|
STAR
Cal STAR — firmware, DAQ server, and engine design
|
Functions | |
| flatten (d, pre="") | |
| int | main () |
Variables | |
| ROOT = Path(__file__).resolve().parent.parent | |
| tuple | CONSEQUENTIAL |
| dict | KNOWN_TRAPS |
Which numbers in this config did anyone actually choose?
EngineDesign configs carry ~900 fields. Most arrive by inheriting configs/default.yaml,
which is a GENERIC template -- its feed system is 3/8" NPT at K0 2.0 with 0.305 m runs,
which is nobody's vehicle in particular. Those values are perfectly valid config, so
nothing in the loader, the optimizer or the UI ever flags them, and a design point can
miss by 20%+ on O/F with every gate reporting a clean reason that points somewhere else.
This prints, for a given config, every physically consequential field that is
BIT-IDENTICAL to default.yaml -- i.e. inherited, not chosen.
python3 scripts/config_provenance_audit.py configs/mine.yaml [--all]
Note the obvious limit: identical-to-default is not the same as wrong. Prandtl 0.7 is
fine for anyone. The point is that nothing distinguishes the fields where that is true
from the fields where it is not, so the list is where to start looking, not a defect list.
| config_provenance_audit.flatten | ( | d, | |
| pre = "" ) |
Definition at line 45 of file config_provenance_audit.py.
References flatten().
| int config_provenance_audit.main | ( | ) |
Definition at line 57 of file config_provenance_audit.py.
References flatten(), and main().
Referenced by main().
| tuple config_provenance_audit.CONSEQUENTIAL |
Definition at line 28 of file config_provenance_audit.py.
| dict config_provenance_audit.KNOWN_TRAPS |
Definition at line 33 of file config_provenance_audit.py.
| config_provenance_audit.ROOT = Path(__file__).resolve().parent.parent |
Definition at line 25 of file config_provenance_audit.py.