| # Restore consistent magnetic-dipole electromagnetic predictions |
|
|
| A geophysical electromagnetic forward-modeling workflow uses a magnetic dipole |
| source and reports magnetic observations at several receiver locations. The |
| calculation completes with finite values, but the result is not consistent with |
| the expected physical definition. This is especially |
| concerning for a forward model, because a finite numerical result can still |
| represent the wrong physical experiment. |
|
|
| Study the complete source snapshot and public reproduction. Trace the physical |
| calculation through the public |
| survey, source, simulation, and receiver interfaces, and repair the source so |
| that valid magnetic-dipole calculations agree with the scientific definition. |
| The repair must generalize to other dipole orientations, magnetic |
| permeabilities, receiver locations, mesh sizes, and supported forward-modeling |
| workflows. |
|
|
| Do not hard-code the supplied observations or replace the solver with a fixed |
| answer. Work only under this task directory. The evaluation environment is |
| offline; all required Python dependencies and source files are provided. |
|
|
| Run the public diagnostic with: |
|
|
| ```bash |
| python reproduce.py |
| ``` |
|
|
| The command returns `0` when the candidate workflow produces finite receiver |
| observations, `1` only when it cannot produce a finite nonempty observation, and |
| `2` for an import, path, dependency, or other runner failure. The public report |
| is diagnostic; the verifier performs the independent physical consistency |
| checks on broader cases. |
|
|
