Closed alecandido closed 2 years ago
@scarlehoff @AleCandido commit eb2af60a69fd9e979a77d822fef904397e46d816 should fix this problem in the general case. Could you please let me know if it works?
@scarlehoff @AleCandido to clarify: with the mentioned commit it should work even if the grid isn't optimized.
General case means I can use it also with the non-optimized ones?
Also, it would be nice to also optimize away the numerical zeros (so that it is exactly equivalent to the NNPDF flavour list)
Also, it would be nice to also optimize away the numerical zeros (so that it is exactly equivalent to the NNPDF flavour list)
That's part of https://github.com/N3PDF/pineappl/issues/45 now.
Yes, it works with some of the unoptimized grids as well.
One problem though is that the convolute
method is not working any more. I'll try to write a MWE and open an new issue.
One problem though is that the
convolute
method is not working any more. I'll try to write a MWE and open an new issue.
I think the problem is commit 4e40743929483f9e670069238e7368b77634ba76, which now requires the PDFs that you pass to PineAPPL to be in the flavour basis, irrespectively of what basis the grid is in.
oh, ok. Then I don't have to use pineko's evol_xfxQ2
. Perfect. Then everything's working.
Great!
As reported by @scarlehoff running the following on the attached FkTable does not work, and it raises.
Most likely we have a
subgrid(0, 0, 0)
empty, and then the subgrid allocated has dimension (#bins, #lumis, 0, 0), and it fails as soon as it iterates over the following subgrids.https://github.com/N3PDF/pineappl/blob/ad6073b15bba36971e8475b7fb637b1f75464d47/pineappl/src/fk_table.rs#L55-L72