[DOC][BUG] Add warning regarding issues with macOS ARM #4010

dainelli98 · 2022-12-28T17:44:15Z

Reference Issues/PRs

What does this implement/fix? Explain your changes.

This PR adds a warning for macOS users using ARM macs of the issues that they may encounter when installing all_extras, related to some dependencies not compatible with that kind of setup.

Does your contribution introduce a new dependency? If yes, which one?

This contribution does not add any new dependency.

What should a reviewer concentrate their feedback on?

Check if the warning is enough, or it would be appropriate to add a link to a separate new article in the documentation explaining in more detail the issue, listing the dependencies that cause the error, in case the user wants to install all the other dependencies in all_extras that do not generate any errors.

PR checklist

For all contributions

I've added myself to the list of contributors.
The PR title starts with either [ENH], [MNT], [DOC], or [BUG] indicating whether the PR topic is related to enhancement, maintenance, documentation, or bug.

Add missing dot at the end of warning regarding pre-existing conda environment with the same name as the one suggested in installation instructions.

@dainelli98

Add myself, @dainelli98, to list of contributors in .all_contributorsrc.

Add warning in installation documentation explaining that Some of the dependencies included in ``all_extras`` do not work on mac ARM-based processors, such as M1, M1Pro, M1Max or M1Ultra.

docs/source/installation.rst

fkiraly

Thanks, great!

If you have a workaround that can be added, that would also be nice, but it´s not blocking.

Add workaround in troubleshooting section explaining which libraries do not work on macOS ARM devices and which ones should be avoided or replaced to keep the maximum amount of dependencies from ``all_extras``.

…ubleshooting

…ncy warning

…ion extensions. Add ``sphinx.ext.autosectionlabel`` to ``conf.py`` extensions to enable the possibility to add links to sections in documentation.

docs/source/installation.rst

fkiraly

Excellent, thanks!
I don't think the troubleshooting paragraph is too long, it is clear and helpful.

No blocking comments from my side.

I left some comments on how you could improve formulations - in tech writing and espceially tech instructions, it's worth being very clear and precise (since the reader will be someone who just got frustrated and possibly confused by a failed install).
Not blocking, up to you if you want to give it another pass to sharpen the formulations (just let me know if not).

fkiraly · 2022-12-31T12:09:56Z

Other question: I do sometimes run into trouble with numba on M1. Did you encounter that at all?

dainelli98 · 2022-12-31T12:20:42Z

Other question: I do sometimes run into trouble with numba on M1. Did you encounter that at all?

To be honest, it has not been that long that I bought my current laptop. As of now, I have not experienced any issue, but that could be due to the fact that I have yet to use a package that leverages numba. Anyway, this also interests me, so I will check. If you want, I can get back to you with what I get.

dainelli98 · 2022-12-31T12:21:54Z

Excellent, thanks! I don't think the troubleshooting paragraph is too long, it is clear and helpful.

No blocking comments from my side.

I left some comments on how you could improve formulations - in tech writing and espceially tech instructions, it's worth being very clear and precise (since the reader will be someone who just got frustrated and possibly confused by a failed install). Not blocking, up to you if you want to give it another pass to sharpen the formulations (just let me know if not).

You are right, thanks for the advice. I will make the change, as I do not see any reason to not be as precise as possible.

Sharpen formulation so that the issue description and the workaround are more clear and tight to the actual actions that need to take place in order to avoid the issue.

dainelli98 · 2022-12-31T12:46:22Z

On the numba issue. As you may know already, Python packages can be distributed as wheels or source distribution. The first comes ready to be used, but the second needs to be built, which can cause issues on an ARM-based processor (The same issue occurs with prophet do to its dependency to pystan, that comes as source distribution).

In the case of numba, the proposed solution is to install it using conda:

conda install numba

Another solution seems to be to install with pip ta version that has a wheel, that is distributed as wheel:

pip install numba==0.55.2

It seems that as of now the last version is provided as a wheel, but this was done lately. So you may have experienced the issue when it was not like this.

It is weird that they sometimes do not provide a wheel for the last versions, but both workarounds seem to work on my mac.

fkiraly · 2022-12-31T12:53:03Z

It is weird that they sometimes do not provide a wheel for the last versions, but both workarounds seem to work on my mac.

Is that worth adding, too?

fkiraly · 2022-12-31T12:53:39Z

Thanks for ths tipps btw! At work, we have a mix of ARM and non-ARM macs, and the two behave differently...

fkiraly · 2022-12-31T12:54:29Z

but that could be due to the fact that I have yet to use a package that leverages numba

sktime has numba as a core dependency, but it is used only in a couple estimators (not in the framework).

Add workaround information on how to avoid issue related to installing packages distributed as source distributions on devices using ARM-based processors.

dainelli98 · 2022-12-31T13:21:25Z

It is weird that they sometimes do not provide a wheel for the last versions, but both workarounds seem to work on my mac.

Is that worth adding, too?

The only reason to add it would be because the issue that the last version of numba that now is fixed could happen in the future with another package if they do a new release without providing the wheel.

As of now, all the packages not mentioned in the workaround do not present this issue, so at this moment this problem should not arise. If you think this is a bit overkill, it is easy to undo the commits.

fkiraly · 2022-12-31T15:16:07Z

ok, makes sense. Let me know when you think you're done, for review and potential merge.

dainelli98 · 2022-12-31T15:22:13Z

ok, makes sense. Let me know when you think you're done, for review and potential merge.

For me it is ready for review as it is.

fkiraly

Excellent, great contribution!

Hope this will be helpful to all the poor souls with ARM macs out there.

fkiraly · 2022-12-31T20:33:42Z

failures unrelated to this PR, see #4033

dainelli98 added 3 commits December 28, 2022 18:00

fix(documentation): add missing dot in warning

20643ad

Add missing dot at the end of warning regarding pre-existing conda environment with the same name as the one suggested in installation instructions.

feat(documentation): add @dainelli98 to contributors

d2c49bf

Add myself, @dainelli98, to list of contributors in .all_contributorsrc.

feat(documentation): add warning about macOS ARM dependency issues

b27fcde

Add warning in installation documentation explaining that Some of the dependencies included in ``all_extras`` do not work on mac ARM-based processors, such as M1, M1Pro, M1Max or M1Ultra.

dainelli98 marked this pull request as ready for review December 28, 2022 17:44

dainelli98 requested review from fkiraly, aiwalter and GuzalBulatova as code owners December 28, 2022 17:44

fkiraly reviewed Dec 28, 2022

View reviewed changes

docs/source/installation.rst Outdated Show resolved Hide resolved

fkiraly previously approved these changes Dec 28, 2022

View reviewed changes

feat(documentation): add workaround for macOS ARM dependency issues

726a1b7

Add workaround in troubleshooting section explaining which libraries do not work on macOS ARM devices and which ones should be avoided or replaced to keep the maximum amount of dependencies from ``all_extras``.

dainelli98 dismissed fkiraly’s stale review via 726a1b7 December 28, 2022 19:15

dainelli98 added 3 commits December 28, 2022 20:45

fix(documentation): add right quotation to packages in dependency tro…

4f956d7

…ubleshooting

feat(documentation): add link to troubleshooting to macOS ARM depende…

db5a709

…ncy warning

fix(documentation): add sphinx.ext.autosectionlabel to documentat…

df82ace

…ion extensions. Add ``sphinx.ext.autosectionlabel`` to ``conf.py`` extensions to enable the possibility to add links to sections in documentation.

fkiraly reviewed Dec 31, 2022

View reviewed changes

docs/source/installation.rst Outdated Show resolved Hide resolved

fkiraly reviewed Dec 31, 2022

View reviewed changes

docs/source/installation.rst Outdated Show resolved Hide resolved

fkiraly reviewed Dec 31, 2022

View reviewed changes

docs/source/installation.rst Outdated Show resolved Hide resolved

fkiraly reviewed Dec 31, 2022

View reviewed changes

docs/source/installation.rst Outdated Show resolved Hide resolved

fkiraly previously approved these changes Dec 31, 2022

View reviewed changes

fix(documentation): sharpen formulations in new documentation

1691b93

Sharpen formulation so that the issue description and the workaround are more clear and tight to the actual actions that need to take place in order to avoid the issue.

dainelli98 dismissed fkiraly’s stale review via 1691b93 December 31, 2022 12:34

dainelli98 added 2 commits December 31, 2022 13:58

feat(documentation): add workaround for source distribution packages

9a107d5

Add workaround information on how to avoid issue related to installing packages distributed as source distributions on devices using ARM-based processors.

fix(documentation): correct typo on source distribution workaround

25e894a

fkiraly approved these changes Dec 31, 2022

View reviewed changes

fkiraly merged commit 8144769 into sktime:main Dec 31, 2022

dainelli98 deleted the feature/document_mac_arm_issues branch January 2, 2023 11:34

hazrulakmal mentioned this pull request Aug 18, 2023

[DOC] Fix make_pipeline, make_reduction, window_summarizer & load_forecasting data docstrings #5065

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[DOC][BUG] Add warning regarding issues with macOS ARM #4010

[DOC][BUG] Add warning regarding issues with macOS ARM #4010

dainelli98 commented Dec 28, 2022

fkiraly left a comment

fkiraly left a comment

fkiraly commented Dec 31, 2022

dainelli98 commented Dec 31, 2022

dainelli98 commented Dec 31, 2022

dainelli98 commented Dec 31, 2022 •

edited

Loading

fkiraly commented Dec 31, 2022

fkiraly commented Dec 31, 2022

fkiraly commented Dec 31, 2022

dainelli98 commented Dec 31, 2022

fkiraly commented Dec 31, 2022

dainelli98 commented Dec 31, 2022 •

edited

Loading

fkiraly left a comment

fkiraly commented Dec 31, 2022

[DOC][BUG] Add warning regarding issues with macOS ARM #4010

[DOC][BUG] Add warning regarding issues with macOS ARM #4010

Conversation

dainelli98 commented Dec 28, 2022

Reference Issues/PRs

What does this implement/fix? Explain your changes.

Does your contribution introduce a new dependency? If yes, which one?

What should a reviewer concentrate their feedback on?

PR checklist

For all contributions

fkiraly left a comment

Choose a reason for hiding this comment

fkiraly left a comment

Choose a reason for hiding this comment

fkiraly commented Dec 31, 2022

dainelli98 commented Dec 31, 2022

dainelli98 commented Dec 31, 2022

dainelli98 commented Dec 31, 2022 • edited Loading

fkiraly commented Dec 31, 2022

fkiraly commented Dec 31, 2022

fkiraly commented Dec 31, 2022

dainelli98 commented Dec 31, 2022

fkiraly commented Dec 31, 2022

dainelli98 commented Dec 31, 2022 • edited Loading

fkiraly left a comment

Choose a reason for hiding this comment

fkiraly commented Dec 31, 2022

dainelli98 commented Dec 31, 2022 •

edited

Loading

dainelli98 commented Dec 31, 2022 •

edited

Loading