Upload app.py

app.py

@@ -0,0 +1,81 @@
+# -*- coding: utf-8 -*-
+"""App.ipynb
+
+Automatically generated by Colab.
+
+Original file is located at
+    https://colab.research.google.com/drive/1VSsT1xPpcetosAuTtw9TavvN4tYuV3Jp
+"""
+
+import gradio as gr
+import whisper
+import os
+from pydub import AudioSegment
+
+# Load Whisper model (can be "small", "medium", "large", etc.)
+model = whisper.load_model("base")
+
+def convert_video_to_audio(video_path):
+    """Convert video file to WAV audio."""
+    audio_path = os.path.splitext(video_path)[0] + ".wav"
+    audio = AudioSegment.from_file(video_path)
+    audio.export(audio_path, format="wav")
+    return audio_path
+
+def chunk_audio(audio_path, chunk_size_mb=25, max_duration_sec=900):
+    """Chunk audio if too large or long."""
+    audio = AudioSegment.from_file(audio_path)
+    chunk_limit_bytes = chunk_size_mb * 1024 * 1024
+    max_duration_ms = min(len(audio), max_duration_sec * 1000)
+
+    if len(audio) > max_duration_ms or audio.frame_count() * audio.frame_width > chunk_limit_bytes:
+        return [audio[:max_duration_ms]]
+    return [audio]
+
+def translate_audio(file):
+    """Translate audio/video to English."""
+    if file is None:
+        return "Error: No file uploaded.", None
+
+    ext = os.path.splitext(file.name)[1].lower()
+    is_video = ext in ['.mp4', '.avi', '.mov']
+
+    audio_path = convert_video_to_audio(file.name) if is_video else file.name
+    chunks = chunk_audio(audio_path)
+
+    translations = []
+    for i, chunk in enumerate(chunks):
+        chunk_file = f"chunk_{i}.wav"
+        chunk.export(chunk_file, format="wav")
+        result = model.transcribe(chunk_file, task="translate")  # 👈 Use task='translate'
+        translations.append(result["text"])
+        os.remove(chunk_file)
+
+    if is_video:
+        os.remove(audio_path)
+
+    full_translation = " ".join(translations)
+    output_txt = os.path.splitext(os.path.basename(file.name))[0] + "_translated.txt"
+
+    with open(output_txt, "w", encoding="utf-8") as f:
+        f.write(full_translation)
+
+    return full_translation, output_txt
+
+# Gradio Interface
+iface = gr.Interface(
+    fn=translate_audio,
+    inputs=gr.File(label="Upload Audio or Video File (Any Language)"),
+    outputs=[
+        gr.Textbox(label="English Translation"),
+        gr.File(label="Download Translation (.txt)")
+    ],
+    title="AutoTranslate AI",
+    description=(
+        "Upload an audio or video file in any language. "
+        "The tool will translate its spoken content to English using OpenAI Whisper. "
+        "Files over 25MB or longer than 15 minutes will be chunked automatically."
+    )
+)
+
+iface.launch(share=True)