munch-group-library 0.1.13

Project description

claude mcp add pubmed --transport http https://pubmed.mcp.claude.com/mcp

Template repository for a library project

Initial set up

pixi run initialize

• In the link above, replace libraryname with your library's name.
• Rename the libraryname folder to the name of your library, do not use hyphens, '-', in the name.
• Replace all instances of libraryname, modulename, and functionname for the names of/in your library (Shift-Cmd-F in VScode).

Development

In addition to any dependencies of your library, the development setup requires quarto the following packages:

conda-forge::jupyter conda-forge::nbconvert conda-forge::quartodoc conda-forge::gh conda-forge::git conda-forge::pip

Version

The only version you need to change is the one in setup.cfg.

Unit tests

Test are put in test/test_modulename.py. Run these and all other test_* files you create by running ./scripts/test.sh.

Creating docs

Use type hits (with the typing module) for code and add this to your settings.json in VScode.

"files.associations": {
    "*.qmd": "quarto"
},
"autoDocstring.docstringFormat": "numpy",
"autoDocstring.startOnNewLine": true,
"autoDocstring.customTemplatePath": "docs/autodoc.mustache",

That lets you can autogenerate doc-strings by typing """ at the top of your functions.

Publishing docs

Docs are published to your github pages site https://{user}.github.io/{libraryname}. The template is set up for the user munch-group, but you can substitute all its instances for your own github user name.

Run all documentation notebooks inplace by running ./docs-run-notebooks.sh. Build and render all documentation by running ./docs-build-render.sh.

Publishing conda package

Once pushed to github the library can be published as a conda package by running ./conda-release.sh.

Publishing the conda package to anaconda requires that you set up a github secret, ANACONDA_TOKEN, that you create on your anaconda.org account.

Library dependencies

Dependencies are managed in pyproject.toml and automatically read by conda-build/meta.yaml.

Entry points

Entry points are defined in pyproject.toml and automatically read by conda-build/meta.yaml.