Hi Troels,
There are still a few things to fix. For example the
Relax_disp.test_hansen_cpmg_data_missing_auto_analysis system test no
longer passes, and I'm not sure why. There is also the following:
- The target_function.relax_disp to lib.dispersion API
is 100% consistent. I.e. R2eff is always returned, or back_calc is
always sent into the function..
For the API it must be one or the other. The 'disp_speed' branch has
both which is not acceptable. Therefore we need to discuss how to
convert all to returning R2eff, or convert all to having back_calc as
a function argument (this second option might be the easiest by using
back_calc[:] = R2eff on the last line). This will clean up the API.
- All test_suite.unit_test._lib._dispersion modules has been validated with
the './devel_scripts/code_validator' script.
The lib/dispersion/m61b.py file has a double empty line remaining (not
found by the script). The rest look good.
- All unit tests have different r20a and r20b values,
when the dispersion function accepts both, to check if the correct one is
returned.
This looks good.
- Checked that the R1rho unit tests are correct.
Running the unit tests with the --numpy-raise flag shows that two of
the test*_rex8*() unit tests fail for high kex values.
I'll look into the branch in more detail tomorrow, and the test
timings. It's looking good though.
Cheers,
Edward
Re: [task #7793] Speed-up of dispersion models