Installation Guide

Before You Start

You need Python installed. If you don't have it, grab the latest 3.x version from python.org — not 2.7, that one's ancient. During the Windows install, tick the "Add Python to PATH" box, otherwise the terminal won't find it later and you'll waste twenty minutes wondering why.

You'll also want a code editor like VS Code, and that's basically it. No database to set up, no API keys, nothing external.

Step-by-Step Setup

1. Unzip the project folder somewhere you'll remember, then open a terminal inside that folder. In VS Code you can just open the folder and hit the terminal shortcut.
2. Install the dependencies. Run:
   pip install -r requirements.txt
   This pulls in Flask, scikit-learn, NLTK, pdfplumber, python-docx, and the rest. Give it a minute.
3. First run downloads NLTK stopwords automatically. The code in utils.py handles that for you, so you don't need to fetch anything manually. Just make sure you're connected to the internet the very first time.
4. (Optional) Retrain the model. If you want to regenerate the model files or train on your own data, open resume_screening.ipynb and run all cells. It rewrites clf.pkl, tfidf.pkl, and encoder.pkl. Skip this step if you're happy using the model that already ships with the project.
5. Start the app. Run:
   python app.py
   You'll see Flask print a local address in the terminal.
6. Open it in your browser. Go to http://127.0.0.1:5000. Upload a resume, optionally paste a job description, and hit Screen Resume.

Common Issues (and Quick Fixes)

"ModuleNotFoundError" when you run app.py. That means the pip install didn't finish, or you ran it in a different folder. Go back, make sure you're inside the project directory, and run the requirements install again.

The PDF upload returns empty text. Some resumes are scanned images, not real text — pdfplumber can only read actual text layers, not pictures of text. Test with a normally exported PDF first. If yours is a scan, that's an OCR problem, not a bug in the app.

How to Know It's Working

Upload any sample resume. If you get a page back with a predicted category, the model is loading and predicting correctly. Now paste a job description and screen again — you should see a match score with a colored Strong, Partial, or Weak label. If both of those show up, everything's wired correctly and you're ready to demo.