Installation updates

Updates the default installation location to the conda environment /bin. This eliminates the need to source bayeswave-user-env.sh, and works with the condor updates that can't use get_env=True any more. User can still specify a specific installation directory if needed; in this case you will still need to source bayeswave-user-env.sh for the python utilities to work.

@james-clark and @sophie.hourihane can one or both of you check whether this makes sense, and maybe see if you can install without any issues?

BayesWaveUtils/bayeswave_pipe/bayeswave_pipe_utils.py

@@ -945,13 +945,20 @@ def condor_job_config(job_type, condor_job, config_parser):
 
     """
     valid_job_types=['bayeswave','bayeswave_post','bayeswave_fpeak','bayeswave_clean_frame','megaplot']
+    # When using default env, need to add capitalization etc in
+    executableNames = {'bayeswave':'BayesWave','bayeswave_post':'BayesWavePost','bayeswave_fpeak':'BayesWaveFpeak','bayeswave_clean_frame':'BayesWaveCleanFrame','megaplot':'megaplot.py'}
     try:
         job_index = valid_job_types.index(job_type)
     except ValueError:
         print("unrecognized job type", file=sys.stderr)
 
     # --- Set executable and choose singularity image
-    executable=config_parser.get('engine',job_type)
+    try:
+        executable=config_parser.get('engine',job_type)
+    except:
+        condaenv=sys.prefix+'/bin/'
+        executable=condaenv+executableNames[job_type]
+    print("Using executable...", executable)
     universe=config_parser.get('condor','universe')
     pipeline.CondorDAGJob.__init__(condor_job,universe,executable)
     pipeline.AnalysisJob.__init__(condor_job,config_parser)