Querying parameters

Access the parameters of past simulations two ways:

• simulationsTable — reads the databases; best for readability.
• getAllParameterValues — reads every XML value; best for programmatic access.

simulationsTable

simulationsTable returns a table of simulation data. By default it shows only varied values and renames columns to be human-readable.

printSimulationsTable is a wrapper that prints the table directly. Use the sink keyword argument to redirect the output, e.g. to a file.

Monad-level: monadsTable

monadsTable is the monad-level analogue of simulationsTable: it returns one row per monad (a group of replicate simulations sharing the same parameters) rather than one row per simulation. Pass it any AbstractTrial (e.g. a Sampling), a vector of monad IDs, or nothing (for all monads):

monadsTable(sampling)                      # one row per monad in the sampling
monadsTable([1, 2, 3]; remove_constants=false)  # by monad ID, keeping constant columns

printMonadsTable prints the table directly, mirroring printSimulationsTable.

getAllParameterValues

getAllParameterValues returns every terminal element in the XML input files for a set of simulations, which must all belong to the same Sampling (i.e. use the same input files). Column names are the XML paths, so splitting one on / gives a path ready for DiscreteVariation:

df = getAllParameterValues(sampling)
col1 = names(df)[1] # get the name of the first column
xml_path = split(col1, "/") # convert to XML path format
dv = DiscreteVariation(xml_path, [0.0, 1.0]) # create a discrete variation using this parameter

columnName and columnNameToXMLPath convert between column names and XML paths. Both are exported by ModelManager and re-exported here, so they need no prefix.

df = getAllParameterValues(sampling)
names_with_temp_id = filter(contains(":temp_id:"), names(df))