tuwien-cms / xprec

Full quadruple precision (double-double) data type for numpy
MIT License
12 stars 4 forks source link

Add conda recipe and deployment #4

Closed shinaoka closed 2 years ago

shinaoka commented 2 years ago

For the sake of convenience of using SparseIR.jl, I want to upload conda packages to my own account on anaconda each time a new tag is created.

The following is how it works.

  1. When a new tag is detected, workers are created using linux/macos/windows for Python 3.6/3.7/3.8/3.9 on GitHub Actions. (I failed to support Python 3.10 for now due to some conflict in building conda package)
  2. In each worker, we build a conda package by parsing setup.py and recipe/meta.yaml.
  3. The build packages are uploaded to the anaconda cloud.

You can publish a new version in the same way as before.

  1. Update the version number in __init__.py and commit the changes.
  2. Create a new version tag.
  3. Push the commit and the new tag to GitHub.
  4. [Then, workers on GitHub Actions will be invoked]

This change is harmless because the resultant packages are uploaded to my own channel/account for now. We can simply merge this PR, and see how it works when a new tag is created.

For the moment, you can see how it works here.

mwallerb commented 2 years ago

Great work! I'm just gonna merge this :)