15  Composing results

15.1 Overview

ems_compose() retrieves and processes results from a solved model run. Results are parsed according to the specified type (variables, coefficients, or all). Data validation and consistency checks are performed during the parsing process.

By default, ems_solve() calls ems_compose() automatically, so this function is primarily used when suppress_outputs = TRUE was passed to ems_solve() or solve_in_situ(), or when re-parsing results from a previous run.

15.2 Arguments

Argument	Default	Description
cmf_path		Path to the CMF file generated by ems_deploy()
type	"all"	Type of data to parse: "all", "variable", or "coefficient"
name	NULL	Character vector to filter results by name, returning a subset of the selected type
minimal	FALSE	Whether to run a minimal number of checks and modifications to output data

15.3 Basic usage

After a standard model run with suppressed outputs:

ems_solve(
  cmf_path = cmf_path,
  matrix_method = "LU",
  solution_method = "Johansen",
  suppress_outputs = TRUE
)

# Parse all results
outputs <- ems_compose(cmf_path = cmf_path)

15.4 Output types

The type argument controls what is parsed from the solver output:

# All model variables and coefficients
all_outputs <- ems_compose(cmf_path = cmf_path, type = "all")

# Percentage change values for model variables only
variables <- ems_compose(cmf_path = cmf_path, type = "variable")

# Values for model coefficients only
coefficients <- ems_compose(cmf_path = cmf_path, type = "coefficient")

15.5 Filtering by name

The name argument filters results to a subset of the selected type:

# Retrieve a single variable
qfd <- ems_compose(cmf_path = cmf_path, type = "variable", name = "qfd")

# Retrieve multiple variables
trade_vars <- ems_compose(cmf_path = cmf_path, type = "variable",
                          name = c("qfd", "qfm", "qgd"))

15.6 Minimal mode

When minimal = TRUE, a reduced number of checks and modifications are applied to the output data:

1. Chronological data is not loaded and the model file is not parsed
2. No post-model set checks are conducted between model sets as written out from the model input and solver set binaries
3. Time steps remain in their model format (integer indices rather than chronological years)

This option must be set to TRUE when using output from solve_in_situ() with writeout = FALSE, since there is no set writeout to validate against:

outputs <- solve_in_situ(
  ...,
  model_file  = "model.tab",
  closure_file = "model.cls",
  shock_file   = "shocks.shf",
  suppress_outputs = TRUE
)

variables <- ems_compose(cmf_path = outputs, type = "variable", minimal = TRUE)