Update 1

app.py

@@ -8,8 +8,6 @@ import docx  # To create Word documents
 from moviepy.video.tools.subtitles import SubtitlesClip
 from moviepy.editor import TextClip
 from pytube import YouTube  # For downloading YouTube videos
-from moviepy.editor import TextClip, CompositeVideoClip
-from moviepy.video.tools.subtitles import SubtitlesClip
 
 # Load Whisper model for speech-to-text (using smaller 'tiny' model for faster performance)
 asr = pipeline("automatic-speech-recognition", model="openai/whisper-tiny")
@@ -48,37 +46,36 @@ def transcribe_audio(chunk):
 
 def add_subtitle(video):
     try:
-        # The video is passed as a file path string, so we use it directly
         video_path = video if isinstance(video, str) else None
         if not video_path:
             return "No video provided!"
         
         video = mp.VideoFileClip(video_path)
         audio = video.audio
-
-        # Use a temporary file for audio extraction
-        with tempfile.NamedTemporaryFile(delete=True, suffix='.wav') as tmp_audio_file:
+        
+        # Extract audio to a temporary file
+        with tempfile.NamedTemporaryFile(delete=False, suffix='.wav') as tmp_audio_file:
             audio.write_audiofile(tmp_audio_file.name, codec='pcm_s16le')
             waveform, sr = librosa.load(tmp_audio_file.name, sr=16000)
 
-            # Transcribe in chunks (parallel)
-            chunk_duration = 15  # seconds
-            chunk_size = sr * chunk_duration
-            chunks = [waveform[i:i + chunk_size] for i in range(0, len(waveform), chunk_size) if len(waveform[i:i + chunk_size]) > 0]
+        # Transcribe in chunks (parallel)
+        chunk_duration = 15  # seconds
+        chunk_size = sr * chunk_duration
+        chunks = [waveform[i:i + chunk_size] for i in range(0, len(waveform), chunk_size)]
 
-            with ThreadPoolExecutor() as executor:
-                transcriptions = list(executor.map(transcribe_audio, chunks))
+        with ThreadPoolExecutor() as executor:
+            transcriptions = list(executor.map(transcribe_audio, chunks))
+
+        full_transcription = " ".join(transcriptions)
+        subtitle_storage["original"] = full_transcription
+        subtitle_storage["video_path"] = video_path
 
-            full_transcription = " ".join(transcriptions)
-            subtitle_storage["original"] = full_transcription  # Store the original subtitle
-            subtitle_storage["video_path"] = video_path  # Store the video path
-            
         return f"Subtitle added: {full_transcription[:100]}..."  # Display first 100 characters
 
     except Exception as e:
         return f"Error in adding subtitle: {e}"
 
-def translate_subtitle(video, language):
+def translate_subtitle(language):
     try:
         # Translate the stored subtitle
         original_subtitle = subtitle_storage.get("original")
@@ -138,20 +135,20 @@ def download_video():
         video_path = subtitle_storage.get("video_path")
         video = mp.VideoFileClip(video_path)
 
-        # Use TextClip with a basic method to avoid ImageMagick
-        generator = lambda txt: TextClip(txt, font="Arial", fontsize=24, color="white", method="caption", size=(video.w - 20, None), align="center")
+        # Generate subtitle text without ImageMagick (using TextClip with a basic font)
+        generator = lambda txt: TextClip(txt, font="Arial-Bold", fontsize=24, color="white", method='label')  # Use 'label'
 
         subs = []
-        subtitle_length = 5  # seconds for each subtitle
+        subtitle_length = 5  # seconds each subtitle will be displayed
         for i in range(0, len(translated_subtitle), 50):
             start_time = (i // 50) * subtitle_length
-            end_time = start_time + subtitle_length
+            end_time = start_time + subtitle_length  # Each subtitle lasts for 'subtitle_length' seconds
             subtitle_text = translated_subtitle[i:i + 50]
             subs.append(((start_time, end_time), subtitle_text))
 
         subtitles = SubtitlesClip(subs, generator)
 
-        subtitled_video = CompositeVideoClip([video, subtitles.set_position(('center', 'bottom'))])
+        subtitled_video = mp.CompositeVideoClip([video, subtitles.set_position(('center', 'bottom'))])
 
         output_video_path = "subtitled_video.mp4"
         subtitled_video.write_videofile(output_video_path)
@@ -161,7 +158,6 @@ def download_video():
     except Exception as e:
         return f"Error in generating subtitled video: {e}"
 
-
 # Gradio UI Interface
 with gr.Blocks() as demo:
     # Title
@@ -189,21 +185,21 @@ with gr.Blocks() as demo:
         translate_button = gr.Button("Translate Subtitle")
         translate_status = gr.Textbox(label="Translation Status")
     
-    translate_button.click(translate_subtitle, inputs=[video_input, language_dropdown], outputs=translate_status)
+    translate_button.click(translate_subtitle, inputs=language_dropdown, outputs=translate_status)  # Update inputs
     
     # Download as Word
     with gr.Row():
-        download_button = gr.Button("Download as Word")
-        download_status = gr.File(label="Download Translated Word File")  # File output for Word download
+        download_word_button = gr.Button("Download as Word")  # Renamed the button to avoid conflict
+        download_word_status = gr.File(label="Download Translated Word File")  # File output for Word download
     
-    download_button.click(download_word, inputs=None, outputs=download_status)
+    download_word_button.click(download_word, inputs=None, outputs=download_word_status)
     
     # Download Subtitled Video
     with gr.Row():
-        download_video_button = gr.Button("Download Subtitled Video")
+        download_video_button2 = gr.Button("Download Subtitled Video")  # Renamed the button to avoid conflict
         download_video_status = gr.Textbox(label="Download Video Status")
     
-    download_video_button.click(download_video, inputs=None, outputs=download_video_status)
+    download_video_button2.click(download_video, inputs=None, outputs=download_video_status)
 
 # Launch the Gradio app
 demo.launch()