Apply memoization to getPhases; this provides a decent perf gain

gsnedders · gsnedders · commit bf3e733bbafd · 2016-05-09T21:49:48.000-07:00
diff --git a/html5lib/html5parser.py b/html5lib/html5parser.py
@@ -418,6 +418,7 @@ def parseRCDataRawtext(self, token, contentType):
         self.phase = self.phases["text"]
 
 
+@utils.memoize
 def getPhases(debug):
     def log(function):
         """Logger that records which phase processes each token"""
diff --git a/html5lib/utils.py b/html5lib/utils.py
@@ -109,3 +109,15 @@ def moduleFactory(baseModule, *args, **kwargs):
             return mod
 
     return moduleFactory
+
+
+def memoize(func):
+    cache = {}
+
+    def wrapped(*args, **kwargs):
+        key = (tuple(args), tuple(kwargs.items()))
+        if key not in cache:
+            cache[key] = func(*args, **kwargs)
+        return cache[key]
+
+    return wrapped
