first commit

stt/Dockerfile

@@ -0,0 +1,21 @@
+FROM nvidia/cuda:12.4.1-runtime-ubuntu22.04
+
+ENV DEBIAN_FRONTEND=noninteractive
+
+RUN apt-get update && apt-get install -y \
+    python3 \
+    python3-pip \
+    ffmpeg \
+    && rm -rf /var/lib/apt/lists/*
+
+WORKDIR /app
+
+COPY requirements.txt .
+
+RUN pip3 install --no-cache-dir -r requirements.txt
+
+COPY app.py .
+
+EXPOSE 5001
+
+CMD ["python3", "app.py"]

stt/app.py

@@ -0,0 +1,78 @@
+# -*- coding: utf-8 -*-
+import os
+import tempfile
+from flask import Flask, request, jsonify
+from faster_whisper import WhisperModel
+
+# ==========================================================
+# Configuration
+# ==========================================================
+MODEL_PATH = os.environ.get("WHISPER_MODEL", "/models/whisper")
+DEVICE = os.environ.get("WHISPER_DEVICE", "cuda")
+COMPUTE_TYPE = "float16" if DEVICE == "cuda" else "int8"
+PORT = int(os.environ.get("PORT", 5001))
+
+app = Flask(__name__)
+
+# ==========================================================
+# Load Whisper Model
+# ==========================================================
+print(f"[stt] Loading model from {MODEL_PATH} on {DEVICE} ({COMPUTE_TYPE})")
+try:
+    model = WhisperModel(MODEL_PATH, device=DEVICE, compute_type=COMPUTE_TYPE)
+    print("[stt] Whisper model loaded successfully")
+except Exception as e:
+    print(f"[stt] Fatal error loading model: {e}")
+    exit(1)
+
+# ==========================================================
+# Speech To Text Endpoint
+# ==========================================================
+@app.route("/transcribe", methods=["POST"])
+def transcribe():
+    if "audio" not in request.files:
+        return jsonify({"error": "audio file missing"}), 400
+
+    audio_file = request.files["audio"]
+    
+    # استفاده از NamedTemporaryFile با مدیریت بهتر برای محیط داکر
+    tmp_file = tempfile.NamedTemporaryFile(delete=False, suffix=".wav")
+    audio_path = tmp_file.name
+    tmp_file.close()
+
+    try:
+        audio_file.save(audio_path)
+
+        # اجرای اینفرنس با تنظیمات بهینه شده
+        segments, info = model.transcribe(
+            audio_path,
+            language="fa",
+            beam_size=10,
+            vad_filter=True,
+            vad_parameters=dict(min_silence_duration_ms=500) # بهبود تشخیص سکوت
+        )
+
+        text = " ".join([segment.text.strip() for segment in segments])
+        
+        return jsonify({
+            "text": text.strip(),
+            "language": info.language
+        })
+
+    except Exception as e:
+        print(f"[stt] Error during transcription: {e}")
+        return jsonify({"error": str(e)}), 500
+
+    finally:
+        if os.path.exists(audio_path):
+            os.remove(audio_path)
+
+# ==========================================================
+# Health Check
+# ==========================================================
+@app.route("/health", methods=["GET"])
+def health():
+    return jsonify({"status": "ok"})
+
+if __name__ == "__main__":
+    app.run(host="0.0.0.0", port=PORT)

stt/requirements.txt

@@ -0,0 +1,5 @@
+flask==3.0.3
+faster-whisper==1.0.3
+ctranslate2
+requests
+numpy