#install and initialize geant4 with -DGEANT4_USE_OPENGL_X11=ON -DGEANT4_USE_QT=ON
git clone https://github.com/rdom/prttools
git clone https://github.com/rdom/eicdirc
cd eicdirc
mkdir build
cd build
cmake ..
make -j4
#neural network
install tensorflow libs, cppflow and run:
cmake -DAI=1 ..
#test event display
./eicdirc -r 0 -theta 30 -x "pi+" -p 6.0 -w 0 -h 11 -c 2031 -l 3 -v 0 -gz 1 -g 1 -ev 0 -e 1
eicdirc [OPTION] [ARGUMENT] ... [OPTION] [ARGUMENT]
example:
./eicdirc -r 0 -theta 30 -x "pi+" -p 6.0 -w 0 -h 11 -c 2031 -l 3 -v 0 -gz 1 -g 1 -ev 0 -e 1
-o output file name
-i input file name
-u look-up file name
-pdf PDFs file name
-nn path to the neural network
-r run type
0 simulation
1 look-up table generation
2 geometrical reconstruction and time imaging
3 likelihood calculation
4 create pdf
5 simulate pdf
20 2 with Cherenkov ring fit
-field field type
0 no field (default)
1 CORE b-field
2 ePIC MARCO 1.7T
3 ePIC MARCO 2.0T
4 solenoidal 3.0T
-g geometry configuration
0 generic
1 ePIC one barbox (default)
2 CORE one barbox
10 generic whole DIRC
11 ePIC whole DIRC
12 CORE whole DIRC
-ev expansion volume type
0 prism with lenses (default)
1 BaBar wedge with focusing prism
3 prism with plate, lens between bars and plate
4 tilted EV
5 prism with plate, lens between plate and EV
6 split prism
7 prism with thicker plate (shifted to center), lens between bars and plate
8 split prism with split plate, lens between plate and EV
9 split prism with split plate, lens between bars and plate
-h number of bars in one radiator box (use with "-g 0")
-c PMT layout
0 4x6 standard MCPs (pixel size == mcp size)
1 4x6 standard MCPs (6.4x6.4 pixels)
3 one MCP covers whole PD plain (3 mm x 3 mm pixels)
4 2x3 LAPD
2031 4x6 MCPs with compact packing and 32x32 pixels (2 mm x 2 mm) (default)
-l focusing system
0 no lens
3 3-layer spherical lens (default)
6 3-layer cylindrical lens
10 ideal lens (thickness = 0, ideal focusing)
-theta polar between particle beam and bar radiator [deg]
25 default
if theta == 0 then theta = [30,160]
-phi azimuth angle between particle beam and bar radiator [deg]
0 default
use [990, 999] (corresponds to bar # [0,9]) to choose which bar should be hit (works also with -field 2, for > ~4 GeV/c)
same use of [990, 999] for LUT generation
-e number of simulated events
-x particle type
"pi+" (default)
"proton"
"kaon+"
...
"opticalphoton"
"mix_pie" 1 pion 1 electron mix
"mix_pimu" 1 pion 1 muon mix
"mix_pik" 1 pion 1 kaon mix
"mix_pip" 1 pion 1 proton mix
"mix_kp" 1 kaon 1 proton mix
-p particle momentum [GeV/c]
-w physical list
0 standard EM (default)
1 without multiple scattering and bremsstrahlung
2 standard EM with HAD (elastic/inelastic/absorption)
3 FTFP_BERT
10 monochromatic Cherenkov light
11 10 + 1
-seed seed number for the random generator
-b batch mode
1 run silent (without GUI)
-d display option
don't use this argument, instead use /Prt/geom/drawHits 2
0 standard (default)
1 display hit occupancy of current run
2 display hit occupancy of occuhits.root (needs to be generated)
-timeres time resolution [ns]
0.1 (default)
-timecut time cut constant [ns]
0.5 (default)
-trackingres tracking resolution [rad], applied at tracking layer
0.0005 (default)
100 use realistic, mapped values
-cor per-pmt correction
0 without correction
1 apply Cherenkov angle correction
2 apply Cherenkov angle and time corrections (default)
5 (re)create corrections
-gx target's X offset
0 default
-gz target's Z offset
0 default
-dn dark noise per PMT
amount of dark noise hits per-pmt [Hz]
-t1 test variable
[double]
-t2 test variable
[double]
-t3 test variable
[double]
cd eicdirc/build
mkdir data
./eicdirc -o data/lut.root -r 1 -g 1 -c 2031 -l 3 -v 0 -ev 0 -x "opticalphoton" -p "3.18 eV" -e 1000000 -b 1
Visualization of 100 events:
./eicdirc -o data/tmp.root -r 1 -g 1 -c 2031 -l 3 -v 0 -ev 0 -x "opticalphoton" -p "3.18 eV" -e 100
LUT averaging:
cd eicdirc/macro
root -q -b lutmean.C'("../build/data/lut.root")'
For geometrical reconstruction one needs to generate ~2000 events (example for 6 GeV/c pion/kaon mix @ 30 degree polar angle):
cd eicdirc/build
./eicdirc -r 0 -o data/sim.root -theta 30 -x "mix_pik" -p 6 -w 0 -g 1 -c 2031 -l 3 -trackingres 0.0005 -e 2000 -b 1
Geometrical reconstruction requires LUT ("lut.avr.root" which was previously created):
./eicdirc -r 2 -i data/sim.root -u data/lut.avr.root -o data/reco.root -trackingres 0.0005 -timeres 0.1 -timecut 0.2 -e 2000 -v 2
after first execution it will create per-pmt corrections "sim.root_corr.root" which will be automatically applied during next executions (use same command):
./eicdirc -r 2 -i data/sim.root -u data/lut.avr.root -o data/reco.root -trackingres 0.0005 -timeres 0.1 -timecut 0.2 -e 2000 -v 3
as a result the program will create figures of relevant parameters (in "data/reco") and store PID information in "data/reco.root"
For time imaging one needs to generate ~25000 events.
./eicdirc -r 0 -o data/sim.root -theta 30 -x "mix_pik" -p 6 -w 0 -g 1 -c 2031 -l 3 -trackingres 0.0005 -e 25000 -b 1
the last 20k events [5000,20000] will be used to generate Probability Density Functions (needed for time imaging):
./eicdirc -r 4 -i data/sim.root -u data/lut.avr.root -trackingres 0.0005 -timeres 0.1 -timecut 0.2 -e 5000
it will create "data/sim.pdf.root". Note: the start event number for PDFs (5000) is hard coded at the moment.
The time imaging reconstruction:
./eicdirc -r 2 -i data/sim.root -u data/lut.avr.root -pdf data/sim.pdf.root -o data/reco.root -trackingres 0.0005 -timeres 0.1 -timecut 0.2 -e 5000 -v 3
Hit pattern:
cd eicdirc/macro
root loadlib.C draw_hp.C'("../build/data/sim.root")'
Example of 1k of 6 GeV/c pions @ 30 degree:
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent