asr_trasncription

.gitignore

@@ -0,0 +1,3 @@
+__pycache__/
+.venv/
+*.pyc

app.py

@@ -1,7 +1,24 @@
 import gradio as gr
+from audio_processing import process_audio, print_results
+def transcribe_audio(audio_file):
+    language_segments, text_segments = process_audio(audio_file)
+    
+    output = "Detected language changes:\n\n"
+    for lang_segment in language_segments:
+        output += f"Language: {lang_segment['language']}\n"
+        output += f"Time: {lang_segment['start']:.2f}s - {lang_segment['end']:.2f}s\n\n"
 
-def greet(name):
-    return "Hello " + name + "!!"
+    output += "Transcription:\n\n"
+    for segment in text_segments:
+        output += f"[{segment['start']:.2f}s - {segment['end']:.2f}s] {segment['text']}\n"
 
-demo = gr.Interface(fn=greet, inputs="text", outputs="text")
-demo.launch()
+    return output
+
+iface = gr.Interface(
+    fn=transcribe_audio,
+    inputs=gr.Audio(type="filepath"),
+    outputs="text",
+    title="WhisperX Audio Transcription"
+)
+
+iface.launch()

flagged/log.csv

@@ -0,0 +1,2 @@
+audio_file,output,flag,username,timestamp
+flagged/audio_file/e6332957799e9944ecd7/audio.wav,,,,2024-08-31 11:46:39.944110

requirements.txt

@@ -1,9 +1,14 @@
-torch
-torchaudio
 transformers
 git+https://github.com/m-bain/whisperx.git
 numpy
 pandas
 pyannote.audio
 pyperclip
-sentencepiece
+sentencepiece
+gradio
+speechbrain
+torch>=2
+torchaudio>=2
+faster-whisper==1.0.0
+setuptools>=65
+nltk