某平台收集题库自动答题：Python+OCR

准备工作：

python3

申请一个百度OCR的appid、appkey、secretkey，不会申请 开通的戳下面链接，视频中有详细说明：

https://www.bilibili.com/video/BV1Xh41117DM

整体思路：

• 1、设置固定微信PC端浏览器位置及窗口大小，以便确定截图坐标范围以及获取各个指定坐标颜色信息；
• 2、确定好截图坐标范围后截图，通过百度文字识别OCR接口上传，进行文字识别，获取题目内容；
• 3、在本地题库搜索该题目，如果没有答案就随机点击前两个答案；
• 4、通过获取点击答案后固定位置出现的✔标志颜色，来判断哪个选项为正确答案；
• 5、获取到正确答案选项后，将题目与答案写入本地题库；
• 6、如果在第3步在本地题库中找到答案，那么就点击答案选项；
• 7、循环控制，通过获取指定位置颜色确定是否开始答题或进入下一局。

完整代码：

import jsonpath
from aip import AipOcr
import pyautogui as ptg
from io import BytesIO
import time
import openpyxl
import os
import random
import win32gui
import win32con

APP_ID = ''
API_KEY = ''
SECRET_KEY = ''

client = AipOcr(APP_ID, API_KEY, SECRET_KEY)

path='./answer.xlsx'
zanswer=[]

if not os.path.exists(path):
    workbook = openpyxl.Workbook()
    workbook.save(path)
else:
    data = openpyxl.load_workbook(path)
    table = data.active
    nrows = table.max_row
    for i in range(1,nrows+1):
        tmp = []
        tmp.append(table.cell(i, 1).value)
        tmp.append(table.cell(i, 2).value)
        zanswer.append(tmp)
    data.close()



def uppic(img):
    client = AipOcr(APP_ID, API_KEY, SECRET_KEY)
    # 识别参数
    options = {}
    options["language_type"] = "CHN_ENG"
    options["detect_direction"] = "false"
    options["detect_language"] = "false"
    options["probability"] = "false"

    # 图片识别
    result = client.basicGeneral(img, options)
    return result

def wxls(qList):
    w=True
    for i in zanswer:
        if qList[0]==i[0]:
            w=False
            print(qList[0],'已存在')
            break
    if w==True:
        print('题库记录:%d,%s'%(len(zanswer),qList))
        data = openpyxl.load_workbook(path)
        table = data.active
        nrows = table.max_row
        for j in range(1, 3):
            table.cell(nrows+1, j).value = qList[j-1]
        data.save(path)
        zanswer.append(qList)

def sshot(a,b,c,d):
    im=ptg.screenshot(region=(a,b,c,d))
    buffered = BytesIO()
    im.save(buffered, format="JPEG")
    img_Bytes = buffered.getvalue()
    return uppic(img_Bytes)

def rgbPro(s):
    strCx = s[-2:]
    if len(s) == 3:
        strCx = '0' + s[2]
    return strCx

def getColor(color):
    a1 = hex(color[0]) #R
    a2 = hex(color[1]) #G
    a3 = hex(color[2]) #B
    s1 = rgbPro(str(a1))
    s2 = rgbPro(str(a2))
    s3 = rgbPro(str(a3))
    strC = str.upper(s3)+  str.upper(s2) + str.upper(s1)
    return strC #BGR

def leftClick(x,y):
    a, b = ptg.position()
    ptg.click(x,y)
    time.sleep(0.1)
    ptg.moveTo(a,b)

def main():
    quitFlag=0
    while True:
        color = getColor(ptg.pixel(150, 90))
        if (color != '5448DC') and (color != '37308A') and (color !='1D1A3F'):     # 判断是否退出
            if quitFlag > 1: exit()
            quitFlag+=1
            time.sleep(1)
        else:
            color = getColor(ptg.pixel(265,768))  # 判断再来一次
            if color == '6A4DFF':
                leftClick(265,768)
                time.sleep(6)
                continue
            quitFlag=0
            try:
                color = getColor(ptg.pixel(84, 845))  # 判断再来一次
                if color == 'FFAA00':
                    leftClick(84+random.randint(0,20), 845+random.randint(0,10))
                    time.sleep(4)
                color = getColor(ptg.pixel(250, 660))  # 判断开始答题
                if color == '00FFFF':
                    leftClick(250, 660)
                    time.sleep(4)

                #题目区域截图上传OCR识别
                rjson = sshot(6, 285, 487, 223)  #获取题目
                rnum = rjson['words_result_num']
                if int(rnum) == 0: continue
                rstr = jsonpath.jsonpath(rjson, '$..words')

                # 拼接题目
                tm =''
                for i in range(0,len(rstr)):
                    tm=tm+rstr[i]

                #特殊字符过滤
                repstr=',."，、。“”[]【】〈〉<>《》？_ '
                for i in repstr:
                    tm=tm.replace(i,'')

                #题库查找是否存在
                w=False
                for i in zanswer:
                    if tm==i[0]:
                        #如果存在就点击对应位置
                        print('找到答案：题目:%s,答案:%s'%(i[0],i[1]))
                        # x=75+random.randint(0,320)
                        # y=560+random.randint(0,30)+(int(i[1])*103)
                        leftClick(75,560+(int(i[1])*103))
                        time.sleep(4)
                        #给一个找到答案的Flag
                        w=True
                        break
                if w==True:
                    #找到答案就进入下一轮
                    time.sleep(4)
                    continue
                else:
                    #没找到答案

                    #答案最多4个:0,1,2,3
                    answer=-1
                    leftClick(75,560+(random.randint(0,1)*103))
                    time.sleep(0.5)
                    try:
                        for i in range(0,4):
                            color = getColor(ptg.pixel(404, 558+(i*104)))
                            if color == '3EC68C': #答案正确
                                answer=i
                                break
                        if answer == -1: continue
                    except:
                        continue

                    wzanswer=[]
                    wzanswer.append(tm)
                    wzanswer.append(answer)

                    #写入EXCEL
                    wxls(wzanswer)
                time.sleep(4)
                    #time.sleep(random.randint(5,10))
            except:
                pass

if __name__ == '__main__':
    hwnd=win32gui.FindWindow('CefWebViewWnd','微信')
    win32gui.SetWindowPos(hwnd, win32con.HWND_TOPMOST, 0,0, 500,1000, win32con.SWP_SHOWWINDOW)
    main()