Open jjoonathan opened 4 years ago
Here is conda info inside the container:
sudo doc(base) jon@jon-Precision-7720:~/hyperspectral$ sudo docker run -it continaconda3
(base) root@68c9563607c9:/# conda info
active environment : base
active env location : /opt/conda
shell level : 1
user config file : /root/.condarc
populated config files :
conda version : 4.8.2
conda-build version : 3.18.11
python version : 3.7.6.final.0
virtual packages : __glibc=2.28
base environment : /opt/conda (writable)
channel URLs : https://repo.anaconda.com/pkgs/main/linux-64
https://repo.anaconda.com/pkgs/main/noarch
https://repo.anaconda.com/pkgs/r/linux-64
https://repo.anaconda.com/pkgs/r/noarch
package cache : /opt/conda/pkgs
/root/.conda/pkgs
envs directories : /opt/conda/envs
/root/.conda/envs
platform : linux-64
user-agent : conda/4.8.2 requests/2.22.0 CPython/3.7.6 Linux/5.3.0-42-generic debian/10 glibc/2.28
UID:GID : 0:0
netrc file : None
offline mode : False
conda install -c pyviz geoviews-core
completes successfully. It's just the full geoviews that conda has issues with. I wish it didn't take so long (~an hour) to fail, so that I could iterate quicker, but that's how conda is.
Update: conda's dependency analysis finally finished after 18 hours! Unfortunately it does not look terribly helpful at first glance. Hopefully it's more helpful to one of you :) conda_conflict_report.txt
Ok, one more possibility eliminated. I don't think it's just the continuumio/anaconda3
docker image.
I manually started with sudo docker run -it ubuntu:latest
, installed anaconda apt update; apt install curl; curl -O https://repo.anaconda.com/archive/Anaconda3-2020.02-Linux-x86_64.sh; bash ./Anaconda3-2020.02-Linux-x86_64.sh
, and ran conda install -c pyviz geoviews
. It failed to Solve Environment.
I've started manually installing combinations of geoviews dependencies. Unfortunately, whether or not they Solve Environment seems to be order-dependent. In each of these cases, I start with a fresh container sudo docker run -it continuumio/anaconda3
:
I can successfully install all of the following (copied from history
, might contain a typo):
conda install -yc pyviz geoviews-core
conda install -yc pyviz bokeh
conda install -yc pyviz datashader
conda install -yc pyviz xarray
conda install -yc pyviz holoviews
conda install -yc pyviz jupyter
conda install -yc pyviz matplotlib
conda install -yc pyviz numpy pandas
conda install -yc pyviz param
conda install -yc pyviz netcdf4
conda install -yc pyviz python
conda install -yc pyviz scipy
conda install -yc pyviz cartopy
conda install -yc pyviz pyct
conda install -yc pyviz shapely
If I subsequently try either of the following, I get a failed Solving Environment:
conda install -c pyviz geopandas
conda install -c pyviz gdal
If I try to install the "working" set of packages all at once, I get a failed Solving Environment:
conda install -c pyviz bokeh cartopy datashader geoviews-core holoviews jupyter matplotlib netcdf4 numpy pandas param pyct python scipy shapely xarray
I am now working on merging install steps in the working set into maximal groups.
Starting, as usual, from sudo docker run -it continuumio/anaconda3
,
This fails at Solving Environment:
conda install -yc pyviz bokeh cartopy
This succeeds:
conda install -yc pyviz geoviews-core bokeh datashader xarray holoviews jupyter matplotlib numpy pandas param netcdf4 python scipy
conda install -yc pyviz cartopy
Huh?
It seems I must have prematurely assumed it would fail to solve once it reported that the initial solve failed. That was incorrect. Starting from sudo docker run -it continuumio/anaconda3
,
This succeeds: conda install -yc pyviz geoviews-core bokeh datashader xarray holoviews jupyter matplotlib numpy pandas param netcdf4 python scipy pyct shapely cartopy
(it is the full list of geoviews dependencies less gdal and geopandas)
It just has to go through a long process:
Collecting package metadata (current_repodata.json): done
Solving environment: failed with initial frozen solve. Retrying with flexible solve.
Solving environment: failed with repodata from current_repodata.json, will retry with next repodata source.
Collecting package metadata (repodata.json): done
Solving environment: done
gdal
fails on its own. Starting from sudo docker run -it continuumio/anaconda3
we get:
(base) root@f3930e25dcba:/# conda install gdal
Collecting package metadata (current_repodata.json): done
Solving environment: failed with initial frozen solve. Retrying with flexible solve.
Solving environment: failed with repodata from current_repodata.json, will retry with next repodata source.
Collecting package metadata (repodata.json): done
Solving environment: failed with initial frozen solve. Retrying with flexible solve.
Solving environment: /
Found conflicts! Looking for incompatible packages.
This can take several minutes. Press CTRL-C to abort.
(... last time this took 18 hours, I'm not waiting that long to post ...)
geopandas
also fails on its own. Starting from sudo docker run -it continuumio/anaconda3
we get:
Collecting package metadata (current_repodata.json): done
Solving environment: failed with initial frozen solve. Retrying with flexible solve.
Solving environment: failed with repodata from current_repodata.json, will retry with next repodata source.
Collecting package metadata (repodata.json): done
Solving environment: failed with initial frozen solve. Retrying with flexible solve.
Solving environment: /
Found conflicts! Looking for incompatible packages.
This can take several minutes. Press CTRL-C to abort.
(... last time this took 18 hours, I'm not waiting that long to post ...)
Starting from sudo docker run -it continuumio/anaconda3
This works, albeit every step is very slow:
conda update anaconda
conda install gdal
conda install geopandas
I'm now going to see if conda update anaconda
fixes conda install -c pyviz geoviews
.
Indeed, it does. Starting from sudo docker run -it continuumio/anaconda3
, this works:
conda update anaconda
conda install -c pyviz geoviews
It seems reasonably clear at this point that this is not a geoviews issue so much as a conda ecosystem issue. I'll close the thread.
Actually, scratch that, should conda update anaconda
be added to the README? As it stands, following the README installation instructions does not work on linux.
Sorry for all the pain you clearly had to go through here. I'd be happy to accept a PR that suggests that as a fix and lists a minimum conda version.
Unfortunately it looks like the anaconda version that doesn't work has a name ("2020.02") while the anaconda "version" that does work does not have a name. The only concise workaround might be running an update command, which in the future may well have as much chance of breaking things as fixing them. See below.
Ok, so immediately after sudo docker run -it continuumio/anaconda3
I get:
(base) root@c1d5e120a425:/# conda list anaconda
anaconda 2020.02 py37_0
anaconda-client 1.7.2 py37_0
anaconda-navigator 1.9.12 py37_0
anaconda-project 0.8.4 py_0
Attempting to install geoviews results in Solving Environment forever.
When I run conda update anaconda
, it gives this warning:
The following packages will be DOWNGRADED:
anaconda 2020.02-py37_0 --> custom-py37_1
although most (all?) of the sub packages that I can see are actually being upgraded. Full logs attached
. I guess anaconda update anaconda
now means approximately what anaconda update --all
used to mean, which leads to an undetermined anaconda version, which leads to poor ordering, which leads to update causing a "downgrade"? In any case, after running that command, conda install -c pyviz geoviews
works and versions are listed as follows:
(base) root@c1d5e120a425:/# conda list anaconda
_anaconda_depends 2019.03 py37_0
anaconda custom py37_1
anaconda-client 1.7.2 py37_0
anaconda-navigator 1.9.12 py37_0
anaconda-project 0.8.4 py_0
I am now testing conda update --all
, the command recommended here for possible inclusion in the README.
Ok, conda update --all
works (that is, after running it, conda install -c pyviz geoviews
completes with issue, albeit is has to do a "flexible solve" that takes a long time). PR incoming.
I just set up a new mac, and after a fresh conda install (Anaconda3-2020.02-MacOSX-x86_64) I had the same exact problem:
(base) jonathandewerd@USXZM1MB6HUMD6R ~ % conda install geoviews
Collecting package metadata (current_repodata.json): done
Solving environment: failed with initial frozen solve. Retrying with flexible solve.
Solving environment: failed with repodata from current_repodata.json, will retry with next repodata source.
Collecting package metadata (repodata.json): done
Solving environment: failed with initial frozen solve. Retrying with flexible solve.
Solving environment: |
Found conflicts! Looking for incompatible packages.
which was solved using the exact same solution (running conda update --all
and then retrying conda install geoviews
).
It still needs conda install -c pyviz geoviews
until we have packages updated on the default channels...
Since I hadn't built heavily on my conda environment I took the opportunity to nuke it (run anaconda-clean and delete /opt/anaconda3) and try again with the correct channel. On top of a fresh Anaconda3-2020.02-MacOSX-x86_64 on Mac OS 10.15.2, it still chokes:
(base) jonathandewerd@USXZM1MB6HUMD6R ~ % conda install -c pyviz geoviews
Collecting package metadata (current_repodata.json): done
Solving environment: failed with initial frozen solve. Retrying with flexible solve.
Solving environment: failed with repodata from current_repodata.json, will retry with next repodata source.
Collecting package metadata (repodata.json): done
Solving environment: failed with initial frozen solve. Retrying with flexible solve.
Solving environment: \
Found conflicts! Looking for incompatible packages.
Since the error looks the same between Mac OS X and linux I went back and dusted off my docker experiments. Here is what I've run so far. "echo AAA-WIN" corresponds to an experiment arbitrarily named AAA that resulted in successful installation while "echo AAF-FAIL" corresponds to an experiment arbitrarily named AAF that resulted in the failure mode seen above.
echo AAA-WIN ; sudo docker run continuumio/anaconda3 conda install -yc pyviz bokeh datashader holoviews jupyter matplotlib numpy pandas param pyct python scipy xarray
echo AAB-WIN ; sudo docker run continuumio/anaconda3 conda install -yc pyviz bokeh cartopy datashader holoviews jupyter matplotlib numpy pandas param pyct python scipy xarray
echo AAC-WIN ; sudo docker run continuumio/anaconda3 conda install -yc pyviz bokeh datashader geoviews-core holoviews jupyter matplotlib numpy pandas param pyct python scipy xarray
echo AAD-WIN ; sudo docker run continuumio/anaconda3 conda install -yc pyviz bokeh datashader holoviews jupyter matplotlib netcdf4 numpy pandas param pyct python scipy xarray
echo AAE-WIN ; sudo docker run continuumio/anaconda3 conda install -yc pyviz bokeh datashader holoviews jupyter matplotlib numpy pandas param pyct python scipy shapely xarray
echo AAF-FAIL; sudo docker run continuumio/anaconda3 conda install -yc pyviz bokeh cartopy datashader gdal geopandas geoviews-core holoviews jupyter matplotlib netcdf4 numpy pandas param pyct python scipy shapely xarray
echo AAG-FAIL; sudo docker run continuumio/anaconda3 conda install -yc pyviz bokeh cartopy datashader geopandas geoviews-core holoviews jupyter matplotlib netcdf4 numpy pandas param pyct python scipy shapely xarray
echo AAH-FAIL; sudo docker run continuumio/anaconda3 conda install -yc pyviz bokeh cartopy datashader gdal geoviews-core holoviews jupyter matplotlib netcdf4 numpy pandas param pyct python scipy shapely xarray
echo AAI-WIN ; sudo docker run continuumio/anaconda3 conda install -yc pyviz bokeh cartopy datashader geoviews-core holoviews jupyter matplotlib netcdf4 numpy pandas param pyct python scipy shapely xarray
Wow, it looks like this isn't even a scrap between gdal/geopandas and the rest, those packages may be directly responsible. EDIT: this is on Mac OS X, dirty environment.
(base) jonathandewerd@USXZM1MB6HUMD6R ~ % conda install gdal geopandas
Collecting package metadata (current_repodata.json): done
Solving environment: failed with initial frozen solve. Retrying with flexible solve.
Solving environment: failed with repodata from current_repodata.json, will retry with next repodata source.
Collecting package metadata (repodata.json): done
Solving environment: failed with initial frozen solve. Retrying with flexible solve.
Solving environment: \
Found conflicts!
Both conda install gdal
and conda install geopandas
fail with "Found Conflicts!" on Mac OS X. Now trying the conda-forge channel.
Linux experiments (updated with final FAIL/SUCCESS results, the "AA*" is just an experiment tag to keep them straight while running a bunch of them):
echo AAJ-FAIL; sudo docker run continuumio/anaconda3 conda install -y gdal geopandas
echo AAK-FAIL; sudo docker run continuumio/anaconda3 conda install -y geopandas
echo AAL-FAIL; sudo docker run continuumio/anaconda3 conda install -y gdal
echo AAM-FAIL; sudo docker run continuumio/anaconda3 conda install -yc conda-forge geopandas
echo AAN-FAIL; sudo docker run continuumio/anaconda3 conda install -yc conda-forge gdal
echo AAO-SUCCESS; sudo docker run continuumio/anaconda3 bash -c 'conda update -y --all; conda install -y -c pyviz geoviews'
New information on Mac OS X:
# Pristine Mac OS X 10.15.2
# Install Anaconda3-2020.02-MacOSX-x86_64.pkg
conda install -y -c pyviz geoviews # Failure
conda update anaconda
conda install -y -c pyviz geoviews # Slow Success (Solving... failed... Solving... failed... Solving... done)
# Pristine Mac OS X 10.15.2
# Install Anaconda3-2020.02-MacOSX-x86_64.pkg
conda install -y -c pyviz geoviews # Failure
conda install -y -c pyviz -c conda-forge geoviews # Success
I can confirm that I can't conda install from the pyviz channel in a completely clean miniconda environment. Here is what conda reports and it looks like a GDAL issue to me.
Note, installing from conda-forge
in a fresh env does work fine for me right now.
FYI got a slightly different error message yesterday regarding install via conda.
ray@ray-MS-7B43:~$ conda install -c pyviz geoviews
Collecting package metadata (current_repodata.json): done
Solving environment: failed with initial frozen solve. Retrying with flexible solve.
Solving environment: failed with repodata from current_repodata.json, will retry with next repodata source.
Collecting package metadata (repodata.json): done
Solving environment: failed with initial frozen solve. Retrying with flexible solve.
Solving environment: |
Found conflicts! Looking for incompatible packages.
This can take several minutes. Press CTRL-C to abort.
failed
UnsatisfiableError: The following specifications were found to be incompatible with each other:
Output in format: Requested package -> Available versionsThe following specifications were found to be incompatible with your CUDA driver:
feature:/linux-64::__cuda==10.2=0
Your installed CUDA driver is: 10.2
Yeah, a new conda version came out last month. I also didn't have success with conda install -c pyviz geoviews
but I did have success with:
conda update --all
conda install -c pyviz -c conda-forge geoviews
@philippjfr Do you know if there are any updates on this? It would be nice to be able to install geoviews from defaults...
Impossible to install geoviews with last Anaconda distrib (Anaconda3-2020.11-Linux-x86_64.sh).
conda update --all
conda install -c pyviz geoviews
runs forever
Did you try including -c conda-forge? In my previous post, -c conda-forge was a necessary component, and I have previously observed lack of convergence using just -c pyviz.
Yes, I tried. The installation works but the release that is installed is the 1.5.0, not the last 1.8.2
Thanks. Does conda install --strict-channel-priority -c pyviz geoviews
work?
Nope... Thus, I am very motivated to explore building of dashboards with geoviews...
$ conda install --strict-channel-priority -c pyviz geoviews
Collecting package metadata (current_repodata.json): done
Solving environment: failed with initial frozen solve. Retrying with flexible solve.
Solving environment: failed with repodata from current_repodata.json, will retry with next repodata source.
Collecting package metadata (repodata.json): done
Solving environment: failed with initial frozen solve. Retrying with flexible solve.
Solving environment: \
Found conflicts! Looking for incompatible packages.
This can take several minutes. Press CTRL-C to abort.
failed
UnsatisfiableError: The following specifications were found to be incompatible with each other:
Output in format: Requested package -> Available versions
Note that strict channel priority may have removed packages required for satisfiability.
Damn. Presumably you can get a working version by starting with Miniconda instead of the Anaconda Distribution, so that it doesn't install so many packages that can conflict, but there's obviously some issues we need to figure out regardless. :-(
Yet a tag from the gitHub source indicates "conda|conda-forge v1.8.2" and should notify that the package is compatible, isn't it ? Let me know by feeding this issue.
Started with a clean environment. Fails with the error:
Collecting package metadata (current_repodata.json): done Solving environment: failed with initial frozen solve. Retrying with flexible solve. Solving environment: failed with repodata from current_repodata.json, will retry with next repodata source. Collecting package metadata (repodata.json): done Solving environment: failed with initial frozen solve. Retrying with flexible solve. Solving environment: - Found conflicts! Looking for incompatible packages. This can take several minutes. Press CTRL-C to abort. failed
UnsatisfiableError: The following specifications were found to be incompatible with each other:
Output in format: Requested package -> Available versions
A pip list of the environment shows:
Package Version
certifi 2020.11.8 pip 20.2.4 setuptools 50.3.1.post20201107 wheel 0.35.1
It works with Anaconda3-2020.02-Linux-x86_64.sh
conda update --prefix /opt/anaconda3 --all
conda install -c pyviz geoviews
Does not work with later distribs (2020.07, 2020.11).
Yes, I'm using 4.9.2 (2020.11). Any word on an update to Geoviews for compatibility with the latest Anaconda versions?
I faced the same issue.
Using conda install -c pyviz geoviews
in clean environment conda can solve environment for Python 3.7, but can't for Python 3.8 and 3.9.
Using conda install -c conda-forge geoviews
in clean environment conda can solve environment for all above.
Edit: I'm using Windows 10.
Thanks for reporting that!
The fact that -c pyviz
doesn't work for 3.8+ suggests to me that certain binary packages aren't available for 3.8+ on defaults (but are available on conda-forge).
Same error here installing holoviz from -c pyviz for arm64 based arch.
=> [3/5] RUN conda config --set unsatisfiable_hints True 3.4s
=> ERROR [4/5] RUN conda install -c pyviz holoviz -v && conda clean --all -f -y && fix-permissions /opt/co 2063.9s
> [4/5] RUN conda install -c pyviz holoviz -v && conda clean --all -f -y && fix-permissions /opt/conda && fix-permissions /home/jovyan:
#7 25.51 Collecting package metadata (current_repodata.json): ...working... Unable to retrieve repodata (response: 404) for https://conda.anaconda.org/pyviz/linux-aarch64/current_repodata.json
#7 27.98
#7 86.21 done
#7 86.21 Solving environment: ...working... failed with initial frozen solve. Retrying with flexible solve.
#7 330.9 Solving environment: ...working... failed with repodata from current_repodata.json, will retry with next repodata source.
#7 549.6 Collecting package metadata (repodata.json): ...working... done
#7 693.8 Solving environment: ...working... failed with initial frozen solve. Retrying with flexible solve.
#7 2061.1 Found conflicts! Looking for incompatible packages.
#7 2061.1 This can take several minutes. Press CTRL-C to abort.
#7 2061.1 failed
Traceback (most recent call last):
#7 2061.1 File "/opt/conda/lib/python3.8/site-packages/conda/cli/install.py", line 261, in install
#7 2061.1 unlink_link_transaction = solver.solve_for_transaction(
#7 2061.1 File "/opt/conda/lib/python3.8/site-packages/conda/core/solve.py", line 114, in solve_for_transaction
#7 2061.1 unlink_precs, link_precs = self.solve_for_diff(update_modifier, deps_modifier,
#7 2061.1 File "/opt/conda/lib/python3.8/site-packages/conda/core/solve.py", line 157, in solve_for_diff
#7 2061.1 final_precs = self.solve_final_state(update_modifier, deps_modifier, prune, ignore_pinned,
#7 2061.1 File "/opt/conda/lib/python3.8/site-packages/conda/core/solve.py", line 281, in solve_final_state
#7 2061.1 ssc = self._run_sat(ssc)
#7 2061.1 File "/opt/conda/lib/python3.8/site-packages/conda/common/io.py", line 88, in decorated
#7 2061.1 return f(*args, **kwds)
#7 2061.1 File "/opt/conda/lib/python3.8/site-packages/conda/core/solve.py", line 815, in _run_sat
#7 2061.1 ssc.solution_precs = ssc.r.solve(tuple(final_environment_specs),
#7 2061.1 File "/opt/conda/lib/python3.8/site-packages/conda/common/io.py", line 88, in decorated
#7 2061.1 return f(*args, **kwds)
#7 2061.1 File "/opt/conda/lib/python3.8/site-packages/conda/resolve.py", line 1320, in solve
#7 2061.1 raise UnsatisfiableError({})
#7 2061.1 conda.exceptions.UnsatisfiableError:
#7 2061.1 Did not find conflicting dependencies. If you would like to know which
#7 2061.1 packages conflict ensure that you have enabled unsatisfiable hints.
#7 2061.1
#7 2061.1 conda config --set unsatisfiable_hints True
#7 2061.1
#7 2061.1
#7 2061.1 During handling of the above exception, another exception occurred:
#7 2061.1
#7 2061.1 Traceback (most recent call last):
#7 2061.1 File "/opt/conda/lib/python3.8/site-packages/conda/exceptions.py", line 1079, in __call__
#7 2061.1 return func(*args, **kwargs)
#7 2061.1 File "/opt/conda/lib/python3.8/site-packages/conda/cli/main.py", line 84, in _main
#7 2061.1 exit_code = do_call(args, p)
#7 2061.1 File "/opt/conda/lib/python3.8/site-packages/conda/cli/conda_argparse.py", line 83, in do_call
#7 2061.1 return getattr(module, func_name)(args, parser)
#7 2061.1 File "/opt/conda/lib/python3.8/site-packages/conda/cli/main_install.py", line 20, in execute
#7 2061.1 install(args, parser, 'install')
#7 2061.1 File "/opt/conda/lib/python3.8/site-packages/conda/cli/install.py", line 308, in install
#7 2061.1 raise e
#7 2061.1 File "/opt/conda/lib/python3.8/site-packages/conda/cli/install.py", line 295, in install
#7 2061.1 unlink_link_transaction = solver.solve_for_transaction(
#7 2061.1 File "/opt/conda/lib/python3.8/site-packages/conda/core/solve.py", line 114, in solve_for_transaction
#7 2061.1 unlink_precs, link_precs = self.solve_for_diff(update_modifier, deps_modifier,
#7 2061.1 File "/opt/conda/lib/python3.8/site-packages/conda/core/solve.py", line 157, in solve_for_diff
#7 2061.1 final_precs = self.solve_final_state(update_modifier, deps_modifier, prune, ignore_pinned,
#7 2061.1 File "/opt/conda/lib/python3.8/site-packages/conda/core/solve.py", line 275, in solve_final_state
#7 2061.1 ssc = self._add_specs(ssc)
#7 2061.1 File "/opt/conda/lib/python3.8/site-packages/conda/core/solve.py", line 704, in _add_specs
#7 2061.1 ssc.r.find_conflicts(spec_set)
#7 2061.1 File "/opt/conda/lib/python3.8/site-packages/conda/resolve.py", line 352, in find_conflicts
#7 2061.1 raise UnsatisfiableError(bad_deps, strict=strict_channel_priority)
#7 2061.1 conda.exceptions.UnsatisfiableError: The following specifications were found
#7 2061.1 to be incompatible with the existing python installation in your environment:
#7 2061.1
#7 2061.1 Specifications:
#7 2061.1
#7 2061.1 - holoviz -> python[version='3.6.12|3.6.12|>=2.7|>=3.6,<3.7.0a0|>=3.8,<3.9.0a0|>=3.7,<3.8.0a0|>=3.9,<3.10.0a0|>=3|3.7.9|3.6.9|3.6.9|3.6.9|3.6.9|>=2.7,<2.8.0a0|>=3.5',build='2_73_pypy|4_73_pypy|5_73_pypy|3_73_pypy|1_73_pypy|0_73_pypy']
#7 2061.1
#7 2061.1 Your python: python=3.8
#7 2061.1
#7 2061.1 If python is on the left-most side of the chain, that's the version you've asked for.
#7 2061.1 When python appears to the right, that indicates that the thing on the left is somehow
#7 2061.1 not available for the python version you are constrained to. Note that conda will not
#7 2061.1 change your python version to a different minor version unless you explicitly specify
#7 2061.1 that.
#7 2061.1
#7 2061.1
#7 2061.1
------
error: failed to solve: rpc error: code = Unknown desc = executor failed running [/bin/bash -o pipefail -c conda install -c pyviz holoviz -v && conda clean --all -f -y && fix-permissions $CONDA_DIR && fix-permissions /home/$NB_USER]: exit code: 1
If you are wondering which conda i am using it's here:
--build-arg conda_version="4.9.2"
--build-arg miniforge_patch_number="5" \
--build-arg miniforge_python="Miniforge3" \
--build-arg miniforge_arch="aarch64" \
Hello, I am also having issues trying to install geoviews with Conda. I am currently using Python 3.9, and I have tried using conda install -c pyviz geoviews
and conda install -c conda-forge geoviews
. I have also tried re-installing and updating conda, and using a py39 environment. However, I always get the same problem.
I would appreciate any help!
Python 3.9 is still very new and I imagine half the geo-stack is not yet available. I'd stick with Python 3.7 or 3.8 for the time being.
I have tried with Conda and Python 3.6, 3.7, 3.8, and 3.9. It always fails.... I only managed to install it using Poetry and fiddling a lot with the versions of the packages. And as soon as new dependencies are added, the project becomes unmanageable....
Honestly, as far as I can tell the geo-dependencies on conda defaults are pretty broken right now. So best thing I can suggest is to install from conda-forge
with conda install -c conda-forge geoviews
or you could try only installing the core dependencies with conda install geoviews-core
.
I have seen issues with geo-dependencies, most typically when mixing defaults
and conda-forge
channels in an environment.
It might be worth trying to create a fresh environment based primarily on the conda-forge
channel, with something like:
conda create -n geo -c conda-forge geoviews
Also, maintaining consistency with channel priority for new packages added to an environment helps reduce the painful issues you might otherwise encounter, so unless conda-forge
is specifically added as a higher priority channel than defaults
in the conda config, it is best to specify using the conda-forge
channel for any new dependencies added.
I couldn't get it to work any of the ways listed above on Ubuntu with conda. It always hangs at solving environment
. I tried conda-forge
and pyviz
channels, as well as via pip
installation. I've tried python=3.5/6/7/8/9
, geoviews-core
, etc. . Interestingly, I couldn't install the dependency cartopy
any way I tried either, nor proj
which it depends on.
However, I did get geoviews
working on a Debian-based docker container.
sudo docker pull continuumio/miniconda3
sudo docker run -i -t continuumio/miniconda3 /bin/bash
conda create -n gv_env -c conda-forge geoviews
Okay, I managed to get it working on my Ubuntu OS (no docker) by uninstalling miniconda3 (version Miniconda3-py39_4.9.2-Linux-x86_64.sh
) and installing the same (outdated) version as in the continuumio/miniconda3
dockerfile.
rm -r ~/miniconda3/
wget https://repo.anaconda.com/miniconda/Miniconda3-4.5.11-Linux-x86_64.sh
bash Miniconda3-4.5.11-Linux-x86_64.sh
Go through the install process then
conda create -n gv_env -c conda-forge geoviews
This works with Python 3.9, so the Python version isn't to blame for this problem, I don't think.
conda install -c pyviz geoviews
currently fails to Solve Environment inside a clean anaconda container launched withsudo docker run -it continuumio/anaconda3
.I've tried the pyviz channel, the default channel, and the conda-forge channel with the same results.
Conda suggests that the problem has to do with conflicts and that it is looking for incompatibilities, but it has been running since yesterday and only reports 90% progress. I'll update this post if it completes.
Found conflicts! Looking for incompatible packages. This can take several minutes. Press CTRL-C to abort. Examining conflict for ...