Steps and methods to implement automated collection tasks using Python and pywinauto

Automated collection tasks based on pywinauto

Implementation technology

This program uses a Python automation library----pywinauto, because the official version has not been updated for a long time, so the python version The highest can only be around Python 3.7, I use Python 3.7.1. I used it to simulate entering words, copying example sentences, getting example sentences, clearing the clipboard, and then repeating this operation. Overall, the implementation is relatively crude. Moreover, in order Simple, I manually switch to the example sentence page, so I don’t have to use a program to switch to the example sentence page.

Code

requirements.txt

pyperclip==1.8.2
pywin32==304
pywinauto==0.6.8

Code

import os
import random
import time
import re
from typing import Dict, List
from pywinauto.application import Application
from pywinauto import mouse
from pywinauto import keyboard
import pyperclip
import json


# 程序处理中的各种路径
dir_path = r"C:/Users/Dick/Desktop/work/DragonEnglish/tools"
input_path = os.path.join(dir_path, r"input.txt")
output_path = os.path.join(dir_path, r"output.json")
error_path = os.path.join(dir_path, r"error.txt")
# 顺序错误的单词
error_words = []
# 有道词典的进程id
processId = 13840


def line_process(content: str) -> str:
    """
    去除所有空行, 再去除前面四行无关内容
    """
    lines = content.split("\r\n")
    # 因为例句开头是 数字. 开头的, 所以先以这个为特点来进行过滤掉多复制的开头
    count = 0
    for i in range(len(lines)):
        if re.match(r"\d+\.", lines[i]):
            count = i
            break

    lines = lines[count:]
    filter_lines = []
    for line in lines:
        if line.strip() != "":  # 过滤空行
            if not line.startswith("youdao") and not \
                    (line.startswith("《") and line.endswith("》")):  # 过滤来源
                filter_lines.append(line)

    if len(filter_lines) % 3 != 0:
        raise Exception("抓取数据错误")

    content = "\n".join(filter_lines) + "\n"  # 补上一个 \n, 不然正则会漏掉一个结果
    return content


def to_list(line: str) -> List[Dict[str, str]]:
    """
    直接生成列表字典对象
    [{
        "no": 1,
        "original": "",
        "translate"
    }]
    """
    sentences = []
    # 正则表达式
    REGEXP = r&#39;(?P<no>\d+?)\.\n(?P<original>.+?)\n(?P<translate>.+?)\n&#39;
    # 编译
    pattern = re.compile(REGEXP)
    # 匹配
    rs = pattern.finditer(line)
    # 组装结果
    for r in rs:
        print(r.groupdict())
        sentences.append(r.groupdict())
    return sentences


if __name__ == "__main__":
    # 连接网易有道词典
    app = Application(backend="uia").connect(process=processId)
    # 获取需要的窗口
    win = app.window(class_name="RICHEDIT50W")

    # 输入词汇列表
    input_words = []
    # 输出词汇对象列表
    output_words = []
    # 打开输入文件，初始化输入词汇列表
    with open(input_path, "r", encoding="utf-8") as input_file:
        input_words = input_file.read().split("\n")

    for word in input_words:
        print("正在抓取单词: %s" % word)
        # 清空剪切板，这步很重要，防止重复复制
        pyperclip.copy("")
        # 将输入数据复制到剪切板
        pyperclip.copy(word)
        # 定位到输入框（采用坐标定位，定位到大致位置即可）
        mouse.click(coords=(2400, 80))
        # 模拟按键操作：全选 删除 粘贴 回车（触发查询）
        keyboard.send_keys("^a{DELETE}^v{ENTER}")
        # 清空剪切板，这步很重要，防止重复复制
        pyperclip.copy("")
        # 鼠标左键点击，这个操作只是为了把鼠标移动到这里
        mouse.click(button="left", coords=(2200, 330))
        # 模拟键盘 CTRL+A CTRL+C，直接全选所有的例句（这里会多选一部分内容，待会再处理）
        keyboard.send_keys("^a^c")
        # 暂停一会儿，不做操作的太快
        time.sleep(random.random() * 2 + 1)
        # pywinauto 复制的内容是在系统的剪切板里面的，所以需要其它库读取
        content = pyperclip.paste()
        # 对内容进行简单的预处理后，加入output_words
        try:
            lines = line_process(content)
        except BaseException as exp:
            print(exp)
            # 如果抓取出现问题，说明被网易抓了现行，直接退出即可。
            break

        sentences = to_list(lines)
        if not sentences:
            print("获取例句为空, 可能是数据格式错误.")
            break
        output_words.append({
            "word": word,
            "example": sentences,
        })
        # 模拟暂停一个较长的随机时间，没有必要追求速度，平稳运行即可。
        time.sleep(random.random() * 3 + 3)
        # 清空剪切板，这步很重要，防止重复复制
        pyperclip.copy("")

    # 抓取完毕一个文件的内容后，然后一次性写入即可。
    # 之前的写法是一个单词写入一次，会造成太多的IO次数，浪费性能！
    with open(output_path, "a+", encoding="utf-8") as output_file:
        output_file.write(json.dumps(
            output_words, ensure_ascii=False, indent=4))

        # 错误单词记录
        with open(error_path, "w", encoding="utf-8") as err_file:
            err_file.writelines("\n".join(error_words))

Demonstration If you want to start this code, it is quite complicated. I will just list the steps required here, hoping to help. Interested students.

1. Modify dir_path and prepare an input.txt file below.

2. Get the ID of the Youdao Dictionary process.

3. Get the coordinates of the word input box and get the coordinates of the copy and paste location.

4. Adjust the Youdao dictionary interface to the example sentence.

To start the project, you need an input.txt file. Here is the file I tested.

##sophisticated

centralization
phenomenon
internationalization
radioactive

I obtained the process pid through the task manager, you can also access it through other methods. Or the simplest is to use Inspect and Spy, I am lazy here , how to save trouble directly.

The coordinates of the word input box, the coordinates of the copy and paste location. The first coordinate is to position the input box, and then the program will Copy the word and press the Enter key, and then the content will be queried. Then move the mouse to the second coordinate, here just move it to the blank space below, and then perform a select-all CTRL A operation. This way All the content of a word has been obtained.

Adjust Youdao to this position, first query a word, select an example sentence, and then keep this interface unchanged.

The last step is the execution of the program. The recorded GIF has been accelerated. In fact, when it is executed, a delay is deliberately added to prevent it from being discovered prematurely. .

Console output

##output.json file

The above is the detailed content of Steps and methods to implement automated collection tasks using Python and pywinauto. For more information, please follow other related articles on the PHP Chinese website!