With synthetic data in mth5 a lot of the tests in aurora (that are actually testing mth5) can be moved there.
This will take some time.
As a start:
[ ] Update how the metadata are provided to make the mth5
Currently, an adhoc scheme is used. This factoring provides a chance to clean up how the metadata are specified.
One approach to this may be to create json versions of the mt_metadata needed to define the synthetic mth5 objects, and archive these in the repo. The mth5 file could then be constructed via explicit calls to initalize station, run and channel objects with metadata defined by these jsons?
another approach would just initialize the mt_metadata obejcts.
Start by defining instances of a Survey, Station, Run, and 5 Channel metadata objects mt_metadata.timeseries.
Populate these with default values, and update any needed attrs.
With synthetic data in mth5 a lot of the tests in aurora (that are actually testing mth5) can be moved there.
This will take some time. As a start:
[ ] Update how the metadata are provided to make the mth5 Currently, an adhoc scheme is used. This factoring provides a chance to clean up how the metadata are specified.
use test_channel_ts.py as a template