Advertisement
Not a member of Pastebin yet?
Sign Up,
it unlocks many cool features!
- from urllib.request import urlopen as uReq
- from bs4 import BeautifulSoup as soup
- # import csv
- # csv_file = open('travelplanet_scrape.csv', 'w', newline = '')
- # csv_writer = csv.writer(csv_file, dialect='excel')
- # csv_writer.writerow(['nazwa', 'procenty', 'cena', 'ocena', 'gwiazdek', 'kraj', 'region', 'url_oferta'])
- for i in range(1, 3):
- print("strona: " + str(i))
- my_url = 'https://www.travelplanet.pl/wczasy/oferty/' + str(i) + '/?wylot=17.03.2019&przylot=07.04.2019&osoby=2&czas=0:100&wyzywienie=1&ocena=3&dojazd=F&sortowanie=1&kolejnosc=up&limit=25'
- uClient = uReq(my_url)
- page_html = uClient.read()
- uClient.close()
- page_soup = soup(page_html, "html.parser")
- oferty = page_soup.findAll("section", {"class": "offer-item offer-item-with-slider"})
- oferta = oferty[0]
- for oferta in oferty:
- nazwa = oferta.find("a", {"class": "oi-h-link"})['title']
- cena = oferta.find("span", {"class":"pb-hp-onepreson"}).text
- oferta_procenty = oferta.find("span", {"class": "oi-ff-pd-discount-precentage-label"})
- if oferta_procenty != None:
- procenty = oferta.find("span", {"class": "oi-ff-pd-discount-precentage-label"}).text
- # if procenty != None:
- # pass
- # cena = oferta.find("span", {"class":"pb-hp-onepreson"}).text
- # url_oferta1 = oferta.find("a", {"class": "imgCont wt"})['href']
- # url_oferta = f'https://www.wakacje.pl{url_oferta1}'
- # kraj1 = oferta.find("span", {"class": "country"}).text
- # kraj = kraj1.split('/')[0]
- # region = kraj1.split('/')[1]
- # ocena = oferta.find("b", {"class": "rating"}).text
- # gwiazdek = oferta.find("span", {"class": "sprite hotelCategory-s"})['data-stars']
- # csv_writer.writerow([nazwa, procenty, cena, ocena, gwiazdek, kraj, region, url_oferta])
- print("nazwa: " + nazwa)
- print("procenty: " + procenty)
- print("cena: " + cena)
- # csv_file.close()
Advertisement
Add Comment
Please, Sign In to add comment
Advertisement