Open maxbriel opened 4 months ago
Although I didn't look into solving this issue, it's somewhat related to #387. Parameters are not initialised properly, despite the data being available. This post explains a workaround + the exact problem:
For the single star matching to work, we require a numeric value for the values:
'mass', 'center_h1', 'log_R', 'he_core_mass'
If such a value is not given when initialising the binary, these values remain None
.
As such the matching crashes by failing on the comparison between None
+ float
with the latter being grid values.
The workaround is to provide these values at the binary initialisation. This can be an actual value, or np.nan
Below is an example to initialise a NS+HMS binary at solar metallicity.
In that case center_h1
and he_core_mass
are known, and only log_R
has to be set to np.nan.
In V1, it was possible to evolve a binary from an arbitrary state, but in V2 you can only start a binary on a grid. For example, the system in the tutorial NS+H-rich_core-h-burning can start on the CO-HMS_RLO grid, but not in the detached step.
This is caused by some values not properly being initialised. I think it should be possible for someone to start a binary with a CO + a fresh star (ZAMS). This is not currently possible.