Make a release, decide versioning and understand Github packaging

[nathanshammah]

• The button "release" on top of the menu makes a Github release.(Also, once it will be public, Github releases maybe can be used to automate conda releases.)
• We need to agree on version number. What about "mitiq v0.1-alpha"? or beta? I have no strong feelings about it.
• There is the possibility to release "packages" hosted by Github, also for private repositories. Still understanding what it means.
• We could add a markdown file with information on how to make a release (it will have pip and conda too, and mention updating the documentation).
• I did not register mitiq on pypi and test-pypi, as some package and related info needs to be uploaded. Not sure if they would allow an empty package to be there.

[willzeng]

• mitiq v0.1-alpha works for me.
• it doesn't look like Github packages yet supports Python packaging: https://help.github.com/en/packages/publishing-and-managing-packages/about-github-packages#supported-clients-and-formats I may have missed something there though
• adding markdown info for how to release will be good
• is it possible to just commit a placeholder release on pypi that then gets replaced? Like a mitiq v0.0 that is just a relatively empty setup.py file?

@nathanshammah

[nathanshammah]

I'll proceed.

[nathanshammah]

I had to modify a bit the setup.py and struggle a bit but now it's on: https://test.pypi.org/project/mitiq/ (the build badge I think doesn't render as the repo is private.) I'll add also on pypi main channel.

[nathanshammah]

By installing with

pip install -i https://test.pypi.org/simple/ mitiq

in an empty python 3.8 miniconda environment, I get this error message

ERROR: Could not find a version that satisfies the requirement numpy>=1.18.2 (from mitiq) (from versions: 1.9.3)
ERROR: No matching distribution found for numpy>=1.18.2 (from mitiq)

although I just updated miniconda. I changed the requirements from numpy==1.18.2 to numpy>=1.18.2, but that should not be an issue.
I don't get it with the pip install ., done locally. However, after installing separately pip install numpy>=1.18.2 (and uninstalling mitiq), re-installing mitiq from test-pypi works.

Numpy v1.18.2 is not bleeding edge, but very very recent (anaconda hosts only v1.18.1, updated 11 days ago). Maybe, if everything works fine, we could downgrade to lower versions if we do not need so recent versions. Still I don't get why the package is not collected, as it should use pip.

By the way, I saw that another way to set requirements is numpy~=1.18.2, which will install an updated micro, but not a new macro, for backward compatibility.

We need to add scipy to dependencies, as it's now called in factories.py

from scipy.optimize import curve_fit

Anyways, this is just to have a placeholder.

[nathanshammah]

It's also registered on pypi, you can check with pip install mitiq==0.0.0.

[willzeng]

@nathanshammah Is there anything left to do on this issue? I'm going to close it but otherwise please do reopen.

[nathanshammah]

It could be nice to do a private Github release as we close milestone 2, cc @willzeng. This way we can also add to the internal report and pre-alpha users that we have a release (albeit not public).

[willzeng]

I'm on board. Let's do an alpha release when milestone 2 closes. (Nothing on pypi yet)

On Tue, Jun 2, 2020, 5:36 AM Nathan Shammah notifications@github.com wrote:

It could be nice to do a private Github release as we close milestone 2, cc @willzeng https://github.com/willzeng. This way we can also add to the internal report and pre-alpha users that we have a release (albeit not public).

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/unitaryfund/mitiq/issues/80#issuecomment-637420389, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABHZDAWF5M27U6WPJR2CJE3RUTBYBANCNFSM4LY6TSDA .