Models
This app only provides the HistogramDataFile
model.
HistogramDataFile
This is a model which describes some generic attributes of the histogram data files, such as:
- The
filepath
where they're located. - Their
filesize
. - Their
contents
, which is aManyToMany
field to theHistogramDataFileContents
table. A single DQM file may contain mutltiple types of histograms (e.g.LumisectionHistograms1D
andLumisectionHistograms2D
).
The DATAFILE_FORMAT_CHOICES
class attribute is also defined, which is used
by the discover_dqm_files
management command.
This management command is, currently, the only way to create HistogramDataFile
entries in the database.
Currently, the choices for available files provided are only refreshed
every time the discover_dqm_files
management command is run. To run it, login to
PaaS, select the ml4dqm-playground
project, go to Administrator
->Pods
, select
the currently running pod, go to Terminal
and run python manage.py discover_dqm_files
Warning
The HistogramDataFile
entries can be deleted without affecting the
Histograms loaded from the deleted files. However, re-reading the
file will NOT update the existing Histogram entries to point to
the newly read file
This is due to the HistogramBase
model having the on_delete=SET_NULL
on
the ForeignKey
to the HistogramDataFile
.
This was chosen so that the large number of extracted lumisections would be
preserved in the DB after deleting an entry to a HistogramDataFile
.
HistogramDataFileContents
- The
granularity
of the data they contain (e.g. Run or Lumisection data) - The
data_dimensionality
of the data contained (1D or 2D).