python检索屏幕文本

import io
from aip import AipOcr
from PIL import ImageGrab

def baiduOCR(string):

#百度文字识别
APP_ID = '18******'
API_KEY = 'Dn***************'
SECRECT_KEY = 'lA************************'
client = AipOcr(APP_ID, API_KEY, SECRECT_KEY)

#截屏
img = ImageGrab.grab()
#字节容器
img_b = io.BytesIO()
#image转换为png
img.save(img_b, format='PNG')
#存入容器
img_b = img_b.getvalue()

options = {"recognize_granularity": "small"}
message = client.general(img_b, options)
location = dict()
for f in message['words_result']:
    n = f['words'].find(string)
    if n != -1:
        l = len(string)
        tops = list()
        heights = list()
        left = f['chars'][n]['location']['left']
        right = f['chars'][n+l-1]['location']['left'] + f['chars'][n+l-1]['location']['width']
        for i in f['chars'][n:n+l]:
            tops.append(i['location']['top'])
            heights.append(i['location']['height'])
        top = min(tops)
        bottom = top + max(heights)
        location = {'top': top, 'bottom': bottom, 'left': left, 'right': right}
        print(f['words'],location)
        break

if name == “main“:
while True:
string = input(‘检索屏幕：’)
baiduOCR(string)