Das deutsche Python-Forum

Code: Alles auswählen

from bs4 import BeautifulSoup
import requests
import csv

url = f'https://de.globalpetrolprices.com/gasoline_prices/'
url2 = f'https://www.finanzen.net/waehrungsrechner/us-dollar_euro'
output_filepath = '/home/Energiepreis.csv'

daten = requests.get(url)
soup = BeautifulSoup(daten.content, 'html.parser')
table = soup.find('div', attrs={'id': 'outsideLinks'})
table2 = soup.find('div', attrs={'id': 'graphic'})
daten2 = requests.get(url2)
soup2a = BeautifulSoup(daten2.content, 'html.parser')
soup2b = BeautifulSoup(daten2.text, 'html.parser')

kurs_name = soup2b.find_all('span', attrs={'id': 'currency-second-display'})
kurs_name2 = soup2a.find_all('span', attrs={'id': 'currency-second-display'})
kurs_zeile = soup2b.find('span', attrs={'id': 'currency-second-display'})
kurs = kurs_zeile.text.split(" ",4)
a = kurs[3].replace(',','.')
summe_1 = float(a)

with open(output_filepath, 'w', encoding='utf-8') as csv_schreib_datei:
    writer = csv.writer(csv_schreib_datei)
    writer.writerow(['Land','USD', 'EUR'])
    myDict0 = {}

    a = 1
    for i in table.find_all('a'):
        link_name = (i.get('href')).replace('gasoline_prices/','').replace('/','')
        list0 = [link_name]
        myDict1 = {a: link_name}
        print(myDict1)
        a = a + 1

    a = 1
    for x in table2.find_all('div', attrs={'style': 'position: absolute; top: 2px; left: 7px; height: 15px; color: #000000;'}):
        myDict2 = {}
        preis_name = (x.text)
        preis_name = str(preis_name)
        preis_eur = float(preis_name)/ float(a)
        list1=[str(preis_name),str(preis_eur)]
        #data1 = [str(link_name), str(preis_name), str(preis_eur)]
        myDict2 = {a: [str(preis_name),str(preis_eur)]}
        a = a + 1

def merge_two_dicts(x, y):
    z = myDict2.copy()  # start with keys and values of x
    z.update(myDict1)  # modifies z with keys and values of y
    return z

print(merge_two_dicts(myDict1, myDict2))

Code: Alles auswählen

#!/usr/bin/env python3
import csv
from pathlib import Path

import requests
from bs4 import BeautifulSoup

PETROL_PRICES_URL = "https://de.globalpetrolprices.com/gasoline_prices/"
CURRENCY_CONVERTER_URL = (
    "https://www.finanzen.net/waehrungsrechner/us-dollar_euro"
)
OUTPUT_FILE_PATH = Path.home() / "Energiepreis.csv"


def get_soup(url):
    response = requests.get(url)
    response.raise_for_status()
    return BeautifulSoup(response.content, "html.parser")


def get_usd_eur_factor(url):
    text = get_soup(url).find("span", id="currency-first-display").text
    one, usd, equals, factor_text, eur = text.split()
    if (one, usd, equals, eur) != ("1", "USD", "=", "EUR"):
        raise ValueError(f"unexpeded text: {text!r}")

    return float(factor_text.replace(",", "."))


def iter_country_names_and_prices(url):
    soup = get_soup(url)

    country_nodes = soup("a", "graph_outside_link")
    dollar_per_litre_nodes = soup.select("#graphic > div > div > div")
    if len(country_nodes) != len(dollar_per_litre_nodes):
        raise ValueError(
            f"#countries ({len(country_nodes)})"
            f" ≠ #prices ({len(dollar_per_litre_nodes)})"
        )

    return zip(
        (node.text.strip() for node in country_nodes),
        (float(node.text) for node in dollar_per_litre_nodes),
    )


def main():
    usd_eur_factor = get_usd_eur_factor(CURRENCY_CONVERTER_URL)

    rows = (
        [country_name, dollar_price, dollar_price * usd_eur_factor]
        for country_name, dollar_price in iter_country_names_and_prices(
            PETROL_PRICES_URL
        )
    )
    with OUTPUT_FILE_PATH.open("w", encoding="utf-8", newline="") as file:
        writer = csv.writer(file)
        writer.writerow(["Land", "USD", "EUR"])
        writer.writerows(rows)


if __name__ == "__main__":
    main()

Code: Alles auswählen

import csv
import requests
from bs4 import BeautifulSoup
from decimal import Decimal, localcontext
from operator import itemgetter
from pathlib import Path


PRICES_URL = 'https://de.globalpetrolprices.com/gasoline_prices/'
EXCHANGE_RATE_URL = 'https://www.finanzen.net/waehrungsrechner/us-dollar_euro'
CSV_FILE_NAME = 'Energiepreis.csv'


def main():

    with localcontext() as ctx:
        # siehe https://docs.python.org/3/library/decimal.html
        ctx.prec = 4

        # Daten einlesen:
        prices_html = requests.get(PRICES_URL)
        exchange_rate_html = requests.get(EXCHANGE_RATE_URL)

        # Preisdaten finden:
        prices_soup = BeautifulSoup(prices_html.content, 'html.parser')
        countries = prices_soup.find('div', attrs={'id': 'outsideLinks'})
        prices = prices_soup.find('div', attrs={'id': 'graphic'})

        # Wechselkurs finden:
        exchange_rate_soup = BeautifulSoup(exchange_rate_html.text, 'html.parser')
        exchange_rate = Decimal(
            exchange_rate_soup
            .find('span', attrs={'id': 'currency-second-display'})
            .text
            .split(" ",4)[3]
            .replace(',','.')
        )

        country_names = []
        prices_usd_raw = []
        price_infos = []

        for a in countries.find_all('a'):
            # siehe https://docs.python.org/3/library/pathlib.html#pathlib.PurePath.parts
            country_names.append(Path(a.get('href')).parts[1])

        style = 'position: absolute; top: 2px; left: 7px; height: 15px; color: #000000;'
        for price_usd_raw in prices.find_all('div', attrs={'style': style}):
            prices_usd_raw.append(price_usd_raw.text)

        for country_name, price_usd_raw in zip(country_names, prices_usd_raw):
            price_usd = Decimal(price_usd_raw)
            price_eur = price_usd / exchange_rate
            price_infos.append({
                'Land': country_name,
                'USD': price_usd,
                'EUR': price_eur,
            })
        with open(CSV_FILE_NAME, 'a', encoding='utf-8') as csv_file:
            # siehe https://docs.python.org/3/library/csv.html#csv.DictWriter
            csv_writer = csv.DictWriter(csv_file, fieldnames=('Land', 'USD', 'EUR'))
            csv_writer.writeheader()
            csv_writer.writerows(sorted(price_infos, key=itemgetter('Land')))


if __name__ == '__main__':
    main()

Code: Alles auswählen

www.GlobalPetrolPrices.com is licensed under a
Creative Commons Attribution-NonCommercial-NoDerivs 3.0

Code: Alles auswählen

countries = [
    anchor.get('href').replace('gasoline_prices/','').replace('/','')
    for anchor in table.find_all('a')
]

prices_us = []
prices_eur = []
for element in table2.find_all('div', attrs={'style': 'position: absolute; top: 2px; left: 7px; height: 15px; color: #000000;'}):
    price_us = float(element.text)
    price_eur = price_us / exchange_rate
    prices_us.append(price_us)
    prices_eur.append(price_eur)

writer.writerows(zip(countries, prices_us, prices_eur))