[Python] 纯文本查看 复制代码
import os
import tkinter as tk
from urllib import parse, request
import json
import keyboard # 监听快捷键
import base64
from io import BytesIO
from zai import ZhipuAiClient # 取消注释
import ctypes
import mss
import mss.tools
from PIL import Image
# ================== 小牛翻译 ==================
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"翻译失败: {res_dict}"
except Exception as e:
return f"翻译接口错误: {str(e)}"
# ================== 框选截图区域 ==================
def select_area():
# 预初始化Tkinter，避免第一次创建窗口时的跳动
root = tk.Tk()
root.withdraw() # 隐藏预初始化窗口
root.update_idletasks()
# 创建实际的全屏窗口
root = tk.Tk()
root.attributes("-fullscreen", True)
root.attributes("-alpha", 0.3)
# 设置窗口位置确保不会影响主界面
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]
# 设置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
# 直接使用坐标，不进行缩放转换
# MSS会自动处理DPI缩放
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 按区域截图，返回 PIL.Image"""
with mss.mss() as sct:
# 确保区域坐标正确
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)
# 保存截图用于调试
#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")
# ================== 主应用 ==================
class TranslationApp:
def __init__(self):
self.root = tk.Tk()
self.root.title("OCR翻译器")
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 截图翻译",
font=('微软雅黑', 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()
# 注册截图快捷键
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 图片转为 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": "请准确提取图片中的文字信息，不增加、删除或改写其它文本内容。"
}
]
}]
)
content = response.choices[0].message.content
if content:
content = content.replace("<|begin_of_box|>", "").replace("<|end_of_box|>", "")
return content
else:
return "未识别到文字"
except Exception as e:
print(f"AI OCR错误: {str(e)}")
# 如果AI OCR失败，尝试使用Tesseract作为备用方案
return self.fallback_ocr(pil_image)
def fallback_ocr(self, pil_image):
"""备用OCR方案（如果AI OCR失败）"""
try:
import pytesseract
# 如果安装了Tesseract
text = pytesseract.image_to_string(pil_image, lang='chi_sim+eng')
return text.strip() if text.strip() else "未识别到文字（备用OCR）"
except:
return "OCR服务不可用，请检查网络连接或AI密钥"
def capture_and_translate(self):
try:
#print("开始截图...")
# 选择截图区域
region = select_area()
#print(f"选择的区域: {region}")
if region[0] == region[2] or region[1] == region[3]:
self.update_display("无效的截图区域")
return
# 使用 mss 截图
screenshot = capture_region(region)
# 保存截图用于调试
#screenshot.save("debug_pil_screenshot.png")
#print("PIL screenshot saved as debug_pil_screenshot.png")
# 显示"识别中..."提示
self.update_display("识别中...")
# 调用 AI OCR，直接传 PIL.Image
text = self.process_image_with_ai(screenshot).strip()
#print(f"识别结果: {text}")
if text and text != "未识别到文字":
self.last_text = text
# 显示"翻译中..."提示
self.update_display(f"原文: {text}\n翻译中...")
# 翻译文本
translation = self.translate_text(text)
#print(f"翻译结果: {translation}")
self.update_display(f"原文: {text}\n\n译文: {translation}")
else:
self.update_display("未识别到文字")
except Exception as e:
import traceback
error_msg = f"错误: {str(e)}\n{traceback.format_exc()}"
print(error_msg)
self.update_display(f"处理错误:\n{str(e)}")
def update_display(self, text):
self.label.config(text=text)
self.root.update_idletasks()
def copy_to_clipboard(self):
"""复制原文和译文到剪贴板"""
if self.last_text:
try:
# 获取当前显示的文本
current_text = self.label.cget("text")
self.root.clipboard_clear()
self.root.clipboard_append(current_text)
# 临时显示复制成功提示
original_text = self.label.cget("text")
self.label.config(text="已复制到剪贴板！")
self.root.after(1500, lambda: self.label.config(text=original_text))
except Exception as e:
print(f"复制失败: {e}")
def zoom_text(self):
"""放大/恢复文本字体"""
if not self.is_zoomed:
# 放大字体
self.font_size = 16
self.label.config(font=('微软雅黑', self.font_size, 'bold'))
self.is_zoomed = True
else:
# 恢复原字体
self.font_size = 12
self.label.config(font=('微软雅黑', 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))
# 左上角放大按钮
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)
# 左下角截图按钮
capture_btn = tk.Button(control_frame, text="截图",
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)
# 右上角复制按钮
copy_btn = tk.Button(control_frame, text="复制",
command=self.copy_to_clipboard,
bg='#9B59B6', fg='white', relief='flat', width=6, font=(10))
copy_btn.pack(side='right', padx=2)
# 快捷键退出
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()
发表于 2025-11-3 14:41
|
发表于 2025-11-3 14:56
