numpy < 2

[smribet]

Changing set-up file (for v 0.14 only) to avoid numpy 2 install errors

[mlhenderson]

@smribet Yes to numpy < 2.0. Is the ncempy restriction also due to numpy?

I answered my own question; yes, ncempy 1.11.2 has numpy<2 in setup.py, and versions beyond will expect numpy >= 2.

[smribet]

Yes I think I made a mistake specifying the version though. Should it be <= 1.11.2?

[sezelt]

Out of curiosity, what specifically was the numpy 2 error?
I made an environment with py4DSTEM that uses numpy 2 and it at least imports successfully.

[mlhenderson]

<= 1.11.2 would be better, yes. There are other changes for 1.11.2 besides the numpy restriction that should be picked up, which you would not get from 1.11.1.

[mlhenderson]

we should also restrict pymatgen to < 2024.8.9, since that is pinned to numpy 2.0.0

[mlhenderson]

I did some local py4DSTEM testing using numpy 2, and emdfile has at least one incompatibility with numpy 2 currently. It makes sense to put some version limits for py4DSTEM numpy install, at least temporarily, until the problems related to dependencies can be addressed.

[bsavitzky]

Thanks Matt - what incompatibility with emdfile did you find?

…

On Thu, Aug 29, 2024 at 5:23 PM Matt Henderson ***@***.***> wrote: I did some local py4DSTEM testing using numpy 2, and emdfile has at least one incompatibility with numpy 2 currently. It makes sense to put some version limits for py4DSTEM numpy install, at least temporarily, until the problems related to dependencies can be addressed. — Reply to this email directly, view it on GitHub <#677 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABQYYNIJXURGX25V7I4SZUDZT6GNNAVCNFSM6AAAAABNK6GNQWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGMJZGAZTINRQGY> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[mlhenderson]

I filed an issue with emdfile, so far one error related to usage of numpy.string_

py4dstem/emdfile#2

[bsavitzky]

Great, thanks - fixed and pushed to a new branch `np2` for numpy 2 compatibility fixes. Will leave the branch open for now while you're testing

…

On Thu, Aug 29, 2024 at 5:58 PM Matt Henderson ***@***.***> wrote: I filed an issue with emdfile, so far one error related to usage of numpy.string_ py4dstem/emdfile#2 <py4dstem/emdfile#2> — Reply to this email directly, view it on GitHub <#677 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABQYYNNM4ZJSMOKMXTDJWQTZT6KQVAVCNFSM6AAAAABNK6GNQWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGMJZGEYDSNBSGQ> . You are receiving this because you commented.Message ID: ***@***.***>

[smribet]

@bsavitzky it sounds like we should pin emdfile as well?

[bsavitzky]

Once we've ensured numpy 2 compatibility I'll merge and version emdfile. Then we could simply bump the version dependency in py4DSTEM but leave the

=. I think this is better than pinning in this situation but am curious -

do you think pinning would be preferable?

…

On Fri, Aug 30, 2024, 11:17 AM Stephanie Ribet ***@***.***> wrote: @bsavitzky <https://github.com/bsavitzky> it sounds like we should pin emdfile as well? — Reply to this email directly, view it on GitHub <#677 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABQYYNPSWLNUOYBOF7KWVV3ZUCEJRAVCNFSM6AAAAABNK6GNQWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGMRRGU4DGMJXGA> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[smribet]

Sounds good!