关于nginx:工具pdf在线转换并保存

import requests
import time

class PDF2Word():
    def __init__(self):
        self.machineid = 'ccc052ee5200088b92342303c4ea9399'
        self.token = ''self.guid =''
        self.keytag = ''
    def produceToken(self):
        url = 'https://app.xunjiepdf.com/api/producetoken'
        headers = {'User-Agent': 'Mozilla/5.0 (Windows NT 6.3; Win64; x64; rv:76.0) Gecko/20100101 Firefox/76.0',
            'Accept': 'application/json, text/javascript, */*; q=0.01',
            'Accept-Language': 'zh-CN,zh;q=0.8,zh-TW;q=0.7,zh-HK;q=0.5,en-US;q=0.3,en;q=0.2',
            'Content-Type': 'application/x-www-form-urlencoded; charset=UTF-8',
            'X-Requested-With': 'XMLHttpRequest',
            'Origin': 'https://app.xunjiepdf.com',
            'Connection': 'keep-alive',
            'Referer': 'https://app.xunjiepdf.com/pdf2word/',}
        data = {'machineid':self.machineid}
        res = requests.post(url,headers=headers,data=data)
        res_json = res.json()
        if res_json['code'] == 10000:
            self.token = res_json['token']
            self.guid = res_json['guid']
            print('胜利获取 token')
            return True
        else:
            return False
    def uploadPDF(self,filepath):
        filename = filepath.split('/')[-1]
        files = {'file': open(filepath,'rb')}
        url = 'https://app.xunjiepdf.com/api/Upload'
        headers = {'User-Agent': 'Mozilla/5.0 (Windows NT 6.3; Win64; x64; rv:76.0) Gecko/20100101 Firefox/76.0',
            'Accept': '*/*',
            'Accept-Language': 'zh-CN,zh;q=0.8,zh-TW;q=0.7,zh-HK;q=0.5,en-US;q=0.3,en;q=0.2',
            'Content-Type': 'application/pdf',
            'Origin': 'https://app.xunjiepdf.com',
            'Connection': 'keep-alive',
            'Referer': 'https://app.xunjiepdf.com/pdf2word/',}
        params = (('tasktype', 'pdf2word'),
            ('phonenumber', ''),
            ('loginkey', ''),
            ('machineid', self.machineid),
            ('token', self.token),
            ('limitsize', '2048'),
            ('pdfname', filename),
            ('queuekey', self.guid),
            ('uploadtime', ''),
            ('filecount', '1'),
            ('fileindex', '1'),
            ('pagerange', 'all'),
            ('picturequality', ''),
            ('outputfileextension', 'docx'),
            ('picturerotate', '0,undefined'),
            ('filesequence', '0,undefined'),
            ('filepwd', ''),
            ('iconsize', ''),
            ('picturetoonepdf', ''),
            ('isshare', '0'),
            ('softname', 'pdfonlineconverter'),
            ('softversion', 'V5.0'),
            ('validpagescount', '20'),
            ('limituse', '1'),
            ('filespwdlist', ''),
            ('fileCountwater', '1'),
            ('languagefrom', ''),
            ('languageto', ''),
            ('cadverchose', ''),
            ('pictureforecolor', ''),
            ('picturebackcolor', ''),
            ('id', 'WU_FILE_1'),
            ('name', filename),
            ('type', 'application/pdf'),
            ('lastModifiedDate', ''),
            ('size', ''),)
        res= requests.post(url,headers=headers,params=params,files=files)
        res_json = res.json()
        if res_json['message'] == '上传胜利':
            self.keytag = res_json['keytag']
            print('胜利上传 PDF')
            return True
        else:
            return False
    def progress(self):
        url = 'https://app.xunjiepdf.com/api/Progress'
        headers = {'User-Agent': 'Mozilla/5.0 (Windows NT 6.3; Win64; x64; rv:76.0) Gecko/20100101 Firefox/76.0',
            'Accept': 'text/plain, */*; q=0.01',
            'Accept-Language': 'zh-CN,zh;q=0.8,zh-TW;q=0.7,zh-HK;q=0.5,en-US;q=0.3,en;q=0.2',
            'Content-Type': 'application/x-www-form-urlencoded; charset=UTF-8',
            'X-Requested-With': 'XMLHttpRequest',
            'Origin': 'https://app.xunjiepdf.com',
            'Connection': 'keep-alive',
            'Referer': 'https://app.xunjiepdf.com/pdf2word/',}
        data = {
            'tasktag': self.keytag,
            'phonenumber': '','loginkey':'',
            'limituse': '1'}
        res= requests.post(url,headers=headers,data=data)
        res_json = res.json()
        if res_json['message'] == '解决胜利':
            print('PDF 解决实现')
            return True
        else:
            print('PDF 解决中')
            return False
    def downloadWord(self,output):
        url = 'https://app.xunjiepdf.com/download/fileid/%s'%self.keytag
        res = requests.get(url)
        with open(output,'wb') as f:
            f.write(res.content)
            print('PDF 下载胜利 ("%s")'%output)
    def convertPDF(self,filepath,outpath):
        filename = filepath.split('/')[-1]
        filename = filename.split('.')[0]+'.docx'
        self.produceToken()
        self.uploadPDF(filepath)
        while True:
            res = self.progress()
            if res == True:
                break
            time.sleep(1)
        self.downloadWord(outpath+filename)

应用：

if __name__=='__main__':
    pdf2word = PDF2Word()
    pdf2word.convertPDF('002.pdf','')

后果：

瑕疵：

如上图，容易把一些段落解析成表格，没有间接应用 wps 来的准确。