How muneebhassan437/vaultguard- works
VaultGuard AI is a transaction-fraud prediction app. A user enters transaction features in the Streamlit form; the UI derives flags and either returns a local non-fraud result for most transaction types or posts the payload to the FastAPI prediction endpoint. The API validates the request, builds a dataframe, and runs the loaded model. The tree also contains DVC-tracked datasets and EDA/modeling notebooks, but their internal behavior and connections are not shown. The graph reflects only relationships visible in the supplied excerpts.
Main entry points: main.py Prediction API · streamlit.py Feature form and routing · validation.py Input schema · logistic_regression_model.joblib.dvc Fraud model artifact
Components 8
User interface
- Feature form and routing
UI/streamlit.py
Prediction service
- Prediction API
app/main.py - Input schema
app/validation.py
ML assets and analysis
- Fraud model artifact
models/logistic_regression_model.joblib.dvc - Transaction dataset
Data/Transactions.csv.dvc - Modeling notebook
Notebooks/Modelling.ipynb - EDA notebook
Notebooks/EDA.ipynb
Other components
- Transaction user
Connections 5
- Transaction user → Feature form and routing: enters features
UI/streamlit.py - Feature form and routing → Prediction API: posts prediction
UI/streamlit.py - Prediction API → Input schema: validates input
app/main.py - Prediction API → Fraud model artifact: loads and predicts
app/main.py - Prediction API → Feature form and routing: returns result
UI/streamlit.py