To build the library and the vtxeval executable follow the instructions in the project's top level ../README.md "README.md". Since in this study we are focusing on evaluation of the vertex finders one can build and install only the relevant libraries:

$ mkdir build/ && cd build/
$ cmake -D CMAKE_INSTALL_PREFIX=../build-install \
        -D BOOST_ROOT=$OPTSTAR \
        -D CMAKE_BUILD_TYPE=Release ../
$ make -j4 star-vertex
$ make install

How to create embedding samples

Use STAR's get_file_list.pl to select a set of desired zerobias files from the database. Here is an example command with some selection constraints:

$ get_file_list.pl -keys path,filename,events \
  -cond filename~st_zerobias_%,runnumber[]14115007-14120000,sanity=1,filetype=online_daq \
  -delim / -limit 0

You may choose to copy the files from hpss to some location on disk. In that case, the paths to files returned by get_file_list.pl can be adjusted to point to the actual file location. The final list of zerobias data files has to be in the following format:

string:"path/to/file.daq run_id num_of_events_to_process"

The contents of your file may look like the following:

$ cat filelist_zerobias.lis
...
string:"/star/data05/daq/2013/embedding/st_zerobias_adc_14115007_raw_1540001.daq 14115007 123"
string:"/star/data05/daq/2013/embedding/st_zerobias_adc_14115007_raw_2440001.daq 14115007 155"
...
string:"/star/data05/daq/2013/embedding/st_zerobias_adc_14115007_raw_3330001.daq 14115007 111"
...

W boson

The script to submit W embedding jobs to condor is called submit_jobs_embedding_wbos.sh and can be found in star-travex/supple/. There is a number of named parameters which user may need to set to appropriate values. The two parameters likely to be changed often are OUT_DIR and INPUT_FILE_LIST. All named variables can be passed to the script in command line like this:

$ OUT_DIR=/tmp/scratch/wbos_embed_results/ \
  INPUT_FILE_LIST=~/star-travex/supple/filelist_zerobias.lis \
  ~/star-travex/supple/submit_jobs_embedding_wbos.sh

Lambda

A command submitting embedding jobs can look like this:

$ FZD_DIR=~/scratch/lambda_fzd/ \
  OUT_DIR=~/scratch/lambda_embed_results/ \
  INPUT_FILE_LIST=~/star-travex/supple/filelist_zerobias_lambda.lis \
  ~/star-travex/supple/submit_jobs_embedding_generic.sh

J/psi

Create a directory from where condor jobs will be submitted and run the script as:

mkdir submit_jpsi && cd submit_jpsi
/path/to/star-travex/supple/submit_jobs_embedding_jpsi.sh 200

If the package was not checked out in the default location of ${HOME}/star-travex/ then you should specify the correct location as:

SOURCE_DIR=/path/to/star-travex/ \
OUTPUT_DIR=/tmp/jpsi_embed \
/path/to/star-travex/supple/submit_jobs_embedding_jpsi.sh 200

There is a number of other script parameters which can be modified if the default values are not appropriate. Search for "NAMED ARGUMENTS" to find these values.

Study the ranking scheme in STAR (deprecated)

This section is outdated but left here for the reference until after all mentioned macros are removed from the repository...

1.- The macro VertexRank.C will read the MuDst sample and create a tree with some vertex information. To run (e.g. 100 events and output file outvertex.root):

$root -l -b -q lMuDst.C 'VertexRank.C(100, "path/to/*.MuDst.root", "outvertex")'

2.- Run TMVA to classify signal and background. A separate set of signal and background must be given, with a define set of variables for the training. Also need to specify the classifier. To run (e.g. with classifiers Fisher and Likelihood):

$root TMVAClassification.C\(\"Fisher,Likelihood\"\)

3.- Apply the results of the classification with ApplyClassifier.C. The output will be the same tree for step (1) but with a new branch with the classifier response. Run:

$root ApplyClassifier.C