set matplotlib cache directory

This MR resolves issues with matplotlib lock files.

Matplotlib uses a cache directory in the user's home directory by default. For a plotting job running in condor, the user's home directory is accessed over the network, which introduces latency and causes lock file errors. The change in this MR uses the previously empty __init__.py script in gstlal.plots to set a special matplotlib environment variable, telling it to use local tmp space for its cache directory.

gstlal/python/plots/__init__.py

 import os
-import sys
-import warnings
 
-if "matplotlib" in sys.modules:
-	warnings.warn("matplotlib is imported before gstlal.plots. This may cause lockfile errors.")
-
-# set matplotlib cache location to tmp directory
+# set matplotlib cache location to local tmp directory
 if os.getenv("_CONDOR_SCRATCH_DIR"):
         os.environ['MPLCONFIGDIR'] = os.getenv("_CONDOR_SCRATCH_DIR")
 else: