higgs_dna.submission package
Submodules
higgs_dna.submission.lxplus module
- class higgs_dna.submission.lxplus.LXPlusVanillaSubmitter(analysis_name, analysis_dict, original_analysis_path, sample_dict, args_string, queue='longlunch', memory='10GB', files_per_job=1, cluster_per_sample=True)[source]
Bases:
objectA class for submitting jobs on the CERN’s LXPlus cluster using HTCondor, one job per file in a sample list of an analysis. All jobs for a given samples are submitted to the same cluster. The constructor creates a directory .higgs_dna_vanilla_lxplus if it does not exist and another one called .higgs_dna_vanilla_lxplus/<analysis_name>. The data and time (YMD_HMS) is appended to the end of <analysis_name> to avoid overwriting previous submissions. Inside this directory two subdirectories called <inputs> and <jobs> will be created. In the former the split JSON files will be stored, in the latter the HTCondor related job files will be stored. To send each job to a separate cluster then set cluster_per_sample=False in the class constructor.
- Parameters:
- param analysis_name:
Name of the analysis.
- type analysis_name:
str
- param analysis_dict:
Dictionary containing the parameters of the analysis.
- type analysis_dict:
dict
- param original_analysis_path:
Path of the original analysis to be replaced with the new ones.
- type original_analysis_path:
str
- param sample_dict:
Dictionary containing the samples and their respective files.
- type sample_dict:
dict
- param args_string:
String containing the command line arguments.
- type args_string:
str
- param queue:
HTCondor queue to submit the job to. Defaults to “longlunch”.
- type queue:
str, optional
- param memory:
Memory request for the job. Defaults to “10GB”.
- type memory:
str, optional
- param cluster_per_sample:
Option to submit each job from a given sample to the same cluster.
- type cluster_per_sample:
bool, optional