We should automate checking if upstream packages have changed and automatically create PRs if that is the case.

We need to move through all the packages and make sure they build on the mac and add cross compile build job.

As a user/developer, I want to install the mrd-storage-server as a conda package.

This will enable us to consume it in the main gadgetron project when we replace the built in storage server with the new one.

In order to switch the mrd-storage-server for Gadgetron (gadgetron/gadgetron#1059), we need a conda package for the storage server (#3) and we will also need a package for libcpr.

Following the merge of the mac builds, gadgetron-python is now being built for Python 3.10 it seems and when installing Gadgetron this is the version that gets pulled in. Consequently, Gadgetron no longer has python support. To reproduce, create an environment.yaml file with:

name: gadgetron-test2
channels:
  - nvidia/label/cuda-11.6.0
  - gadgetron
  - conda-forge
  - bioconda
  - defaults
  - intel
dependencies:
  - gadgetron>=4.1.4
  - siemens_to_ismrmrd>=1.0.0

Create the environment with:

conda env create -f environment.yaml

Then after activating the environment with:

conda activate gadgetron-test2

$ gadgetron --info
02-20 17:35:29.372 DEBUG [gadgetron_paths.cpp:107] Executable path: "/anaconda/envs/gadgetron-test2/bin/gadgetron"
02-20 17:35:29.372 DEBUG [gadgetron_paths.cpp:113] Default Gadgetron home: "/anaconda/envs/gadgetron-test2"
02-20 17:35:29.372 WARNING [initialization.cpp:38] Environment variable 'OMP_WAIT_POLICY' not set to 'PASSIVE'.
02-20 17:35:29.372 WARNING [initialization.cpp:39] Gadgetron may experience serious performance issues under heavy load (multiple simultaneous reconstructions, etc.)
Gadgetron Version Info
  -- Version            : 4.1.4
  -- Git SHA1           : 4619ff9277758ed18787df23edc5b9fe40d38f95
  -- System Memory size : 112705 MB
  -- Python Support     : NO
  -- Julia Support      : NO
  -- Matlab Support     : NO
CUDA DEVICE COUNT 1 and error number 0
  -- CUDA Support       : YES
  -- NVCC Flags         : -gencode arch=compute_60,code=sm_60;-gencode arch=compute_61,code=sm_61;-gencode arch=compute_70,code=sm_70 --std=c++17
    * Number of CUDA capable devices: 1
      - Device 0: Tesla P100-PCIE-16GB
         + CUDA Driver Version / Runtime Version: 11.6/11.6
         + CUDA Capability Major / Minor version number: 6.0
         + Total amount of global GPU memory: 16280 MB

If you try to do import the gadgetron module:

Python 3.9.10 | packaged by conda-forge | (main, Feb  1 2022, 21:24:11)
[GCC 9.4.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import gadgetron
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
ModuleNotFoundError: No module named 'gadgetron'
>>>

We are missing metadata in all the packages.

We use junitparser in pipelines, so we should consider having a conda package for that. It would be good to add with #7

We should make a conda package that installs the Gadgetron test scripts so that it is easy to test an installation and craft new test cases.

Is there any interest from the gadgetron devs in collaborating to make some or all of these packages available through conda-forge (http://conda-forge.github.io/)? The main benefit provided would be the use of continuous integration to verify package builds for each platform. Also, you will likely be able to get additional maintainers on some of the recipes that are of broader interest to the community (e.g. the link in the armadillo recipe here currently seems to be expired).

I can help start the conversions if you are interested. fftw is already on conda-forge, but I think the rest would be new. I have found the developers there to be very responsive and helpful with recipes I have submitted in the past.