How codepoison/vendorbender works
Vendor Invoice Intelligence provides two invoice-scoring workflows: estimate freight cost and classify invoices for manual review. The README describes a SQLite-to-feature-engineering-to-scaled-model pipeline; the dashboard also accepts invoice features and presents predictions. Training, inference, and dashboard components are shown separately. Some source was not sampled, so links for documented but unverified runtime wiring are omitted; dashed edges mark unverified relationships.
Main entry points: app.py Streamlit dashboard · predict_freight.py Freight inference · predict_invoice_flag.py Invoice flag inference · train_freight.py Freight training
Components 19
Dashboard
- Streamlit dashboard
app.py - Freight inputs
app.py - Risk inputs
app.py - Freight result
app.py - Risk result
app.py
Inference
- Freight inference
Inference/predict_freight.py - Invoice flag inference
Inference/predict_invoice_flag.py - Freight model
models/predict_freight_cost.pkl - Freight scaler
models/freight_scaler.pkl - Risk classifier
models/predict_flag_invoice.pkl - Risk scaler
models/scaler.pkl
Data preparation
- Invoice features
invoice_flagging/data_preprocessing.py - Freight data prep
freight_cost_prediction/data_preprocessing.py - SQLite invoice data
Model training
- Freight training
invoice_flagging/train_freight.py - Freight model evaluation
freight_cost_prediction/model_evaluation.py - Risk model evaluation
invoice_flagging/modeling_evaluation.py - Freight training workflow
freight_cost_prediction/train.py
Other components
- Finance user
Connections 20
- Finance user → Streamlit dashboard: uses
README.md - Streamlit dashboard → Freight inputs: collects
app.py - Streamlit dashboard → Risk inputs: collects
app.py - Streamlit dashboard → Freight result: presents
README.md - Streamlit dashboard → Risk result: presents
app.py - Freight inputs → Freight inference: scores with
- Risk inputs → Invoice flag inference: scores with
- Freight inference → Freight model: loads
Inference/predict_freight.py - Freight inference → Freight scaler: loads
Inference/predict_freight.py - Invoice flag inference → Risk classifier: loads
Inference/predict_invoice_flag.py - Invoice flag inference → Risk scaler: loads
Inference/predict_invoice_flag.py - Freight inference → Freight result: returns prediction
- Invoice flag inference → Risk result: returns flag
- SQLite invoice data → Invoice features: supplies records
README.md - SQLite invoice data → Freight data prep: supplies records
README.md - Freight training → Invoice features: loads data through
invoice_flagging/train_freight.py - Freight training → Risk model evaluation: trains with
invoice_flagging/train_freight.py - Freight training workflow → Freight data prep: loads data through
freight_cost_prediction/train.py - Freight training workflow → Freight model evaluation: trains with
freight_cost_prediction/train.py - Invoice features → Freight training: prepares training data
invoice_flagging/train_freight.py