Still in development. Python scripts in the folder serve as examples to program your own displays.
Some examples:
Using this script to play around. You can load in a larlite
file and an larcv
file.
This script plots most of the reconstruction quantities in the DL reco. chain.
Runs on the merged_dlreco
files produced by the production at FNAL (and now Tufts).
This shows how one setups menu buttons.
Example for plotting CRT.
Follow this page to install plotly dash
.
You also need to build EITHER ubdl
or dllee_unified
version of the DL LEE software.
The ubdl
repo.
The dllee_unified
repo.
First log into uboonebuild02
with port forwarding (setup using the -L
flag in the command below).
Right now this is the only machine that will work, as it runs SL7.
In the future, will make an SL6 solution for the UB gpvms.
ssh -XY -L 8005:localhost:8005 [email protected]
Note that the port I use, 8005
, can be changed. And only one person can use a port at a time.
When you log in and see a message of a port already in use, try another one. By convention use something above and around 8000
.
You only need to do this once
source /cvmfs/uboone.opensciencegrid.org/products/setup_uboone.sh
setup dllee_unified v1_0_3 -q e17:prof
pip install --user plotly
pip install --user dash==1.8.0
The --user
is important. So is setting up dllee_unified
before running pip
.
This makes sure you are using the right version of python
/pip
.
Now get the lardly
code.
git clone https://github.com/larbys/lardly
Note, you always have to setup port forwarding (see above) and you have to setup cvmfs
everytime you start a new login.
source /cvmfs/uboone.opensciencegrid.org/products/setup_uboone.sh
You also have to setup dllee_unified
for each new login as well.
setup dllee_unified v1_0_3 -q e17:prof
Now go back to the lardly
folder you cloned.
The you can start an event viewer.
python test_dlmerged.py --input-file [dlmerged root file] -e 0 -p 8005
Note the port you used. dash
will use 8050
by default, so you need to set the port to the one you used in your ssh
command.
Then open a browser on your local machine (e.g. laptop) and go to https://localhost:8005
. Again, use your port.
The event should load -- it'll take awhile as the image data is large. But once loaded, navigation "should" be fairly fast.
This is just a development script. Play around with it! There are flags to turn certain things on and off. There are other scripts that may or may not work.