Publishing Conda libraries

This repository template is set up to publish a Conda library into Foundry. The build.gradle file configures the publish task to only run when the repository is tagged. You can create a new tag from the "Branches" tab.

By default, the repository's name at creation time is used as the name for the Conda package. It is possible to change the name of the package by updating the condaPackageName variable in the gradle.properties file. Note that since this is a hidden file, you will need to enable "Show hidden files and folders".

Important: underscores in the repository name are rewritten to dash. For example, if your repository is named my_library, then the library will be published as my-library.

Consuming Conda Libraries

Consumers will require read access on this repository to be able to consume the libraries it publishes. They can search for them in the Libraries section on the left-hand side in the consuming code repository. This will automatically add the dependency to meta.yaml and configure the appropriate Artifacts backing repositories.

Adding a library to your project will install packages from the source directory. The source directory defaults to src/ and we recommend not changing this. You still need to import packages before you can use them in your module. Be aware that you have to import package name and not library name (in this template, the package name is myproject).

Example

Let's say your library structure is:

conda_recipe/
src/
  examplepkg/
    __init__.py
    mymodule.py
  otherpkg/
    __init__.py
    utils.py
  setup.cfg
  setup.py

And in gradle.properties, the value of condaPackageName is mylibrary.

When consuming this library, the consuming repository's conda_recipe/meta.yaml file will contain:

requirements:
  run:
    - mylibrary

Then the packages, which in this example are examplepkg and otherpkg, can be imported as follows:

import examplepkg as E
from examplepkg import mymodule
from otherpkg.utils import some_function_in_utils

Note that the import will fail if the package does not include a file named __init__.py