Skip to content

Latest commit

 

History

History
63 lines (36 loc) · 3.5 KB

README.md

File metadata and controls

63 lines (36 loc) · 3.5 KB
Raw

linkedin-job-analizer-and-easy-apply

Project that checks in Linkedin (using NLP) if you have the Language, Experience and Technology requirements for a chosen position and country. If you have them it applies using EasyApply if it is possible. All the information is saved to a PostgreSQL database or a json file.

The user can choose different options in configfile.ini

Playwright framework is used to search for a position and country and if you want applies a filter to find the jobs that have "EasyApply". Then scraps each job position of the list and afer perform the following tasks:

  • If the job description is in other language than in English it translates it with the googletrans library.
  • Cleans the job description.
  • Analyze job descriptions with spaCy, an NLP (Natural Language Processing) library.

With the options that you choose in the configfile.ini it will decide if you should apply to the position or not based on the language you are fluent, your experience and the technologies that you know. The apply decision is recorded to the job instance together with the reasons not to apply to the job and a list of word tags that are present in the description.

The words that spaCy check are entities that are in ./data/data.json. They were added in the context of an IT job search.

If easy_apply = True in configfile.ini then it applies for the position. It has a dictionary in data/easy_apply_questions_answers.json with the EasyApply questions and answers. When it has to answer and it doesn't find an answer it saves the missing questions in another json file in ./data folder.

The webscrapped data can be saved to a json file and/or to a PostgreSQL Database

Steps to use it

  1. Create a virtual environment for example with pipenv. Open the terminal and write:

     pipenv shell
 pipenv install -r requirements.txt

  2. Install the playwright browsers:

     playwright install

  3. Download spaCy large NLP model:

     python -m spacy download en_core_web_lg

  4. Generate cookies to login to linkedin:

    Run the following code:

     playwright codegen linkedin.com --save-storage=auth.json

    Login manually to linkedin in the browser. After login close the browser, it will save all the information to a file auth.json. DON'T SHARE THIS FILE (I added to the gitignore file of the repo).

  5. Choose the options that you want and the words to use in spaCy in configfile.ini

  6. Fill the ./data/work_and_education_history.json file with the Work Experience and Education that you want to fill in EasyApply tab for the position that you are applying (in case that is required).

  7. If you choose to use a PostgreSQL Database create a .env file with the required env variables:

     POSTGRESQL_HOSTNAME = 'localhost'
 POSTGRESQL_USERNAME = "username"
 POSTGRESQL_PASSWORD = 'password'
 POSTGRESQL_DATABASE = 'db_name'

  8. Run in the terminal to run the script:

     python linkedin_job_analyzer.py

Added support to CVs in different languages

Now the script support multiple languages CVs (for now Spanish, Italian and English). The script will choose the CV language according to the job description language. If the description is in Spanish or Italian it will use the CVs in these languages, otherwise it will apply using the CV in English.

You have to upload the CVs to https://www.linkedin.com/jobs/application-settings/ with the languages names in each CV file name. For example: "CV - Name - Espanol" or "CV - Name - Italiano" or "CV - Name - English"