From 50a65338c645685f8b17c5ddb6fe0cd61f104568 Mon Sep 17 00:00:00 2001
From: Laurent Mercadier <laurent.mercadier@xfel.eu>
Date: Mon, 15 Jul 2024 13:21:57 +0200
Subject: [PATCH] Modify load() to look at both raw and proc folders
---
src/toolbox_scs/load.py | 15 +++++++--------
1 file changed, 7 insertions(+), 8 deletions(-)
diff --git a/src/toolbox_scs/load.py b/src/toolbox_scs/load.py
index 99ef299..3aa0c0d 100644
--- a/src/toolbox_scs/load.py
+++ b/src/toolbox_scs/load.py
@@ -39,7 +39,6 @@ log = logging.getLogger(__name__)
def load(proposalNB=None, runNB=None,
fields=None,
- subFolder='raw',
display=False,
validate=False,
subset=None,
@@ -48,6 +47,7 @@ def load(proposalNB=None, runNB=None,
extract_xgm=True,
extract_bam=True,
bunchPattern='sase3',
+ parallelize=True,
):
"""
Load a run and extract the data. Output is an xarray with aligned
@@ -66,10 +66,6 @@ def load(proposalNB=None, runNB=None,
{"extra": {'source': 'SCS_CDIFFT_MAG/SUPPLY/CURRENT',
'key': 'actual_current.value',
'dim': None}}
- subFolder: str
- 'raw', 'proc' (processed) or 'all' (both 'raw' and 'proc') to access
- data from either or both of those folders. If 'all' is used, sources
- present in 'proc' overwrite those in 'raw'. The default is 'raw'.
display: bool
whether to show the run.info or not
validate: bool
@@ -97,10 +93,13 @@ def load(proposalNB=None, runNB=None,
bunchPattern: str
bunch pattern used to extract the Fast ADC pulses.
A string or a dict as in::
-
{'FFT_PD2': 'sase3', 'ILH_I0': 'scs_ppl'}
Ignored if extract_digitizers=False.
+ parallelize: bool
+ from EXtra-Data: enable or disable opening files in parallel.
+ Particularly useful if creating child processes is not allowed
+ (e.g. in a daemonized multiprocessing.Process).
Returns
-------
@@ -114,8 +113,8 @@ def load(proposalNB=None, runNB=None,
>>> run, data = tb.load(2212, 208, ['SCS_SA3', 'MCP2apd', 'nrj'])
"""
- runFolder = find_run_path(proposalNB, runNB, subFolder)
- run = ed.RunDirectory(runFolder)
+ runFolder = find_run_path(proposalNB, runNB, 'raw')
+ run = ed.open_run(proposalNB, runNB, data='all', parallelize=parallelize)
if subset is not None:
run = run.select_trains(subset)
if fields is None:
--
GitLab