# felipesidooski/creditscore architecture

> Architecture diagram of the GitHub repository felipesidooski/creditscore, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-11-17). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/felipesidooski/creditscore
- Repository: https://github.com/felipesidooski/creditscore

## Overview

This repository is a credit-default modeling project, with a reusable champion-model training and inference path alongside historical experiments. The principal user workflow loads raw client rows, applies deterministic preprocessing and feature engineering, scores them with a saved model artifact, and writes predictions. The README documents additional exploration, validation, resampling, ensembling, feature selection, tuning, stacking, and calibration capabilities. Some experimental implementations are unsampled; their relationships are included only where README statements or indexed imports support them.

Main entry points: [`legacy/archive/credit_score_experiments_full.py`](https://github.com/felipesidooski/creditscore/blob/main/legacy/archive/credit_score_experiments_full.py) (Historical experiment engine), [`src/creditscore/training.py`](https://github.com/felipesidooski/creditscore/blob/main/src/creditscore/training.py) (Champion trainer), [`src/creditscore/inference.py`](https://github.com/felipesidooski/creditscore/blob/main/src/creditscore/inference.py) (Risk predictor), [`legacy/eda.py`](https://github.com/felipesidooski/creditscore/blob/main/legacy/eda.py) (Exploratory analysis)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_data["Data preparation"]
  node_dataset["Training data<br/>[train.csv]"]
  node_legacy_preprocessing["Legacy preprocessing<br/>[preprocessing.py]"]
  node_config["Experiment settings<br/>[config.py]"]
  node_data_pipeline["Preparation pipeline<br/>[preprocessing.py]"]
end

subgraph group_experiments["Model experiments"]
  node_eda["Exploratory analysis<br/>[eda.py]"]
  node_pipeline_builders["Model pipelines<br/>[pipelines.py]"]
  node_decision_validation["Decision validation"]
  node_original_models["KNN and trees<br/>[original_models.py]"]
  node_enhanced_models["Enhanced models<br/>[enhanced_models.py]"]
  node_advanced_models["Calibration and thresholds<br/>[advanced_models.py]"]
  node_resampling["Resampling<br/>[resampling.py]"]
  node_ensembles["Boosting and ensembles"]
  node_feature_selection["Feature selection"]
  node_optuna["Optuna tuning<br/>[optuna_xgboost.py]"]
  node_stacking["Stacking<br/>[stacking.py]"]
  node_target_encoding["Target encoding<br/>[target_encoding.py]"]
  node_evaluation["Evaluation<br/>[evaluation.py]"]
  node_experiment_engine["Historical experiment engine"]
end

subgraph group_runtime["Champion runtime"]
  node_train_cli["Training command"]
  node_trainer["Champion trainer<br/>[training.py]"]
  node_inference_cli["Prediction command"]
  node_predictor["Risk predictor<br/>[inference.py]"]
end

subgraph group_outputs["Artifacts and results"]
  node_models[("Saved models")]
  node_reports["Metrics and predictions"]
end

node_user(("Analyst"))

node_user -->|"starts training"| node_train_cli
node_user -->|"requests scores"| node_inference_cli
node_dataset -->|"examined by"| node_eda
node_dataset -->|"loaded for training"| node_trainer
node_train_cli -->|"invokes"| node_trainer
node_trainer -->|"prepares rows"| node_data_pipeline
node_data_pipeline -->|"uses settings"| node_config
node_trainer -->|"uses settings"| node_config
node_trainer -->|"saves artifacts"| node_models
node_models -->|"loaded by"| node_predictor
node_inference_cli -->|"invokes"| node_predictor
node_inference_cli -->|"reads input rows"| node_dataset
node_predictor -->|"prepares rows"| node_data_pipeline
node_predictor -->|"uses settings"| node_config
node_predictor -->|"writes predictions"| node_reports
node_eda -->|"produces analysis"| node_reports
node_legacy_preprocessing -->|"delegates to"| node_experiment_engine
node_pipeline_builders -->|"delegates to"| node_experiment_engine
node_evaluation -->|"delegates to"| node_experiment_engine
node_eda -->|"delegates to"| node_experiment_engine
node_ensembles -->|"delegates to"| node_experiment_engine
node_original_models -->|"delegates to"| node_experiment_engine
node_enhanced_models -->|"delegates to"| node_experiment_engine
node_advanced_models -->|"delegates to"| node_experiment_engine
node_decision_validation -->|"delegates to"| node_experiment_engine
node_resampling -->|"delegates to"| node_experiment_engine
node_feature_selection -->|"delegates to"| node_experiment_engine
node_optuna -->|"delegates to"| node_experiment_engine
node_stacking -->|"delegates to"| node_experiment_engine
node_target_encoding -->|"delegates to"| node_experiment_engine
node_experiment_engine -->|"records experiment results"| node_reports
node_experiment_engine -.->|"uses sklearn pipelines"| node_pipeline_builders

click node_dataset "https://github.com/felipesidooski/creditscore/blob/main/data/train.csv"
click node_eda "https://github.com/felipesidooski/creditscore/blob/main/legacy/eda.py"
click node_legacy_preprocessing "https://github.com/felipesidooski/creditscore/blob/main/legacy/preprocessing.py"
click node_pipeline_builders "https://github.com/felipesidooski/creditscore/blob/main/legacy/pipelines.py"
click node_decision_validation "https://github.com/felipesidooski/creditscore/blob/main/legacy/decision_validation.py"
click node_original_models "https://github.com/felipesidooski/creditscore/blob/main/legacy/original_models.py"
click node_enhanced_models "https://github.com/felipesidooski/creditscore/blob/main/legacy/enhanced_models.py"
click node_advanced_models "https://github.com/felipesidooski/creditscore/blob/main/legacy/advanced_models.py"
click node_resampling "https://github.com/felipesidooski/creditscore/blob/main/legacy/resampling.py"
click node_ensembles "https://github.com/felipesidooski/creditscore/blob/main/legacy/ensemble_benchmarks.py"
click node_feature_selection "https://github.com/felipesidooski/creditscore/blob/main/legacy/feature_selection.py"
click node_optuna "https://github.com/felipesidooski/creditscore/blob/main/legacy/optuna_xgboost.py"
click node_stacking "https://github.com/felipesidooski/creditscore/blob/main/legacy/stacking.py"
click node_target_encoding "https://github.com/felipesidooski/creditscore/blob/main/legacy/target_encoding.py"
click node_evaluation "https://github.com/felipesidooski/creditscore/blob/main/legacy/evaluation.py"
click node_experiment_engine "https://github.com/felipesidooski/creditscore/blob/main/legacy/archive/credit_score_experiments_full.py"
click node_config "https://github.com/felipesidooski/creditscore/blob/main/src/creditscore/config.py"
click node_data_pipeline "https://github.com/felipesidooski/creditscore/blob/main/src/creditscore/preprocessing.py"
click node_train_cli "https://github.com/felipesidooski/creditscore/blob/main/scripts/train_champion_models.py"
click node_trainer "https://github.com/felipesidooski/creditscore/blob/main/src/creditscore/training.py"
click node_inference_cli "https://github.com/felipesidooski/creditscore/blob/main/scripts/predict_with_artifacts.py"
click node_predictor "https://github.com/felipesidooski/creditscore/blob/main/src/creditscore/inference.py"
click node_models "https://github.com/felipesidooski/creditscore/tree/main/models"
click node_reports "https://github.com/felipesidooski/creditscore/tree/main/reports"

classDef toneNeutral fill:#f8fafc,stroke:#334155,stroke-width:1.5px,color:#0f172a
classDef toneBlue fill:#dbeafe,stroke:#2563eb,stroke-width:1.5px,color:#172554
classDef toneAmber fill:#fef3c7,stroke:#d97706,stroke-width:1.5px,color:#78350f
classDef toneMint fill:#dcfce7,stroke:#16a34a,stroke-width:1.5px,color:#14532d
classDef toneRose fill:#ffe4e6,stroke:#e11d48,stroke-width:1.5px,color:#881337
classDef toneIndigo fill:#e0e7ff,stroke:#4f46e5,stroke-width:1.5px,color:#312e81
classDef toneTeal fill:#ccfbf1,stroke:#0f766e,stroke-width:1.5px,color:#134e4a
class node_dataset,node_legacy_preprocessing,node_config,node_data_pipeline toneBlue
class node_eda,node_pipeline_builders,node_decision_validation,node_original_models,node_enhanced_models,node_advanced_models,node_resampling,node_ensembles,node_feature_selection,node_optuna,node_stacking,node_target_encoding,node_evaluation,node_experiment_engine toneAmber
class node_train_cli,node_trainer,node_inference_cli,node_predictor toneMint
class node_models,node_reports toneRose
class node_user toneIndigo
```

## Components

### Data preparation

- [Training data](https://github.com/felipesidooski/creditscore/blob/main/data/train.csv): `data/train.csv`
- [Legacy preprocessing](https://github.com/felipesidooski/creditscore/blob/main/legacy/preprocessing.py): `legacy/preprocessing.py`
- [Experiment settings](https://github.com/felipesidooski/creditscore/blob/main/src/creditscore/config.py): `src/creditscore/config.py`
- [Preparation pipeline](https://github.com/felipesidooski/creditscore/blob/main/src/creditscore/preprocessing.py): `src/creditscore/preprocessing.py`

### Model experiments

- [Exploratory analysis](https://github.com/felipesidooski/creditscore/blob/main/legacy/eda.py): `legacy/eda.py`
- [Model pipelines](https://github.com/felipesidooski/creditscore/blob/main/legacy/pipelines.py): `legacy/pipelines.py`
- [Decision validation](https://github.com/felipesidooski/creditscore/blob/main/legacy/decision_validation.py): `legacy/decision_validation.py`
- [KNN and trees](https://github.com/felipesidooski/creditscore/blob/main/legacy/original_models.py): `legacy/original_models.py`
- [Enhanced models](https://github.com/felipesidooski/creditscore/blob/main/legacy/enhanced_models.py): `legacy/enhanced_models.py`
- [Calibration and thresholds](https://github.com/felipesidooski/creditscore/blob/main/legacy/advanced_models.py): `legacy/advanced_models.py`
- [Resampling](https://github.com/felipesidooski/creditscore/blob/main/legacy/resampling.py): `legacy/resampling.py`
- [Boosting and ensembles](https://github.com/felipesidooski/creditscore/blob/main/legacy/ensemble_benchmarks.py): `legacy/ensemble_benchmarks.py`
- [Feature selection](https://github.com/felipesidooski/creditscore/blob/main/legacy/feature_selection.py): `legacy/feature_selection.py`
- [Optuna tuning](https://github.com/felipesidooski/creditscore/blob/main/legacy/optuna_xgboost.py): `legacy/optuna_xgboost.py`
- [Stacking](https://github.com/felipesidooski/creditscore/blob/main/legacy/stacking.py): `legacy/stacking.py`
- [Target encoding](https://github.com/felipesidooski/creditscore/blob/main/legacy/target_encoding.py): `legacy/target_encoding.py`
- [Evaluation](https://github.com/felipesidooski/creditscore/blob/main/legacy/evaluation.py): `legacy/evaluation.py`
- [Historical experiment engine](https://github.com/felipesidooski/creditscore/blob/main/legacy/archive/credit_score_experiments_full.py): `legacy/archive/credit_score_experiments_full.py`

### Champion runtime

- [Training command](https://github.com/felipesidooski/creditscore/blob/main/scripts/train_champion_models.py): `scripts/train_champion_models.py`
- [Champion trainer](https://github.com/felipesidooski/creditscore/blob/main/src/creditscore/training.py): `src/creditscore/training.py`
- [Prediction command](https://github.com/felipesidooski/creditscore/blob/main/scripts/predict_with_artifacts.py): `scripts/predict_with_artifacts.py`
- [Risk predictor](https://github.com/felipesidooski/creditscore/blob/main/src/creditscore/inference.py): `src/creditscore/inference.py`

### Artifacts and results

- [Saved models](https://github.com/felipesidooski/creditscore/tree/main/models): `models`
- [Metrics and predictions](https://github.com/felipesidooski/creditscore/tree/main/reports): `reports`

### Other components

- **Analyst**

## Connections

- Analyst → Training command: starts training (evidence: `README.md`)
- Analyst → Prediction command: requests scores (evidence: `README.md`)
- Training data → Exploratory analysis: examined by (evidence: `README.md`)
- Training data → Champion trainer: loaded for training (evidence: `src/creditscore/training.py`)
- Training command → Champion trainer: invokes (evidence: `scripts/train_champion_models.py`)
- Champion trainer → Preparation pipeline: prepares rows (evidence: `src/creditscore/training.py`)
- Preparation pipeline → Experiment settings: uses settings (evidence: `src/creditscore/preprocessing.py`)
- Champion trainer → Experiment settings: uses settings (evidence: `src/creditscore/training.py`)
- Champion trainer → Saved models: saves artifacts (evidence: `README.md`)
- Saved models → Risk predictor: loaded by (evidence: `src/creditscore/inference.py`)
- Prediction command → Risk predictor: invokes (evidence: `scripts/predict_with_artifacts.py`)
- Prediction command → Training data: reads input rows (evidence: `README.md`)
- Risk predictor → Preparation pipeline: prepares rows (evidence: `src/creditscore/inference.py`)
- Risk predictor → Experiment settings: uses settings (evidence: `src/creditscore/inference.py`)
- Risk predictor → Metrics and predictions: writes predictions (evidence: `README.md`)
- Exploratory analysis → Metrics and predictions: produces analysis (evidence: `README.md`)
- Legacy preprocessing → Historical experiment engine: delegates to (evidence: `legacy/preprocessing.py`)
- Model pipelines → Historical experiment engine: delegates to (evidence: `legacy/pipelines.py`)
- Evaluation → Historical experiment engine: delegates to (evidence: `legacy/evaluation.py`)
- Exploratory analysis → Historical experiment engine: delegates to (evidence: `legacy/eda.py`)
- Boosting and ensembles → Historical experiment engine: delegates to (evidence: `legacy/ensemble_benchmarks.py`)
- KNN and trees → Historical experiment engine: delegates to (evidence: `legacy/original_models.py`)
- Enhanced models → Historical experiment engine: delegates to (evidence: `legacy/enhanced_models.py`)
- Calibration and thresholds → Historical experiment engine: delegates to (evidence: `legacy/advanced_models.py`)
- Decision validation → Historical experiment engine: delegates to (evidence: `legacy/decision_validation.py`)
- Resampling → Historical experiment engine: delegates to (evidence: `legacy/resampling.py`)
- Feature selection → Historical experiment engine: delegates to (evidence: `legacy/feature_selection.py`)
- Optuna tuning → Historical experiment engine: delegates to (evidence: `legacy/optuna_xgboost.py`)
- Stacking → Historical experiment engine: delegates to (evidence: `legacy/stacking.py`)
- Target encoding → Historical experiment engine: delegates to (evidence: `legacy/target_encoding.py`)
- Historical experiment engine → Metrics and predictions: records experiment results (evidence: `README.md`)
- Historical experiment engine → Model pipelines: uses sklearn pipelines (evidence: `legacy/pipelines.py`)
