Zotero PDF翻译插件自定义翻译功能实现

一 需求

在使用Zotero PDF翻译插件的时候，感觉总是受到一些限制。Google翻译在国内无法使用，百度、deepl等翻译，都需要key。所以很大程度上，无法直接进行划词翻译。最常见的操作就是复制到浏览器中进行翻译。现在的需求就是：
** 能不能将复制这个动作自动化？**

我的思路就是：
对Zotero 插件进行修改，使其能够支持自定义的服务器访问。然后将需要翻译的内容，在服务器上，利用Pytho的爬虫功能，实现自动化的操作浏览器，并将结果返回到Zotero。

二、实现演示

三、Zetero翻译插件的功能定制

3.1 开发环境

nodejs

3.2 开发步骤

依据官网的介绍，开发的具体步骤如下：

插件是基于Zotero插件模板.

可以运行如下命令进行构建：

git clone https://github.com/windingwind/zotero-pdf-translate.git
cd zotero-pdf-translate
npm install
npm run build

插件生成在： ./builds/*.xpi.

添加新的翻译服务：

1. 添加新的服务配置到： src/utils/config.ts > SERVICES;
2. 添加一个与其他服务格式相同的新的任务处理过程到src/modules/services/${serviceId}.ts 。 如果>程序运行成功，导出函数将翻译的结果放在 data.result 中，如果运行失败，则抛出一个异常;
3. 在src/modules/services.ts中导入任务处理函数
4. 在addon/chrome/locale/${lang}/addon.properties中添加本地语言的服务名称： service.${serviceId}
5. 构建与测试

3.3 插件开发实现

3.3.1 添加服务

在这里相当于要用建立一个自定义服务的对象

3.3.2 编写任务处理程序

文件位置如下图所示：

图中代码如下：

import { TranslateTask, TranslateTaskProcessor } from "../../utils/translate";

export default <TranslateTaskProcessor>async function (data) {
    const url = data.secret;
    const reqBody = `data=${encodeURIComponent(data.raw)}&from=${encodeURIComponent(data.langfrom)}&to=${encodeURIComponent(data.langto)}`;
    const xhr = await Zotero.HTTP.request("POST", url, {
    responseType: "json",
    body: reqBody,
  });
  if (xhr?.status !== 200) {
    throw `Request error: ${xhr?.status}`;
  }
  data.result = xhr.response.data;
};

3.3.3 导入任务处理程序

3.3.4 添加服务名称

3.3.5 构建

构建命令如下：

npm run build

构建后的插件目录：

插件安装后的效果：

四、服务器的开发

4.1 环境

python:

1. flask作为服务器
2. selenium:作为自动化工具

4.2 代码实现

from flask import Flask, jsonify,request
import logging
from urllib.parse import urlparse
import seleniumwire.undetected_chromedriver as uc
from  selenium import webdriver
from selenium.webdriver.common.by import By
from selenium.common.exceptions import TimeoutException
from selenium.webdriver.support.wait import WebDriverWait
from selenium.webdriver.support import expected_conditions as EC
import pyperclip
import time
import os
app = Flask(__name__)

class Translate():
    def __init__(self,options=None,sw_options={},isUC=False) -> None:

        self.opened_domain = {}

        if options is None:
            options = uc.ChromeOptions()
        assert isinstance(options,uc.ChromeOptions)

        self.options = options
        self.sw_options = sw_options
        if isUC:
            self.driver = uc.Chrome(
                options=self.options,
                seleniumwire_options=self.sw_options
            )
        else:
            self.driver = webdriver.Chrome(options=self.options)
    
    def openUrl(self,url):
        domain = urlparse(url).netloc
        handles = self.driver.window_handles
        # 判断当前浏览器是否打开
        if len(self.opened_domain) < 1:
            self.driver.get(url)
            self.opened_domain[domain] = self.driver.current_window_handle
        elif domain not in  self.opened_domain.keys():
            self.driver.execute_script(f"window.open('{url}', '_blank')")
            wait = WebDriverWait(self.driver, timeout=10, poll_frequency=1, ignored_exceptions=[TimeoutException])    
            wait.until(EC.new_window_is_opened(handles))
            new_handles = self.driver.window_handles
            new_handle = [handle for handle in new_handles if handle not in handles]
            assert new_handle is not []
            self.opened_domain[domain] = new_handle[0]
        self.driver.switch_to.window(self.opened_domain[domain])        
    def translate(self,url:str,data:str,transFun:callable):
        self.openUrl(url)
        if transFun is None:
            print("翻译函数为空！！！")
            return
        result = transFun(data)
        # data = {'data': result}
        return result
        # return jsonify(data)
    
    def deepl(self,data):
        if EC.visibility_of_element_located((By.XPATH, '//*[@id="panelTranslateText"]/div[1]/div[2]/section[1]/div[3]/div[2]/d-textarea/div'))(self.driver):
            clear_button = self.driver.find_element(By.XPATH, '//*[@id="panelTranslateText"]/div[1]/div[2]/section[1]/div[3]/div[2]/d-textarea/div')
            clear_button.click()
        lang = self.driver.find_element(By.TAG_NAME,"html").get_attribute('lang')
        if lang == "en":
            tag_text = "Like translation"
            copy_text = "Copy to clipboard"
        elif lang =="zh":
            tag_text = "喜欢该翻译"
            copy_text = "复制到剪贴板"
        
        el_input = WebDriverWait(self.driver, 10).until(EC.presence_of_element_located((By.XPATH, '//*[@id="panelTranslateText"]/div[1]/div[2]/section[1]/div[3]/div[2]/d-textarea/div')))
        el_input.clear()
        el_input.click()
        el_input.send_keys(data)
        WebDriverWait(self.driver, timeout=10).until(lambda d: d.find_element(By.CSS_SELECTOR,f'button[aria-label="{tag_text}"]'))
        time.sleep(2)
        copy_button = self.driver.find_element(By.CSS_SELECTOR,f'button[data-testid="translator-target-toolbar-copy"][aria-label="{copy_text}"]')                
        copy_button.click()
        result = pyperclip.paste()
        return result
    
    def deeplFY(self,input,_from="en",_to="zh"):
        base_url  = f"https://www.deepl.com/translator#{_from}/{_to}/"
        result = self.translate(base_url,input,self.deepl)
        return result
        
    def baidu(self,data):
        submit = WebDriverWait(self.driver, 10).until(EC.presence_of_element_located((By.XPATH, '//*[@id="baidu_translate_input"]')))
        submit.clear()
        submit.send_keys(data)
        el_output = WebDriverWait(self.driver, timeout=10).until(lambda d: d.find_element(By.XPATH,'//*[@id="main-outer"]/div/div/div[1]/div[2]/div[1]/div[2]/div/div/div[1]/p[2]'))
        result = el_output.text
        return result
    
    def bdFY(self,data:str="Hello",_from:str="en",_to:str="zh"):
        url = f"https://fanyi.baidu.com/translate#{_from}/{_to}"
        return self.translate(url,data,self.baidu)


transDriver = Translate()
log_file_name = 'logger-' + time.strftime('%Y-%m-%d', time.localtime(time.time())) + '.log'
log_file_path = os.path.join(os.path.abspath(os.curdir),log_file_name)

handler = logging.FileHandler(log_file_path,encoding='UTF-8')
logging_format = logging.Formatter(
    '%(asctime)s-%(message)s')
handler.setFormatter(logging_format)
app.logger.addHandler(handler)
print(f"log location:{log_file_path}")



@app.route('/', methods=['POST','GET'])
def root():
    rq_data = request.form.get("data")
    print(rq_data)
    # 在这里编写API逻辑
    translators = {
        "BAIDUI":'/bdFY',
        "DEEPL":'/deeplFY'
    }
    data = {'data': translators}
    return jsonify(data)

@app.route('/bdFY',methods=['POST'])
def bdFY():
    _data = request.form.get("data")
    _from = request.form.get("from")
    _to = request.form.get("to")
    app.logger.info(f"data:{_data}")
    try:
        result = transDriver.bdFY(_data)
    except Exception as e:
        result = str(e)
    app.logger.info(f"result:{result}")
    return jsonify({'data':result})

@app.route('/deeplFY',methods=['POST'])
def deeplFY():
    _data = request.form.get("data")
    _from = request.form.get("from")
    _to = request.form.get("to")
    app.logger.info(f"data:{_data}")
    try:
        result = transDriver.deeplFY(_data)
    except Exception as e:
        result = str(e)
    app.logger.info(f"result:{result}")
    return jsonify({'data':result})

if __name__ == '__main__':
    
    app.run(port=80)