Skip to content

cefege/cartuseria

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 

Repository files navigation

Cartuseria Crawler

This is a Python script that extracts data from the HTML of a given URL using the Playwright and BeautifulSoup libraries. The script is designed to extract information about a specific type of product from a Romanian eCommerce website called Cartuseria.ro. The extracted information includes whether or not a Frequently Asked Questions (FAQ) section exists, whether or not a product category exists, the number of internal links within the category description, and the total number of products available in the category.

Requirements

  • Python 3.x
  • playwright library
  • bs4 library

Usage

  • Install playwright and bs4 libraries using the following command: pip install playwright bs4
  • Clone the repository containing the script
  • Run the script with the following command: python script_name.py url
    • Replace script_name.py with the name of the script file
    • Replace url with the URL of the page to extract data from

Output

The script outputs the following data:

  • Whether or not a FAQ section exists
  • Whether or not a product category exists
  • The number of internal links within the category description
  • The total number of products available in the category

The output is displayed in the console.

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages