Metadata-Version: 2.1
Name: sktmls
Version: 2020.12.16rc41
Summary: MLS SDK
Home-page: https://github.com/sktaiflow/mls-sdk
Author: SKTMLS
Author-email: mls@sktai.io
License: UNKNOWN
Platform: UNKNOWN
Classifier: Programming Language :: Python :: 3
Classifier: License :: OSI Approved :: MIT License
Classifier: Operating System :: OS Independent
Requires-Python: >=3.6
Description-Content-Type: text/markdown
Requires-Dist: numpy (<1.20,>=1.19.4)
Requires-Dist: catboost (<0.25.0,>=0.24.3)
Requires-Dist: joblib
Requires-Dist: lightgbm (<2.4.0,>=2.3.1)
Requires-Dist: pandas (<1.2.0,>=1.1.1)
Requires-Dist: pytz
Requires-Dist: requests
Requires-Dist: scikit-learn (<0.24,>=0.23.2)
Requires-Dist: scipy (<1.5.0,>=1.4.1)
Requires-Dist: xgboost (<1.3,>=1.2.1)
Requires-Dist: python-dateutil
Requires-Dist: matplotlib (<3.3,>=3.2.1)
Requires-Dist: tqdm (>=4.51.0)
Requires-Dist: boto3
Requires-Dist: ConfigSpace (<=0.4.10)
Requires-Dist: cryptography (>=3.2.1)
Requires-Dist: cython
Requires-Dist: dask (>=2.6.0)
Requires-Dist: distributed (>=2.6.0)
Requires-Dist: fastparquet (==0.4.1)
Requires-Dist: gluoncv (<0.9.0,>=0.5.0)
Requires-Dist: graphviz (<0.9.0,>=0.8.1)
Requires-Dist: mxnet (<2.0.0)
Requires-Dist: networkx (<3.0,>=2.3)
Requires-Dist: paramiko (>=2.4)
Requires-Dist: Pillow (>=8.0.1)
Requires-Dist: psutil (>=5.7.3)
Requires-Dist: pyarrow (>=2.0.0)
Requires-Dist: scikit-optimize
Requires-Dist: tornado (>=6.1)

# mls-model-registry (sktmls)

## Contents

- [Description](#description)
- [How to use](#how-to-use)
- [Development](#development)
  - [Requirements for development](#requirements-for-development)
  - [Local model registry](#local-model-registry)
  - [Python environment](#python-environment)
  - [Documents generation](#documents-generation)
- [Version](#version)

## Description

A Python package for MLS model registry.

This python package includes
- Customized prediction pipelines inheriting MLSModel
- Model uploader to AWS S3 for meta management and online prediction

## Installation

Installation is automatically done by training containers in YE. If you want to install manually for local machines,

```bash
# develop
pip install --index-url https://test.pypi.org/simple/ --no-deps sktmls

# production
pip install sktmls
```

## How to use

- MLS Docs: https://ab.sktmls.com/docs/model-registry
- sktmls Docs: https://sktaiflow.github.io/mls-sdk/sktmls

## Development

### Requirements for development
- Python 3.6
- requirements.txt
- requirements-dev.txt

### Local model registry

To enable all model related features in local environment, you need to create a directory `models` in your home directory.

```bash
$ cd ~/
$ mkdir models
```

### Python environment

First you need to do the followings

```bash
$ python -V # Check if the version is 3.6.
$ python -m venv env # Create a virtualenv.
$ . env/bin/activate # Activate the env.
$ pip install "numpy>=1.19.4,<1.20" # Install numpy to avoid a requirement error.
$ pip install -r requirements.txt # Install required packages.
$ pip install -r requirements-dev.txt # Install required dev packages.
```

### Documents generation

Before a commit, generate documents if any docstring has been changed

```bash
rm -rf docs
pdoc --html --config show_source_code=False -f -o ./docs sktmls
```

### Version
`sktmls` package version is automatically genereated followd by a production release on format `YY.MM.DD`  
We use [Calendar Versioning](https://calver.org). For version available, see the [tags on this repository](https://github.com/sktaiflow/mls-model-registry/releases).  


