remove(audio_path)
         if os.path.exists(f"temp_video_{session_id}.mp4"):
             os.remove(f"temp_video_{session_id}.mp4")
 
@@ -195,7 +184,7 @@ def create_interface():
     voices = get_voices()
     
     with gr.Blocks() as app:
-        gr.Markdown("# JSON Train")
+        gr.Markdown("# Lipsync Video Generator")
         with gr.Row():
             with gr.Column():
                 voice_dropdown = gr.Dropdown(choices=[v[0] for v in voices], label="Select Voice", value=voices[0][0] if voices else None)