API reference

Conatus has two layers. Most scripts only need the first.

• The user layer, import conatus as cn: named geometries (cn.Box, cn.Channel, cn.Disk, …), the model in which equations are written as on paper (cn.Model), the objects it returns (simulation, linearisation, compiled problem), and the utilities (cn.problems, cn.runs, cn.doctor(), cn.from_dedalus, cn.parse_equation). See User layer: import conatus as cn.

• The expert layer, conatus.expert: the mathematical intermediate representation (fields, bases, operators, problems written term by term) and every toolkit that works on it (solvers, time stepping, adjoints, invariant solutions, continuation, bifurcations, Floquet–Hill, non-modal analysis, chaos, data-driven methods, telemetry, provenance). See Expert layer: conatus.expert.

The two layers meet in one place: m.compile(N).problem is an ordinary expert-layer problem, and sim.solver is an ordinary expert-layer time stepper, so every toolkit applies to a native model.

• User layer: import conatus as cn
  • Geometries
  • Wall conditions on a vector unknown
  • The model
  • Ready problems: cn.problems
  • The run registry: cn.runs
  • Quantities of a run
  • Utilities
  • Regions of the complex plane: cn.box, cn.contour
• Expert layer: conatus.expert
  • explain()
  • Expert layer by module