Skip to content

Commit

Permalink
Update documentation
Browse files Browse the repository at this point in the history
  • Loading branch information
makortel committed Jun 16, 2017
1 parent 6011997 commit 3e6dc7e
Showing 1 changed file with 32 additions and 1 deletion.
33 changes: 32 additions & 1 deletion Validation/RecoTrack/README.md
Original file line number Diff line number Diff line change
Expand Up @@ -18,7 +18,13 @@ There is also an ntuple-version of MultiTrackValidator, called
`cmsDriver.py`-generated workflow containing `VALIDATION` sequence by
including
`--customise Validation/RecoTrack/customiseTrackingNtuple.customiseTrackingNtuple`
argument to the `cmsDriver.py`. The customise function disables all
argument to the `cmsDriver.py`, or adding the following near the end
of an existing job configuration file
```
from Validation.RecoTrack.customiseTrackingNtuple import customiseTrackingNtuple
process = customiseTrackingNtuple(process)
```
The customise function disables all
output modules and replaces the validation sequence with a sequence
producing the ntuple in `trackingNtuple.root` file. If ran without
RECO, it needs both RECO and DIGI files as an input.
Expand Down Expand Up @@ -64,6 +70,30 @@ more information see
If the "playback mode" is not enabled, an exception will be thrown in
the C++ code for a missing SimHit.

### Using tracks from a single iteration as an input

The ntuple can be used to study the tracks produced by a single
iteration (without going throught generalTracks) as well. Below is an
example (for highPtTripletStep) what modifications need to be done in
the job configuration file (after the `customiseTrackingNtuple()` call)
```
process.trackingNtuple.tracks = "highPtTripletStepTracks"
# following are needed only if _includeSeeds is True
process.trackingNtuple.seedTracks = ["seedTrackshighPtTripletStepSeeds"]
process.trackingNtuple.trackCandidates = ["highPtTripletStepTrackCandidates"]
# following is needed only if _includeMVA is True
# it also shows how to add MVA output of multiple classifiers
process.trackingNtuple.trackMVAs = ["highPtTripletStepClassifier1",
"highPtTripletStepClassifier2"]
# Especially for iteration-specific MVA studies, it is important to
# use firstStepPrimaryVertices instead of offlinePrimaryVertices as
# that is the collection used in the MVA input variable calculations
process.trackingNtuple.vertices = "firstStepPrimaryVertices"
```

### Applications

Use `--help` to check out the parameters.
Expand All @@ -77,6 +107,7 @@ Use `--help` to check out the parameters.
* [`analyseDuplicateFake.py`](test/analyseDuplicateFake.py) examples of printouts
* [`fakeAnalysis/main.py`](test/fakeAnalysis/main.py) complete analysis code for fake tracks


### Caveats

#### RecHit <-> SimHit links
Expand Down

0 comments on commit 3e6dc7e

Please sign in to comment.