In this article, we are going to see how to build a web scraping bot in Python.
Web Scraping is a process of extracting data from websites. A Bot is a piece of code that will automate our task. Therefore, A web scraping bot is a program that will automatically scrape a website for data, based on our requirements.
Module needed
- bs4: Beautiful Soup(bs4) is a Python library for pulling data out of HTML and XML files. This module does not come built-in with Python. To install this type the below command in the terminal.
pip install bs4
- requests: Request allows you to send HTTP/1.1 requests extremely easily. This module also does not come built-in with Python. To install this type the below command in the terminal.
pip install requests
- Selenium: Selenium is one of the most popular automation testing tools. It can be used to automate browsers like Chrome, Firefox, Safari, etc.
pip install selenium
Method 1: Using Selenium
We need to install a chrome driver to automate using selenium, our task is to create a bot that will be continuously scraping the google news website and display all the headlines every 10mins.
Stepwise implementation:
Step 1: First we will import some required modules.
Python3
# These are the imports to be madeimport timefrom selenium import webdriverfrom datetime import datetime |
Step 2: The next step is to open the required website.
Python3
# path of the chromedriver we have just downloadedPATH = r"D:\chromedriver"driver = webdriver.Chrome(PATH) # to open the browser # url of google news website # to open the url in the browserdriver.get(url) |
Output:
Step 3: Extracting the news title from the webpage, to extract a specific part of the page, we need its XPath, which can be accessed by right-clicking on the required element and selecting Inspect in the dropdown bar.
After clicking Inspect a window appears. From there, we have to copy the elements full XPath to access it:
Note: You might not always get the exact element that you want by inspecting (depends on the structure of the website), so you may have to surf the HTML code for a while to get the exact element you want. And now, just copy that path and paste that into your code. After running all these lines of code, you will get the title of the first heading printed on your terminal.
Python3
# Xpath you just copiednews_path = '/html/body/c-wiz/div/div[2]/div[2]/\div/main/c-wiz/div[1]/div[3]/div/div/article/h3/a' # to get that elementlink = driver.find_element_by_xpath(news_path) # to read the text from that elementprint(link.text) |
Output:
‘Attack on Afghan territory’: Taliban on US airstrike that killed 2 ISIS-K men
Step 4: Now, the target is to get the X_Paths of all the headlines present.
One way is that we can copy all the XPaths of all the headlines (about 6 headlines will be there in google news every time) and we can fetch all those, but that method is not suited if there are a large number of things to be scrapped. So, the elegant way is to find the pattern of the XPaths of the titles which will make our tasks way easier and efficient. Below are the XPaths of all the headlines on the website, and let’s figure out the pattern.
/html/body/c-wiz/div/div[2]/div[2]/div/main/c-wiz/div[1]/div[3]/div/div/article/h3/a
/html/body/c-wiz/div/div[2]/div[2]/div/main/c-wiz/div[1]/div[4]/div/div/article/h3/a
/html/body/c-wiz/div/div[2]/div[2]/div/main/c-wiz/div[1]/div[5]/div/div/article/h3/a
/html/body/c-wiz/div/div[2]/div[2]/div/main/c-wiz/div[1]/div[6]/div/div/article/h3/a
/html/body/c-wiz/div/div[2]/div[2]/div/main/c-wiz/div[1]/div[7]/div/div/article/h3/a
/html/body/c-wiz/div/div[2]/div[2]/div/main/c-wiz/div[1]/div[8]/div/div/article/h3/a
So, by seeing these XPath’s, we can see that only the 5th div is changing (bolded ones). So based upon this, we can generate the XPaths of all the headlines. We will get all the titles from the page by accessing them with their XPath. So to extract all these, we have the code as
Python3
# I have used f-strings to format the stringc = 1for x in range(3, 9): print(f"Heading {c}: ") c += 1 curr_path = f'/html/body/c-wiz/div/div[2]/div[2]/div/main\ /c-wiz/div[1]/div[{x}]/div/div/article/h3/a' title = driver.find_element_by_xpath(curr_path) print(title.text) |
Output:
Now, the code is almost complete, the last thing we have to do is that the code should get headlines for every 10 mins. So we will run a while loop and sleep for 10 mins after getting all the headlines.
Below is the full implementation
Python3
import timefrom selenium import webdriverfrom datetime import datetime PATH = r"D:\chromedriver" driver = webdriver.Chrome(PATH) driver.get(url) while(True): now = datetime.now() # this is just to get the time at the time of # web scraping current_time = now.strftime("%H:%M:%S") print(f'At time : {current_time} IST') c = 1 for x in range(3, 9): curr_path = '' # Exception handling to handle unexpected changes # in the structure of the website try: curr_path = f'/html/body/c-wiz/div/div[2]/div[2]/\ div/main/c-wiz/div[1]/div[{x}]/div/div/article/h3/a' title = driver.find_element_by_xpath(curr_path) except: continue print(f"Heading {c}: ") c += 1 print(title.text) # to stop the running of code for 10 mins time.sleep(600) |
Output:
Method 2: Using Requests and BeautifulSoup
The requests module gets the raw HTML data from websites and beautiful soup is used to parse that information clearly to get the exact data we require. Unlike Selenium, there is no browser installation involved and it is even lighter because it directly accesses the web without the help of a browser.
Stepwise implementation:
Step 1: Import module.
Python3
import requestsfrom bs4 import BeautifulSoupimport time |
Step 2: The next thing to do is to get the URL data and then parse the HTML code
Python3
response = requests.get(url)text = response.textdata = BeautifulSoup(text, 'html.parser') |
Step 3: First, we shall get all the headings from the table.
Python3
# since, headings are the first row of the tableheadings = data.find_all('tr')[0]headings_list = [] # list to store all headings for x in headings: headings_list.append(x.text)# since, we require only the first ten columnsheadings_list = headings_list[:10] print('Headings are: ')for column in headings_list: print(column) |
Output:
Step 4: In the same way, all the values in each row can be obtained
Python3
# since we need only first five coinsfor x in range(1, 6): table = data.find_all('tr')[x] c = table.find_all('td') for x in c: print(x.text, end=' ') print('') |
Output:
Below is the full implementation:
Python3
import requestsfrom bs4 import BeautifulSoupfrom datetime import datetimeimport time while(True): now = datetime.now() # this is just to get the time at the time of # web scraping current_time = now.strftime("%H:%M:%S") print(f'At time : {current_time} IST') text = response.text html_data = BeautifulSoup(text, 'html.parser') headings = html_data.find_all('tr')[0] headings_list = [] for x in headings: headings_list.append(x.text) headings_list = headings_list[:10] data = [] for x in range(1, 6): row = html_data.find_all('tr')[x] column_value = row.find_all('td') dict = {} for i in range(10): dict[headings_list[i]] = column_value[i].text data.append(dict) for coin in data: print(coin) print('') time.sleep(600) |
Output:
Hosting the Bot
This is a specific method, used to run the bot continuously online without the need for any human intervention. replit.com is an online compiler, where we will be running the code. We will be creating a mini webserver with the help of a flask module in python that helps in the continuous running of the code. Please create an account on that website and create a new repl.
After creating the repl, Create two files, one to run the bot code and the other to create the web server using flask.
Code for cryptotracker.py:
Python3
import requestsfrom bs4 import BeautifulSoupfrom datetime import datetimeimport time # keep_alive function, that maintains continuous # running of the code.from keep_alive import keep_aliveimport pytz # to start the threadkeep_alive()while(True): tz_NY = pytz.timezone('Asia/Kolkata') datetime_NY = datetime.now(tz_NY) # this is just to get the time at the time of web scraping current_time = datetime_NY.strftime("%H:%M:%S - (%d/%m)") print(f'At time : {current_time} IST') text = response.text html_data = BeautifulSoup(text, 'html.parser') headings = html_data.find_all('tr')[0] headings_list = [] for x in headings: headings_list.append(x.text) headings_list = headings_list[:10] data = [] for x in range(1, 6): row = html_data.find_all('tr')[x] column_value = row.find_all('td') dict = {} for i in range(10): dict[headings_list[i]] = column_value[i].text data.append(dict) for coin in data: print(coin) time.sleep(60) |
Code for the keep_alive.py (webserver):
Python3
from flask import Flaskfrom threading import Thread app = Flask('') @app.route('/')def home(): return "Hello. the bot is alive!" def run(): app.run(host='0.0.0.0',port=8080) def keep_alive(): t = Thread(target=run) t.start() |
Keep-alive is a method in networking that is used to prevent a certain link from breaking. Here the purpose of the keep-alive code is to create a web server using flask, that will keep the thread of the code (crypto-tracker code) to be active so that it can give the updates continuously.
Now, we have a web server create, and now, we need something to ping it continuously so that the server does not go down and the code keeps on running continuously. There is a website uptimerobot.com that does this job. Create an account in it
Running the Crypto tracker code in Replit. Thus, We have successfully created a web scraping bot that will scrap the particular website continuously for every 10 mins and print the data to the terminal.
