Downloading files by crawling sub-URLs in python

user_13386866 0
Viewed 90

I am trying to download documents (mainly in pdf) from a large number of web links like the following:

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

However, the pdf files are not directly accessible from these links. One needs to click on sub-URLs to access the pdfs. Is there any way to crawl the sub-URLs and download all the related files from them? I am trying it with the following codes but have not had any success so far specifically for these URLs listed here.

Please let me know if you need any further clarifications. I would be happy to do so. Thank you.

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
python-requests scrapy web-scraping web-crawler python-3.x
1 Answers

There's an API endpoint that contains the entire response you see on the web-site along with... the URL to the document pdf. :D

So, you can query the API, get the URLS, and finally fetch the documents.

Here's how:

import requests

pids = ["P167897", "P173997", "P166309"]

for pid in pids:
    end_point = f"https://search.worldbank.org/api/v2/wds?" \
                f"format=json&includepublicdocs=1&" \
                f"fl=docna,lang,docty,repnb,docdt,doc_authr,available_in&" \
                f"os=0&rows=20&proid={pid}&apilang=en"
    documents = requests.get(end_point).json()["documents"]
    for document_data in documents.values():
        try:
            pdf_url = document_data["pdfurl"]
            print(f"Fetching: {pdf_url}")
            with open(pdf_url.rsplit("/")[-1], "wb") as pdf:
                pdf.write(requests.get(pdf_url).content)
        except KeyError:
            continue

Output: (fully downloaded .pdf files)

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 ...
user_61067910
Related