meta.yaml

#
# This is how to do the build (and some other useful commands)
#
# To build requires access to the conda-forge, pytorchm and anaconda channels
# can be used with --debug flag
# conda build . -c pytorch -c conda-forge -c anaconda -c uncbiag
#
# After the package has build, copy it in your local directory (subdirectory osx64, for example).
# Then index it with this command (needs to be run for each subdirectory / architecture: noarch, linux-64, osx-64):
# conda index /Users/mn/PycharmProjects/tst_channel
#
# Once it is indexed it can be installed from there (ideally into a new, empty virtual environment as follows)
# conda list env
# conda activate [your_env] or conda create --name my_new_env pip python=3.7
# conda install -vv -c file://Users/mn/PycharmProjects/tst_channel/ -c pytorch -c conda-forge -c anaconda -c uncbiag mermaid=0.2.0 
# conda list
#
# Here are also some descriptions on how to host a private conda repository (as used above)
# https://stackoverflow.com/questions/35359147/how-can-i-host-my-own-private-conda-repository
#
# to upload to the cloud
# anaconda login
# anaconda upload --user uncbiag mermaid-0.2.0-0.tar.bz2
#
# Combined with travis, to auto-create conda packages and upload them to the conda cloud, tag them.

{% set name = "mermaid" %}

package:
  name: '{{ name|lower }}'
  version: {{ GIT_DESCRIBE_TAG }}

source:
  - git_url: .

requirements:
  build:
    - python ==3.7
    - setuptools
    - pip

  run:
    - python ==3.7
    - ipython
    - jupyter
    - ipykernel
    - numpy >=1.17
    - future
    - termcolor
    - cython
    - scikit-image
    - tabulate
    - cffi
    - itk
    - pytorch
    - pynrrd
    - torchvision
    - pandas
    - matplotlib
    - scipy
    - openpyxl
    - sphinx-gallery ==0.3.1
    - sphinx
    - nbsphinx
    - pandoc
    - pypandoc
    - graphviz
    - statsmodels
    - progressbar2

test:
  imports:
    - mermaid

about:
  home: https://github.com/uncbiag/mermaid
  license: BSD License
  license_family: BSD
  license_file: LICENSE
  summary: 'Image registration toolbox in pyTorch'
  description: 'This image registration toolbox provides various image registration algorithms (mostly non-parametric ones). It is written in pytorch allowing for easy prototyping of new approaches. To install include the conda-forge and the anaconda channels by executing: conda install -c pytorch -c conda-forge -c anaconda -c uncbiag mermaid'
  doc_url: 'https://mermaid.readthedocs.io/en/latest/'
  dev_url: 'https://github.com/uncbiag/mermaid'

extra:
  recipe-maintainers:
    - marcniethammer
    - rkwitt
    - hbgtjxzbbx