Ultimate Resume Scanner - NLP

A live resume scanning and ranking tool built using Python, Streamlit, and NLP. The app parses resumes (PDF/DOCX), extracts skills, experience, and education, ranks candidates against a job description, and generates highlighted resumes and full PDF reports.

---

Features

• Upload multiple resumes (PDF / DOCX)

• Job Description Input with key skills

• Live scoring of candidates:

  • Weighted matching score
  • Semantic similarity score

• Resume preview with highlighted skills

• Analytics Dashboard:

  • Candidate skill distribution
  • Experience levels
  • Top skills

• Skill Gap Heatmap: Compare candidate skills vs job requirements

• Export Options:

  • Highlighted resumes in HTML & PDF
  • Full PDF report with all candidates and analytics

• Filters:

  • Minimum score
  • Years of experience
  • Education levels

---

Project Structure

resume_scanner_nlp/
├── app.py                     # Main Streamlit app
├── config.py                  # Configuration (paths, API keys, etc.)
├── data/
│   ├── job_descriptions/      # Store sample or real job descriptions
│   ├── resumes/               # Upload raw resumes here
│   ├── highlighted_resumes/   # Generated highlighted resumes (HTML & PDF)
│   └── reports/               # Full reports PDFs
├── modules/                   # Core modules
│   ├── __init__.py
│   ├── dashboard.py
│   ├── experience_level.py
│   ├── highlight.py
│   ├── job_analysis.py
│   ├── matching.py
│   ├── ner_extraction.py
│   ├── parser.py
│   ├── preprocessing.py
│   ├── ranking.py
│   ├── report.py
│   ├── resume_scanner.py
│   ├── semantic.py
│   └── skill_extraction.py
├── README.md
├── requirements.txt
└── venv/                      # Virtual environment

---

Installation

1. Clone the repository

git clone <your-repo-url>
cd resume_scanner_nlp

2. Create a virtual environment

python3 -m venv .venv
source .venv/bin/activate  # macOS/Linux
.venv\Scripts\activate     # Windows

3. Install dependencies

pip install -r requirements.txt

4. Install wkhtmltopdf (required for PDF export)

• macOS (Homebrew):

brew install wkhtmltopdf

• Ubuntu/Debian:

sudo apt-get install wkhtmltopdf

---

Usage

1. Run Streamlit app

streamlit run app.py

2. Open in browser

http://localhost:8502

3. Steps in the app:

• Paste the job description
• Enter key skills (comma-separated)
• Upload resumes (PDF or DOCX)
• Adjust filters in the sidebar (Score, Experience, Education)
• View candidate list, resume preview, analytics dashboard, and skill gap heatmap
• Export highlighted resumes and full PDF report

---

Output Directories

• data/highlighted_resumes/ → Highlighted HTML & PDF resumes
• data/reports/ → Full PDF report with all candidates and analytics

---

Example Data

• Place resumes in data/resumes/
• Place job descriptions in data/job_descriptions/

---

Notes

• Make sure wkhtmltopdf is installed and path is correctly set in app.py:

config = pdfkit.configuration(wkhtmltopdf="/usr/local/bin/wkhtmltopdf")

• Streamlit may give warnings about use_container_width. Use width='stretch' instead.
• HuggingFace tokenizers warning is normal when using multiprocessing; can be ignored.

---

Dependencies

• Python ≥ 3.9
• Streamlit
• Pandas
• PDFKit
• PyPDF2
• python-docx
• spaCy (NER extraction)
• matplotlib / seaborn (for dashboards)
• HuggingFace Transformers (semantic similarity)
• scikit-learn

pip install -r requirements.txt

---

License

MIT License – Open source