Create main.py

main.py

@@ -0,0 +1,100 @@
+from flask import Flask, request, jsonify
+from datetime import datetime
+import whisper
+import os
+import time
+
+
+app = Flask(__name__)
+
+# Load the Whisper model
+print("Loading Whisper model...")
+model = whisper.load_model("tiny")
+print("Whisper model loaded.")
+
+def transcribe(audio_path):
+    print(f"Transcribing audio from: {audio_path}")
+
+    # Load audio and pad/trim it to fit 30 seconds
+    print("Loading and processing audio...")
+    audio = whisper.load_audio(audio_path)
+    audio = whisper.pad_or_trim(audio)
+
+    # Make log-Mel spectrogram and move to the same device as the model
+    print("Creating log-Mel spectrogram...")
+    mel = whisper.log_mel_spectrogram(audio).to(model.device)
+
+    # Detect the spoken language
+    print("Detecting language...")
+    _, probs = model.detect_language(mel)
+    language = max(probs, key=probs.get)
+    print(f"Detected language: {language}")
+
+    # Decode the audio
+    print("Decoding audio...")
+    options = whisper.DecodingOptions(fp16=False)
+    result = whisper.decode(model, mel, options)
+    
+    print("Transcription complete.")
+    return result.text, language
+
+@app.route('/transcribe', methods=['POST'])
+def transcribe_audio():
+    # Record the time when the request was received
+    request_received_time = datetime.now()
+    print(f"Received request at /transcribe at {request_received_time}")
+
+    if 'audio' not in request.files:
+        print("Error: No audio file provided")
+        return jsonify({"error": "No audio file provided"}), 400
+    
+    audio_file = request.files['audio']
+    audio_file_size = len(audio_file.read())  # Calculate the size of the file in bytes
+    audio_file.seek(0)  # Reset the file pointer after reading
+    
+    # Save the uploaded audio file
+    audio_path = os.path.join("temp_audio", audio_file.filename)
+    os.makedirs("temp_audio", exist_ok=True)
+    audio_file.save(audio_path)
+    print(f"Audio file saved to: {audio_path} (Size: {audio_file_size} bytes)")
+    
+    # Record the time before starting transcription
+    transcription_start_time = time.time()
+    
+    # Transcribe the audio
+    try:
+        transcription, language = transcribe(audio_path)
+    except Exception as e:
+        print(f"Error during transcription: {str(e)}")
+        return jsonify({"error": f"An error occurred: {str(e)}"}), 500
+    
+    # Calculate the time taken for transcription
+    transcription_end_time = time.time()
+    transcription_duration = transcription_end_time - transcription_start_time
+    
+    # Clean up the saved file
+    os.remove(audio_path)
+    print(f"Audio file removed from: {audio_path}")
+    
+    # Record the time when the response is being sent
+    response_sent_time = datetime.now()
+    
+    # Return the transcription, detected language, and timing information
+    print(f"Transcription: {transcription}, Language: {language}")
+    return jsonify({
+        "transcription": transcription,
+        "language": language,
+        "request_received_time": request_received_time.isoformat(),
+        "transcription_duration": transcription_duration,
+        "response_sent_time": response_sent_time.isoformat(),
+        "audio_file_size_bytes": audio_file_size
+    }), 200
+
+@app.route('/healthcheck', methods=['GET'])
+def healthcheck():
+    print("Received request at /healthcheck")
+    return jsonify({"status": "API is running"}), 200
+
+if __name__ == '__main__':
+    print("Starting Flask app...")
+    app.run(host="0.0.0.0", port=5000)