Updated

Dockerfile

@@ -2,7 +2,7 @@ FROM python:3.9-slim
 
 WORKDIR /app
 
-# ffmpeg install karo — pydub ke liye zaroori
+# ffmpeg — pydub ke liye zaroori
 RUN apt-get update && apt-get install -y ffmpeg && rm -rf /var/lib/apt/lists/*
 
 COPY requirements.txt .

app.py

@@ -1,49 +1,77 @@
-from flask import Flask, request, send_file
+"""
+ESP32 TTS Server — Hugging Face Space
+WAV: 44100Hz, Stereo, 16-bit, Normalized + Boosted
+"""
+
+from flask import Flask, request, Response
 import os
+import io
+import hashlib
 from gtts import gTTS
 from pydub import AudioSegment
+from pydub.effects import normalize
 
 app = Flask(__name__)
 
-BASE_DIR = os.path.dirname(os.path.abspath(__file__))
-TMP_DIR  = os.path.join(BASE_DIR, "tts_tmp")
-os.makedirs(TMP_DIR, exist_ok=True)
-
+BASE_DIR   = os.path.dirname(os.path.abspath(__file__))
+CACHE_DIR  = os.path.join(BASE_DIR, "tts_cache")
+os.makedirs(CACHE_DIR, exist_ok=True)
 
-def text_to_wav(text, lang, wav_path):
-    mp3_path = os.path.join(TMP_DIR, "temp.mp3")
 
-    # gTTS se MP3
+def text_to_wav_bytes(text, lang):
+    # Step 1: gTTS → MP3 in memory (no disk write)
+    mp3_buf = io.BytesIO()
     tts = gTTS(text, lang=lang)
-    tts.save(mp3_path)
-    print(f"[gTTS] MP3 ready")
+    tts.write_to_fp(mp3_buf)
+    mp3_buf.seek(0)
 
-    # pydub se convert — 44100Hz, Stereo, 16-bit (original jaisa exactly)
-    audio = AudioSegment.from_mp3(mp3_path)
+    # Step 2: Load + convert — 44100Hz, Stereo, 16-bit
+    audio = AudioSegment.from_file(mp3_buf, format="mp3")
     audio = audio.set_frame_rate(44100).set_channels(2).set_sample_width(2)
-    audio.export(wav_path, format="wav")
-    print(f"[WAV] 44100Hz Stereo 16-bit ready")
 
-    if os.path.exists(mp3_path):
-        os.remove(mp3_path)
+    # Step 3: Normalize — consistent volume
+    audio = normalize(audio)
+
+    # Step 4: +3dB boost — DAC output thoda weak hota hai
+    audio = audio + 3
+
+    # Step 5: WAV bytes in memory
+    wav_buf = io.BytesIO()
+    audio.export(wav_buf, format="wav")
+    wav_buf.seek(0)
+    return wav_buf.read()
+
+
+def get_wav(text, lang):
+    # Cache check — same text pe instantly serve
+    key        = hashlib.md5((text + lang).encode()).hexdigest()
+    cache_path = os.path.join(CACHE_DIR, key + ".wav")
+
+    if os.path.exists(cache_path):
+        print(f'[Cache HIT] "{text}"')
+        with open(cache_path, "rb") as f:
+            return f.read()
 
-    return wav_path
+    print(f'[Generating] "{text}" | {lang}')
+    wav_data = text_to_wav_bytes(text, lang)
+
+    with open(cache_path, "wb") as f:
+        f.write(wav_data)
+
+    return wav_data
 
 
 @app.route("/tts", methods=["GET"])
 def tts_get():
-    text = request.args.get("text", "Hello")
+    text = request.args.get("text", "Hello").replace("+", " ")
     lang = request.args.get("lang", "en")
-    print(f"\n[TTS] Text: \"{text}\" | Lang: {lang}")
-
-    wav_path = os.path.join(TMP_DIR, "response.wav")
     try:
-        text_to_wav(text, lang, wav_path)
+        wav_data = get_wav(text, lang)
     except Exception as e:
         print(f"[ERROR] {e}")
         return f"Error: {e}", 500
-
-    return send_file(wav_path, mimetype="audio/wav")
+    return Response(wav_data, mimetype="audio/wav",
+                    headers={"Content-Length": str(len(wav_data))})
 
 
 @app.route("/tts", methods=["POST"])
@@ -51,16 +79,13 @@ def tts_post():
     data = request.get_json(silent=True) or {}
     text = data.get("text", "Hello")
     lang = data.get("lang", "en")
-    print(f"\n[TTS] Text: \"{text}\" | Lang: {lang}")
-
-    wav_path = os.path.join(TMP_DIR, "response.wav")
     try:
-        text_to_wav(text, lang, wav_path)
+        wav_data = get_wav(text, lang)
     except Exception as e:
         print(f"[ERROR] {e}")
         return f"Error: {e}", 500
-
-    return send_file(wav_path, mimetype="audio/wav")
+    return Response(wav_data, mimetype="audio/wav",
+                    headers={"Content-Length": str(len(wav_data))})
 
 
 @app.route("/ping")