dmqclib 0.2.5

This package aims to offer helper functions that simplify model building and evaluation

DMQCLib

The DMQCLib package offers helper functions and classes that simplify model building and evaluation for the AIQC project.

Installation

The package is indexed on PyPI and Anaconda.org, allowing you to install it using either pip or conda.

Using pip:

pip install dmqclib

Using conda:

conda install takayasaito::dmqclib 

Contribution

Package Manager

You can create a new environment using any package management system, such as conda and mamba.

Additionally, using uv is recommended when contributing modifications to the package.

• uv

After the installation of uv, running uv sync inside the project will create the environment.

Example of Environment Setup

For example, the following commands create a new conda environment with mamba and set up the library environment with uv:

mamba create -n aiqc -c conda-forge python=3.12
mamba activate aiqc
mamba install uv

cd /your/path/to/dmqclib
uv sync

Unit Test

You may need to install the library in editable mode at least once before running unit tests.

uv pip install -e .

After the library installation, you can run unit tests with pytest.

uv run pytest -v

Python Linter

To lint the code under the src folder with ruff, use the following command:

uvx ruff check src

and the unit test code under the tests folder:

uvx ruff check tests

Alternatively, to lint the code with flake8, use the following command:

uvx flake8 src --max-line-length=100

Code Formatter

To format the code under the src folder with ruff, use the following command:

uvx ruff format src

and the unit test code under the tests folder:

uvx ruff format tests

Alternatively, to format the code with black, use the following command:

uvx black src

Deployment

Release to PyPI

The GitHub Action (.github/workflows/publish_to_pypi.yaml) automatically publishes the package to PyPI whenever a GitHub release is created.

Alternatively, you can manually publish the package to PyPI:

uv build
uv publish --token pypi-xxxx-xxxx-xxxx-xxxx

Release to Anaconda.org

Unlike using a GitHub Action for PyPI, publishing to Anaconda.org is a manual process.

You’ll need the following tools:

• conda-build
• anaconda-client
• grayskull

Install them (preferably in a dedicated environment):

conda install -c conda-forge conda-build anaconda-client grayskull

1. Generate the conda recipe with Grayskull

From the project root, run:

grayskull pypi dmqclib

This creates a meta.yaml file in the dmqclib/ directory.

2. Build the package

cd dmqclib
conda build .
cd ..

This creates a .conda package in your local conda-bld directory (e.g., ~/miniconda3/conda-bld/noarch/).

3. Upload to Anaconda.org

anaconda login
anaconda upload /full/path/to/conda-bld/noarch/dmqclib-<version>-<build>.conda

4. Keep the recipe under version control

cp dmqclib/meta.yaml conda/meta.yaml
rm -r dmqclib