NNPDF/nnpdf: 4.1.5

The last bug is in captivity, the galaxy is at peace. This is the last of the 4.1 releases in which results/function calls are expected to change. From here to 4.1 we only expect polishing of the code and documentation (and new data implementations). Of course, there will be new bugs to squash. Most of the changes introduced to 4.1.5 don't really change anything but make fits and tests more robust: New features A hash of the fktable used in the grid is saved as a .txt file for future reference #2476 New datasets #2468 #2411 Check md5 integrity against vp-setupfit in n3fit #2462 Added HT theory covmat #2126 A new 4.1 runcard has been aded #2477 Bugfixes Dataset bugfixes #2465 #2469 #2475 #2489 #2491 Pseudodata in the diagonal basis is now saved correctly #2455 Removed graveyard of functions #2472 The commondata test workflow is now slightly more robust #2474 Fit regressions are much more robust now #2480 Shifted plots now appear correctly (failures were due to some of the dataset bugfixes above) #2484 Full Changelog: https://github.com/NNPDF/nnpdf/compare/4.1.4...4.1.5