Reactor

Reactor is a user-facing scenario object that loads a reactor.yaml specification, holds variable values and relation-selection settings, and can build a RelationSystem for numeric evaluation.

Reactor Data Layout

Each reactor lives in one of these forms:

• reactors/<reactor_id>/reactor.yaml
• reactors/<reactor_id>.yaml

High-level structure:

• metadata: id/name/year/country/source information
• tags: labels used for relation filtering
• solver_tags: solve mode, verbosity, optional solving order
• variables: scalar/profile variable values and optional solver hints

The generated reactor YAML reference is built from the current files under reactors/.

Loading and Solving

from fusdb import Reactor

reactor = Reactor.from_yaml("reactors/ARC_V0")
result = reactor.run()

Core fields - name, organization, country, year, doi, notes - tags: tuple of strings used to filter relations - mode: one of "verify", "reconcile", "optimize", or "ordered" - variables: mapping of loaded Variable instances keyed by canonical name - relation_include, relation_exclude, relation_order: relation selection controls - constraints: optional system-level constraint expressions - grid_size: optional shared profile size inferred from the reactor grid block - verbose: runtime verbosity flag

Example

from fusdb import Reactor, Variable

reactor = Reactor(
  name="Example Reactor",
  tags=("tokamak", "hmode"),
  variables={
    "R": Variable("R", value=3.2, unit="m"),
    "a": Variable("a", value=1.1, unit="m"),
  },
)

Methods - from_yaml(path) loads reactor.yaml, parses variables (including numeric profile files), and applies registry defaults. - selected_relations(...) returns the relation list after applying includes, excludes, tags, and ordering. - relation_system(...) builds and returns a RelationSystem bound to the reactor's current variables and selected relations. - run(**kwargs) convenience wrapper that builds a RelationSystem and runs the configured mode.

Profiles and variables - Variables are declared with value (scalar or 1D profile), optional unit, rel_tol, fixed, size, and constraints entries. Profile files may be supplied via file: path.csv or value: path.csv in YAML; simple numeric strings are left to variable coercion. - When selected relations require variables that were not provided, relation_system will create placeholder Variable objects using registry defaults (including profile size from grid.size when available).

Reactor YAML (example)

metadata:
  id: REACTOR_ID
  name: Reactor Name

tags:
  - device_tag
  - regime_tag

solver_tags:
  mode: verify         # verify | reconcile | optimize | ordered
  verbosity: false

grid:
  size: 46

variables:
  R: 3.2
  a:
    value: 1.1
    unit: m
    rel_tol: 0.02
    fixed: false
  n_e:
    file: profiles/ne.csv
    delimiter: ","
    skiprows: 1