Extract all cryptocurrencies data from coinmarketcap.com using Python 3 and Request

coinmarketcap_extract.py

#!/usr/bin/python3
# coding: utf-8

import requests
from bs4 import BeautifulSoup
from scrapy import Selector
import csv
import datetime


def extract(url):
    """
    Export all cryptodata from cryptomarketcap.com
    website
    Arguments:
         url (str):
            url of the aimed Coinmarketcap page
    Return:
        .csv file
    """

    # INITIALISATION
    r = requests.session()
    start = datetime.datetime.now()

    # COLLECTE DU CODE SOURCE
    for retry in range(10):  # ajout d'un retrier en cas de site inaccessible
        response = r.get(url=url)
        print(response.headers)
        print("-- STATUS CODE --")
        print(response.status_code)

        # PARSING ET CREATION DU CSV
        if response.status_code == 200:
            with open("/path/to/coinmarketcap/cryptocurrencies_{}.csv".format(str(datetime.date.today())), "w") as f:
                fieldnames = ['Nom', 'Symbole', 'Cap. marché', 'Prix', 'Offre en circulation', 'Volume (24h)', '% 1h', '% 24h', '7 j']
                writer = csv.DictWriter(f, fieldnames=fieldnames, delimiter='\t')
                writer.writeheader()

                soup = BeautifulSoup(response.text, features='html.parser')
                sel = Selector(text=soup.prettify())
                cryptos = sel.xpath("//tr[contains(@id, 'id-')]").extract()
                print(cryptos)
                for crypto in cryptos:
                    soup = BeautifulSoup(crypto, features='html.parser')
                    sel = Selector(text=soup.prettify())

                    nom = sel.xpath("//td[contains(@class, 'currency-name')]/@data-sort").extract_first()
                    symbole = sel.xpath("//td[contains(@class, 'col-symbol')]/text()").extract_first()
                    cap_marche = sel.xpath("//td[contains(@class, 'market-cap')]/text()").extract_first()
                    prix = sel.xpath("//a[@class='price']/@data-usd").extract_first()
                    offre_circulation = sel.xpath("//a[@class='volume']/@data-usd").extract_first()
                    volume = sel.xpath("//td[contains(@class, 'circulating-supply')]/@data-sort").extract_first()
                    percent_1h = sel.xpath("//td[@data-timespan='1h']/@data-sort").extract_first()
                    percent_24h = sel.xpath("//td[@data-timespan='24h']/@data-sort").extract_first()
                    percent_7j = sel.xpath("//td[@data-timespan='7d']/@data-sort").extract_first()

                    clean_values = []
                    values = [nom, symbole, cap_marche, prix, offre_circulation, volume, percent_1h, percent_24h, percent_7j]
                    for value in values:
                        if value:
                            value = value.strip().replace('\n', '')
                        clean_values.append(value)

                    print(', '.join(clean_values))
                    dict_row = dict(zip(fieldnames, clean_values))
                    writer.writerow(dict_row)

            # TEMPS PASSE
            end = datetime.datetime.now()
            time_elapsed = str(end - start)
            print('\n')
            print('-- TIME ELAPSED --')
            print(time_elapsed)
            break

        elif response.status_code == 404:
            print("Page indisponible")
            break

        else:
            print("Impossible d'accéder à la page")
            return []


def main():
    url = "https://coinmarketcap.com/fr/all/views/all/"
    extract(url)


if __name__ == '__main__':
    main()

Same issue here:( is there any thing to do to fix it ?

same here

The code is correct in principle but the problem is the server side has such screen scraping against terms of service, and so obfuscates the html and xml class tags and id tags, which causes a failure because the HTML has changed format and the selectors don't find anything, Then the beautifulsoup loop returns no matching criteria.

Here's some hints to hack the breath of life back into it:

cryptos = sel.xpath("//tr[contains(@class, 'cmc-table-row')]").extract()

and

nom = sel.xpath("//a[contains(@class, 'cmc-table__column-name--name')]/text()").extract_first()

Works for me at the moment, but by the time you read this, it will have probably changed again. Better to reach into your wallet and pay cash-money for the unlimited coinbase API.

See: https://gist.github.com/sentientmachine/820e0c47b07fe90580e4c758d9c5df97

Update: scraping coinmarketcap is too much work. Sustainable solution is to get an account at an exchange like coinbase, get a developer API key and get that way. Mostly plug and play, much easier than playing calvinball with this html javascript https://developers.coinbase.com/docs/wallet/guides/price-data

@sentientmachine thanks