Conversation with Matt/Ric/Riccardo/cpo on April 4, 2024. For complex area detectors.
WE CAN REQUEST CHANGES FROM TID
You could consider running Ric's scripts and create a dummy epixM in
configdb (don't touch the official one).
In epixuhr_config_store can we create a Root object and iterate over
it to create the json? Worth exploring. Might require the hardware
to exist? Or we could use "simulation" somehow?
**********************************************************************
- look in yaml files
- translate those yaml files into epixuhr_config_store.py
(do with python?)
o another possibility: rogue command to dump out the registers?
pick out the read-write ones. might also include the type info
- all these registers should go into the "expert" section
o there is also a "user" section with a simplified user variables
which get "translated" by code into the official "expert" structure.
Ric says there are "helper" variables in the user section,
e.g. charge-injection patterns, that aren't necessarily modified
by the user. The charge-injection script modifies those variables.
Ric says perhaps this should be in a third section? (not user/expert?)
- need to find out the types by looking at devGui or firmware/*.py files
that define devGui
- set most values to 0 then tweak critical ones to non-zero
- a command for the next step:
(first command creates confidb schema)
o python epixm320_config_store.py --prod --user tstopr --inst tst --alias BEAM --name epixm --segm 0 --id epixm320_serial1234 --dir /cds/home/c/cpo/git/epix-hr-m-320k/
(second command fills in values)
o python epixhr_config_from_yaml.py --prod --user tstopr --inst tst --alias BEAM --name epixm --segm 0 --id epixm320_serial1234 --yaml Root:/cds/home/c/cpo/git/epix-hr-m-320k/software/config/ePixHRM320k_75000018efb4ab01_ASIC_u1.yml
"--dir" points to csv files in the project that are useful?
top = epixm320_cdict(args.dir+'/software/config')
top.setInfo('epixm320hw', args.name, args.segm, args.id, 'No comment')
these are the csv file-loading lines:
top.set(base+'_250_MHz', np.loadtxt(prjCfg+'/EPixHRM320KPllConfig250Mhz.csv', dtype='uint16', delimiter=',', skiprows=1, converters=conv))
top.set(base+'_125_MHz', np.loadtxt(prjCfg+'/EPixHRM320KPllConfig125Mhz.csv', dtype='uint16', delimiter=',', skiprows=1, converters=conv))
csv are passed to the root.init_asic(csv_file_index_into_list_of_csv_files)
run epixhr_config_from_yaml.py with different arguments to update the values
in the schema. epixm has 5 or 6 yaml files.
new yaml files (with a new "schema") can be provided by the TID
group as time goes on. Ric has a mechanism to update the schema
but preserve the values (!) (see https://confluence.slac.stanford.edu/display/PSDMInternal/Debugging+DAQ#DebuggingDAQ-MakingSchemaUpdatesinconfigdb).
Can add/drop/modify fields as necessary with Ric's sheme.
in epixm_config.py reads the configdb and re-creates the .yaml files
in /tmp. Also creates Root.filename<yamlname> variables. yamlname is
the third argument here (PowerSupply):
toYaml('App',['PowerControl'],'PowerSupply')
then this initasic gets the /tmp/yamlname inserting a filename into
a variable of this Root class (editing the class): (filename is not passed in)
(arg has the index that points to an entry in a list that
has clock frequencies, .csv files, etc.)
cbase.fnInitAsicScript(None,None,arg)
goal might be to get the above script to work
in future perhaps consider adding arguments instead of modifying class?
**********************************************************************
copy lines from devGui, e.g.
cbase = ePixM.Root(top_level = '/tmp',
dev = '/dev/datadev_0',
pollEn = False,
initRead = False,
justCtrl = True,
fullRateDataReceiverEn = False)
cbase.saveAddressMap("filename")
parse with python to generate a schema?
Root python class knows the schema, .yaml has the values
Matt's epixhr_config_from_yaml_set.py iterates over multiple yaml files
**********************************************************************
Watch out for:
- enum keywords have [] () that are invalid python names. can request
to have them changed, or need to "translate"
- previously debugging those errors in JsonToXtc was difficult, but
this may be improved with recent changes from Gabriel. Ric thinks
this may not cover much of the phase-space for errors, but we can
consider improving it by talking with Gabriel. e.g. a new
epixm release removed a variable from the yaml, but was in the
schema and yaml. Got a default value of zero and JsonToXtc was unhappy.
Overview
Content Tools