pygcam.mcs.gcamdata

The gcamdata module allows the GCAM data system to be run for each Monte Carlo trial to generate XML files that correspond to some stochastic alteration to data system input (CSV) files. See Running the GCAM data system for more information.

API

class pygcam.mcs.gcamdata.GcamDataSystem(mapper: SimFileMapper, renv_dir=None, xml_modifier=None)

Class to run the GCAM data system within a Monte Carlo simulation

activate_renv()

Activate the “renv” at directory self.renv_dir.

Returns

none

load_gcamdata()

Call devtools::load_all(self.ref_gcamdata_dir).

Returns

none

move_modified_xml_files(src_dir, dst_dir)

Move modified XML files from the temporary source directory src_dir to the destination directory dst_dir.

Parameters

  • src_dir – (str) the source directory

  • dst_dir – (str) the destination directory

Returns

(list of str) the paths to the modified XML files in the destination directory

run_data_system(trials, user_modifications, delete=True)

Run the GCAM data system using “driver_drake” to generate XML files with user_modifications.

Parameters

  • trials – (“PATH” or a comma-delimited list of ints or hyphen-separated numbers indicating the trials to run. (Example: “1,4,7-11,2,21-50”)

  • user_modifications – (str or list of str) names of “user modification” R functions to insert into the GCAM data system for use with drake. If None, just run the data system in the reference workspace, without any modifications to create the baesline.

  • delete – (bool) whether to delete the temporary directory in which the data system is run. Can be helpful to set to False for debugging.

Returns

none

run_drake(gcamdata_dir)

Run the gcamdata (R) function “driver_drake” in the given workspace. This is used to create the baseline drake information before running it with user modifications, and thus before copying the reference workspace to temporary trial sandboxes.

Parameters

gcamdata_dir – (str) the pathname of the gcamdata dir in which to run “driver_drake”.

Returns

none

run_drake_with_mods(func_names)

Run the gcamdata (R) function “driver_drake”, passing in the user modification functions in func_names

Parameters

func_names – (str or list(str)) the function name or list of function names to pass as the “user_modifications” to “driver_drake”.

Returns

none

trial_func(trial_num)

A function to call for each Monte Carlo trial. Does nothing in GcamDataSystem; subclass can define this if needed.

Parameters

trial_num – (int) the trial being run

Returns

none

trial_sandbox(trial_num, delete=True)

Return the path to a temporary sandbox directory in which to build dependent XML files. A cache of these is maintained so on subsequent calls with the same trial_num, the cached path is returned and no other action is taken.

Parameters

  • trial_num – (int) the number of the trial being run

  • delete – (bool) whether to delete the temporary directory created by this function at program exit.

Returns

(str) the trial’s sandbox dir

update_scenario_config(rel_dir: Path, abs_paths)

Update the configuration XML file for the simId and scenario found in the saved mapper, substituting the path to the modified XML for the file ending with the same basename. This assumes basenames within the config file are unique.

Parameters

  • rel_dir – (Path) the directory that holds the modified XML files, relative to the “exe” directory for the given scenario, suitable for insertion into the configuration XML file.)

  • abs_paths – (list of Path) the absolute paths of the modified XML files.

Returns

none

pygcam.mcs.gcamdata.load_R_code(code_str)

Loads a string containing R code into the R interpreter running within Python.

Parameters

code_str – (str) R code string

Returns

none