Conventions:
- The "python" command on your lxplus should be python 2.7
- An example script is given in mystudy.py the important points are:
- from python.batching import Settings, submit_job
- create a Settings object, by calling Settings with your project name as an argument. Adjust the settings from default as needed.
- submit your study to the cluster with submit_job(settings)
- By default output is written to /eos/USER_INITIAL/USERNAME/madxBatchData/PROJECTNAME, but a different disk can be chosen when creating the Settings object with the optional argument disk=...:
- 'afsprivate' : /afs/cern.ch/work/USER_INITIAL/USERNAME/private/madxBatchData/PROJECTNAME
- 'afspublic' : /afs/cern.ch/work/USER_INITIAL/USERNAME/public/madxBatchData/PROJECTNAME
- Aditionally you can override this directory by using settings.datadir=DESIRED_OUTPUT_DIR in your script after the creation of the settings object, but be carefull that the directory will no longer automatically change with set_name if you do that!
- By default pycollimate is on, it can be turned off by setting settings.pycollimate=False in your script
- By default the madxBatch folder and the pycollimate folder are assumed to be located in the same folder. This can be changed by setting settings.pycolldir=PYCOLL_DIR in your script.
Other reminders:
- If pycollimate is turned off, madx_dev is used. This means that in case of a large MAD-X update the code could be broken until it is updated.