Closed cschwan closed 2 years ago
Note that not every useful grid has an LHAID. Specifically the internal results from the fit do not have that.
That's a good point; so this is even more important than I thought!
Implemented in commit 1104374ae1b6721b9353e50095428dce5559226b.
@Radonirinaunimi you might be interested in that as well (or might have liked to have it a few days ago :wink: ...).
@Radonirinaunimi you might be interested in that as well (or might have liked to have it a few days ago 😉 ...).
Thanks a lot @cschwan for this. It works perfectly fine! I am sure I will need this often in the future.
This is now documented in https://github.com/N3PDF/pineappl/blob/master/docs/cli-reference.md#pdfset-specifying-pdfs-or-pdf-sets.
When we call
pineappl convolute <grid> <pdfset>
, thepdfset
field is allowed to be either:331100
for NNPDF4.0 NNLO as=0.1180NNPDF40_nnlo_as_01180
for the case above.Sometimes we'd like to convolute with very specific replicas, which is only possible using the LHAID to which the replica ID must be added. For instance, the first (uncertainty) replica would be
331101
. LHAPDF offers another possibility, which we should also support:NNPDF40_nnlo_as_01180/1
. This requires adding another method in the LHAPDF rust interface first though.