app.py

# KC Robot AI V4.1 – Cloud Brain Intelligent Assistant
# Flask server: Chat (HF), TTS, STT, Telegram poller, REST API cho ESP32
# Features: bilingual greetings, radar detect, OLED lines, TTS/STT, Telegram, HuggingFace brain

import os, io, time, json, threading, logging, requests
from typing import Optional, List, Tuple
from flask import Flask, request, jsonify, send_file, render_template_string

logging.basicConfig(level=logging.INFO)
logger = logging.getLogger("kcrobot.v4.1")
app = Flask(__name__)

# ====== Config from env / Secrets ======
HF_API_TOKEN = os.getenv("HF_API_TOKEN", "")
HF_MODEL = os.getenv("HF_MODEL", "google/flan-t5-large")
HF_TTS_MODEL = os.getenv("HF_TTS_MODEL", "facebook/tts_transformer-en-ljspeech")
HF_STT_MODEL = os.getenv("HF_STT_MODEL", "openai/whisper-small")
TELEGRAM_TOKEN = os.getenv("TELEGRAM_TOKEN", "")
PORT = int(os.getenv("PORT", 7860))

HF_HEADERS = {"Authorization": f"Bearer {HF_API_TOKEN}"} if HF_API_TOKEN else {}
if not HF_API_TOKEN:
    logger.warning("⚠️ HF_API_TOKEN not set. Put HF_API_TOKEN in Secrets.")

# ====== In-memory storage ======
CONV: List[Tuple[str, str]] = []
DISPLAY_LINES: List[str] = []

def push_display(line: str, limit=6):
    global DISPLAY_LINES
    DISPLAY_LINES.append(line)
    if len(DISPLAY_LINES) > limit:
        DISPLAY_LINES = DISPLAY_LINES[-limit:]

# ====== Hugging Face API helpers ======
def hf_text_generate(prompt: str, model: Optional[str] = None, max_new_tokens: int = 256) -> str:
    model = model or HF_MODEL
    url = f"https://api-inference.huggingface.co/models/{model}"
    payload = {
        "inputs": prompt,
        "parameters": {"max_new_tokens": max_new_tokens, "temperature": 0.7},
        "options": {"wait_for_model": True}
    }
    r = requests.post(url, headers=HF_HEADERS, json=payload, timeout=120)
    if r.status_code != 200:
        raise RuntimeError(f"HF text generation failed: {r.status_code}: {r.text}")
    data = r.json()
    if isinstance(data, list) and len(data) and isinstance(data[0], dict):
        return data[0].get("generated_text", "")
    if isinstance(data, dict) and "generated_text" in data:
        return data["generated_text"]
    return str(data)

def hf_tts_get_mp3(text: str, model: Optional[str] = None) -> bytes:
    model = model or HF_TTS_MODEL
    url = f"https://api-inference.huggingface.co/models/{model}"
    payload = {"inputs": text}
    headers = dict(HF_HEADERS)
    headers["Content-Type"] = "application/json"
    r = requests.post(url, headers=headers, json=payload, stream=True, timeout=120)
    if r.status_code != 200:
        raise RuntimeError(f"HF TTS failed: {r.status_code}: {r.text}")
    return r.content

def hf_stt_from_bytes(audio_bytes: bytes, model: Optional[str] = None) -> str:
    model = model or HF_STT_MODEL
    url = f"https://api-inference.huggingface.co/models/{model}"
    headers = dict(HF_HEADERS)
    headers["Content-Type"] = "application/octet-stream"
    r = requests.post(url, headers=headers, data=audio_bytes, timeout=180)
    if r.status_code != 200:
        raise RuntimeError(f"HF STT failed: {r.status_code}: {r.text}")
    j = r.json()
    return j.get("text", str(j))

# ====== API cho ESP32 ======
@app.route("/ask", methods=["POST"])
def api_ask():
    data = request.get_json(force=True)
    text = data.get("text", "").strip()
    lang = data.get("lang", "auto")
    if not text:
        return jsonify({"error": "no text"}), 400
    if lang == "vi":
        prompt = "Bạn là trợ lý thông minh, trả lời bằng tiếng Việt:\n" + text
    elif lang == "en":
        prompt = "You are a helpful assistant. Answer in English:\n" + text
    else:
        prompt = "Bạn là trợ lý song ngữ Việt-Anh, trả lời theo ngôn ngữ người dùng:\n" + text
    try:
        ans = hf_text_generate(prompt)
    except Exception as e:
        return jsonify({"error": str(e)}), 500
    CONV.append((text, ans))
    push_display("YOU: " + text[:40])
    push_display("BOT: " + ans[:40])
    return jsonify({"answer": ans})

@app.route("/presence", methods=["POST"])
def api_presence():
    data = request.get_json(force=True)
    note = data.get("note", "Có người tới gần!")
    greeting_vi = f"Xin chào! {note}"
    greeting_en = "Hello there! Nice to see you."
    combined = f"{greeting_vi}\n{greeting_en}"
    CONV.append(("__presence__", combined))
    push_display("RADAR: " + note[:40])
    if TELEGRAM_TOKEN:
        try:
            send_telegram_message(f"⚠️ Phát hiện có người: {note}")
        except Exception as e:
            logger.error("Telegram send failed: %s", e)
    return jsonify({"greeting": combined})

@app.route("/display", methods=["GET"])
def api_display():
    return jsonify({"lines": DISPLAY_LINES, "conv_len": len(CONV)})

# ====== Web UI (simple) ======
@app.route("/")
def index():
    return render_template_string("<h3>🤖 KC Robot AI V4.1 - Cloud Brain Running</h3><p>Bilingual & Smart Connected to ESP32</p>")

# ====== Telegram ======
def send_telegram_message(text: str):
    if not TELEGRAM_TOKEN:
        return
    chat_id = os.getenv("TELEGRAM_CHATID", "")
    url = f"https://api.telegram.org/bot{TELEGRAM_TOKEN}/sendMessage"
    requests.post(url, json={"chat_id": chat_id, "text": text}, timeout=10)

def telegram_poll_loop():
    if not TELEGRAM_TOKEN: return
    logger.info("Starting Telegram poller...")
    offset = None
    base = f"https://api.telegram.org/bot{TELEGRAM_TOKEN}"
    while True:
        try:
            params = {"timeout": 30}
            if offset: params["offset"] = offset
            r = requests.get(base + "/getUpdates", params=params, timeout=35)
            j = r.json()
            for u in j.get("result", []):
                offset = u["update_id"] + 1
                msg = u.get("message", {})
                text = msg.get("text", "")
                chat_id = msg.get("chat", {}).get("id")
                if text.lower().startswith("/ask "):
                    q = text[5:].strip()
                    ans = hf_text_generate(q)
                    requests.post(base + "/sendMessage", json={"chat_id": chat_id, "text": ans})
                elif text.lower().startswith("/say "):
                    t = text[5:].strip()
                    mp3 = hf_tts_get_mp3(t)
                    files = {"audio": ("robot.mp3", mp3, "audio/mpeg")}
                    requests.post(base + "/sendAudio", files=files, data={"chat_id": chat_id})
        except Exception as e:
            logger.error("TG poll error: %s", e)
            time.sleep(3)

def start_background():
    if TELEGRAM_TOKEN:
        threading.Thread(target=telegram_poll_loop, daemon=True).start()

@app.before_first_request
def _startup():
    start_background()

if __name__ == "__main__":
    start_background()
    logger.info(f"🚀 KC Robot AI V4.1 running on port {PORT}")
    print("Xin chào chủ nhân! Em là KC Robot — rất vui được gặp bạn.\nHello master! I’m KC Robot, your smart assistant.")
    app.run(host="0.0.0.0", port=PORT)