teems

Trade and Environment Equilibrium Modeling System --- an open-source R package for reproducible Computable General Equilibrium (CGE) model runs.

TEEMS provides a complete pipeline from raw GTAP data through model solution: data loading, aggregation, model parsing, shock specification, and Docker-based solving --- all from R.

Overview

Function	Purpose
ems_data()	Load and prepare data inputs
ems_model()	Parse and modify model and closure files
ems_uniform_shock()/ems_custom_shock()/ems_scenario_shock()	Prepare shocks
ems_swap()	Prepare closure swaps
ems_deploy()	Validate and write all model inputs
ems_solve()	Configure and solve model
ems_compose()	Compose model results into structured data objects
ems_option_set()/ems_option_get()/ems_option_reset()	Get, set, and reset advanced package options
ems_example()	Write teems example models and scripts to file
solve_in_situ()	Configure and solve model in-situ
GTAP_convert()	Convert GTAP HAR files and data formats

Installation

Prerequisites

• R (>= 4.3.0)
• Docker --- required for ems_solve() (see teems-solver)
• Solver build --- required for ems_solve() (see teems-solver)

Install from CRAN

# install package
install.packages("teems")

teems 1.0.0 and later require teems-solver 1.1.0 or later and check the image's version before every solve. If you built the solver image for an earlier release, rebuild it (see "Updating an Existing Installation" in the teems-solver README) and restart R.

If the installation fails citing a missing package, install that package first and retry.

Quick start

library(teems)

# 1. Load and aggregate data
dat <- ems_data(
  dat_input = "v7_data/gsdfdat.har",
  par_input = "v7_data/gsdfpar.har",
  set_input = "v7_data/gsdfset.har",
  REG  = "AR5",
  ACTS = "macro_sector",
  ENDW = "labor_agg",
  time_steps = c(0, 1, 2, 3)
)

# 2. Retrieve a model
model_files <- ems_example("GTAP-RE", tempdir())

# 3. Load model and closure
model <- ems_model(model_file = model_files[["model_file"]],
                   closure_file = model_files[["closure_file"]])

# 4. Define shocks
shock <- ems_uniform_shock(var = "pop",
                           REGr = "lam",
                           value = 1)

# 5. Validate and write solver inputs
cmf_path <- ems_deploy(.data = dat,
                       model = model,
                       shock = shock)

# 6. Solve
results <- ems_solve(cmf_path = cmf_path,
                     solution_method = "Gragg",
                     matrix_method = "SBBD",
                     n_subintervals = 2,
                     n_tasks = 2)

Data requirements

TEEMS works with GTAP FlexAgg data.

Documentation

Full documentation is available at teemsphere.github.io.

Related repositories

Repository	Description
teems-models	Vetted models and their standard closures
teems-solver	C/Fortran optimization solver (Docker)
teems-mappings	Regional and sectoral aggregation mappings
teems-manual	Full package documentation

License

AGPL-3.0 --- see LICENSE.md for details.

Contact

Matthew Cantele --- matthew.cantele@protonmail.com