How to add UserData to the output D3PD: adding new branches and filling them with user-defined custom formulas


We provide below two examples of how to include UsedData, i.e. user-defined quantities, in the output D3PD.

Adding UserData to the D3PD means defining two new branches and filling them with user-defined formulas.

ALERT! It does not mean dumping information from AOD, for example. For that you should refere to ATLASWatchManHowToDumpInfoFromContainers

How to add UserData: examples

In our example we define two new branches, to store Meff (Effective Mass) and Sphericity.

The first one is filled with a user-defined custom formula, while the latter is filled with a buil-in formula. (You can find the formula in ATLASWatchMan/python/

userD3PDBranchesToFill = {
                          'meffCSC4j0lep' : {'label': 'meff4j', 'type': 'float', 'formula': 'meff'},
                          'sphericityCSC' : {'label': 'sphCSC', 'type': 'float', 'formula': 'sphericity'},

In userD3PDBranchesToFill we set our definitions for the new branches. In 'formula' we specify the name of the formula. That name will be searched first in the userFormula dictionary (details below) and then, if not found, in the built-in library of pre-defined furmulas (you can find the formulas in ATLASWatchMan/python/

So the branch 'sphericityCSC' will be filled with the formula 'sphericity' from the built-in library, while 'meffCSC4j0lep' in this example is a user-defined formula and it has been defined like here below, in the userFormula dictionary:

userFormula = {
meff = 0.
for i,jet in enumerate(candidates['jet']):
    meff += candidates['jet'][i].pt()
meff += candidates['met'].et()
return meff

The triple quotes """   """ are the Python character to declare an indented multi-lines string.

So you can actually write your own formula in a very simple way, and you can use the candidates dict to access particles, or everything you put in collections dictionary setting them as 'select': True (for details see ).

For example to access the jet.Pt() of the 3rd jet, in your formula you can use candidates['jet'][2].pt() , and the same for met.Phi(): candidates['met'].phi().

Every branch you specify here in userD3PDBranchesToFill will be written in the output D3PD as two branches, in our example:

  • meffCSC4j0lepValues

  • meffCSC4j0lepChannels

The first one storing the actual values from your computation, and the latter storing a vector of strings which say to you, for each value, wich sets of candidates have been used to compute that value.

This because each channel and/or analysis you define in your steering file can have a custom object selection and/or overlap removal, and so each channel with custom selection cuts "owns" a different set of candidates, depending of those having passed the custom cuts.

And so the quantity you define as UserData is computed with all these sets of candidates, the different values are stored in Values branch, and the actual channels name whose set of candidates was used to compute that value is stored in Channels branch.

So in the end you have only to plot meffCSC4jolepValues branch, asking if 'myChannel' is in meffCSC4j0lepChannels[i] as we explain here: ATLASWatchManReadD3PD

-- RiccardoMariaBianchi - 11 Mar 2009

Edit | Attach | Watch | Print version | History: r2 < r1 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r2 - 2009-03-23 - RiccardoMariaBianchi
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    Main All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2021 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
or Ideas, requests, problems regarding TWiki? use Discourse or Send feedback