prism-auto: Add basic filtering of benchmarks using models.csv metadata.
New switch --filter-models X which restricts the list of models used from
a directory to those that match the filter X. Currently, this can refer to
the number of states and/or the model type. Examples:
* prism-auto . --filter-models "states>100 and states<10000"
* prism-auto . --filter-models "model_type=='DTMC'"
* prism-auto . --filter-models "'MC' in model_type"
The model metadata is by default read in from a models.csv file (as found,
for example, in the PRISM benchmark suite). The name of the file used can
be changed with --models-info FILE.
parser.add_option("--echo-full", action="store_true", dest="echoFull", default=False, help="An expanded version of -e/--echo")
parser.add_option("--models-filename", dest="modelsFilename", metavar="X", default="models", help="Read in list of models/parameters for a directory from file X, if present [default=models]")
parser.add_option("--models-info", dest="modelsInfoFilename", metavar="X", default="models.csv", help="Read model details from CSV file X, if present [default=models.csv]")