Unable to install azureml-fsspec

[melmatlis]

Running pip install azureml-fsspec fails with the following error:

Collecting azureml-fsspec
  Using cached azureml_fsspec-0.1.0b1-py3-none-any.whl (6.2 kB)
ERROR: Cannot install azureml-fsspec because these package versions have conflicting dependencies.

The conflict is caused by:
    azureml-dataprep 4.5.7 depends on azureml-dataprep-rslex~=2.11.0dev0
    azureml-dataprep 4.5.6 depends on azureml-dataprep-rslex~=2.11.0dev0
    azureml-dataprep 4.5.5 depends on azureml-dataprep-rslex~=2.11.0dev0
    azureml-dataprep 4.5.4 depends on azureml-dataprep-rslex~=2.11.0dev0
    azureml-dataprep 4.5.3 depends on azureml-dataprep-rslex~=2.11.0dev0
    azureml-dataprep 4.5.1 depends on azureml-dataprep-rslex~=2.11.0dev0
    azureml-dataprep 4.5.0 depends on azureml-dataprep-rslex~=2.11.0dev0

To fix this you could try to:
1. loosen the range of package versions you've specified
2. remove package versions to allow pip attempt to solve the dependency conflict

ERROR: ResolutionImpossible: for help visit https://pip.pypa.io/en/latest/topics/dependency-resolution/#dealing-with-dependency-conflicts

The installation is done in a clean environment. OS - iOS Python 3.9 When checking pip list, the libraby azureml-dataprep isn't installed

[QianqianNie]

Hi Can you try latest version azureml-fsspec? and the error you are hitting is related to a version of azrueml-dataprep you already have in your environment is conflicted, please try to resolve it to the version that azureml-fsspec expects as the error message shows.

[fdroessler]

I had the same issue. For me it seems like an issue with iOS. The same works under linux just fine but under iOS with conda it does not.

[melmatlis]

@QianqianNie still not working. azureml-dataprep is not installed in the environment so there are no conflicts.

'azureml-dataprep' in sys.modules returns: False

pip list returns:

Please advise, is there another way to identify the potential conflict?

[mitkel]

Same issue on macOS Ventura 13.1, tried with python3.8, 3.9 and 3.10

[QianqianNie]

Hi there is a known gap that azureml-dataprep does not work on M1/M2 CPU.

You can try to work around it by forcing it to use x86 packages, example like this: CONDA_SUBDIR=osx-64 conda create -n myenv_x86 python=3.9 conda activate myenv_x86 conda config --env --set subdir osx-64

[Ilia-Kosenkov]

Hey, I am hitting a similar issue. I am building a custom environment for AzureML notebook, and while running pip install azureml-fsspec, I get similar errors. This time, however, I suspect this is the root cause

Collecting azureml-dataprep<4.11.0a,>=4.10.0a

If I am reading this correctly, one of the dependencies requires azureml-datarep >= 4.11.0a, but the latest release of azureml-datarep is 4.10.6, so this requirement is always unmet.

[natwille1]

Hi there is a known gap that azureml-dataprep does not work on M1/M2 CPU.

You can try to work around it by forcing it to use x86 packages, example like this: CONDA_SUBDIR=osx-64 conda create -n myenv_x86 python=3.9 conda activate myenv_x86 conda config --env --set subdir osx-64

When can we expect M1/M2 CPU to be supported using AML packages? It's been out for 3 years now..

[korommatyi]

Same issue with 1.0.0. Installing azureml-fsspec in a fresh virtualenv produces the following error.

% pip install azureml-fsspec
Collecting azureml-fsspec
  Using cached azureml_fsspec-1.0.0-py3-none-any.whl (11 kB)
Collecting azureml-core (from azureml-fsspec)
  Using cached azureml_core-1.51.0-py3-none-any.whl (3.3 MB)
Collecting azureml-dataprep<4.11.0a,>=4.10.0a (from azureml-fsspec)
  Using cached azureml_dataprep-4.10.7-py3-none-any.whl (38.2 MB)
Collecting fsspec>=2021.6.1 (from azureml-fsspec)
  Using cached fsspec-2023.5.0-py3-none-any.whl (160 kB)
INFO: pip is looking at multiple versions of azureml-dataprep to determine which version is compatible with other requirements. This could take a while.
Collecting azureml-dataprep<4.11.0a,>=4.10.0a (from azureml-fsspec)
  Using cached azureml_dataprep-4.10.6-py3-none-any.whl (38.2 MB)
  Using cached azureml_dataprep-4.10.5-py3-none-any.whl (38.2 MB)
  Using cached azureml_dataprep-4.10.4-py3-none-any.whl (38.2 MB)
  Using cached azureml_dataprep-4.10.1-py3-none-any.whl (38.4 MB)
  Using cached azureml_dataprep-4.10.0-py3-none-any.whl (43.5 MB)
ERROR: Cannot install azureml-fsspec because these package versions have conflicting dependencies.

The conflict is caused by:
    azureml-dataprep 4.10.7 depends on dotnetcore2<4.0.0 and >=3.0.0
    azureml-dataprep 4.10.6 depends on dotnetcore2<4.0.0 and >=3.0.0
    azureml-dataprep 4.10.5 depends on dotnetcore2<4.0.0 and >=3.0.0
    azureml-dataprep 4.10.4 depends on dotnetcore2<4.0.0 and >=3.0.0
    azureml-dataprep 4.10.1 depends on dotnetcore2<4.0.0 and >=3.0.0
    azureml-dataprep 4.10.0 depends on dotnetcore2<4.0.0 and >=3.0.0

To fix this you could try to:
1. loosen the range of package versions you've specified
2. remove package versions to allow pip attempt to solve the dependency conflict

ERROR: ResolutionImpossible: for help visit https://pip.pypa.io/en/latest/topics/dependency-resolution/#dealing-with-dependency-conflicts

[Carl132]

Is there any update on this? It's crazy that we still can't utilise all of Microsoft's own packages.

[tawan0109]

@Carl132 did you hit the same issue when installing azureml-fsspec on apple silicon chips? we have been working to remove the dependency on dotnetcore and now we are very close, please stay tuned.

[Carl132]

@Carl132 did you hit the same issue when installing azureml-fsspec on apple silicon chips? we have been working to remove the dependency on dotnetcore and now we are very close, please stay tuned.

@tawan0109 yes that's exactly it - this is exciting to hear!

[feeser]

hey @tawan0109 is there any update regarding the dotnetcore dependency? I'm facing the same issue

[SamAbdelbadie]

Any updates?, I am also getting the same issue

[Gunnvant]

Getting the same error.

[jzm0144]

Getting the same error on Mac M1. A lot of people cannot switch to another machine just because of this issue. Can Microsoft team provide a temporary way around this until the proper solution becomes available?

[gasiorowskir]

Encountered the same issue today and it's a blocker. Any progress on the resolution? :-(

[utkarshtrehan]

@gasiorowskir @jzm0144 The quick work around is that download intel version of anaconda, works 100% 👍🏻

https://repo.anaconda.com/archive/Anaconda3-2024.02-1-MacOSX-x86_64.pkg

[Darcos923]

@utkarshtrehan Could you please provide the complete process (step by step) to use Anaconda for this case?