Fix no central orders and completely empty grids in scale variations#248
Fix no central orders and completely empty grids in scale variations#248Radonirinaunimi wants to merge 13 commits into
Conversation
|
@kamillaurent Could you please check with this branch (say for one of the datasets and sub-theories that failed before) that it is no longer an issue? |
|
@Radonirinaunimi I checked and seems the problem is no longer there. However, for these datasets: I still encounter the second error: |
Radonirinaunimi
changed the title
|
Could you please try now @kamillaurent ? (Fingers crossed 🤞🏼 it won't break the nFONLL combinations) |
|
@Radonirinaunimi the fix worked and I could compute the remaining ekos and FK Tables. |
| fktable.set_metadata("pineko_version", version.__version__) | ||
| fktable.set_metadata("theory_card", json.dumps(theory_meta)) | ||
| fktable.write_lz4(str(fktable_path)) | ||
| return fktable |
There was a problem hiding this comment.
Why is this an internal function? I'd say this should be useful enough to be outside and called taking a grid as an input. Something like construct_empty_fktable_like(grid, **otherinfo), similar to zeros_like from numpy.
There was a problem hiding this comment.
Simply promoted as a global function in 50e9b68.
|
AFAIC, this is ready to go. |
scarlehoff
left a comment
scarlehoff
left a comment
There was a problem hiding this comment.
Thank you for this. I mostly get the gist of it now that I have the issue in mind but I think I will have no idea why we are returning empty/none things in a year 😅
| @@ -174,7 +176,14 @@ def contains_sv( | |||
| ) -> Tuple[AvailableAtMax, int]: | |||
| """Check whether renormalization scale-variations are available in the pineappl grid.""" | |||
There was a problem hiding this comment.
Could you explain in the docstr of this function what are we exactly returning and why it makes sense to do return AvailableAtMax.BOTH, 0 if the grid doesn't contain anything?
| min_as: None or int | ||
| minimum power of strong coupling | ||
| """ | ||
| # A grid with no orders is a valid input for some workflows. In that case we |
There was a problem hiding this comment.
Just as before, could you add here the what are we returning and why does it make sense to return the grid, and empty fktable and None
4 participants
This PR fixes two issues arising from specific grids during scale variations, namely:
PS: It also bumps the version of
pineappltov1.4.0a1for the tests.