Use headless ("Agg") matplotlib backend in CI (#1133)

Fixes #1132

This also drops internal matplotlib.use() calls, leaving the
choice of backend in the hands of the user/environment.
We believe these calls were counterproductive and redundant, given
that, from https://matplotlib.org/stable/users/explain/backends.html

> Without a backend explicitly set, Matplotlib automatically detects a
> usable backend based on what is available on your system

Co-authored-by: Joshua Newton <joshuacwnewton@gmail.com>

.github/workflows/run_tests.yml

@@ -19,6 +19,11 @@ env:
   # (TravisCI quietly defined this on all their platforms, but we have to give it manually on GithubCI.)
   DEBIAN_FRONTEND: 'noninteractive'
   HDF5_USE_FILE_LOCKING: 'FALSE'
+  # Skip to the headless matplotlib renderer, which is less
+  # bug-prone in the constrained environment of CI
+  # Tip from a matplotlib dev: https://github.com/spinalcordtoolbox/spinalcordtoolbox/issues/3388#issuecomment-846091012
+  # Ref: https://matplotlib.org/stable/users/explain/backends.html
+  MPLBACKEND: 'Agg'
 
 jobs:
   ultra_matrix_test:

ivadomed/scripts/visualize_and_compare_testing_models.py

@@ -9,6 +9,7 @@
 ###########################################################################################################
 
 import matplotlib
+from matplotlib import pyplot as plt
 import pandas as pd
 import numpy as np
 import itertools
@@ -21,21 +22,7 @@
 
 matplotlib.rcParams['toolbar'] = 'None'  # Remove buttons
 
-gui_env = ['TKAgg', 'GTKAgg', 'Qt4Agg', 'WXAgg']
-selected_gui_env = []
-for gui in gui_env:
-    try:
-        matplotlib.use(gui)
-        from matplotlib import pyplot as plt
-
-        selected_gui_env = gui
-        break
-    except:
-        continue
-# If none works
-if selected_gui_env == []:
-    from matplotlib import pyplot as plt
-
+if matplotlib.get_backend() == "agg":
     logger.warning("No backend can be used - Visualization will fail")
 else:
     logger.info(f"Using: {matplotlib.get_backend()}  gui")

ivadomed/utils.py

@@ -222,7 +222,6 @@ def plot_transformed_sample(before, after, list_title=None, fname_out="", cmap="
     if fname_out:
         plt.savefig(fname_out)
     else:
-        matplotlib.use('TkAgg')
         plt.show()