Metadata-Version: 2.4
Name: ramentruck
Version: 0.1.0
Summary: A deep, layered ML/AI toolkit — the machine learning companion to ThaiTruck.
License: MIT
Requires-Python: >=3.9
Description-Content-Type: text/markdown

# ramentruck

**RamenTruck** is a deep, layered ML/AI toolkit — the expert-crafted machine learning companion to [ThaiTruck](https://pypi.org/project/thaitruck/).

> Ramen takes time, precision, and expertise. So does good ML.

## Planned Modules

| Module | Role |
|---|---|
| `broth` | Base model training/fitting wrapper — the foundation |
| `tare` | Hyperparameter tuning (GridSearchCV, RandomizedSearchCV, Optuna) |
| `soft_boiled_egg` | Cross-validation (k-fold, stratified, time-series splits) |
| `chashu` | Model serialization & persistence with versioning metadata |
| `nori` | Explainability — SHAP values, feature importance, PDP |
| `miso` | Experiment tracking — MLflow / W&B wrapper |
| `tonkotsu` | Deep learning interface — PyTorch / TensorFlow architectures |

## Install

```bash
pip install ramentruck                  # core (sklearn + numpy only)
pip install ramentruck[deep]            # + tonkotsu (torch/tensorflow)
pip install ramentruck[explain]         # + nori (shap)
pip install ramentruck[tracking]        # + miso (mlflow/wandb)
```

## Fleet

RamenTruck is part of the food truck fleet:

- **ThaiTruck** — batch DataFrame cleaning & processing
- **RamenTruck** — ML/AI toolkit *(this package, in development)*
- **TacoTruck** — streaming ingestion & API connectors *(planned)*

---

*Package name secured. Full release coming soon.*
