Integrate the knowledge ASAP.

If we do this before the parsing we can leverage that semantics
in the matching phase.

naive-nlu/tree_nlu/knowledge_base.py

@@ -24,6 +24,13 @@ class KnowledgeBase(object):
         # Parse everything
         parsed_examples = []
         for example in examples:
+            # If there's parsed data, leverage it ASAP
+            if 'parsed' in example:
+                result = knowledge_evaluation.integrate_information(self.knowledge, {
+                    "parsed": example['parsed'],
+                })
+                self.act_upon(result)
+
             logging.info("\x1b[7;32m> {} \x1b[0m".format(example))
             tokens, decomposition, inferred_tree = parsing.integrate_language(self, example)
             logging.info(tokens)

naive-nlu/tree_nlu/knowledge_evaluation.py

@@ -123,12 +123,17 @@ def pertenence_to_group(knowledge_base, elements, subj, group):
     if "groups" not in knowledge_base[subj]:
         knowledge_base[subj]["groups"] = set()
 
-    return modifiable_element_for_existance_in_set(
-        container=knowledge_base[subj],
-        set_name="groups",
-        element=group
-    )
+    if group not in knowledge_base:
+        knowledge_base[group] = {}
 
+    if "groups" not in knowledge_base[group]:
+        knowledge_base[group]["groups"] = set()
+
+    return modifiable_element_for_existance_in_group(
+        container=knowledge_base[subj],
+        element=group,
+        backlink=knowledge_base[group],
+    )
 
 def has_capacity(knowledge_base, elements, subj, capacity):
     subj = resolve(knowledge_base, elements, subj)