"""
OCR模块 - 提供基于PaddleOCR的文本识别和定位功能
"""
import os
import time
from typing import List, Dict, Tuple, Optional
import numpy as np
from paddleocr import PaddleOCR
import cv2
from .android import get_device
class OCRManager:
_instance = None
_ocr = None
_last_screen = None
_last_result = None
_last_time = 0
_cache_timeout = 1 # 缓存超时时间(秒)
def __new__(cls):
if cls._instance is None:
cls._instance = super(OCRManager, cls).__new__(cls)
return cls._instance
def __init__(self):
if self._ocr is None:
self._ocr = PaddleOCR(use_angle_cls=True, lang="ch", show_log=False)
def _get_screenshot(self) -> np.ndarray:
"""获取当前屏幕截图"""
device = get_device()
screenshot = device.screenshot(format='opencv')
return screenshot
def _should_refresh_cache(self) -> bool:
"""判断是否需要刷新缓存"""
return (time.time() - self._last_time) > self._cache_timeout
def ocr_screen(self, use_cache: bool = True) -> List[Dict]:
"""
对当前屏幕进行OCR识别
Args:
use_cache: 是否使用缓存结果
Returns:
List[Dict]: OCR识别结果列表，每个Dict包含:
- text: 识别的文本
- confidence: 置信度
- position: 文本框坐标 ((x1,y1), (x2,y2), (x3,y3), (x4,y4))
"""
if use_cache and not self._should_refresh_cache() and self._last_result is not None:
return self._last_result
# 获取屏幕截图
self._last_screen = self._get_screenshot()
# 执行OCR识别
result = self._ocr.ocr(self._last_screen, cls=True)
# 格式化结果
formatted_result = []
if result:
for line in result[0]:
position, (text, confidence) = line
formatted_result.append({
'text': text,
'confidence': confidence,
'position': position
})
# 更新缓存
self._last_result = formatted_result
self._last_time = time.time()
return formatted_result
def find_text_position(self, text: str, confidence_threshold: float = 0.5) -> Optional[Tuple[int, int]]:
"""
查找文本位置
Args:
text: 要查找的文本
confidence_threshold: 置信度阈值
Returns:
Tuple[int, int]: 文本中心点坐标，如果未找到返回None
"""
results = self.ocr_screen()
# 查找匹配的文本
for item in results:
if (text in item['text'] and
item['confidence'] >= confidence_threshold):
# 计算中心点坐标
position = item['position']
center_x = int(sum(x for x, _ in position) / 4)
center_y = int(sum(y for _, y in position) / 4)
return (center_x, center_y)
return None
def click_text(self, text: str, confidence_threshold: float = 0.5) -> bool:
"""
点击指定文本
Args:
text: 要点击的文本
confidence_threshold: 置信度阈值
Returns:
bool: 是否点击成功
"""
position = self.find_text_position(text, confidence_threshold)
if position:
return self.click_position(*position)
return False
def click_position(self, x: int, y: int) -> bool:
"""
点击指定坐标
Args:
x: X坐标
y: Y坐标
Returns:
bool: 是否点击成功
"""
try:
device = get_device()
device.click(x, y)
return True
except Exception:
return False
# 创建全局OCR管理器实例
ocr_manager = OCRManager()
