Open data

Simplified load function to get rid of the tedious task of finding the semester number.

In the process, the variables semester, topic, and runpath are not used anymore. The later is replaced by subFolder. @lleguy : was there any other purpose of customizing runpath than selecting 'raw' or 'proc' subfolders?

Load.py

@@ -6,8 +6,10 @@
     Copyright (2019) SCS Team.
 """
 import numpy as np
-from karabo_data import RunDirectory, by_index
+from karabo_data import by_index, RunDirectory
+from karabo_data.read_machinery import find_proposal
 import xarray as xr
+import os
 
 mnemonics = {
     # Machine
@@ -362,23 +364,20 @@ mnemonics = {
                     'dim': ['gott_pId','pixelId']}
 }
 
-def load(fields, runNB, proposalNB, semesterNB, topic='SCS', display=False,
-    validate=False, runpath='/gpfs/exfel/exp/{}/{}/{}/raw/r{:04d}/',
-    subset=by_index[:], rois={}):
+def load(fields, runNB, proposalNB, subFolder='raw', display=False, validate=False,
+         subset=by_index[:], rois={}):
     """ Load a run and extract the data. Output is an xarray with aligned trainIds
 
         Inputs:
             fields: list of mnemonic strings to load specific data such as "fastccd", "SCS_XGM",
                 or dictionnaries defining a custom mnemonic such as
                 {"extra": {'SCS_CDIFFT_MAG/SUPPLY/CURRENT', 'actual_current.value', None}}
-            runNB: run number as integer
-            proposalNB: string or int of the proposal number e.g. 'p002252' or 2252
-            semesterNB: string or int of the semester number where the proposal data are saved
-            topic: string of the topic
-            display: boolean, whether to show the run.info or not
-            validate: boolean, whether to run karabo-data-validate or not
-            runpath: a string to fromat the run folder path with topic,
-                semesterNB, proposalNB and runNB
+            runNB: (str, int) run number as integer
+            proposalNB: (str, int) of the proposal number e.g. 'p002252' or 2252
+            subFolder: (str) sub-folder from which to load the data. Use 'raw' for raw
+                data or 'proc' for processed data.
+            display: (bool) whether to show the run.info or not
+            validate: (bool) whether to run karabo-data-validate or not
             subset: a subset of train that can be load with by_index[:5] for the
                 first 5 trains
             rois: a dictionnary of mnemonics with a list of rois definition and the desired
@@ -389,15 +388,18 @@ def load(fields, runNB, proposalNB, semesterNB, topic='SCS', display=False,
         Outputs:
             res: an xarray DataSet with aligned trainIds
     """
-
-    if type(proposalNB)==int:
+    
+    if isinstance(runNB, int):
+        runNB = 'r{:04d}'.format(runNB)
+    if isinstance(proposalNB,int):
         proposalNB = 'p{:06d}'.format(proposalNB)
-    runFolder = runpath.format(topic, semesterNB, proposalNB, runNB)
+    runFolder = os.path.join(find_proposal(proposalNB), subFolder, runNB)
     run = RunDirectory(runFolder).select_trains(subset)
 
     if validate:
         get_ipython().system('karabo-data-validate ' + runFolder)
     if display:
+        print('Loading data from {}'.format(runFolder))
         run.info()
 
     keys = []