Domain __eq__: only cache 10 domains

Author: markotoplak

Orange/data/domain.py

@@ -170,7 +170,7 @@ def __init__(self, attributes, class_vars=None, metas=None, source=None):
         self.anonymous = False
 
         self._hash = None  # cache for __hash__()
-        self._eq_cache = IDWeakrefCache({})  # cache for __eq__()
+        self._eq_cache = IDWeakrefCache(_LRS10Dict())  # cache for __eq__()
 
     def _ensure_indices(self):
         if self._indices is None:
@@ -536,3 +536,12 @@ def __hash__(self):
         if self._hash is None:
             self._hash = hash(self.attributes) ^ hash(self.class_vars) ^ hash(self.metas)
         return self._hash
+
+
+class _LRS10Dict(dict):
+    """ A small "least recently stored" (not LRU) dict """
+
+    def __setitem__(self, key, value):
+        if len(self) >= 10:
+            del self[next(iter(self))]
+        super().__setitem__(key, value)

Orange/tests/test_domain.py

@@ -488,6 +488,27 @@ def __hash__(self):
         domain1._eq_cache[(domain2,)] = False  # pylint: disable=protected-access
         self.assertFalse(domain1 == domain2)
 
+    def test_eq_cache_not_grow(self):
+        var = ContinuousVariable('var')
+        domain = Domain([var])
+        domains = [Domain([var]) for _ in range(10)]
+        for d in domains:
+            self.assertTrue(domain == d)
+
+        # pylint: disable=protected-access,pointless-statement
+
+        # __eq__ results to all ten domains should be cached
+        for d in domains:
+            domain._eq_cache[(d,)]
+
+        dn = Domain([var])
+        self.assertTrue(domain == dn)
+        # the last compared domain should be cached
+        domain._eq_cache[(dn,)]
+        # but the first compared should be lost in cache
+        with self.assertRaises(KeyError):
+            domain._eq_cache[(domains[0],)]
+
     def test_domain_conversion_is_fast_enough(self):
         attrs = [ContinuousVariable("f%i" % i) for i in range(10000)]
         class_vars = [ContinuousVariable("c%i" % i) for i in range(10)]