Python 3.x 在python中通过抓取子URL下载文件

我正在尝试从大量的web链接下载文档（主要是pdf格式），如下所示：

但是，这些链接无法直接访问pdf文件。您需要单击子URL来访问PDF。有没有办法抓取子URL并从中下载所有相关文件？我正在用下面的代码进行尝试，但到目前为止还没有任何成功，特别是对于这里列出的这些URL

如果您需要进一步澄清，请告诉我。我很乐意这样做。多谢各位

from simplified_scrapy import Spider, SimplifiedDoc, SimplifiedMain, utils

class MySpider(Spider):
    name = 'download_pdf'
    allowed_domains = ["www.worldbank.org"]
    start_urls = [
        "https://projects.worldbank.org/en/projects-operations/document-detail/P167897?type=projects",
        "https://projects.worldbank.org/en/projects-operations/document-detail/P173997?type=projects",
        "https://projects.worldbank.org/en/projects-operations/document-detail/P166309?type=projects"
    ]  # Entry page

    def afterResponse(self, response, url, error=None, extra=None):
        if not extra:
            print ("The version of library simplified_scrapy is too old, please update.")
            SimplifiedMain.setRunFlag(False)
            return
        try:
            path = './pdfs'
            # create folder start
            srcUrl = extra.get('srcUrl')
            if srcUrl:
                index = srcUrl.find('year/')
                year = ''
                if index > 0:
                    year = srcUrl[index + 5:]
                    index = year.find('?')
                    if index>0:
                        path = path + year[:index]
                        utils.createDir(path)
            # create folder end

            path = path + url[url.rindex('/'):]
            index = path.find('?')
            if index > 0: path = path[:index]
            flag = utils.saveResponseAsFile(response, path, fileType="pdf")
            if flag:
                return None
            else:  # If it's not a pdf, leave it to the frame
                return Spider.afterResponse(self, response, url, error, extra)
        except Exception as err:
            print(err)

    def extract(self, url, html, models, modelNames):
        doc = SimplifiedDoc(html)
        lst = doc.selects('div.list >a').contains("documents/", attr="href")
        if not lst:
            lst = doc.selects('div.hidden-md hidden-lg >a')
        urls = []
        for a in lst:
            a["url"] = utils.absoluteUrl(url.url, a["href"])
            # Set root url start
            a["srcUrl"] = url.get('srcUrl')
            if not a['srcUrl']:
                a["srcUrl"] = url.url
            # Set root url end
            urls.append(a)

        return {"Urls": urls}

    # Download again by resetting the URL. Called when you want to download again.
    def resetUrl(self):
        Spider.clearUrl(self)
        Spider.resetUrlsTest(self)

SimplifiedMain.startThread(MySpider())  # Start download

---

有一个API端点包含您在网站上看到的整个响应以及。。。文档的URL

pdf

：D

因此，您可以查询API，获取URL，最后获取文档

以下是方法：

导入请求
pids=[“P167897”、“P173997”、“P166309”]
对于pid中的pid：
终点=f“https://search.worldbank.org/api/v2/wds?" \
f“format=json&includepublicdocs=1&”\
f“fl=docna，lang，docty，repnb，docdt，doc\u authr，可在&”\
f“os=0&rows=20&proid={pid}&apilang=en”
documents=requests.get（end_point）.json（）[“documents”]
对于documents.values（）中的document_数据：
尝试：
pdf\u url=文件\u数据[“pdfurl”]
打印（f“获取：{pdf_url}”）
以pdf格式打开（pdf_url.rsplit（“/”[-1]，“wb”）：
write（requests.get（pdf_url.content）
除KeyError外：
持续

输出：（完全下载的.pdf文件）

---

谢谢这个@baduker。我以前试过使用他们的API，但问题是下载会因请求大量文档而中断。是否有一种方法可以从中断发生的地方重新启动，否则所有内容都需要从头开始重新下载？当然有。它被称为重试逻辑，但您必须自己实现。提示：使用while循环和

try execpt

块，在请求之间暂停。

Fetching: http://documents.worldbank.org/curated/en/106981614570591392/pdf/Official-Documents-Grant-Agreement-for-Additional-Financing-Grant-TF0B4694.pdf
Fetching: http://documents.worldbank.org/curated/en/331341614570579132/pdf/Official-Documents-First-Restatement-to-the-Disbursement-Letter-for-Grant-D6810-SL-and-for-Additional-Financing-Grant-TF0B4694.pdf
Fetching: http://documents.worldbank.org/curated/en/387211614570564353/pdf/Official-Documents-Amendment-to-the-Financing-Agreement-for-Grant-D6810-SL.pdf
Fetching: http://documents.worldbank.org/curated/en/799541612993594209/pdf/Sierra-Leone-AFRICA-WEST-P167897-Sierra-Leone-Free-Education-Project-Procurement-Plan.pdf
Fetching: http://documents.worldbank.org/curated/en/310641612199201329/pdf/Disclosable-Version-of-the-ISR-Sierra-Leone-Free-Education-Project-P167897-Sequence-No-02.pdf

and more ...