jobGPT

Description

This project scrapes Job websites like stepstone.de or indeed.de for job listings matching the entered search criteria using Selenium and generates a Cover letter for them. The scraped information is then saved to a .csv file. You can select the jobs you want to have a cover letter generated to and the program then generates a cover letter for them with the job data. If a CV / Resume .pdf/.txt was submitted, it also extracts the personal information and therefore personalizes the cover letter texts.

Installation

This program requires tkinter and Python 3.6+

1. Clone this repository
2. Install the required packages using the command pip install -r requirements.txt.
3. Rename the file config_default.cfg to config.cfg and insert your OpenAI API key under api_key. Also change the used GPT-Model to the one you like (e.g. gpt-3.5-turbo [default], gpt-4)

Usage

Run the main.py script to start the project and input the needed information to start the query. If a CV/Resume should be used to extract the personal information, select it. Using the Dropdown-Menu you can choose the website to scrape. Scraping can take some time due to selenium being quite slow, so grab a coffee and come back later :) Sometimes the program breaks because of spam protection captchas. The output will be in the output/ folder containing the single .csv-file and several cover letters named after the job corresponding to the .csv-file.

Modifying

This tool scrapes german websites and also generates german cover letters. If you want this tool to use your native language feel free to edit the starting-prompt at src/gpt.py and the website Scraping Classes at src/scraper.py.

You can also add other not yet supported websites by creating a new class in src/scraper.py that extends the abstract class Scraper. To do that implement the following methods:

• BASE_URL -> class-attribute which is the main domain of your website (e.g. https://www.stepstone.de)
• get_display_url -> method for calculating the url of all joblistings for each page-index
• get_job_urls -> method for retrieving link to the job descriptions for each job
• extract_job -> method for retrieving the job information like company name, location, personal profile, etc...
• click_to_enter [optional] -> method for handling cookie-accept banners or pop-up messages