From 9aad161711dfe64ad3225dc66233cc4a1ee699e6 Mon Sep 17 00:00:00 2001
From: Laurent Mercadier <laurent.mercadier@xfel.eu>
Date: Wed, 17 Mar 2021 16:20:59 +0100
Subject: [PATCH] Changed first arguments of get_peaks() to accept extra_data
DataCollection
---
src/toolbox_scs/detectors/digitizers.py | 26 +++++++-------
src/toolbox_scs/detectors/xgm.py | 47 ++++++++++++++++---------
2 files changed, 44 insertions(+), 29 deletions(-)
diff --git a/src/toolbox_scs/detectors/digitizers.py b/src/toolbox_scs/detectors/digitizers.py
index 4d28e69..0bb2662 100644
--- a/src/toolbox_scs/detectors/digitizers.py
+++ b/src/toolbox_scs/detectors/digitizers.py
@@ -92,8 +92,7 @@ def peaks_from_raw_trace(traces,
return results
-def get_peaks(proposal,
- runNB,
+def get_peaks(run,
data=None,
source=None,
key=None,
@@ -110,10 +109,8 @@ def get_peaks(proposal,
Parameters
----------
- proposal: int or str
- proposal number
- runNB: int or str
- run number
+ run: extra_data.DataCollection
+ DataCollection containing the digitizer data
data: xarray DataArray or str
array containing the raw traces or peak-integrated values from the
digitizer. If str, must be one of the ToolBox mnemonics. If None,
@@ -145,7 +142,8 @@ def get_peaks(proposal,
pattern can be provided, e.g. {'source':'SCS_RR_UTC/TSYS/TIMESERVER',
'key':'bunchPatternTable.value', 'pattern':'sase3'}
extra_dim: str
- Name given to the dimension along the peaks.
+ Name given to the dimension along the peaks. If None, the name is given
+ according to the bunchPattern.
indices: array, slice
indices from the peak-integrated data to retrieve. Only required
when bunch pattern is missing and useRaw is False.
@@ -158,7 +156,6 @@ def get_peaks(proposal,
raise ValueError('At least data or source + key arguments ' +
'are required.')
# load data
- run = ed.open_run(proposal, runNB)
if data is None:
arr = run.get_array(source, key)
elif type(data) is str:
@@ -273,7 +270,7 @@ def get_peaks(proposal,
if autoFind:
integParams = find_integ_params(arr.mean(dim='trainId'),
min_distance=min_distance)
- log.debug('Auto find peaks result:', integParams)
+ log.debug(f'Auto find peaks result: {integParams}')
# 2.1. No bunch pattern provided
if bpt is None:
@@ -361,7 +358,7 @@ def find_integ_params(trace, min_distance=1, height=None, width=1):
intstop = np.round(peaks['right_ips'][0]
+ 0.5*np.mean(peaks['widths'])).astype(int)
bkgstop = intstart - int(0.5*np.mean(peaks['widths']))
- bkgstart = bkgstop - 100
+ bkgstart = np.max([bkgstop - min_distance, 0])
result = {'intstart':intstart, 'intstop':intstop,
'bkgstart':bkgstart, 'bkgstop':bkgstop,
'period':period, 'npulses':npulses}
@@ -380,7 +377,8 @@ def get_tim_peaks(data, bunchPattern='sase3',
data: xarray Dataset containing TIM data
bunchPattern: str
'sase1' or 'sase3' or 'scs_ppl', bunch pattern
- used to extract peaks.
+ used to extract peaks. The pulse ID dimension will be named
+ 'sa1_pId', 'sa3_pId' or 'ol_pId', respectively.
integParams: dict
dictionnary for raw trace integration, e.g.
{'intstart':100, 'intstop':200, 'bkgstart':50,
@@ -401,6 +399,7 @@ def get_tim_peaks(data, bunchPattern='sase3',
proposal = data.attrs['runFolder'].split('/')[-3]
runNB = data.attrs['runFolder'].split('/')[-1]
+ run = ed.open_run(proposal, runNB)
peakDict = {}
toRemove = []
autoFind=True
@@ -416,7 +415,7 @@ def get_tim_peaks(data, bunchPattern='sase3',
continue
log.debug(f'Retrieving TIM peaks from {key}...')
mnemonic = _mnemonics[key]
- peaks = get_peaks(proposal, runNB, data[key],
+ peaks = get_peaks(run, data[key],
source=mnemonic['source'],
key=mnemonic['key'],
digitizer='ADQ412',
@@ -467,6 +466,7 @@ def get_fast_adc_peaks(data, bunchPattern='scs_ppl',
proposal = data.attrs['runFolder'].split('/')[-3]
runNB = data.attrs['runFolder'].split('/')[-1]
+ run = ed.open_run(proposal, runNB)
peakDict = {}
toRemove = []
autoFind=True
@@ -482,7 +482,7 @@ def get_fast_adc_peaks(data, bunchPattern='scs_ppl',
continue
log.debug('Retrieving Fast ADC peaks from {key}...')
mnemonic = _mnemonics[key]
- peaks = get_peaks(proposal, runNB, data[key],
+ peaks = get_peaks(run, data[key],
source=mnemonic['source'],
key=mnemonic['key'],
digitizer='FastADC',
diff --git a/src/toolbox_scs/detectors/xgm.py b/src/toolbox_scs/detectors/xgm.py
index 76e0215..c0869d6 100644
--- a/src/toolbox_scs/detectors/xgm.py
+++ b/src/toolbox_scs/detectors/xgm.py
@@ -23,11 +23,11 @@ log = logging.getLogger(__name__)
def load_xgm(run, xgm_mnemonic='SCS_SA3'):
"""
- Loads XGM data from karabo_data.DataCollection
+ Loads XGM data from extra_data.DataCollection
Parameters
----------
- run: karabo_data.DataCollection
+ run: extra_data.DataCollection
Returns
-------
@@ -71,24 +71,35 @@ def cleanXGMdata(data, npulses=None, sase3First=True, keepAllSase=False):
dimension names 'sa1_pId' and 'sa3_pId'
'''
if 'XGMbunchId' not in data.dims:
- log.debug('Dataset does not contain XGMbunchId dimension. Aborting.')
+ log.info('Dataset does not contain XGMbunchId '+
+ 'dimension. Aborting.')
return data
proposal = data.attrs['runFolder'].split('/')[-3]
runNB = data.attrs['runFolder'].split('/')[-1]
run = ed.open_run(proposal, runNB)
load_sa1 = True
+ load_sa3 = True
#check bunch pattern table
if _mnemonics['bunchPatternTable']['source'] in run.all_sources:
- bpt = run.get_array(*_mnemonics['bunchPatternTable'].values())
+ if 'bunchPatternTable' in data:
+ bpt = data['bunchPatternTable']
+ else:
+ bpt = run.get_array(*_mnemonics['bunchPatternTable'].values())
mask_sa1 = is_sase_1(bpt).sel(trainId=data.trainId)
mask_sa3 = is_sase_3(bpt).sel(trainId=data.trainId)
- valid_tid_sa1 = mask_sa1.where(mask_sa1.sum(dim='pulse_slot')>0,
+ tid_sa1 = mask_sa1.where(mask_sa1.sum(dim='pulse_slot')>0,
drop=True).trainId
- valid_tid_sa3 = mask_sa3.where(mask_sa3.sum(dim='pulse_slot')>0,
+ tid_sa3 = mask_sa3.where(mask_sa3.sum(dim='pulse_slot')>0,
drop=True).trainId
- if valid_tid_sa1.size==0:
- load_sa1 = False
- if np.intersect1d(valid_tid_sa1, valid_tid_sa3,
+ if tid_sa1.size==0 or tid_sa3.size==0:
+ keepAllSase = False
+ if tid_sa1.size==0:
+ log.info('No SASE 1 pulses in the data.')
+ load_sa1 = False
+ if tid_sa3.size==0:
+ log.info('No SASE 3 pulses in the data.')
+ load_sa3 = False
+ elif np.intersect1d(tid_sa1, tid_sa3,
assume_unique=True).size==0:
load_sa1 = keepAllSase
if keepAllSase:
@@ -98,7 +109,7 @@ def cleanXGMdata(data, npulses=None, sase3First=True, keepAllSase=False):
log.info('Dedicated trains, only loading SASE 3.')
isBunchPattern = True
else:
- print('Missing bunch pattern info!')
+ log.info('Missing bunch pattern info!')
bpt = None
isBunchPattern = False
dropList = []
@@ -117,7 +128,7 @@ def cleanXGMdata(data, npulses=None, sase3First=True, keepAllSase=False):
for key in keys_all_sase:
if isBunchPattern is False:
if npulses is None:
- raise TypeError('npulses argument is required '
+ raise TypeError('npulses argument is required '+
'when retrieving non-dedicated XGM arrays '+
'without bunch pattern info. Consider using '+
'SASE 1 or SASE 3 dedicated data arrays.')
@@ -137,23 +148,27 @@ def cleanXGMdata(data, npulses=None, sase3First=True, keepAllSase=False):
{'XGMbunchId':'sa1_pId'})
sa1 = sa1.rename(key.replace('XGM', 'SA1'))
mergeList.append(sa1)
- sa3 = selectSASEinXGM(data, xgm=key, sase='sase3',
+ if load_sa3:
+ sa3 = selectSASEinXGM(data, xgm=key, sase='sase3',
bpt=bpt, npulses=npulses,
sase3First=sase3First).rename(
{'XGMbunchId':'sa3_pId'})
- sa3 = sa3.rename(key.replace('XGM', 'SA3'))
- mergeList.append(sa3)
+ sa3 = sa3.rename(key.replace('XGM', 'SA3'))
+ mergeList.append(sa3)
dropList.append(key)
#dedicated sase data
for key in keys_dedicated:
if "sa3" in key.lower():
+ if load_sa3 is False:
+ dropList.append(key)
+ continue
sase = 'sa3_'
else:
- sase = 'sa1_'
if load_sa1 is False:
dropList.append(key)
continue
+ sase = 'sa1_'
res = data[key].where(data[key] != 1.0, drop=True).rename(
{'XGMbunchId':'{}pId'.format(sase)}).rename(key)
dropList.append(key)
@@ -179,7 +194,7 @@ def cleanXGMdata(data, npulses=None, sase3First=True, keepAllSase=False):
2700-ds[key].sizes[sase])))).flatten()
xgm_flat_arg = np.argwhere(xgm_flat!=1.0)
if(xgm_flat_arg.shape != mask_flat_arg.shape):
- log.warning('The XGM data does not match the
+ log.warning('The XGM data does not match the '+
'bunch pattern! Cannot assign pulse '+
'id coordinates.')
break
--
GitLab