Read dependencies from one file for setup.py and travis CI (#81)

* Read dependencies from one file for setup.py and use it in .travis.yml for setting up conda env

* delete obsolete visualization stuff and remove folium

* remove keras which should now be a submodule of tensorflow

* test multiple python versions in travis CI

.travis.yml

@@ -3,19 +3,8 @@ sudo: false
 matrix:
     include:
     - python: "3.7"
-      env: DEPS="numpy
-                 scipy
-                 pandas>=0.22
-                 matplotlib
-                 numba
-                 folium
-                 h5py==2.10.0
-                 xarray
-                 shapely
-                 pyproj
-                 tqdm
-                 pykrige
-                 tensorflow==1.14"
+    - python: "3.8"
+    - python: "3.9"
 
 # setup adapted from https://github.com/bsmurphy/PyKrige/blob/master/.travis.yml
 before_install:
@@ -28,15 +17,15 @@ before_install:
     - conda info -a
 
 install:
-    - conda create -q -n pycomlink-env python=${TRAVIS_PYTHON_VERSION} --yes
+    - conda create -q -n pycomlink-env python=${TRAVIS_PYTHON_VERSION} --file requirements.txt --channel conda-forge --yes
     - source activate pycomlink-env
-    - conda install --yes $DEPS pip
-    - conda install -c conda-forge --yes codecov
+    - conda install -c conda-forge --yes pip codecov
     #- pip install coveralls
     - pip install pytest pytest-cov
     - python -V
     #- python setup.py build_ext --inplace
-    - python setup.py install
+    #- python setup.py install
+    - python setup.py develop --no-deps
     - pwd
     - du -h
 

environment.yml

@@ -7,18 +7,15 @@ dependencies:
  - pandas>=0.22
  - matplotlib
  - numba
- - folium
- - h5py==2.10.0
+ - h5py
  - xarray
  - netcdf4
  - shapely
  - pyproj
  - tqdm
  - pykrige
- - tensorflow==1.14
- - keras
+ - tensorflow
  - future
- - pytest
  - pip
  - pip:
    - git+https://github.com/pycomlink/pycomlink.git#egg=pycomlink

environment_dev.yml

@@ -7,15 +7,14 @@ dependencies:
  - pandas>=0.22
  - matplotlib
  - numba
- - folium
- - h5py==2.10.0
+ - h5py
  - xarray
+ - netcdf4
  - shapely
  - pyproj
  - tqdm
  - pykrige
- - tensorflow==1.14
- - keras
+ - tensorflow
  - future
  - jupyterlab
  - notebook

pycomlink/__init__.py

@@ -5,7 +5,6 @@
 
 from . import processing
 from . import io
-from . import vis
 from . import spatial
 from . import validation
 from . import util

requirements.txt

@@ -3,7 +3,6 @@ scipy
 pandas>=0.22
 matplotlib
 numba
-folium
 h5py
 xarray
 netcdf4

setup.py

@@ -8,6 +8,9 @@
 import os
 from setuptools import setup, find_packages
 
+with open("requirements.txt", "r") as f:
+    INSTALL_REQUIRES = [rq for rq in f.read().split("\n") if rq != ""]
+
 # Utility function to read the README file.
 # Used for the long_description.  It's nice, because now 1) we have a top level
 # README file and 2) it's easier to type in the README file than to put a raw
@@ -33,23 +36,11 @@ def read(fname):
         "Development Status :: 3 - Alpha",
         "Topic :: Scientific/Engineering :: Atmospheric Science",
         "License :: OSI Approved :: BSD License",
-        'Programming Language :: Python :: 3.6'
+        'Programming Language :: Python :: 3.7',
+        'Programming Language :: Python :: 3.8',
+        'Programming Language :: Python :: 3.9',
     ],
     # A list of all available classifiers can be found at 
     # https://pypi.python.org/pypi?%3Aaction=list_classifiers
-    install_requires=[
-        'numpy',
-        'scipy',
-        'pandas>=0.22',
-        'matplotlib',
-        'numba',
-        'folium',
-        'h5py',
-        'xarray',
-        'netcdf4',
-        'shapely',
-        'pyproj',
-        'tqdm',
-        'pykrige',
-        'future'],
+    install_requires=INSTALL_REQUIRES,
 )