How manojmeena0001/job_skills_extractor works
Project Type and Purpose
• This is a lightweight “full‑stack” Python web application using Streamlit as the front‑end/UI layer and pure Python modules for the back‑end processing pipeline.
• Purpose: Accept a job description via a web interface, run NLP‑based preprocessing, extract skills from text against a predefined list, categorize them, and render results back in the browser.
File Structure
• Top‑level
• .streamlit/ – Streamlit configuration for theming, credentials, etc.
• data/ – JSON files: skill_dictionary.json (full skill list) and skill_categories.json (mapping skills to categories).
• app.py – Streamlit application entry point, UI layout, input/output.
• preprocessing.py – Text cleaning, tokenization, stopword removal.
• skill_extractor.py – Core extraction logic matching tokens to skills.
• skill_categorizer.py – Logic to group extracted skills into categories.
• utils.py – Helper functions (logging, file I/O, common routines).
• pyproject.toml & uv.lock – Dependency management.
• No separate “database”—data is loaded from JSON. No microservices; single monolithic service.
Insights from README
• Data Flow Steps:
1. Input via Streamlit UI (app.py)
2. Preprocessing (preprocessing.py)
3. Skill Extraction (skill_extractor.py)
4. Skill Categorization (skill_categorizer.py)
5. Render in UI
• Tech Stack: Python, Streamlit, NLTK, pandas
• Ready for extension into larger ATS/HR platforms, so the pipeline is modular.
• Use clearly labeled rectangles for each module/component.
• UI layer in blue
• Processing pipeline in green
• Data store in orange
• External libraries in gray