Web Scraping

 Web Scraping

Some websites don't have lovely APIs for us to interface with.

If we want data from these pages, we have to use a tecnique called scraping. This means downloading the whole webpage and poking at it until we can find the information we want.

You're going to use scraping to get the top ten restaurants near you.

Get started

👉 Go to a website like Yelp and search for the top 10 reastaurants in your location. Copy the URL.

 

url = "https://www.yelp.co.uk/search?find_desc=Restaurants&find_loc=San+Francisco%2C+CA%2C+United+States"  

Import libraries

👉 Import your libraries. Beautiful soup is a specialist library for extracting the contents of HTML and helping us parse them. Run the Repl once your imports are sorted because we want the Beautiful Soup library to be installed (it'll run quicker this way).

import requests

from bs4 import BeautifulSoup

url = "https://www.yelp.co.uk/search?find_desc=Restaurants&find_loc=San+Francisco%2C+CA%2C+United+States"  

Webpage as text

👉 Use requests to get the webpage as text. When printing the html, we get back just how much info there is contained in the page.

import requests

from bs4 import BeautifulSoup

url = "https://www.yelp.co.uk/search?find_desc=Restaurants&find_loc=San+Francisco%2C+CA%2C+United+States"

response = requests.get(url)

html = response.text

print(html)  

More Scraping

👉 Next, we pass the html to Beautiful Soup to make more sense of it.

html.parser will scan through the HTML recognizing tokens in the text and breaking it down into something more meaningful.  

import requests

from bs4 import BeautifulSoup

url = "https://www.yelp.co.uk/search?find_desc=Restaurants&find_loc=San+Francisco%2C+CA%2C+United+States"

response = requests.get(url)

html = response.text

soup = BeautifulSoup(html, 'html.parser')  

Inspect

👉 Back on Yelp, right click the first (non-sponsored) restaurant on the list and inspect it.

Right click and copy element for the whole URL, then paste it into your repl as a temporary measure.  

import requests

from bs4 import BeautifulSoup

url = "https://www.yelp.co.uk/search?find_desc=Restaurants&find_loc=San+Francisco%2C+CA%2C+United+States"

response = requests.get(url)

html = response.text

soup = BeautifulSoup(html, 'html.parser')

<a href="/biz/marufuku-ramen-san-francisco-5?osq=Restaurants" class="css-1m051bw" target="_blank" name="Marufuku Ramen" rel="noopener">Marufuku Ramen</a>  

Inspecting the link gives us clues about what we want beautiful soup to look for. In this case, I want it to look for <a> tags and the class css-1m051bw.

Store results

👉 I've created a new variable to store the result of the beautiful soup search.

find_all takes two arguments. The first is the a tag. The second is a dictionary that tells it what class to search for. This effectively says 'find me all the a tags with this class in them.'

I've printed the len of those results to see how many I get back.

import requests

from bs4 import BeautifulSoup

url = "https://www.yelp.co.uk/search?find_desc=Restaurants&find_loc=San+Francisco%2C+CA%2C+United+States"

response = requests.get(url)

html = response.text

soup = BeautifulSoup(html, 'html.parser')

myLinks = soup.find_all("a", {"class":"css-1m051bw"})

print(len(myLinks))

Loop it

Now I'll use a loop to output all the links.  

import requests

from bs4 import BeautifulSoup

url = "https://www.yelp.co.uk/search?find_desc=Restaurants&find_loc=San+Francisco%2C+CA%2C+United+States"

response = requests.get(url)

html = response.text

soup = BeautifulSoup(html, 'html.parser')

myLinks = soup.find_all("a", {"class":"css-1m051bw"})

print(len(myLinks))

for link in myLinks:

    print(link.text)  

👉 You'll see that the same tag has been used in the info about the location and category, so they're included in the results.

I'm going to use a loop counter to start the print loop at item 3 to leave them off the output.

I'm also going to include the link to the restaurant in the output by using a dictionary.  

 

counter = 0

for link in myLinks:

  if counter > 1:

    print(link.text)

    print(link["href"])

  counter +=1  

Add an fstring

The web links are local, they aren't relative to the site I'm using, so I've formatted the print(link["href"]) as an fString to add the relative address (that I found in the Yelp inspect code).

counter = 0

for link in myLinks:

  if counter > 1:

    print(link.text)

    print(f"""https://www.yelp.com{link["href"]}""")

  counter +=1