glue
glue copied to clipboard
Latest release not in conda?
Describe the bug
I can no install the latest release using conda
To Reproduce The latest release in https://anaconda.org/glueviz/glueviz is almost a year old (1.2.0), even though there is a release from two months ago (1.5.0). Even worse, when I try to install the package I get version 1.0.0 (or 1.0.1?)
$ conda install -c glueviz glueviz
Collecting package metadata (current_repodata.json): done
Solving environment: done
## Package Plan ##
environment location: /home/gabriel/miniconda3/envs/py3
added / updated specs:
- glueviz
The following NEW packages will be INSTALLED:
asttokens pkgs/main/noarch::asttokens-2.0.5-pyhd3eb1b0_0
backcall pkgs/main/noarch::backcall-0.2.0-pyhd3eb1b0_0
blosc pkgs/main/linux-64::blosc-1.21.0-h8c45485_0
bottleneck pkgs/main/linux-64::bottleneck-1.3.5-py310ha9d4c09_0
brunsli pkgs/main/linux-64::brunsli-0.1-h2531618_0
cfitsio astroconda/linux-64::cfitsio-3.470-1
charls pkgs/main/linux-64::charls-2.2.0-h2531618_0
cloudpickle pkgs/main/noarch::cloudpickle-2.0.0-pyhd3eb1b0_0
cytoolz pkgs/main/linux-64::cytoolz-0.11.0-py310h7f8727e_0
dask-core pkgs/main/linux-64::dask-core-2022.7.0-py310h06a4308_0
debugpy pkgs/main/linux-64::debugpy-1.5.1-py310h295c915_0
decorator pkgs/main/noarch::decorator-5.1.1-pyhd3eb1b0_0
dill pkgs/main/noarch::dill-0.3.4-pyhd3eb1b0_0
echo glueviz/noarch::echo-0.5-pyh9f0ad1d_0
executing pkgs/main/noarch::executing-0.8.3-pyhd3eb1b0_0
fast-histogram glueviz/linux-64::fast-histogram-0.10-py310h96516ba_1
font-ttf-dejavu-s~ pkgs/main/noarch::font-ttf-dejavu-sans-mono-2.37-hd3eb1b0_0
font-ttf-inconsol~ pkgs/main/noarch::font-ttf-inconsolata-2.001-hcb22688_0
font-ttf-source-c~ pkgs/main/noarch::font-ttf-source-code-pro-2.030-hd3eb1b0_0
font-ttf-ubuntu pkgs/main/noarch::font-ttf-ubuntu-0.83-h8b1ccd4_0
fonts-anaconda pkgs/main/noarch::fonts-anaconda-1-h8fa9717_0
fonts-conda-ecosy~ pkgs/main/noarch::fonts-conda-ecosystem-1-hd3eb1b0_0
freetype-py glueviz/noarch::freetype-py-2.2.0-pyh9f0ad1d_0
fsspec pkgs/main/linux-64::fsspec-2022.3.0-py310h06a4308_0
glue-core pkgs/main/linux-64::glue-core-1.0.1-py310h06a4308_1
glue-vispy-viewers glueviz/noarch::glue-vispy-viewers-1.0.5-py_1
glueviz glueviz/linux-64::glueviz-1.0.0-0
h5py pkgs/main/linux-64::h5py-3.7.0-py310he06866b_0
hdf5 pkgs/main/linux-64::hdf5-1.10.6-hb1b8bf9_0
hsluv pkgs/main/noarch::hsluv-5.0.2-pyhd3eb1b0_0
imagecodecs pkgs/main/linux-64::imagecodecs-2021.8.26-py310h2a2ad71_0
imageio pkgs/main/linux-64::imageio-2.19.3-py310h06a4308_0
ipykernel pkgs/main/linux-64::ipykernel-6.9.1-py310h06a4308_0
ipython pkgs/main/linux-64::ipython-8.4.0-py310h06a4308_0
ipython_genutils pkgs/main/noarch::ipython_genutils-0.2.0-pyhd3eb1b0_1
jedi pkgs/main/linux-64::jedi-0.18.1-py310h06a4308_1
jupyter_client pkgs/main/noarch::jupyter_client-6.1.12-pyhd3eb1b0_0
jupyter_core pkgs/main/linux-64::jupyter_core-4.10.0-py310h06a4308_0
jxrlib pkgs/main/linux-64::jxrlib-1.1-h7b6447c_2
lerc pkgs/main/linux-64::lerc-3.0-h295c915_0
libaec pkgs/main/linux-64::libaec-1.0.4-he6710b0_1
libdeflate pkgs/main/linux-64::libdeflate-1.8-h7f8727e_5
libgcc pkgs/main/linux-64::libgcc-7.2.0-h69d50b8_2
libsodium pkgs/main/linux-64::libsodium-1.0.18-h7b6447c_0
libzopfli pkgs/main/linux-64::libzopfli-1.0.3-he6710b0_0
locket pkgs/main/linux-64::locket-1.0.0-py310h06a4308_0
matplotlib-inline pkgs/main/noarch::matplotlib-inline-0.1.2-pyhd3eb1b0_2
mpl-scatter-densi~ glueviz/noarch::mpl-scatter-density-0.7-py_0
nest-asyncio pkgs/main/linux-64::nest-asyncio-1.5.5-py310h06a4308_0
networkx pkgs/main/linux-64::networkx-2.8.4-py310h06a4308_0
numexpr pkgs/main/linux-64::numexpr-2.8.3-py310hcea2de6_0
openjpeg pkgs/main/linux-64::openjpeg-2.4.0-h3ad879b_0
pandas pkgs/main/linux-64::pandas-1.4.3-py310h6a678d5_0
parso pkgs/main/noarch::parso-0.8.3-pyhd3eb1b0_0
partd pkgs/main/noarch::partd-1.2.0-pyhd3eb1b0_1
pexpect pkgs/main/noarch::pexpect-4.8.0-pyhd3eb1b0_3
pickleshare pkgs/main/noarch::pickleshare-0.7.5-pyhd3eb1b0_1003
plotly pkgs/main/linux-64::plotly-5.9.0-py310h06a4308_0
prompt-toolkit pkgs/main/noarch::prompt-toolkit-3.0.20-pyhd3eb1b0_0
ptyprocess pkgs/main/noarch::ptyprocess-0.7.0-pyhd3eb1b0_2
pure_eval pkgs/main/noarch::pure_eval-0.2.2-pyhd3eb1b0_0
pygments pkgs/main/noarch::pygments-2.11.2-pyhd3eb1b0_0
pyopengl glueviz/noarch::pyopengl-3.1.5-py_0
python_abi glueviz/linux-64::python_abi-3.10-2_cp310
pytz pkgs/main/linux-64::pytz-2022.1-py310h06a4308_0
pywavelets pkgs/main/linux-64::pywavelets-1.3.0-py310h7f8727e_0
pyzmq pkgs/main/linux-64::pyzmq-23.2.0-py310h6a678d5_0
qt pkgs/main/linux-64::qt-5.9.7-h5867ecd_1
qtconsole pkgs/main/linux-64::qtconsole-5.3.1-py310h06a4308_1
qtpy pkgs/main/noarch::qtpy-2.0.1-pyhd3eb1b0_0
scikit-image pkgs/main/linux-64::scikit-image-0.19.2-py310h00e6091_0
snappy pkgs/main/linux-64::snappy-1.1.9-h295c915_0
stack_data pkgs/main/noarch::stack_data-0.2.0-pyhd3eb1b0_0
tenacity pkgs/main/linux-64::tenacity-8.0.1-py310h06a4308_1
tifffile pkgs/main/noarch::tifffile-2021.7.2-pyhd3eb1b0_2
toolz pkgs/main/noarch::toolz-0.11.2-pyhd3eb1b0_0
traitlets pkgs/main/noarch::traitlets-5.1.1-pyhd3eb1b0_0
vispy glueviz/linux-64::vispy-0.9.3-py310h0b38e52_1
wcwidth pkgs/main/noarch::wcwidth-0.2.5-pyhd3eb1b0_0
xlrd pkgs/main/noarch::xlrd-2.0.1-pyhd3eb1b0_0
zeromq pkgs/main/linux-64::zeromq-4.3.4-h2531618_0
zfp pkgs/main/linux-64::zfp-0.5.5-h295c915_6
The following packages will be REMOVED:
qt-main-5.15.2-h327a75a_7
qt-webengine-5.15.9-hd2b0992_4
qtwebkit-5.212-h4eab89a_4
The following packages will be DOWNGRADED:
libtiff 4.2.0-h2818925_1 --> 4.2.0-h85742a9_0
pyqt 5.15.7-py310h6a678d5_1 --> 5.9.2-py310h295c915_6
sip 6.6.2-py310h6a678d5_0 --> 4.19.13-py310h295c915_0
zstd 1.5.2-ha4553b6_0 --> 1.4.9-haebb681_0
Expected behavior I should be able to install the latest release
Screenshots If applicable, add screenshots to help explain your problem.
Details:
- Operating System: elementary OS (based on Ubuntu 18.04)
- Python version (
python --version): 3.10.4 - Glue version (
glue --version): [e.g. 0.14.5] - How you installed glue: [e.g. conda or pip]
Additional context Add any other context about the problem here.
The problem seems to be that the Python version I'm using is too new. But this only addresses why I'm not getting v1.2.0, there's still the issue that the glueviz channel in Anaconda needs updating.
@pllim you mentioned that the description should be pulled from https://github.com/conda-forge/glue-core-feedstock/blob/f6c13ce840345640ed451766ca20694e6598a450/recipe/meta.yaml#L10-L11 which indeed points to the correct source package for 1.5.0 – any idea why conda would still not find it?
Does your glueviz channel grab automatically from conda-forge ? Why do you have your own channel?
Don't know, maybe historical reasons when conda-forge did not get all the required packages?
But conda install -c conda-forge glue-core even with Python3.9 only gives me 1.0.1 as well, which is from
glue-core pkgs/main/noarch::glue-core-1.0.1-pyhd3eb1b0_1
maybe this is an osx-arm64-specific issue in my case.
You can try ping the conda-forge people on your recipe.
I made some more tests under macOS and found that the major issue here is indeed running natively on arm64.
Running under Rosetta as x86_64 I am able in an Anaconda3 env, as well as with new Miniforge and Mambaforge installations, to install glue-viz with glue-core 1.5.0, both for Python 3.9 and 3.10, but only using the default or conda-forge channels.
conda install -c glueviz --update-all glueviz apparently still pulls most of the packages from conda-forge but is first finding glue-core 1.2.2 in glueviz`, like
glib-tools conda-forge/osx-64::glib-tools-2.72.1-h2292cb8_0 glue-core glueviz/osx-64::glue-core-1.2.2-py39h6e9494a_0 glue-vispy-viewers glueviz/noarch::glue-vispy-viewers-1.0.5-py_1 glueviz glueviz/osx-64::glueviz-1.2.0-py39h6e9494a_0
As osx-arm64 is still considered somewhat experimental and unsupported, I guess there is no immediate remedy to that, although I can get a fully working mamba installation of almost all dependencies and pip-install the arm64 glueviz packages, of course performing much better than the x86_64 versions.
On other platforms it seems the best solution for the moment is indeed to install from conda-forge instead.
@Gabriel-p could you test if conda install -c conda-forge --update-all glueviz works for installing the current versions, and otherwise try to update your Anaconda or Miniconda installation?
The osx-arm64 distribution has been fixed in conda-forge/glue-core-feedstock#53; don't know if the issues on Linux can still be reproduced.