scossa/scrapeje
1
1
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Update 'BozzaScrapeJE.py'

Browse Source
This commit is contained in:
scossa
2022-09-30 23:27:10 +02:00
parent f366d10a9f
commit 141f3cbcde
1 changed files with 19 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files

30
BozzaScrapeJE.py
View File

@@ -20,10 +20,13 @@ with open('aleppo.html', 'rb') as f:
soup = BeautifulSoup(page, "html.parser") soup = BeautifulSoup(page, "html.parser")
menu = soup.find(attrs={"data-test-id": "menu-item"}) menu = soup.find(attrs={"data-test-id": "menu-item"})
#Stampa nome ristorante
print("\nRISTORANTE:",soup.title.text)
# alla riga 870 dell html c'è "allergenPhoneNumber":"3389529446"
for menu in soup.find_all(attrs={"data-test-id": "menu-item"}): for menu in soup.find_all(attrs={"data-test-id": "menu-item"}):
att=menu att=menu
#riempie la lista "nome" #riempie la lista "nome"
for att in menu.find(attrs={"data-test-id": "menu-item-name"}): for att in menu.find(attrs={"data-test-id": "menu-item-name"}):
if att != " ": if att != " ":
@@ -37,24 +40,29 @@ for menu in soup.find_all(attrs={"data-test-id": "menu-item"}):
else: else:
desc.append(None) desc.append(None)
# #riempie la lista "npezzi"
# for att in menu.find_next(attrs={"data-test-id": "menu-item-description"}):
# npezzi.append(att.lstrip().splitlines()[0])
#riempie la lista "prezzo" #riempie la lista "prezzo"
for att in menu.find("p", class_="c-menuItems-price notranslate"): for att in menu.find("p", class_="c-menuItems-price notranslate"):
prezzo.append(att.lstrip().splitlines()[0]) prezzo.append(att.lstrip().splitlines()[0])
# #riempie la lista "npezzi"
att=menu.find_all(attrs={"data-test-id": "menu-item-description"})
if att != None:
if menu.text.find("pezzo") > 0 or menu.text.find("pezzi") > 0:
npezzi.append(menu.text.splitlines()[7].lstrip())
else:
npezzi.append(None)
continue
#stampa liste #stampa liste
for x in range(len(nome)): for x in range(len(nome)):
print("\n") print("\n")
print(nome[x]) print(nome[x])
print(desc[x]) print(desc[x])
# print(npezzi[x]) print(npezzi[x])
print(prezzo[x]) print(prezzo[x])
#stampa lunghezza liste #stampa lunghezza liste
print("\nlista nomi:",len(nome)) print("\nlista nomi:",len(nome))
print("lista desc:",len(desc)) print("lista desc:",len(desc))
print("lista npezzi:",len(npezzi)) print("lista npezzi:",len(npezzi))
print("lista prezzi:",len(prezzo)) print("lista prezzi:",len(prezzo))