How chaimaebouyarmane/job-recommendation-system-ml-nlp works
Main Components
• Web Application (Frontend):
• Templates folder: Contains HTML files (index.html, index_cv.html, etc.) that render the user interface.
• Static assets: CSS (including Bootstrap and SCSS files), JavaScript, and images that style and add interactivity.
• Backend Application (Flask-based Server):
• app.py: The entry point for the Flask web application.
• Login.py, GetCvData.py, GetOffreData.py, and processing.py: Components handling requests, authentication, data retrieval, and processing.
• Machine Learning & NLP Modules:
• Models directory: Contains modules such as CV.py, OffreEmploi.py, User.py, similarityOffre.py, skills.py, plus test and training scripts.
• training.py: Script for model training.
• Database & External Services:
• README points to MongoDB for data storage (connection parameters mentioned in config.py as per instructions).
• External libraries: Spacy, Scikit-learn, BeautifulSoup for NLP/ML and data extraction.
• Supporting Files:
• requirements.txt: Lists project dependencies.
• CV directory: Stores a collection of CV PDFs used as input data.
Relationships and Interactions
• User Interaction:
• Users interact via the web interface (HTML pages, forms) rendered by Flask.
• Request Handling:
• The Flask server (app.py) routes user requests to specific endpoints, which may involve login, data fetching (for CVs or job offers), and recommendations.
• Data Processing:
• Upon receiving a request, the backend calls functions/scripts (e.g., processing.py and modules within Models/) that perform NLP processing and use machine learning algorithms to determine skill similarity.
• Data Storage:
• The system likely logs user data, CV information, and job offers in a MongoDB database.
• Model Training:
• The training.py script is used to train the underlying ML model that computes recommendations by analyzing the content of CVs and job offers.
• Asset Delivery:
• Static assets (CSS, JavaScript) are served to the client to ensure a responsive UI.
Architectural Patterns & Design Principles
• Separation of Concerns:
• The project clearly separates the user interface (templates, static assets) from the backend processing (Flask server, ML modules) and the data/modeling layer.
• MVC-Like Structure:
• Though not a strict MVC architecture, the organization suggests a division between the view (templates), controller (Flask routing, app.py, Login.py), and model (ML models in Models/).
• Modular Design:
• Individual Python modules for database connections (Connexion.py), data extraction, and similarity computation ensure each component has a focused responsibility.
• Reusability and Extensibility:
• The system could be extended by adding more ML models or additional endpoints due to its modular layout.