create_model

Build one SBML model from SPARCED input tables. Optionally split it afterwards into stochastic and deterministic SBML files.

Requires Python 3.11 or later.

Install

pip install create_model

From a clone of this repository:

pip install -e ".[dev]"

After either install, the create_model command is available on your PATH. If your environment does not put Scripts on PATH, python -m create_model is equivalent.

Build

create_model --config tests/data/config.yaml

Writes SPARCED_I.xml (One4All) to tests/data/output/ by default.

Partition

After a successful build:

create_model --config tests/data/config.yaml --partition-sbml

Writes:

• stochastic-gene-expression.xml
• deterministic-interactions.xml

Reactions are split by the species table solver column: any reaction with a Stochastic species as a reactant or product goes to the stochastic model; modifier-only links do not. The rest go to the deterministic model (disjoint). Each partition is then reduced to the species and parameters that participate in its reactions. Species that appear in no reaction are still kept, assigned by their solver label (stochastic-labeled unused species go to the stochastic model; remaining unused species go deterministic), so the union of partition species covers the One4All model. Model @id attributes are sanitized to valid SBML SIds (hyphens → underscores); output filenames keep the configured names.

Verifies reaction/species/global coverage against the One4All model by default (--no-verify to skip).

Useful flags

Config

tests/data/config.yaml is an example. Paths are relative to the config file unless absolute.

name: "SPARCED-I"
version: "1.4"
description: "Configuration for loading model input tables and writing SBML"

compilation:
  directory: "."
  files:
    compartments: "SPARCED-Compartments.tsv"
    ratelaws: "SPARCED-Ratelaws.tsv"
    species: "SPARCED-Species.tsv"
    parameters: "SPARCED-Parameters.tsv"
    annotations: "SPARCED-Annotations.tsv"

output:
  directory: "output"
  keep_antimony: true

partition:
  stochastic_model_id: "stochastic-gene-expression"
  deterministic_model_id: "deterministic-interactions"

compilation.files lists the TSV tables, output controls where the One4All SBML (and optional Antimony) is written, and partition sets the split model IDs.

Releasing

Publishing runs on GitHub Release publish (publish.yml). It builds an sdist and wheel and uploads them to PyPI with trusted publishing. Before the first release, create a pypi GitHub Environment and register this repository as a trusted publisher on PyPI for the create_model project.