gh-575: bring back inv_triangle_number #577
Conversation
Saransh-cpp
added
maintenance
glass/fields.py
Outdated
| @@ -51,15 +51,25 @@ def wrapper(*args: _P.args, **kwargs: _P.kwargs) -> _R: | |||
| return decorator | |||
|
|
|||
|
|
|||
| def _is_inv_triangle_number(triangle_number: int) -> tuple[int, bool]: | |||
There was a problem hiding this comment.
I'm not a big fan of this return signature. The signature minimises repetitions but I am open to any better implementations.
There was a problem hiding this comment.
Just have _inv_triangle_number() throw an exception if it's not -- both downstream functions can catch and rethrow it
There was a problem hiding this comment.
Thanks! I have updated the implementation.
glass/fields.py
Outdated
| def nfields_from_nspectra(nspectra: int) -> int: | ||
| r""" | ||
| Returns the number of fields for a number of spectra. | ||
|
|
||
| Given the number of spectra *nspectra*, returns the number of | ||
| fields *n* such that ``n * (n + 1) // 2 == nspectra``. | ||
| """ | ||
| n = math.floor(math.sqrt(2 * nspectra)) | ||
| if n * (n + 1) // 2 != nspectra: | ||
| n, is_inv_triangle = _is_inv_triangle_number(nspectra) |
There was a problem hiding this comment.
_is_inv_triangle_number will similarly be used in a new function (to be added in #533) which will be specific to alm.
|
The docstrings are fixed now. |
Co-authored-by: Patrick J. Roddy <[email protected]>
There was a problem hiding this comment.
We're getting test failures now.
I have recently got numeric errors here https://github.com/astro-informatics/sleplet/actions/runs/13656527580 so I wonder if there has been a recent NumPy change or something
Co-authored-by: Patrick J. Roddy <[email protected]>
Description
Add
_is_inv_triangle_numberto check for triangle numbers and use it internally innfields_from_nspectra. The refactor might not look very useful right now, but it paves the way for adding annfields_from_nspectra-equivalent function foralmin #533 without repeating the core logic.Closes: #575
Checks