feat: reintroduce invalid_value in bisect #33
Conversation
Co-authored-by: Halvor Lund <[email protected]>
linerate/solver.py
Outdated
| @@ -142,4 +148,4 @@ def compute_conductor_ampacity( | |||
| """ | |||
| f = partial(heat_balance, max_conductor_temperature) | |||
|
|
|||
| return bisect(f, min_ampacity, max_ampacity, tolerance) | |||
| return bisect(f, min_ampacity, max_ampacity, tolerance, invalid_value=0) | |||
There was a problem hiding this comment.
It doesn't make sense to me to set this value to 0, what is the use case? I think it would be better to set to nan by default, and provide invalid_value as an input argument in compute_conductor_ampacity to be able to override. Does that solve the issue?
There was a problem hiding this comment.
The use case, in my understanding, is calculating the conductor ampacity for a span for a large array of weather parameters with one function call. A span with a strict temperature limit will have an ampacity that is frequently 0 (i.e. can't be used in many weather conditions). From bisect's perspective, that means the heat equation never equals zero.
If this is a common usage of compute_conductor_ampacity, I think returning 0-ampacity as a proxy for "this line can't be used under these conditions" makes sense. That way, the returned array can be used without post-processing.
There was a problem hiding this comment.
Current proposed solution is what @gunnhildsp suggested.
Due to unforeseen complications in using
bisectwithout aninvalid_valueparameter, I suggest reverting previous changes and re-introducing theinvalid_valueinbisectand by extensioncompute_conductor_ampacity