Scraping Web - Awesome Facts

Scraping Web - Awesome Facts

Subscribe to my newsletter and never miss my upcoming articles

Want some awesome facts to start your day or just want to learn something new or out here for some fun !

Well, in this blog I will be sharing a script that I did over a weekend just for fun (and for a side project 😉).

What I will not discuss here:

What is Web Scraping ?

How to use BeautifulSoup ?

How to Scrap ? That we will discuss !!


TechStack

  • Python
  • BeautifulSoup

Final Result (Before diving into the steps)

We will be saving all the scraped quotes in a text file. That would look something like this

result_1.png

Let's Do Scraping

Analyse The Website To Scrape

As we want to scrape the awesome facts, we are going to get that from Computer Facts

A look at webpage

web_page.png

Now, we want facts like 1. The first electronic computer ENIAC weighed more than 27 tons and took up 1800 square feet., so we should take a look at its source page first and find where is this present, to get the pattern (we will use that pattern later to scrape these quotes).

page_source.png

As we can see on page source, the facts start from <p> tag on line 475 and there are several <p> tags that contains these facts. If we look further, we would find that there are a lot of <p> tags being used within the HTML. How would we differentiate our facts with other <p> tags ?

If you looked hard, then you know ! There is a unique descriptor where facts are present, each fact precedes a with number (0-9) followed by (.) and space. Hence, we have our pattern 😀

We can ignore all other <p> tags and take only that matches a particular pattern we identified. To match pattern, we will use regex module of python, re.

Hence, the pattern we identified is

  • All facts are inside <p> tags.
  • There are other contents inside <p> tag as well.
  • To get facts, we will use regex to extract required content, leaving other <p> tags as it is !

Import Required Libraries And Define Constants

Before doing the task of scraping, we need to import libraries and define some constants.

# Standard Imports

import os
import re
from typing import List
from urllib.request import Request, urlopen
from bs4 import BeautifulSoup
from bs4.element import Tag
  • import os will import python standard library, used to make directory and other OS stuff.
  • re will be used to extract a pattern using some regex.
  • Then we import List from typing, to explicitly type hint some arguments passed to some functions.
  • Request and urlopen from standard library urllib. They would help us to get HTML data we want.
  • BeautifulSoup from third-party library bs4, this creates a easy to manage BeautifulSoup object, that represents HTML content.
  • Tag from bs4.element, again to be used for type hinting.
QUOTES_FILENAME = "/funfacts.txt"
QUOTES_TXT_PATH = os.getcwd() + "/funfacts"
QUOTES_FILE_PATH = QUOTES_TXT_PATH + QUOTES_FILENAME
QUOTES_URL = "https://gotechug.com/interesting-facts-about-computers-you-didnt-know/"
  • QUOTES_FILENAME represents the filename of the text file where the facts will be stored.
  • QUOTES_TXT_PATH represents the folder where the funfacts.txt will reside.
  • QUOTES_FILE_PATH file path, to be used further to store data.
  • QUOTES_URL url to webpage we want to scrape.

Why name the variables as QUOTES ? I had another script to scrape some great quotes and reused that same for scraping awesome facts. Need to know more about scraping quotes ? Visit this link.

Create BeautifulSoup Object

Now, we have some constants to play with, we will use QUOTES_URL to get the HTML of the page and create a BeautifulSoup object to filter our facts.

def get_bs4_obj(url: str) -> BeautifulSoup:
    '''
    Get BeautifulSoup object for given QUOTES_URL.
    '''
    req = Request(url, headers={'User-Agent': 'Mozilla/5.0'})
    html = urlopen(req).read()
    bs4Obj = BeautifulSoup(html, 'html.parser')
    return bs4Obj

We create a function get_bs4_obj that takes url a string type object and returns the required object bs4Obj.

  • Request helps us to bypass the security that is blocking the use of urllib based on the user agent. Hence, we need to give Mozilla or any other browser user agent to let us access their source page.
  • Then we do urlopen with the defined user agent in Request and read it's content.
  • At last we create BeautifulSoup object using html.parser to parse the html contents and return the required object.

Steps I used to check if we are allowed to scrape the site or not

import requests #pip install requests
req = requests.get(QUOTES_URL)
print(req.status_code) #should be 200, other than 200 means scraping not/partially allowed

Filter Tags

With having bs4Obj, now it is possible to filter the HTML on <p> tags, we can achieve that using below functions

def filter_p_tags(ptag: Tag) -> bool:
    '''
    Get required p tags only.
    '''
    return re.match(r"[0-9]+.", ptag.get_text())


def get_p_tags(bs4Obj: BeautifulSoup) -> List[Tag]:
    '''
    Get all p tags from the BeautifulSoup obj.
    Note: It is the requirement for given QUOTES_URL, it shall be different for different URL to scrap.
    '''
    allP = bs4Obj.find_all('p')
    allReleventP = list(filter(filter_p_tags, allP))
    return allReleventP

See how we used the List and Tag for type hints and BeautifulSoup as well. Type hints are not mandatory and are not checked by python, these are just for developers and know that Python will always be dynamically typed !

allP list contains all <p> tags available in HTML. We filter them using python's inbuilt filter function on the condition that, if text of any <p> tag does not have a pattern matching regex [0-9]+. then filter them out from allP list.

Pattern matching regex, we define that in function filter_p_tags, where we use re.match to match the occurence of our pattern in the given text. You can do that in get_p_tags itself, but I think it is better to have such task done in pure functions instead of making them in single function.

Store all relevent tags in variable allReleventP and return that (to be used by another function !)

Get Facts

Now, we will create a generator to yield the facts we need to save in a text file. We will be using allReleventP variable passed to the new function.

def get_all_facts(ptags: List[Tag]):
    '''
    Yield all facts present in p tags.
    '''
    for p in ptags:
        fact = re.sub(r"[0-9]+\. ", "", p.get_text())
        yield fact + "\n"

Here we are extracting text present in <p> tag. But wait !! We don't need preceding number, right ? That's why we are using re.sub to substitute it with empty string. Regex we used "[0-9]+\. ", it will take any number followed by . and a space.

We will use this generator in next section, inside another function.

Save Facts

Here is the function

def save_fun_facts(ptags: List[Tag]):
    '''
    Save extracted facts in a text file, create a new folder if not already present
    '''
    global QUOTES_TXT_PATH, QUOTES_FILE_PATH
    if not os.path.exists(QUOTES_TXT_PATH):
        os.mkdir(QUOTES_TXT_PATH)

    with open(QUOTES_FILE_PATH, 'w') as file:
        for txt in get_all_facts(ptags):
            file.write(txt)

    print(f'All Fun Facts written to file: {QUOTES_FILE_PATH}')

We take global constants QUOTES_TXT_PATH and QUOTES_FILE_PATH to be used for writing to the file. We check if the directory exists or not (the folder where we will save our file). Then we open the file in the created directory (if it not exists).

Now here comes the use of generator, we call generator on each <p> tag that yields text present in it.

Main

Now the script to execute the code

if __name__ == "__main__":
    bs4Obj = get_bs4_obj(QUOTES_URL)
    allP = get_p_tags(bs4Obj)
    save_fun_facts(allP)

If you check now, the facts should be appearing in the file created funfacts.txt inside the folder /funfacts.


Well, that was it, if you followed this then you now know the gist of general process for web scraping and how to scrape Awesome Facts !

It was just a drop in the ocean full of different functionalities you can achieve with web scraping, but the general idea remains the same:

  1. Get the website we want to scrape.
  2. Analyse the required components we need and find the pattern.
  3. Create BeautifulSoup object to ease the process and access HTML without much problem.
  4. Create small functions to get the task done.
  5. Finally integrate all function and see the magic of the script !

Full python script can be found here.

GitHub action where this script is used. Would love to hear your feedback on this as well.

Just starting your Open Source Journey ? Don't forget to check out Hello Open Source

Want to make a simple and awesome game from scratch ? Check PongPong

Till next time !

Namaste 🙏

Interested in reading more such articles from Siddharth Chandra?

Support the author by donating an amount of your choice.

Recent sponsors
 
Share this