Improve performance of generating histograms.xml

This CL implements IterElementsWithTag to avoid not needed recursion.

On local machine it saved around 1s (from 7.6s to 6.6s).

Bug: 1016281
Change-Id: I0fea65892bc52574fd4e6fd3fd2455ae3ec646e0
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1872052
Commit-Queue: Alexei Svitkine <asvitkine@chromium.org>
Reviewed-by: Alexei Svitkine <asvitkine@chromium.org>
Cr-Commit-Position: refs/heads/master@{#709564}

tools/metrics/histograms/expand_owners.py

@@ -303,9 +303,10 @@ def ExpandHistogramsOWNERS(histograms):
     Error: Raised if the OWNERS file with the given path does not exist.
   """
   email_pattern = re.compile(_EMAIL_PATTERN)
+  iter_matches = extract_histograms.IterElementsWithTag
 
-  for histogram in histograms.getElementsByTagName('histogram'):
-    owners = histogram.getElementsByTagName('owner')
+  for histogram in iter_matches(histograms, 'histogram'):
+    owners = [owner for owner in iter_matches(histogram, 'owner', 1)]
 
     # owner is a DOM Element with a single child, which is a DOM Text Node.
     emails_with_dom_elements = set([
@@ -316,12 +317,10 @@ def ExpandHistogramsOWNERS(histograms):
     # component is a DOM Element with a single child, which is a DOM Text Node.
     components_with_dom_elements = set([
       extract_histograms.NormalizeString(component.childNodes[0].data)
-      for component in histogram.getElementsByTagName('component')])
+      for component in iter_matches(histogram, 'component', 1)])
 
-    for index in range(len(owners)):
-      owner = owners[index]
+    for index, owner in enumerate(owners):
       owner_text = owner.childNodes[0].data
-
       name = histogram.getAttribute('name')
       if _IsEmailOrPlaceholder(index == 0, owner_text, name):
         continue

tools/metrics/histograms/merge_xml.py

@@ -6,9 +6,11 @@
 """A script to merge multiple source xml files into a single histograms.xml."""
 
 import argparse
-import expand_owners
 import xml.dom.minidom
 
+import expand_owners
+import extract_histograms
+
 
 def GetElementsByTagName(trees, tag):
   """Gets all elements with the specified tag from a set of DOM trees.
@@ -19,7 +21,8 @@ def GetElementsByTagName(trees, tag):
   Returns:
     A list of DOM nodes with the specified tag.
   """
-  return [e for t in trees for e in t.getElementsByTagName(tag)]
+  iterator = extract_histograms.IterElementsWithTag
+  return list(e for t in trees for e in iterator(t, tag, 2))
 
 
 def MakeNodeWithChildren(doc, tag, children):