import os
import tkinter as tk
from urllib import parse, request
import json
import keyboard  # 監(jiān)聽(tīng)快捷鍵
import base64
from io import BytesIO
from zai import ZhipuAiClient  # 取消注釋
import ctypes
import mss
import mss.tools
from PIL import Image
 
# ================== 小牛翻譯 ==================
apikey = "d99fc0600b16efa6e310f53579593c57"
 
def niutrans_translate(sentence, src_lan="auto", tgt_lan="zh"):
    try:
        url = 'http://api.niutrans.com/NiuTransServer/translation?'
        data = {"from": src_lan, "to": tgt_lan, "apikey": apikey, "src_text": sentence}
        data_en = parse.urlencode(data)
        req = url + "&" + data_en
        res = request.urlopen(req, timeout=5)
        res_dict = json.loads(res.read())
        if "tgt_text" in res_dict:
            return res_dict['tgt_text']
        else:
            return f"翻譯失敗: {res_dict}"
    except Exception as e:
        return f"翻譯接口錯(cuò)誤: {str(e)}"
 
# ================== 框選截圖區(qū)域 ==================
def select_area():
    # 預(yù)初始化Tkinter，避免第一次創(chuàng)建窗口時(shí)的跳動(dòng)
    root = tk.Tk()
    root.withdraw()  # 隱藏預(yù)初始化窗口
    root.update_idletasks()
     
    # 創(chuàng)建實(shí)際的全屏窗口
    root = tk.Tk()
    root.attributes("-fullscreen", True)
    root.attributes("-alpha", 0.3)
    # 設(shè)置窗口位置確保不會(huì)影響主界面
    root.geometry("0x0+0+0")
    root.update_idletasks()
     
    canvas = tk.Canvas(root, cursor="cross")
    canvas.pack(fill="both", expand=True)
 
    rect = None
    start_x = start_y = 0
    area = [0, 0, 0, 0]
 
    # 設(shè)置DPI感知
    try:
        ctypes.windll.user32.SetProcessDPIAware()
    except:
        pass
 
    def on_mouse_down(event):
        nonlocal start_x, start_y, rect
        start_x, start_y = event.x, event.y
        rect = canvas.create_rectangle(start_x, start_y, start_x, start_y, outline="red", width=3)
 
    def on_mouse_move(event):
        nonlocal rect
        if rect:
            canvas.coords(rect, start_x, start_y, event.x, event.y)
 
    def on_mouse_up(event):
        nonlocal area
        x1, y1, x2, y2 = start_x, start_y, event.x, event.y
         
        # 直接使用坐標(biāo)，不進(jìn)行縮放轉(zhuǎn)換
        # MSS會(huì)自動(dòng)處理DPI縮放
        area = [int(min(x1, x2)), int(min(y1, y2)),
                int(max(x1, x2)), int(max(y1, y2))]
         
        #print(f"Selected area: {area}")
        root.withdraw()
        root.quit()
 
    canvas.bind("<ButtonPress-1>", on_mouse_down)
    canvas.bind("<B1-Motion>", on_mouse_move)
    canvas.bind("<ButtonRelease-1>", on_mouse_up)
 
    root.mainloop()
    root.destroy()
    return tuple(area)
 
def capture_region(region):
    """用 mss 按區(qū)域截圖，返回 PIL.Image"""
    with mss.mss() as sct:
        # 確保區(qū)域坐標(biāo)正確
        left, top, right, bottom = region
        width = max(1, right - left)
        height = max(1, bottom - top)
         
        monitor = {
            "top": top, 
            "left": left,
            "width": width,
            "height": height
        }
         
        #print(f"Capturing region: {monitor}")
        sct_img = sct.grab(monitor)
         
        # 保存截圖用于調(diào)試
        #mss.tools.to_png(sct_img.rgb, sct_img.size, output="debug_mss_screenshot.png")
        #print("MSS screenshot saved as debug_mss_screenshot.png")
         
        return Image.frombytes("RGB", sct_img.size, sct_img.bgra, "raw", "BGRX")
 
# ================== 主應(yīng)用 ==================
class TranslationApp:
    def __init__(self):
        self.root = tk.Tk()
        self.root.title("OCR翻譯器")
        self.root.overrideredirect(False)
        self.root.attributes('-topmost', True)
        self.root.geometry("400x200")
        self.font_size = 12
        self.is_zoomed = False
        self.frame = tk.Frame(self.root, bg='#2C3E50')
        self.frame.pack(padx=5, pady=5, fill="both", expand=True)
 
        self.label = tk.Label(self.frame, text="按 Ctrl+Alt+A 截圖翻譯",
                              font=('微軟雅黑', 12),
                              bg='#2C3E50', fg='#ECF0F1',
                              wraplength=400, justify='left',
                              padx=10, pady=5)
        self.label.pack(fill="both", expand=True)
 
        self.running = True
        self.last_text = ""
 
        self.setup_ui()
 
        # 注冊(cè)截圖快捷鍵
        keyboard.add_hotkey("ctrl+alt+a", self.capture_and_translate)
 
    def translate_text(self, text):
        is_chinese = any('\u4e00' <= c <= '\u9fa5' for c in text)
        from_lang = "zh" if is_chinese else "en"
        to_lang = "en" if is_chinese else "zh"
        return niutrans_translate(text, from_lang, to_lang)
 
    def process_image_with_ai(self, pil_image):
        # 把 PIL 圖片轉(zhuǎn)為 base64
        buffered = BytesIO()
        pil_image.save(buffered, format="JPEG")
        encoded_string = base64.b64encode(buffered.getvalue()).decode('utf-8')
 
        try:
            client = ZhipuAiClient(api_key="101fb0dca2d148cc937a73dc61f73368.HfIMCncTJPNibjYl")
 
            response = client.chat.completions.create(
                model="glm-4.1v-thinking-flash",
                messages=[{
                    "role": "user",
                    "content": [
                        {
                            "type": "image_url",
                            "image_url": {
                                "url": f"data:image/jpeg;base64,{encoded_string}"
                            }
                        },
                        {
                            "type": "text",
                            "text": "請(qǐng)準(zhǔn)確提取圖片中的文字信息，不增加、刪除或改寫(xiě)其它文本內(nèi)容。"
                        }
                    ]
                }]
            )
 
            content = response.choices[0].message.content
            if content:
                content = content.replace("<|begin_of_box|>", "").replace("<|end_of_box|>", "")
                return content
            else:
                return "未識(shí)別到文字"
                 
        except Exception as e:
            print(f"AI OCR錯(cuò)誤: {str(e)}")
            # 如果AI OCR失敗，嘗試使用Tesseract作為備用方案
            return self.fallback_ocr(pil_image)
 
    def fallback_ocr(self, pil_image):
        """備用OCR方案（如果AI OCR失敗）"""
        try:
            import pytesseract
            # 如果安裝了Tesseract
            text = pytesseract.image_to_string(pil_image, lang='chi_sim+eng')
            return text.strip() if text.strip() else "未識(shí)別到文字（備用OCR）"
        except:
            return "OCR服務(wù)不可用，請(qǐng)檢查網(wǎng)絡(luò)連接或AI密鑰"
 
    def capture_and_translate(self):
        try:
            #print("開(kāi)始截圖...")
            # 選擇截圖區(qū)域
            region = select_area()
            #print(f"選擇的區(qū)域: {region}")
             
            if region[0] == region[2] or region[1] == region[3]:
                self.update_display("無(wú)效的截圖區(qū)域")
                return
 
            # 使用 mss 截圖
            screenshot = capture_region(region)
 
            # 保存截圖用于調(diào)試
            #screenshot.save("debug_pil_screenshot.png")
            #print("PIL screenshot saved as debug_pil_screenshot.png")
 
            # 顯示"識(shí)別中..."提示
            self.update_display("識(shí)別中...")
 
            # 調(diào)用 AI OCR，直接傳 PIL.Image
            text = self.process_image_with_ai(screenshot).strip()
            #print(f"識(shí)別結(jié)果: {text}")
 
            if text and text != "未識(shí)別到文字":
                self.last_text = text
                # 顯示"翻譯中..."提示
                self.update_display(f"原文: {text}\n翻譯中...")
                 
                # 翻譯文本
                translation = self.translate_text(text)
                #print(f"翻譯結(jié)果: {translation}")
                 
                self.update_display(f"原文: {text}\n\n譯文: {translation}")
            else:
                self.update_display("未識(shí)別到文字")
                 
        except Exception as e:
            import traceback
            error_msg = f"錯(cuò)誤: {str(e)}\n{traceback.format_exc()}"
            print(error_msg)
            self.update_display(f"處理錯(cuò)誤:\n{str(e)}")
 
    def update_display(self, text):
        self.label.config(text=text)
        self.root.update_idletasks()
    def copy_to_clipboard(self):
        """復(fù)制原文和譯文到剪貼板"""
        if self.last_text:
            try:
                # 獲取當(dāng)前顯示的文本
                current_text = self.label.cget("text")
                self.root.clipboard_clear()
                self.root.clipboard_append(current_text)
                # 臨時(shí)顯示復(fù)制成功提示
                original_text = self.label.cget("text")
                self.label.config(text="已復(fù)制到剪貼板！")
                self.root.after(1500, lambda: self.label.config(text=original_text))
            except Exception as e:
                print(f"復(fù)制失敗: {e}")
 
    def zoom_text(self):
        """放大/恢復(fù)文本字體"""
        if not self.is_zoomed:
            # 放大字體
            self.font_size = 16
            self.label.config(font=('微軟雅黑', self.font_size, 'bold'))
            self.is_zoomed = True
        else:
            # 恢復(fù)原字體
            self.font_size = 12
            self.label.config(font=('微軟雅黑', self.font_size))
            self.is_zoomed = False
    def setup_ui(self):
        control_frame = tk.Frame(self.frame, bg='#2C3E50')
        control_frame.pack(fill='x', pady=(5,0))
         
        # 左上角放大按鈕
        zoom_btn = tk.Button(control_frame, text="放大",
                            command=self.zoom_text,
                            bg='#27AE60', fg='white', relief='flat', width=6, font=(10))
        zoom_btn.pack(side='left', padx=2)
         
        # 左下角截圖按鈕
        capture_btn = tk.Button(control_frame, text="截圖",
                               command=self.capture_and_translate,
                               bg='#3498DB', fg='white', relief='flat', width=6, font=(10))
        capture_btn.pack(side='left', padx=2)
     
        exit_btn = tk.Button(control_frame, text="退出",
                             command=self.exit_app,
                             bg='#E74C3C', fg='white', relief='flat', width=6, font=(10))
        exit_btn.pack(side='right', padx=2)
 
 
        # 右上角復(fù)制按鈕
        copy_btn = tk.Button(control_frame, text="復(fù)制",
                            command=self.copy_to_clipboard,
                            bg='#9B59B6', fg='white', relief='flat', width=6, font=(10))
        copy_btn.pack(side='right', padx=2)
         
         
        # 快捷鍵退出
        self.root.bind("<Control-Shift-Q>", lambda e: self.exit_app())
 
    def exit_app(self):
        self.running = False
        self.root.destroy()
 
    def run(self):
        self.root.mainloop()
 
if __name__ == "__main__":
    app = TranslationApp()
    app.run()