DatacardTask: cache datacard resolving

this brings:

• consistency: (sub)tasks do not become inconsistent with each other if globbing for datacards changes due to new/removed files
• massive speed up: instancing workflow branch tasks (e.g. in UpperLimits) would previously repeatedly resolve the datacards parameter slowing down the task-tree building & workflow starting by a factor upto 100x (depending on storage load; this was enough to bring down eos directories/nodes that contained the datacards)

dhi/tasks/combine.py

@@ -15,6 +15,7 @@ import functools
 from abc import abstractproperty
 from collections import defaultdict, OrderedDict
 
+from backports.functools_lru_cache import lru_cache
 import law
 import luigi
 import six
@@ -452,6 +453,7 @@ class DatacardTask(HHModelTask):
         return _path, bin_name, inject, store_dir
 
     @classmethod
+    @lru_cache(maxsize=None)
     def resolve_datacards(cls, patterns):
         paths = []
         bin_names = []