Scale Variations

[felixhekhorn]

Closes #6

[alecandido]

Remember to add to the documentation (see PR #18) considerations about schemes: pineko is the best place, since both EKO and yadism see only a part of it.

Moreover, you'll need most likely to patch something even in EKO. Indeed, if you apply factorization scale variations on the DGLAP side, and you use the expanded mode [1]:

1. fact_to_ren (still looking for a better name) has to be set different from 1, but I'd say it should really be xif; indeed, as for Eqs. 3.34 - 3.35 of our beloved arXiv:1906.10698v2, you'll evolve the PDF up to t + k = log(Q^2 * xiF^2) = log(muF^2), so the renormalization scale plays no role in here
2. fact_to_ren is not passed to main evolution step, following Eq. 3.35: we are now computing f(t + k) := fbar(as(t + k), 0), so the end-point is changed, but no correction is applied here (to the "resummation" scale, according to the nomenclature of a recent popular talk ^^)
3. fact_to_ren is used in computing the expanded correction, explicitly quoted in Eq. 3.35

If ModSv = exponentiated is chosen, then no expanded correction has to be computed, but fact_to_ren has to be passed in step 2. to the shift the point in which anomalous dimensions are evaluated (i.e. to shift the resummation scale wrt the factorization scale in which we're evolving), that corresponds to Eq. 3.34.

@andreab1997 please include all of this (possibly even improved with links and better descriptions) in the docs.

[1]: default for scheme B, even though not the one implemented by APFEL (that use exponentiated, even though it should be almost the same thanks to the adoption of TRN solution)

[felixhekhorn]

Closes #23

[codecov]

Codecov Report

Merging #19 (25ed076) into main (e12c648) will increase coverage by 1.60%.
The diff coverage is 48.52%.

@@            Coverage Diff             @@
##             main      #19      +/-   ##
==========================================
+ Coverage   37.64%   39.24%   +1.60%     
==========================================
  Files          15       15              
  Lines         441      479      +38     
==========================================
+ Hits          166      188      +22     
- Misses        275      291      +16     

Flag	Coverage Δ
unittests	39.24% <48.52%> (+1.60%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
src/pineko/cli/_base.py	100.00% <ø> (ø)
src/pineko/evolve.py	28.84% <20.00%> (-2.27%)	⬇️
src/pineko/theory.py	20.51% <25.00%> (+0.79%)	⬆️
src/pineko/cli/theory_.py	80.00% <33.33%> (+0.40%)	⬆️
src/pineko/theory_card.py	19.23% <40.00%> (+3.23%)	⬆️
src/pineko/check.py	25.00% <66.66%> (+5.00%)	⬆️
src/pineko/cli/check.py	63.15% <75.00%> (+4.33%)	⬆️
src/pineko/cli/compare.py	83.33% <80.00%> (+3.33%)	⬆️
src/pineko/__init__.py	100.00% <100.00%> (ø)
src/pineko/cli/__init__.py	100.00% <100.00%> (ø)
... and 5 more

[felixhekhorn]

let's start by merging this - a significant amount of the changes is due to #23