Installation issue using pip

[skoudoro]

Hi Gandlf Team,

I tried to install gandlf via pip install gandlf and it fails with the following error:

❯ pip install gandlf
Collecting gandlf
  Downloading GANDLF-0.0.1720231112-py3-none-any.whl.metadata (8.7 kB)
INFO: pip is looking at multiple versions of gandlf to determine which version is compatible with other requirements. This could take a while.
  Downloading GANDLF-0.0.17-py3-none-any.whl.metadata (8.7 kB)
  Downloading GANDLF-0.0.16-py3-none-any.whl.metadata (8.6 kB)
ERROR: Cannot install gandlf==0.0.16, gandlf==0.0.17 and gandlf==0.0.1720231112 because these package versions have conflicting dependencies.

The conflict is caused by:
    gandlf 0.0.1720231112 depends on torch==1.13.1
    gandlf 0.0.17 depends on torch==1.13.1
    gandlf 0.0.16 depends on torch==1.13.1

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

So I decided restart from a fresh virtual env and clone the repo. I did pip install . from the source and it still fails.

❯ pip install .
Processing /Users/skoudoro/devel/GaNDLF
  Installing build dependencies ... done
  Getting requirements to build wheel ... done
  Preparing metadata (pyproject.toml) ... done
INFO: pip is looking at multiple versions of gandlf to determine which version is compatible with other requirements. This could take a while.
ERROR: Could not find a version that satisfies the requirement torch==1.13.1 (from gandlf) (from versions: 2.0.0, 2.0.1, 2.1.0)
ERROR: No matching distribution found for torch==1.13.1

I do not plan to use docker and I did not try the conda package. Is there a way to install GaNDLF using pip?

Also, Does it support a recent version of Numpy and pytorch ?

EDIT: use python 3.11

[sarthakpati]

Thanks for your interest in GaNDLF, @skoudoro!

We are currently working on updating the dependencies [ref]. In the meantime, I will suggest you follow the installation instructions as given in the guide [ref]. The reason why we have such a tight requirement for various libraries (including python) is because we do an extensive suite of tests [ref], which will be neigh impossible to run for all permutations and combinations of libraries that we depend on 😄.

Please respond to this thread or open a new one if you run into any issues!

EDIT:

use python 3.11

This is specifically not supported right now [ref].

[skoudoro]

ok, I created a new python 3.10 environment, and the installation seems to be fine now. thanks