This repository holds tools for centralized updating and fetching of cross sections at MIT. Analyses throughout CMS have a variety of formats for storing their Monte Carlos cross sections. This is an attempt to store these cross sections in a clean, machine-friendly, and documented way.
To access the Python tools, place CrossSecDB/python in your $PYTHONPATH.
To access CLIs provided, place CrossSecDB/bin in your $PATH.
In addition, the MySQLdb Python package must be installed on the machine you are using.
If it's not installed, you can always add it with:
pip install -r CrossSecDB/requirements.txt
If you are on the Tier-3, MySQLdb should already be installed on your machine.
You can easily add an existing installation to your path by running the setup.sh inside the location.
Note, all the python executables use the system Python, /usr/bin/python, in the shebang.
This way, you will be able to access the command line tools while using CMSSW.
However, using the Python modules within your own script that is using the CMSSW may fail since CMSSW does not have MySQLdb installed.
For reading the database, make the following replacement in the examples:
from CrossSecDB.reader import get_xsec with from CrossSecDB.reader_cmssw import get_xsec
The main motivation for this repository is to provide an exceptionally lazy tool to find cross sections.
For example, if I don't know the most recent cross section for WJetsToLNu_TuneCUETP8M1_13TeV-amcatnloFXFX-pythia8,
instead of clicking around Sid's GitHub repos, I would like to type:
$ get_xs.py WJetsToLNu_TuneCUETP8M1_13TeV-amcatnloFXFX-pythia8
61527.0
(Result is according to PandaCore at the time of writing. Also, all units are assumed to be pb.)
TODO: Add optional unit definition for input and output.
In addition to a Python Command Line Interface, functions in Python modules are provided for use. A curl interface is also available for anyone who wants their analysis code to access information from outside the Tier-3.
A C++ header interface is planned as well for anyone who wants their analysis code to access information on the fly.
Cross sections can be retrieved from the central database very easily once this package is installed:
from CrossSecDB.reader import get_xsec
print get_xsec('WJetsToLNu_TuneCUETP8M1_13TeV-amcatnloFXFX-pythia8')
print get_xsec(['WJetsToLNu_TuneCUETP8M1_13TeV-amcatnloFXFX-pythia8', 'ST_tW_top_5f_inclusiveDecays_13TeV-powheg-pythia8_TuneCUETP8M1'])
This would give an output like:
61527.0
[61527.0, 35.85]
You can always
print get_xsec.__doc__
for addition, up to date documentation.
For those that like dumping things with system calls or just checking interactively, a command line interface is also available. It's used just as desired in the motivation section above:
$ get_xs.py WJetsToLNu_TuneCUETP8M1_13TeV-amcatnloFXFX-pythia8
61527.0
$ get_xs.py WJetsToLNu_TuneCUETP8M1_13TeV-amcatnloFXFX-pythia8 ST_tW_top_5f_inclusiveDecays_13TeV-powheg-pythia8_TuneCUETP8M1
61527.0
35.85
More usage information (like how to access alternate energies) can be gathered by
calling the script without any arguments or with -h or --help as the first argument.
Centralized cross sections can also be accessed from anywhere over the internet. A webpage is available at the link at the top of the central repository.
In addition, a machine-friendly curl interface has been implimented.
This is used the same way as get_xs.py:
$ web_get_xs.sh WJetsToLNu_TuneCUETP8M1_13TeV-amcatnloFXFX-pythia8
61527
$ web_get_xs.sh WJetsToLNu_TuneCUETP8M1_13TeV-amcatnloFXFX-pythia8 ST_tW_top_5f_inclusiveDecays_13TeV-powheg-pythia8_TuneCUETP8M1
61527
35.85
Note that when compared to get_xs.py, this interface truncates trailing 0s after the decimal.
More usage information (like how to access alternate energies) can be gathered by
calling the script without any arguments or with -h or --help as the first argument.
TODO: Create C++ header and tests
To encourage contribution to the centralized database, scripts for sending updates are also fun and easy to use. These are provided only through Python interfaces. Using the Python module would look something like this:
from CrossSecDB.inserter import put_xsec
print put_xsec.__doc__
put_xsec('WJetsToLNu_TuneCUETP8M1_13TeV-amcatnloFXFX-pythia8', 61527.0, source='README from this repo')
From the docstring, you might learn that parallel lists work as well:
import process
from CrossSecDB.inserter import put_xsec
samples = []
xs = []
for key, val in process.processes.items():
if 'mad' not in key and 'pythia' not in key:
continue
samples.append(key)
xs.append(val[2])
put_xsec(samples, xs, '/home/dhsu/CMSSW_8_0_26_patch1/src/PandaCore/Tools/python/processes.py')
Easy!
There's also a command line interface that can be used the following way:
put_xs.py "Source is README from this repo" WJetsToLNu_TuneCUETP8M1_13TeV-amcatnloFXFX-pythia8 61527.0
More usage information (like how to access alternate energies) can be gathered by
calling the script without any arguments or with -h or --help as the first argument.
Cross sections can also be "updated" by reverting to old cross sections.
Use revert_xs.py and follow the instructions on the screen.
Here are some examples on how to call it.
revert_xs.py WJetsToLNu_TuneCUETP8M1_13TeV-amcatnloFXFX-pythia8
ENERGY=8 revert_xs.py WJetsToLNu_TuneCUETP8M1_13TeV-amcatnloFXFX-pythia8
revert_xs.py WJetsToLNu_TuneCUETP8M1_13TeV-amcatnloFXFX-pythia8 ST_tW_top_5f_inclusiveDecays_13TeV-powheg-pythia8_TuneCUETP8M1
revert_xs.py --like 'ST_%'
Immediate improvements should be found the following way:
grep -R "TODO" .
Try to format your comments accordingly. Please write tests to prove that what you wrote works.