Make it easier to add many weights in a binned dataframe, e.g. wildcarding

[benkrikler]

At the moment, each weight needs to be specified in advance to build the output dataframe. This makes the configuration files very long in many cases, and can lead to bugs if an important weight is missed out by mistake. It would be good to add a way to specify an expression which automatically identifies which branches should be used for weights, e.g. via a wildcard.

Risks / challenges

• What is the right way to include and exclude weights via an expression
• This potentially introduces a new "mini-language" if we need to be able to match some weights but not others
• When is this list expanded?
  • If it is expanded at run-time, this breaks the static DAG workflow model
  • If it is expanded at launch-time, this requires a possible pre-expansion step and can only be done if 1) all files have the same branches, and 2) all branches are known in advance.