How s-m-quadri/toon-suggest works
Project type and purpose
• This is a standalone, desktop GUI application (not a web or server-based system).
• Purpose: provide an offline, content-based recommender for animation movies. It trains a TF-IDF model on IMDb .tsv.gz files and then lets users query similar titles via a Tkinter GUI.
File structure
• Top-level:
• main.py – launches the home GUI
• model.pkl – persisted TF-IDF model
• requirements.txt – Python dependencies
• setup.spec – PyInstaller configuration
• src/ directory:
• gui_home.py – main recommendation window (view + controller)
• gui_train.py – training window (view + controller)
• train.py – data loading, preprocessing, TF-IDF vectorization, and saving model
• recommend.py – loading model, computing similarities, returning recommendations
• Build and CI config: .github/, .gitignore, LICENSE
README for additional insights
• Confirms two distinct flows: model training vs. model usage (recommendation).
• Built with Python, Tkinter, Pandas, and scikit-learn’s TF-IDF.
• Packaged into a single .exe via PyInstaller (setup.spec).
Essential details
• Both the training and recommendation subsystems as separate but related paths
• PyInstaller packaging step to show how all code and dependencies become a single deployable .exe