v2

Dockerfile

@@ -1,19 +1,46 @@
-FROM python:3.10-slim
+# Base image
+FROM ubuntu:22.04
 
-# Install dependencies
+# Install system dependencies
 RUN apt-get update && apt-get install -y \
-    build-essential \
+    python3 \
+    python3-pip \
+    git \
+    cmake \
+    make \
+    g++ \
     ffmpeg \
     curl \
- && rm -rf /var/lib/apt/lists/*
+    && rm -rf /var/lib/apt/lists/*
 
+# Set working directory
 WORKDIR /app
 
+# Clone whisper.cpp and compile
+RUN git clone https://github.com/ggerganov/whisper.cpp.git && \
+    cd whisper.cpp && \
+    make
+
+# Download the small.en model
+RUN cd whisper.cpp/models && \
+    curl -L "https://huggingface.co/ggerganov/whisper.cpp/resolve/main/ggml-small.en.bin" --output ggml-small.en.bin
+
+# Install Python dependencies
 COPY requirements.txt .
 RUN pip install --no-cache-dir -r requirements.txt
 
-# Copy app and model
-COPY app ./app
+# Copy API code
+COPY app.py .
 
+# Create uploads directory
+RUN mkdir -p uploads
+
+# Environment variables (adjust paths if needed)
+ENV WHISPER_CPP_PATH="/app/whisper.cpp/main"
+ENV MODEL_PATH="/app/whisper.cpp/models/ggml-small.en.bin"
+
+# Expose Flask port
 EXPOSE 4002
-CMD ["uvicorn", "app.main:app", "--host", "0.0.0.0", "--port", "4002"]
+
+# Run the API
+CMD ["python3", "app.py"]

app/main.py

@@ -1,19 +1,59 @@
-from fastapi import FastAPI, File, UploadFile
-import shutil
+from flask import Flask, request, jsonify
 import os
-from whispercpp import Whisper
+import subprocess
+from werkzeug.utils import secure_filename
 
-app = FastAPI()
+app = Flask(__name__)
 
-# Load the local model without Hugging Face token
-whisper = Whisper.from_pretrained("./app/model/ggml-tiny.en.bin")
+# Configure upload folder (adjust as needed)
+UPLOAD_FOLDER = "./uploads"
+os.makedirs(UPLOAD_FOLDER, exist_ok=True)
+app.config['UPLOAD_FOLDER'] = UPLOAD_FOLDER
 
-@app.post("/transcribe")
-async def transcribe_audio(audio: UploadFile = File(...)):
-    temp_file = f"temp_{audio.filename}"
-    with open(temp_file, "wb") as buffer:
-        shutil.copyfileobj(audio.file, buffer)
+# Path to whisper.cpp executable & model (update paths as needed)
+WHISPER_CPP_PATH = "./whisper.cpp/main"  # Path to compiled whisper.cpp binary
+MODEL_PATH = "./whisper.cpp/models/ggml-small.en.bin"  # Path to model
 
-    result = whisper.transcribe(temp_file)
-    os.remove(temp_file)
-    return {"text": result}
+@app.route('/transcribe', methods=['POST'])
+def transcribe_audio():
+    if 'audio' not in request.files:
+        return jsonify({"error": "No audio file provided"}), 400
+
+    audio_file = request.files['audio']
+    if audio_file.filename == '':
+        return jsonify({"error": "Empty filename"}), 400
+
+    # Save the uploaded file
+    filename = secure_filename(audio_file.filename)
+    filepath = os.path.join(app.config['UPLOAD_FOLDER'], filename)
+    audio_file.save(filepath)
+
+    try:
+        # Run whisper.cpp to transcribe
+        command = [
+            WHISPER_CPP_PATH,
+            "-m", MODEL_PATH,
+            "-f", filepath,
+            "--output-txt"  # Output as text
+        ]
+        result = subprocess.run(command, capture_output=True, text=True)
+
+        if result.returncode != 0:
+            return jsonify({"error": "Transcription failed", "details": result.stderr}), 500
+
+        # Read the transcription
+        output_txt = filepath + ".txt"
+        with open(output_txt, 'r') as f:
+            transcription = f.read()
+
+        # Clean up files
+        os.remove(filepath)
+        os.remove(output_txt)
+
+        return jsonify({"text": transcription.strip()})
+
+    except Exception as e:
+        return jsonify({"error": str(e)}), 500
+
+if __name__ == '__main__':
+    app.run(host='0.0.0.0', port=4002, debug=True)

requirements.txt

@@ -1,4 +1,3 @@
-fastapi
-uvicorn
-python-multipart
-whispercpp
+flask==2.3.2
+python-dotenv==1.0.0
+werkzeug==2.3.6