MELITE — Multi-Model Classifier Evaluator

Description

MELITE is a Python package and command-line tool for evaluating and comparing classifiers on numeric tabular datasets. It separates hyperparameter tuning from classifier evaluation, preserves the evidence used for selection, and exports the selected model as a reusable artifact for downstream inference.

MELITE operates at the tabular modeling level. Its learning algorithms consume numeric feature matrices (X) and target labels (y), regardless of how those features were produced. Inputs may therefore originate from fingerprints, descriptors, dimensionality-reduction methods, clinical variables, experimental measurements, industrial features, or other numeric representations.

Purpose

MELITE is designed to make classifier comparison and selection explicit, reproducible, and auditable. Its workflow separates stages that are often mixed together in small classification workflows:

• hyperparameter tuning;
• classifier evaluation;
• comparison and selection;
• final fitting on all available data;
• model export and inference.

This separation ensures that, within each outer cross-validation split, the data used to evaluate a tuned classifier are held out from the hyperparameter search that produced it, while preserving the evidence needed to understand how the competing classifiers performed.

Why Use MELITE?

• Controlled evaluation. Hyperparameter tuning is kept separate from the evidence used to compare classifiers.
• Evidence preservation. Aggregate and fold-level evaluation results are retained for every evaluated classifier, not only for the selected one.
• Explicit selection. Classifier selection follows a predefined criterion based on cross-validation evidence rather than an informal choice after training.
• Domain-agnostic inputs. MELITE works with numeric tabular data without assuming how the features were generated.
• Reusable artifacts. After selection, the chosen classifier can be fitted on all available data and saved as a model artifact for prediction.
• CLI and Python interfaces. MELITE can be used through its command-line workflow and through a focused public Python API.

What MELITE Does

Evaluation Contract

For a registered dataset, MELITE follows the contract below:

1. X is a two-dimensional numeric feature matrix and y provides the target labels for the same samples.
2. Each active classifier is evaluated under the configured outer cross-validation design.
3. For tunable classifiers, hyperparameter search occurs only within the training portion of each outer split.
4. Evaluation evidence is obtained from the held-out folds of repeated stratified outer cross-validation.
5. Mean outer-CV F1-macro is used to select the best active classifier for each dataset.
6. Aggregate and per-fold evidence are preserved for every evaluated classifier.
7. After selection, MELITE performs the final full-data fitting stage. For a tunable classifier, melite run first performs a final full-data hyperparameter search and records the resulting parameters in results.csv; Stacking is fitted directly.
8. melite export reconstructs the selected classifier from the persisted result, fits it on all available data, and serializes the final model artifact. It performs no additional hyperparameter search, cross-validation, or classifier selection.
9. Smoke mode is intended for fast execution checks, not final classifier selection.

Installation

Package Users

Install MELITE in a supported Python environment:

python -m pip install melite

Verify the installation:

melite --version

Contributors and Developers

Clone the repository and install in editable mode with development dependencies:

git clone https://github.com/NanoBiostructuresRG/melite.git
cd melite
conda create -n melite_env python=3.11
conda activate melite_env
python -m pip install -e ".[dev]"

To build the documentation locally, install the docs extra as well:

python -m pip install -e ".[dev,docs]"
mkdocs serve

Quick Start

The bundled example creates a ready-to-run synthetic numeric CSV dataset and example configuration for a short evaluation workflow, so you can verify that MELITE is installed and working. Cloning the repository is not required.

melite example
melite run --smoke --config melite_example/config.toml

melite example creates ./melite_example/ in the current directory.

Next steps

• Prepare and use your own datasets.
• Configure MELITE.
• Run a full evaluation.
• Export a model and use the Python prediction API.

Workflow

CLI Workflow

The command-line interface provides the canonical end-to-end MELITE workflow:

1. Register one or more numeric datasets in a TOML configuration file.
2. Choose the active classifiers.
3. Run melite run to generate evaluation evidence and selected results.
4. Inspect results.csv, evaluations.csv, evaluation_folds.csv, optimization_searches.csv, optimization_provenance.json, and the dataset-level F1-macro evidence figures.
5. Run melite export for the selected result you want to preserve as a model artifact.
6. Use the exported .pkl artifact through melite.predict() for inference.

Python Workflow

The Python API is intentionally component-oriented. It exposes configuration, dataset loading, evaluation-evidence plotting, artifact-based prediction, and version metadata as public symbols.

MELITE does not expose the full evaluation orchestration as a stable high-level Python workflow API. For reproducible end-to-end execution, use the CLI and a version-controlled TOML configuration.

Supported Classifiers

MELITE currently supports four classifier keys:

MELITE v0.3.0 supports this fixed set of four classifier keys and does not expose public registration of custom classifiers.

The default configuration is:

[classifiers]
active = ["svc", "rf", "xgb"]

Add "stack" to evaluate Stacking alongside the default classifiers.

Standalone SVC is evaluated as a StandardScaler -> SVC pipeline, with probability fitting disabled during standalone evaluation. Exported SVC artifacts retain probability support for inference. Random Forest and XGBoost remain unscaled. The opt-in Stacking classifier combines a scaled probabilistic SVC with Random Forest and XGBoost base estimators and uses logistic regression as the final estimator.

Tunable classifiers use Optuna with its TPE sampler for hyperparameter optimization. The public n_trials setting explicitly declares the budget per search, with a normal default of 100 trials. Searches are sequential and seeded from the canonical RANDOM_STATE, supporting reproducible optimization under the same configuration and software environment.

Input Format

Dataset Registry

Datasets are registered under user-defined [datasets.<dataset_id>] entries. For example:

[datasets.morgan_r2_2048]
path = "data/morgan_r2_2048.npz"
label_path = "raw/labels.npy"
family = "fingerprints"
method = "Morgan"
variant = "r2_2048"
description = "Morgan fingerprints, radius 2, 2048 bits"

Here, family is optional dataset metadata used to describe the feature representation. It is unrelated to the classifier selected or evaluated by MELITE.

Each dataset must define path and label_path. Optional metadata fields such as family, method, variant, level, and description are preserved for traceability and do not trigger dataset-specific execution logic.

The legacy [benchmark] configuration section remains supported for backward compatibility. New configurations should use the dataset registry.

Array Requirements

A registered .npz dataset must contain an X array. MELITE validates that:

• X is two-dimensional;
• X is numeric;
• the number of rows in X matches the number of labels in y;
• an embedded y array, when present, matches the configured label vector.

A typical input layout is:

raw/
└── labels.npy

data/
├── morgan_r2_2048.npz
├── rdkit_descriptors.npz
├── PCA85.npz
└── UMAP90.npz

The filenames and feature families are examples only. MELITE does not require PCA, UMAP, fingerprints, descriptors, or any other specific feature-generation method.

Main Outputs

A standard MELITE workflow produces evaluation artifacts and, when requested, a final exported model:

output/
├── results.txt
├── results.csv
├── evaluations.csv
├── evaluation_folds.csv
├── optimization_searches.csv
├── optimization_provenance.json
├── figures/
│   └── evaluation_f1_macro_<dataset>.png
└── Model_<classifier>_<dataset>.pkl

The artifacts have distinct roles:

• results.txt — human-readable summary of the selected results.
• results.csv — selected classifier result for each dataset and the persisted parameter source used by melite export.
• evaluations.csv — aggregate evaluation evidence for every active classifier.
• evaluation_folds.csv — outer-CV evidence for every dataset, classifier, and outer split.
• optimization_searches.csv — one row per completed outer or final optimization search.
• optimization_provenance.json — the effective optimization and evaluation contract for the run.
• figures/evaluation_f1_macro_<dataset>.png — visualization of the outer-CV F1-macro evidence used for classifier selection.
• Model_<classifier>_<dataset>.pkl — final full-data fitted model created by melite export.

See the Output Data Contract for the column-level schemas and semantics.

The evaluation figure is generated from already-computed outer-CV evidence. It does not trigger additional fitting, tuning, cross-validation, or selection.

Configuration

MELITE uses TOML configuration files to keep execution choices explicit and reproducible. Configuration controls, among other settings:

• registered datasets and their metadata;
• active classifiers;
• random state;
• inner and outer cross-validation settings;
• optimization trial budget (n_trials);
• input and output paths.

Use --config to supply a project-specific configuration:

melite run --config my_config.toml
melite export --config my_config.toml --row 0

Smoke mode can be requested independently from the configuration:

melite run --smoke --config my_config.toml

See the full configuration reference in the project documentation.

Development

Running Tests

Run the test suite:

python -m pytest tests -q

Build the documentation in strict mode:

mkdocs build --strict

Build and check the distributions:

python -m build --no-isolation
python -m twine check dist/*

Run the installed-wheel smoke test:

python scripts/smoke_install_wheel.py

Contributing

Contributions are welcome. Please open an issue before submitting a pull request. Follow the existing code style: NumPy-style docstrings, type hints, and SPDX license headers in all source files.

See CONTRIBUTING.md for full guidelines, including the development setup and the pull request target branch. Please also read our Code of Conduct.

Documentation

The full documentation is published at:

https://nanobiostructuresrg.github.io/melite/

Citation

If you use MELITE in your research, please cite it using the metadata in CITATION.cff or the format below:

Contreras-Torres, F. F., & Murrieta, A. C. (2026). MELITE — Multi-Model Classifier Evaluator. Zenodo. https://doi.org/10.5281/zenodo.20382752

Authors

• Flavio F. Contreras-Torres — Tecnológico de Monterrey
• Ana C. Murrieta — Tecnológico de Monterrey

License

This project is licensed under the terms of the GNU Lesser General Public License v3.0 or later. SPDX identifier: LGPL-3.0-or-later.