Input data: model execution

The execution of the model is orchestrated using Snakemake, as defined in the project’s Snakefile located in the root directory. Snakemake acts as the workflow engine, coordinating individual rules and chaining them into a single, automated execution process. This allows for scalable, parallel computation across multiple cores or threads.

Execution of Snakemake workflow

Below is a simplified representation, showing the workflow for the years 2025-2050 with the power sector only.

Rules and functions

The table below explains the main Snakemake rules and helper functions used in the PyPSA-SPICE workflow:

Name	Type	Description
add_baseyear	Snakemake rule	First step of building up the model after input data preparation is done. This rule imports all input data in the base year into the model. The base year is defined in config.yaml (see new-model or model-builder-configuration).
add_brownfield	Snakemake rule	Imports all output data from the previous year, using the previous_year_outputs and solved_previous_year functions. The definition of the different years is described in config.yaml (see new-model or model-builder-configuration).
make_summary	Snakemake rule	Generates output CSVs and summary of a specific year.
combine_summaries	Snakemake rule	Last step of the whole workflow aside from solve_all_networks rule. It generates summary of all years after all the networks from different years are solved.
solve_network	Snakemake rule	Solves the model builder for a specific year. This rule is only triggered after either add_baseyear or add_brownfield. Optimisation settings are defined in config.yaml (see model-builder-configuration).
solve_all_networks	Snakemake rule	Executes the entire model builder workflow at once.
previous_year_outputs	Python Function	Reads the output CSVs of the previous year.
solved_previous_year	Python Function	Reads the network from the output of the previous year.

Running the model builder

Run the entire workflow at once

To execute the entire workflow, modify the configfile parameter in the snakemake to point to your scenario config_scenario.yaml file. After this, you can run the whole workflow using:

Run the entire model builder workflow at once.

snakemake -j1 -c4 solve_all_networks #(1)!

1. -j1 means running only 1 job in parallel at a time and -c4 means allowing each job to use up to 4 CPU threads. See the Snakemake CLI documentation to customise cores, threads, and parallel execution.

or

Running the entire model builder workflow using this call command

snakemake -call

Run a single year

If you want to run a specific year instead of running multiple years, You can do it using the following command:

Option 2: Run a single output file.

snakemake -j1 -c4 post-solve/elec_{SECTOR}_{YEAR}.nc

Since a particular year run needs previous year outputs (except base year), snakemake will run the all the required worksflows to generate the output for specified year.