Replacing GMT/PyGMT backend by pure Matplotlib backend #266

	name: Python application

	on:
	push:
	branches: [ master ]
	pull_request:
	branches: [ master ]

	jobs:
	build:

	# we find that CPU/memory limits are less often exceeded using this
	# specific version of Ubuntu
	runs-on: ubuntu-20.04

	#runs-on: ubuntu-latest


	steps:
	- uses: actions/checkout@v2
	- name: Set up Python 3.9
	uses: actions/setup-python@v2
	with:
	python-version: 3.9

	- name: pip installation
	# test pip installation first (much faster than conda)
	run: \|
	pip install --upgrade pip
	pip install -e .
	- name: Basic tests
	run: \|
	python tests/check_import.py
	python tests/check_entry_points.py
	bash tests/check_examples.bash
	- name: Advanced tests
	run: \|
	bash data/examples/unpack.bash
	bash data/tests/unpack.bash
	python tests/benchmark_cap_vs_mtuq.py --no_figures
	python tests/test_grid_search_mt.py --no_figures
	python tests/test_grid_search_mt_depth.py --no_figures
	python tests/test_greens_SPECFEM3D_SAC.py --no_figures
	python tests/test_time_shifts.py --no_figures

	# unfortunately, these Conda installation tests exceed the resource limits
	# for GitHub workflows
	# (https://github.com/actions/runner-images/issues/6680)
	# (https://github.com/conda/conda/issues/8650)
	#- name: conda installation
	# timeout-minutes: 60
	# run: \|
	# bash tests/conda_install.bash

Provide feedback