Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
* Remove all backwards incomp. stuff (#186) * Improve Survey (#187) - Survey: - `frequencies` is also a dict. - Sources and receivers must be tuples or dicts; NO lists. - `name` is optional (and keyword only). - Optional `info` and `date` parameters. - No `Survey.observed`-attribute any longer. - No more `fixed` surveys. - `noise_floor` and `relative_error` are new stored as data array if they are not floats. - `data` can be a dict containing many data set. - Dipole - No `name` any longer. - Simulation: - `name` is optional (and keyword only). - Optional `info` parameter. * Sphinx theme (#188) * Re-work Core (#189) * Re-work Solver (#190) Solver: * Start solver overhaul * solve() does not take grid any longer; new signature, new `plain` * New verbosity levels Other * Model is aware of grid; reduced TensorMesh * _Map -> BaseMap * _TensorMesh -> BaseMesh * Field knows its grid (#191) Documentation: * Lots of documentation adjustements * IPython and matplotlib sphinx-directives (#192) * Re-work Model (#194) - Not set properties (`property_y`, `property_z`, `mu_r`, `epsilon_r`) all return None if called. - Not set properties cannot be set later. - `model.case` are strings instead of numbers: `['isotropic', 'HTI', 'VTI', 'triaxial'] * Separate API docs per module (#195) * Re-work Maps (#196) * Rework entire interpolation stuff * Cleaned-up maps * Re-work meshes (#197) * PEC to prolongation (#198) * Re-work Fields Part I (#199) * Remove SourceField * Fields is not a subclassed ndarray any longer * Inherit discretize conditionally * Add setup.cfg for F401 on __init__ * Fix bug src-dep gridding in Simulation * Do not (de-)serialize Maps * Rework fields (#200) * Cleaned module fields * Add Win/MacOs; Python 3.9 * New fct emg3d.solver.solve_source * New module electrodes; includes electric wire source * Fix windows pytest failure (#202) `pytest test_fields.py::TestGetReceiver::test_basics` fails on Windows, most likely with `emg3d.construct_mesh` - No idea why (`int32` issue?). * Re-work electrodes (#203) * Rework Surveys, Simulations, Optimize (#204) * Magnetic field lives on faces again * Outsource gallery again * Do not store sfield, simplify rfield * Add 'relative' to receivers * New electrode convenience fcts * Enable relative receivers in Survey/Simulation * magnetic field use volume avg for mu again * Simulation; no more grid * to emsig.xyz and emg3d.emsig.xyz * dict instead of tuples possible for construct_mesh * Rework CLI/utils/io (#205) * Rework io, utils, time * master->main * Move Fourier to new module `time` * Fixes and docs for rc1 (#207) * Improve docs * Script to update old file format * Fix rel. receivers in Simulation class * Fixes to v1.0.0-rc.1 (#209) - Fix adding noise to observed data in simulation if noise_floor is an xarray (related to point below) - Improve `survey.{noise_floor;relative_error}` * fix workflow sed * Further improvements towards 1.0 (#211) * Further pep8 names in `core` * Change discretize from master to main * Re-enable `sphinx.ext.viewcode` * Shorten doc urls (manual/api/dev) * `max_prime` to `max_lowest` in `good_mg_cell_nr` * Add CPU/RAM script to doc * Fix fields.interpolate * Update logo
- Loading branch information