import os
import sys
import time
import json
import base64
import requests
import importlib
from datetime import datetime
from dotenv import load_dotenv

# Try importing SpeechRecognition.
try:
    import speech_recognition as sr
except ImportError:
    print("⚠️ 偵測到尚未安裝 `SpeechRecognition` 套件。")
    print("請先執行以下指令安裝：")
    print("  pip install SpeechRecognition")
    sr = None

# Determine directories and load environment variables
SCRIPT_DIR = os.path.dirname(os.path.abspath(__file__))
load_dotenv(os.path.join(SCRIPT_DIR, '.env'))
load_dotenv(os.path.join(SCRIPT_DIR, '../.env'))
load_dotenv(os.path.join(SCRIPT_DIR, '../sales-bot/.env'))

if SCRIPT_DIR not in sys.path:
    sys.path.insert(0, SCRIPT_DIR)

# Camp AI Control flag
CAMP_AI_RUNNING = False

def get_server_ip():
    ip = os.getenv("server_ip")
    if ip:
        return ip
    try:
        env_path = os.path.join(SCRIPT_DIR, '.env')
        if os.path.exists(env_path):
            with open(env_path, 'r', encoding='utf-8') as f:
                for line in f:
                    line = line.strip()
                    if line and not line.startswith('#') and '=' in line:
                        parts = line.split('=', 1)
                        key = parts[0].strip()
                        val = parts[1].strip().strip('\'"')
                        if key == "server_ip":
                            return val
    except Exception:
        pass
    return "10.112.5.79"

def call_whisper_audio(audio_bytes):
    """使用 requests 呼叫 Whisper ASR 服務"""
    server_ip = get_server_ip()
    url = f"http://{server_ip}:4002/v1/audio/transcriptions"
    
    import io
    try:
        files = {"file": ("audio.wav", io.BytesIO(audio_bytes), "audio/wav")}
        response = requests.post(url, files=files, timeout=30)
        response.raise_for_status()
        result = response.json()
        return result.get("text", "").strip()
    except Exception as e:
        print(f"\n❌ 呼叫 Whisper 失敗: {e}")
        return None

def process_and_save(audio_data, timestamp, duration):
    """處理音訊：上傳給 Whisper、儲存轉錄的文字為 JSON、轉換並保存語音檔案，並返回轉錄的文字結果"""
    try:
        wav_bytes = audio_data.get_wav_data()
        
        # 呼叫 Whisper 進行轉錄
        result = call_whisper_audio(wav_bytes)
        
        if result:
            record_dir = os.path.join(SCRIPT_DIR, "record")
            os.makedirs(record_dir, exist_ok=True)
            
            # 檔名固定為：latest_message.txt
            filename = os.path.join(record_dir, "latest_message.txt")
            
            # 建立要儲存的單筆 JSON 資料結構
            record_data = {
                "timestamp": timestamp,
                "duration_time": round(duration, 1),
                "text": result,
                "status": "not_processed"
            }
            
            # 寫入文字檔 (內容為 JSON)
            with open(filename, "w", encoding="utf-8") as f:
                json.dump(record_data, f, ensure_ascii=False, indent=4)
            
            print(f"\n✨ [轉錄完成並儲存] -> record/latest_message.txt")
            print(f"👉 \"{result}\"")
            
            # 保存音訊檔案
            import subprocess
            wav_filename = os.path.join(record_dir, "latest_voice.wav")
            mp3_filename = os.path.join(record_dir, "latest_voice.mp3")
            
            # 寫入暫存的 wav 檔
            with open(wav_filename, "wb") as f:
                f.write(wav_bytes)
                
            converted = False
            # 嘗試使用 pydub 轉檔為 MP3
            try:
                from pydub import AudioSegment
                import io
                audio_segment = AudioSegment.from_wav(io.BytesIO(wav_bytes))
                audio_segment.export(mp3_filename, format="mp3")
                converted = True
            except Exception:
                pass
                
            # 嘗試呼叫系統 ffmpeg 指令轉檔為 MP3
            if not converted:
                try:
                    subprocess.run(
                        ["ffmpeg", "-y", "-i", wav_filename, "-codec:a", "libmp3lame", "-qscale:a", "2", mp3_filename],
                        stdout=subprocess.DEVNULL,
                        stderr=subprocess.DEVNULL,
                        check=True
                    )
                    converted = True
                except Exception:
                    pass
            
            if converted:
                # 刪除暫存的 wav 檔
                if os.path.exists(wav_filename):
                    os.remove(wav_filename)
                print(f"✨ [聲音儲存成功] -> record/latest_voice.mp3")
            else:
                # 若無法轉成 mp3，保留原始 wav 檔案
                print(f"⚠️ [聲音轉換 MP3 失敗] 已保留原始格式為 record/latest_voice.wav")
                print("  提示：若要啟用 MP3 轉換，請安裝 ffmpeg (例如: brew install ffmpeg) 與 pydub (pip install pydub)")
            
            return result
        else:
            print(f"\n⚠️ [轉錄失敗] 時間: {timestamp} (時長: {duration:.1f} 秒) - Whisper 未能成功回傳文字")
            return None
            
    except Exception as e:
        print(f"\n❌ [處理錯誤] 在處理音訊時發生異常: {e}")
        return None

def run_command_script(script_to_run):
    """動態載入並執行指定的 Python 程式碼模組"""
    script_path = os.path.join(SCRIPT_DIR, script_to_run)
    if os.path.exists(script_path):
        print(f"\n🚀 啟動程式碼: {script_to_run} (直接載入執行)...", flush=True)
        try:
            module_name = script_to_run.replace(".py", "")
            if module_name in sys.modules:
                module = importlib.reload(sys.modules[module_name])
            else:
                module = importlib.import_module(module_name)
            
            if hasattr(module, "main"):
                module.main()
            else:
                print(f"⚠️ 模組 {module_name} 沒有 main() 函式，直接執行模組內頂層程式碼可能已在載入時完成。")
            print(f"✅ {script_to_run} 執行完畢。", flush=True)
        except Exception as e:
            import traceback
            print(f"❌ 執行 {script_to_run} 時出錯:", flush=True)
            traceback.print_exc()
    else:
        print(f"❌ 找不到對應的程式碼檔案: {script_to_run}", flush=True)

def main():
    if sr is None:
        print("❌ 請安裝 SpeechRecognition 後再重新執行此腳本。")
        return

    # 確保 record 資料夾存在
    record_dir = os.path.join(SCRIPT_DIR, "record")
    os.makedirs(record_dir, exist_ok=True)

    recognizer = sr.Recognizer()
    
    # === 語音辨識靈敏度與斷句設定 ===
    # 能量閥值：數字越大越不靈敏。預設 300，建議設 800 - 1500 避開小雜音與呼吸聲。
    recognizer.energy_threshold = 1200
    
    # 是否開啟動態自動調節：True 會自動調，但安靜房間常會降得太低變極度靈敏；推薦設為 False。
    recognizer.dynamic_energy_threshold = False
    
    # 靜音判定秒數：說完話後，停頓超過此秒數即判定這句話結束。
    recognizer.pause_threshold = 1.2 
    
    print("\n" + "="*60)
    print("🤖 Reachy Mini 語音控制與排程服務已啟動！")
    print("📂 所有轉錄與語音檔案將儲存在：record/ 目錄下")
    print("🎙️ 持續監聽語音中，若有觸發關鍵字將會同步執行對應的任務...")
    print("🛑 請按 Ctrl+C 可隨時終止程式")
    print("="*60)
    
    global CAMP_AI_RUNNING
    CAMP_AI_RUNNING = True
    
    try:
        with sr.Microphone() as source:
            print("⚡ 正在適應周圍環境噪音 (1秒)...請先保持安靜...")
            recognizer.adjust_for_ambient_noise(source, duration=1.0)
            print(f"🟢 系統準備就緒！(目前能量門檻設定為: {recognizer.energy_threshold})")
            print("🎤 隨時可以開始說話！")
    except Exception as e:
        print(f"❌ 麥克風初始化失敗: {e}")
        CAMP_AI_RUNNING = False
        return

    while CAMP_AI_RUNNING:
        try:
            with sr.Microphone() as source:
                print("\n🎤 正在聆聽...")
                # listen 加上 timeout=2.0，每 2 秒檢查一次是否已關閉服務
                audio_data = recognizer.listen(source, timeout=2.0, phrase_time_limit=30)
                
            # 計算錄音時間戳與時長
            timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
            duration = len(audio_data.frame_data) / (audio_data.sample_rate * audio_data.sample_width)
            
            # 過濾低於 0.8 秒的短雜音
            if duration < 0.8:
                continue
                
            print(f"📝 偵測到語音，時長約 {duration:.1f} 秒，正在處理與上傳...")
            result = process_and_save(audio_data, timestamp, duration)
            
            if result:
                text_lower = result.lower()
                script_to_run = None
                
                # 關鍵字判定 (支援常見的語音辨識同音/近音字，不分大小寫)
                if "what is this" in text_lower or "what's this" in text_lower:
                    script_to_run = "look_that.py"
                elif "suggestion" in text_lower or "suggestions" in text_lower:
                    script_to_run = "ask.py"
                elif "guard mode" in text_lower or "god mode" in text_lower or "grad mode" in text_lower:
                    script_to_run = "guard.py"
                elif "location" in text_lower:
                    script_to_run = "location.py"
                
                if script_to_run:
                    run_command_script(script_to_run)
                else:
                    print("ℹ️ 語音內容未包含觸發指令的關鍵字。")
                    
        except sr.WaitTimeoutError:
            # 逾時未偵測到聲音，回到迴圈頂部以檢查 CAMP_AI_RUNNING 狀態
            continue
        except KeyboardInterrupt:
            print("\n👋 偵測到終止指令，正在關閉程式...")
            CAMP_AI_RUNNING = False
            break
        except Exception as e:
            print(f"\n❌ 監聽或處理過程發生異常: {e}")
            time.sleep(1)

if __name__ == "__main__":
    main()