Migration from Rivet v2 to Rivet v3

Rivet 3.x is a major new release series, in which we’ve made some big changes that affect the programming interface (or “API”) exposed to analysis authors. Fear not: the changes are simple, and all beneficial to routine writers… it’s just the need for transition that’s a bit unfortunate. Here is a very quick guide for updating your analysis code from v2 to v3:

• The addProjection(...) function no longer exists: replace uses of this with declare(...).
• Similarly, applyProjection() should now be apply(). Fewer characters to type: yay!
• There is also a change in the histogram-booking syntax from e.g. _h_xxx = bookHisto1D(...) to book(_h_xxx, ...)
• You should no longer attempt to access “the” event weight, since that’s now done automatically and for all weights in the systematics vector: _h_xxx->fill(x, event.weight()) becomes _h_xxx->fill(x).

We recommend that anyone writing a new analysis should start by running the rivet-mkanalysis script rather than directly copying an old analysis code: we update the templates generated by the script as new API features and ideas occur, and it always represents best-practice. You can still copy-paste chunks from your old analysis as you need them, modulo the unavoidable API changes above.