Closed mmacata closed 1 year ago
Added github actions to auto-update version in pyproject.toml.
https://test.pypi.org/project/actinia-core 4.8.1 is somehow broken...
pip install -i https://test.pypi.org/simple/ actinia-core
Workflows were tested via mmacata fork - I created a release there to test the publishing to pypi and the version update in pyproject.toml. pip3 install actinia-core
was successful. I now deleted the resource again to not interfere with this actinia-org remote.
One question is left to @neteler and @anikaweinmann :
In this PR I suggest to rename some scripts, e.g. rq_custom_worker
can now simply be called with actinia-worker
but also actinia_user
-> actinia-user
. I thought this is a good idea as we renamed actinia_core in general to actinia-core.
But would it be an API breaking change? Do you think it is a good idea, especially as we have actinia_user in all start.sh
scripts?
Workflows were tested via mmacata fork - I created a release there to test the publishing to pypi and the version update in pyproject.toml.
pip3 install actinia-core
was successful. I now deleted the resource again to not interfere with this actinia-org remote.One question is left to @neteler and @anikaweinmann : In this PR I suggest to rename some scripts, e.g.
rq_custom_worker
can now simply be called withactinia-worker
but alsoactinia_user
->actinia-user
. I thought this is a good idea as we renamed actinia_core in general to actinia-core. But would it be an API breaking change? Do you think it is a good idea, especially as we have actinia_user in allstart.sh
scripts?
I think it is in general a good idea. Can we do it like other libraries and have first both actinia_user
and actinia-user
and actinia_user
prints a warning, that it will be depricated soon?
Workflows were tested via mmacata fork - I created a release there to test the publishing to pypi and the version update in pyproject.toml.
pip3 install actinia-core
was successful. I now deleted the resource again to not interfere with this actinia-org remote. One question is left to @neteler and @anikaweinmann : In this PR I suggest to rename some scripts, e.g.rq_custom_worker
can now simply be called withactinia-worker
but alsoactinia_user
->actinia-user
. I thought this is a good idea as we renamed actinia_core in general to actinia-core. But would it be an API breaking change? Do you think it is a good idea, especially as we have actinia_user in allstart.sh
scripts?I think it is in general a good idea. Can we do it like other libraries and have first both
actinia_user
andactinia-user
andactinia_user
prints a warning, that it will be depricated soon?
Yes I added this. It now looks like this:
/src/actinia_core # actinia-user list
[2023-05-09 14:57:58,717] INFO : actinia-core.logging_interface -Log level is set to 'INFO' [in /usr/lib/python3.9/site-packages/actinia_core/core/logging_interface.py:254]
[b'actinia-gdi']
['actinia-gdi']
/src/actinia_core # actinia_user list
[2023-05-09 14:58:06,922] INFO : actinia-core.logging_interface -Log level is set to 'INFO' [in /usr/lib/python3.9/site-packages/actinia_core/core/logging_interface.py:254]
[2023-05-09 14:58:06,927] WARNING : actinia-core.actinia_user -The call "actinia_user" is deprecated and will be removed soon. User "actinia-user" instead! [in /usr/lib/python3.9/site-packages/actinia_core/cli/actinia_user.py:143]
[b'actinia-gdi]
['actinia-gdi']
Cool! :-) Just nitpicking: may I suggest "command" rather than "call" since it is on command line (?) in https://github.com/actinia-org/actinia-core/pull/373/commits/38944000cf6f6973d826671cb4b6aa1e4e4ab7c2#diff-26f6ed23fc8ccf9f2532e41ffac1f25448937225ec0f973e23d6921a3da072b8R56
Workflows were tested via mmacata fork - I created a release there to test the publishing to pypi and the version update in pyproject.toml.
pip3 install actinia-core
was successful. I now deleted the resource again to not interfere with this actinia-org remote. One question is left to @neteler and @anikaweinmann : In this PR I suggest to rename some scripts, e.g.rq_custom_worker
can now simply be called withactinia-worker
but alsoactinia_user
->actinia-user
. I thought this is a good idea as we renamed actinia_core in general to actinia-core. But would it be an API breaking change? Do you think it is a good idea, especially as we have actinia_user in allstart.sh
scripts?I think it is in general a good idea. Can we do it like other libraries and have first both
actinia_user
andactinia-user
andactinia_user
prints a warning, that it will be depricated soon?Yes I added this. It now looks like this:
/src/actinia_core # actinia-user list [2023-05-09 14:57:58,717] INFO : actinia-core.logging_interface -Log level is set to 'INFO' [in /usr/lib/python3.9/site-packages/actinia_core/core/logging_interface.py:254] [b'actinia-gdi'] ['actinia-gdi'] /src/actinia_core # actinia_user list [2023-05-09 14:58:06,922] INFO : actinia-core.logging_interface -Log level is set to 'INFO' [in /usr/lib/python3.9/site-packages/actinia_core/core/logging_interface.py:254] [2023-05-09 14:58:06,927] WARNING : actinia-core.actinia_user -The call "actinia_user" is deprecated and will be removed soon. User "actinia-user" instead! [in /usr/lib/python3.9/site-packages/actinia_core/cli/actinia_user.py:143] [b'actinia-gdi] ['actinia-gdi']
I oversaw that I renamed the files and the commands didn't change for 3/4 of them :see_no_evil:
Definitely not good working such a long time on one PR :upside_down_face:
I will undo the warnings for all except rq_custom_worker
.
As this is not widely used up to now - do you still see the need to do a major version increase @anikaweinmann ?
Workflows were tested via mmacata fork - I created a release there to test the publishing to pypi and the version update in pyproject.toml.
pip3 install actinia-core
was successful. I now deleted the resource again to not interfere with this actinia-org remote. One question is left to @neteler and @anikaweinmann : In this PR I suggest to rename some scripts, e.g.rq_custom_worker
can now simply be called withactinia-worker
but alsoactinia_user
->actinia-user
. I thought this is a good idea as we renamed actinia_core in general to actinia-core. But would it be an API breaking change? Do you think it is a good idea, especially as we have actinia_user in allstart.sh
scripts?I think it is in general a good idea. Can we do it like other libraries and have first both
actinia_user
andactinia-user
andactinia_user
prints a warning, that it will be depricated soon?Yes I added this. It now looks like this:
/src/actinia_core # actinia-user list [2023-05-09 14:57:58,717] INFO : actinia-core.logging_interface -Log level is set to 'INFO' [in /usr/lib/python3.9/site-packages/actinia_core/core/logging_interface.py:254] [b'actinia-gdi'] ['actinia-gdi'] /src/actinia_core # actinia_user list [2023-05-09 14:58:06,922] INFO : actinia-core.logging_interface -Log level is set to 'INFO' [in /usr/lib/python3.9/site-packages/actinia_core/core/logging_interface.py:254] [2023-05-09 14:58:06,927] WARNING : actinia-core.actinia_user -The call "actinia_user" is deprecated and will be removed soon. User "actinia-user" instead! [in /usr/lib/python3.9/site-packages/actinia_core/cli/actinia_user.py:143] [b'actinia-gdi] ['actinia-gdi']
I oversaw that I renamed the files and the commands didn't change for 3/4 of them :see_no_evil:
Definitely not good working such a long time on one PR :upside_down_face:
I will undo the warnings for all except rq_custom_worker
.
As this is not widely used up to now - do you still see the need to do a major version increase @anikaweinmann ?
This PR adresses #370
rq_custom_worker
can now simply be called withactinia-worker
TODOs:
the hard version pinning is not best practise, but won't work otherwise due to dependency conflicts like #345clear duplicate config in setup.py, setup.cfg and pyproject.tomlERROR HTTPError: 400 Bad Request from https://test.pypi.org/legacy/ Invalid value for requires_dist. Error: Can't have direct dependency: 'actinia-api @ https://github.com/mundialis/actinia-api/releases/download/3.2.0/actinia_api-3.2.0-py3-none-any.whl'
)~~add GHA