- Generated by
1.15.0
|
STAR
Cal STAR — firmware, DAQ server, and engine design
|
Classes | |
| class | EngineImportError |
Functions | |
| Any | _require (Mapping[str, Any] config, str path, str source) |
| float | _optional (Mapping[str, Any] config, str path, float default=0.0) |
| DischargeModel | _discharge (Mapping[str, Any]|None block) |
| EngineDesign | engine_from_config (Mapping[str, Any] config, *, str name="engine") |
| EngineDesign | load_engine (str|Path path) |
Reading an EngineDesign Layer-1 config into a feed-system boundary. The contract this module keeps is narrow and worth stating: **an engine design is imported, never re-typed, and never edited on the way in.** Whatever Layer 1 optimised to is what gets fired. If a number is missing the import says which one and stops; it does not substitute a plausible default, because a plausible default here produces a run that looks right and is not the engine anybody designed. What is read, and what is ignored --------------------------------- Read: ``injector.type`` and ``injector.geometry``, ``discharge.*``, ``fluids.*.name`` and ``.temperature``, ``chamber_geometry.A_throat``, ``expansion_ratio``, ``volume``, ``design_pressure``, ``design_MR`` and ``design_thrust``. Ignored, on purpose: ablative cooling, spray SMD correlations, optimizer weights, structural margins, CEA cache paths, stability scores. None of them is a feed-system boundary condition, and reading them would create a coupling that has to be maintained for no benefit. The file is YAML because EngineDesign writes YAML. That is the one place this library touches it -- and it is read with ``safe_load`` into plain dicts, so an engine config cannot execute anything on import.
|
protected |
One side's discharge block, read literally. EngineDesign does not read it literally. Its ``load_config`` re-stamps the injector type's baseline over the Cd fields on every load (``engine.pipeline.config_switch.apply_injector_type``; impinging: Cd_inf 0.60, a_Re 0.18, Cd_min 0.35, geometry Cd on, small-hole exponent 0.20), keeping only a declared ``use_geometry_cd`` and the correction settings. So this agrees with EngineDesign only where the block *is* that baseline -- every shipped ethalox config and LE4's engine -- and not on the older methalox configs, nor on ``default.yaml``'s exponent of 0.0 (which the ``or 0.20`` below happens to read as EngineDesign does). Matching EngineDesign's schema defaults instead was tried and is wrong for the same reason (2026-10-08). The engine card carries EngineDesign's own Cd; this is the simplified engine's.
Definition at line 86 of file importer.py.
Referenced by engine_from_config().
|
protected |
Definition at line 77 of file importer.py.
Referenced by engine_from_config().
|
protected |
Fetch a dotted path, or fail naming exactly what was missing.
Definition at line 55 of file importer.py.
Referenced by engine_from_config().
| EngineDesign feedtwin.engine.importer.engine_from_config | ( | Mapping[str, Any] | config, |
| * | , | ||
| str | name = "engine" ) |
Build a feed-system boundary from an already-parsed engine config. Separated from :func:`load_engine` so a config that arrived over HTTP, or one assembled in a test, imports through exactly the same path as one read off disk. The parser is not the interesting part; the extraction is.
Definition at line 126 of file importer.py.
References _discharge(), _optional(), _require(), and max.
Referenced by load_engine().
| EngineDesign feedtwin.engine.importer.load_engine | ( | str | Path | path | ) |
Import a Layer-1 engine config from disk.
The whole Phase 08 promise in one call: point it at what the optimizer
wrote and get a feed-system boundary, with nothing retyped.
Raises:
EngineImportError: the file is unreadable as an engine boundary, naming
the key that was missing and where it looked.
Definition at line 331 of file importer.py.
References engine_from_config().
1.15.0