Dmitriy Zub ☀️

Posted on Jul 9, 2021 • Edited on Dec 4, 2021

Scrape Google Recipe Results with Python

#python #webscraping #tutorial #datascience

Intro

This blog post is a continuation of Google's web scraping series. Here you'll see how to scrape Google Recipe Results from Organic Results using Python with beautifulsoup, requests, lxml libraries. An alternative API solution will be shown.

Imports

import requests, lxml
from bs4 import BeautifulSoup
from serpapi import GoogleSearch

What will be scraped

Process

SelectorGadget Chrome extension was used to grab CSS selectors.

Selecting Container, Title, Rating, Reviews, Time to cook, Ingredients, Source

Selecting URL

Code

import requests, lxml
from bs4 import BeautifulSoup

headers = {
    "User-Agent":
    "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/70.0.3538.102 Safari/537.36 Edge/18.19582"
}

params = {"q": "lasagna recipe", "hl": "en", 'gl': 'us'}

response = requests.get("https://www.google.com/search", params=params, headers=headers)
soup = BeautifulSoup(response.text, 'lxml')

for result in soup.select('.cv2VAd'):
    try:
        title = result.select_one('.hfac6d').text
    except: title = None

    source = result.select_one('.KuNgxf').text
    total_time = result.select_one('.z8gr9e').text

    # returns a list() if need to extract certain ingredient
    ingredients = result.select_one('.LDr9cf').text.split(',')
    rating = result.select_one('.YDIN4c').text
    reviews = result.select_one('.HypWnf').text.replace('(', '').replace(')', '')

    print(f'{title}\n{source}\n{total_time}\n{rating}\n{reviews}\n{ingredients}\n')

--------------
'''
World's Best Lasagna
Allrecipes
3 hr 15 min
4.8
19K
['Sweet italian sausage', ' lean ground beef', ' ricotta cheese', ' tomato sauce', ' lasagna noodles']
'''

Extracting all recipes

If you want to extend this program and extract every recipe that Google provides, then it's time for Selenium.

I believe that there're several ways to make it work, I'll show one approach.

To get actual clicking until there's nothing to click on we need to use a while loop in combination with .is_displayed() method which returns True or False.

The logic is that it clicks on the Show more button and once there's nothing to click on it breaks out of the while loop and begin to scrape all recipes that where loaded during clicking:

While loop snippet explanation:

while True:
    # locates show more button element
    show_more_button = driver.find_element_by_xpath('//*[@id="isl_13"]/div[5]/div[2]').is_displayed()
    time.sleep(1)
    try:
        # clicks on show more button
       WebDriverWait(driver, 20).until(EC.element_to_be_clickable((By.XPATH, '//*[@id="isl_13"]/div[5]/div[2]'))).click()
    except:
        # pass were used to ignore exception error and continue execution
        pass

    # if show_more_button element becomes False, breaks the loop
    if show_more_button == False:
        break

# under the hood it looks like this
'''
True
True
True
True
True
True
True
True
True
True
True
True
False # breaks from the loop
'''

Full code:

import time
from selenium import webdriver
from selenium.webdriver.chrome.options import Options
from selenium.webdriver.support.wait import WebDriverWait
from selenium.webdriver.support import expected_conditions as EC
from selenium.webdriver.common.by import By

driver = webdriver.Chrome(executable_path='path/to/chromedriver.exe')
driver.get('https://www.google.com/search?q=lasagna+recipe&hl=en')

# buffer for everything to load
time.sleep(10)

while True:

    # returns True or False. If element is not displayed (False), breaks out of the while loop

    # pardon my french (xpath)
    show_more_button = driver.find_element_by_xpath('/html/body/div[7]/div/div[10]/div[1]/div/div[2]/div[2]/div/div/div[1]/div/div/g-section-with-header/div[2]/g-expandable-container/div/div/div/div[5]/div[2]').is_displayed()
    print(show_more_button) # just for debug
    time.sleep(1)
    try:
        WebDriverWait(driver, 5).until(EC.element_to_be_clickable((By.XPATH, '/html/body/div[7]/div/div[10]/div[1]/div/div[2]/div[2]/div/div/div[1]/div/div/g-section-with-header/div[2]/g-expandable-container/div/div/div/div[5]/div[2]'))).click()
    except:
        pass

    # if show_more_button element becomes False, break the loop
    if show_more_button == False:
        break

for index, result in enumerate(driver.find_elements_by_css_selector('.cv2VAd')):
    try:
        title = result.find_element_by_css_selector('.hfac6d').text 
    except: title = None

    link = result.find_element_by_css_selector('.cv2VAd .v1uiFd a').get_attribute('href') 
    source = result.find_element_by_css_selector('.KuNgxf').text

    try:
        total_time = result.find_element_by_css_selector('.wHYlTd').text
    except:
        total_time = None

    try:
        # stays the list if need to extract certain ingredient
        ingredients = result.find_element_by_css_selector('.LDr9cf').text.split(',')
    except:
        ingredients = None

    try:
        rating = result.find_element_by_css_selector('.YDIN4c').text
    except:
        rating = None

    try:
        reviews = result.find_element_by_css_selector('.HypWnf').text.replace('(', '').replace(')', '')
    except:
        reviews = None

    print(f'{index + 1}\n{title}\n{link}\n{source}\n{total_time}\n{ingredients}\n{rating}\n{reviews}\n')

driver.quit()


-----------
'''
1 # first element
World's Best Lasagna
https://www.allrecipes.com/recipe/23600/worlds-best-lasagna/
Allrecipes
3 hr 15 min
['Sweet italian sausage', ' lean ground beef', ' ricotta cheese', ' tomato sauce', ' lasagna noodles']
4.8
19K

...

102 # last element
Last Minute Red Lasagna
https://www.101cookbooks.com/last-minute-red-lasagna-recipe/
101 Cookbooks
40 min
['Whole wheat no', ' red lentils', ' red pepper flakes', ' fresh pasta sheets', ' mozzarella cheese']
4.4
29
'''

Note that GIF is sped-up

Using Google Recipes Results API

SerpApi is a paid API with a free plan.

The difference is that you get structured JSON with correctly formatted data, for example, 19K reviews will be 19000 which is useful if you're doing some sort of analysis and don't have to waste time on making conversion. Also, you get access to thumbnails data if you need them.

Note: it scrapes only 3 results, just like with beautifulsoup.

All recipe results is currently under development (check out in the docs if it's already implemented)

import json # just for pretty output
from serpapi import GoogleSearch

params = {
  "api_key": "YOUR_API_KEY",
  "engine": "google",
  "q": "lasagna recipe",
  "gl": "us",
  "hl": "en"
}

search = GoogleSearch(params)
results = search.get_dict()

for result in results['recipes_results']:
    print(json.dumps(result, indent=2, ensure_ascii=False))

-------------
'''
{
  "title": "World's Best Lasagna",
  "link": "https://www.allrecipes.com/recipe/23600/worlds-best-lasagna/",
  "source": "Allrecipes",
  "rating": 4.8,
  "reviews": 19000,
  "total_time": "3 hr 15 min",
  "ingredients": [
    "Sweet italian sausage",
    "lean ground beef",
    "ricotta cheese",
    "tomato sauce",
    "lasagna noodles"
  ],
  "thumbnail": "https://serpapi.com/searches/60e58864ad7fa97ae27832be/images/b6c5f341384d29417e5d9a3b87dba8a0a0c34837ef00120357584c23f98de567.jpeg"
}
'''

Acknowledgements

Evan (Evaggelos-EvanE) for showing repl code issue and his solution.

Links

Code in the online IDE • Google Recipes Results API

Outro

If you have any questions or something isn't working correctly or you want to write something else, feel free to drop a comment in the comment section or via Twitter at @serp_api.

Yours,
Dimitry, and the rest of SerpApi Team.

Top comments (2)

Hasanul Islam • Jul 5 '22

Do you think selecting by css selector is a good idea? Because, it changes frequently.

Dmitriy Zub ☀️ • Jul 6 '22

Hi @mhihasan, thank you for the comment🙂

It is not a particulary good idea unless you want to fix erros in the parser on regular basis when CSS selector is changed🙂

In case of recipe results, there's no other option to parse data from i.e there's no way to parse data from inline <script> tags or from XHR or other type of request. Only from CSS selectors.

We can can use jscontroller/jsname/jsaction attributes if possible instead since they're changing less frequently or not at all.

DEV Community