.travis.yml

# This following script attempts build dependible conda environments on linux
# numpy, pandas and matplotlib are required dependencies; care is needed
# in improving build time and reproducibilty.

sudo: false                                                # works w/Docker
language: python
python:
  - "2.7"
  - "2.7.12"
  - "3.3"
  - "3.4"
  - "3.5"
  - "3.5.2"

notifications:
  email:
    on_success: never
    on_failure: always

# Necessary packages for building numpy; turned off for now since conda adds it
#cache: apt
#addons:
#  apt:
#    packages:
#    - libatlas-dev
#    - libatlas-base-dev
#    - liblapack-dev
#    - gfortran

# Setup Miniconda
before_install:
  # Do this conditionally since it saves download time if the version is the same.
  - if [[ "$TRAVIS_PYTHON_VERSION" == "2.7" ]]; then
      wget https://repo.continuum.io/miniconda/Miniconda-latest-Linux-x86_64.sh -O miniconda.sh;
    else
      wget https://repo.continuum.io/miniconda/Miniconda3-latest-Linux-x86_64.sh -O miniconda.sh;
    fi
    ##- chmod +x miniconda.sh                              # release permissions
  - bash miniconda.sh -b -p $HOME/miniconda                # assign conda program to HOME dir
  - export PATH="$HOME/miniconda/bin:$PATH"                # enable `conda` command
  - hash -r                                                # reset links
  - conda config --set always_yes yes --set changeps1 no   # say yes to prompts
  - conda update -q conda                                  # update since mini lags behind conda; say "yes"
    # Useful for debugging any issues with conda
  - conda info -a
  # The next couple lines fix a crash with multiprocessing on Travis and are not specific to  Miniconda
  ##- rm -rf /dev/shm
  ##- ln -s /run/shm /dev/shm

install:
  # Pre-build major dependencies with conda (much faster and easier)
  # Ensure `pip` and `nose` will install within the conda env
  - conda create -q -n test-environment python=$TRAVIS_PYTHON_VERSION pip nose numpy matplotlib pandas
  - source activate test-environment
  - pip install coverage codecov nose-cov                  # code coverage; see other projects on how to install w/Miniconda
  ##- python setup.py install
  - conda list
  # Uses conda's `pip` instead of Travis' `pip`
  - pip install -r dev_requirements.txt                    # pre-install pinned dependencies; includes reg. test deps
  - pip install .                                          # install source and build/install wheels
  #- "pip install wheel"
  #- "python setup.py bdist_wheel"
  #- "pip install dist/*.whl"
  - pip list

# Commands to run tests, coverage and nbval regression tests
script:
  - nosetests --with-cov --cov lamana
  - ls -al
  - coverage combine
  - coverage xml
  - which coverage
#  - cd docs
#  - py.test --nbval _demo_pinned.ipynb --sanitize-with nbval_sanitize.cfg

# Interact with codecov
after_success:
  - codecov

# Whitelist
# branches:
#   only:
#    - master
#     - develop


# References
# ----- ---------                           -------------
# (001) Conda and coverage on Travis        http://dan-blanchard.roughdraft.io/7045057-quicker-travis-builds-that-rely-on-numpy-and-scipy-using-miniconda
# (002) Travis and conda                    http://conda.pydata.org/docs/travis.html
# (003) Conda on Travis w/details           http://lmjohns3.com/2015/06/using-travis-ci-with-miniconda-scipy-and-nose.html
# (004) Example yaml                        https://github.com/lmjohns3/theanets/blob/master/.travis.yml
# (005) Example 2 yaml                      http://samueljackson.me/2015/12/12/better-python-travis-config.html
# (006) Sample codecov w/miniconda          https://github.com/JIC-CSB/jicbioimage.illustrate/blob/master/.travis.yml#L31