Create app.py

app.py

@@ -0,0 +1,79 @@
+import gradio as gr
+import assemblyai as aai
+import re
+import os
+import tempfile
+
+# Set your AssemblyAI API key
+aai.settings.api_key = os.genenv('ASSEMBLYAI_API_KEY')
+
+def create_assembly_transcript(audio_file):
+    transcriber = aai.Transcriber()
+    transcript = transcriber.transcribe(
+        audio_file.name, config=aai.TranscriptionConfig(speaker_labels=True)
+    )
+    return transcript
+
+def transcript_to_string(transcript):
+    output = ""
+    for utterance in transcript.utterances:
+        name = f"SPEAKER {utterance.speaker}"
+        start_time = format_time(utterance.start)
+        output += f"{name} {start_time}\n{utterance.text}\n\n"
+    return output
+
+def format_time(milliseconds):
+    seconds = milliseconds // 1000
+    hours, seconds = divmod(seconds, 3600)
+    minutes, seconds = divmod(seconds, 60)
+    return f"{int(hours):02}:{int(minutes):02}:{int(seconds):02}"
+
+def format_transcript(transcript_string):
+    # Regex pattern for speaker labels
+    speaker_label_pattern = r"^(.+?)(?=\s\d{2}:\d{2}:\d{2})"
+    # Regex pattern for timestamps
+    timestamp_pattern = r"(\d{2}:\d{2}:\d{2})"
+    # Replace speaker labels with bold syntax
+    formatted_transcript = re.sub(
+        speaker_label_pattern, r"**\1**", transcript_string, flags=re.MULTILINE
+    )
+    # Replace timestamps with italicized syntax
+    formatted_transcript = re.sub(
+        timestamp_pattern, r"_\1_", formatted_transcript, flags=re.MULTILINE
+    )
+    return formatted_transcript
+
+def transcribe_audio(audio_file):
+    if audio_file is None:
+        return "Please upload an audio file."
+    
+    transcript = create_assembly_transcript(audio_file)
+    
+    if transcript.error:
+        return f"An error occurred: {transcript.error}"
+    
+    transcript_string = transcript_to_string(transcript)
+    md_transcript = format_transcript(transcript_string)
+    
+    # Save the markdown transcript to a temporary file
+    with tempfile.NamedTemporaryFile(mode='w', delete=False, suffix='.md') as temp_file:
+        temp_file.write(md_transcript)
+        temp_file_path = temp_file.name
+    
+    return md_transcript, temp_file_path
+
+def launch_app():
+    iface = gr.Interface(
+        fn=transcribe_audio,
+        inputs=gr.Audio(type="filepath", label="Upload Audio File"),
+        outputs=[
+            gr.Textbox(label="Transcript Preview", lines=10),
+            gr.File(label="Download Transcript")
+        ],
+        title="Audio Transcription App",
+        description="Upload an audio file to get a transcription with speaker labels."
+    )
+    iface.launch()
+
+if __name__ == "__main__":
+    launch_app()