ebroecker
commented
4 months ago Hi @xRowe ,
We could also try to cache short-names, if we do this during fill_caches with an additional cache this could help. I'd try something like below. (Completely untested, may not work!)
--- a/src/canmatrix/formats/arxml.py
+++ b/src/canmatrix/formats/arxml.py
@@ -57,12 +57,14 @@ class Earxml:
def __init__(self):
self.xml_element_cache = dict() # type: typing.Dict[str, _Element]
self.path_cache = {}
+ self.sn_cache = {}
def fill_caches(self, start_element=None, ar_path=""):
if start_element is None:
start_element = self.root
self.path_cache = {}
if start_element.tag == self.ns + "SHORT-NAME":
+ self.sn_cache[start_element.getparent()] = start_element.text
return start_element.text
for sub_element in start_element:
text = sub_element.text
@@ -153,6 +155,7 @@ class Earxml:
def get_short_name(self, element):
# type: (_Element, str) -> str
+ return self.sn_cache.get(element, "")
"""Get element short name."""
if element is None:
return ""
xRowe
I have two arxml files including both CAN and Ethernet PDU/message
one is v1.arxml, size 23.4M, loading CAN takes about 2s
another is v2.arxml, size 33.2M, loading CAN takes about 10s
below are log provide by the cProfile 23_4.log vs 33_2.log
Seems that the METHOD get_short_name() takes too much time.
I am working on it but without progress, Could you help on it?
Thanks, Rowe