Merge branch 'ConsiderClassifyingProperties' of https://github.com/Wikidata-lib/PropertySuggester-Python into ConsiderClassifyingProperties

Conflicts:
	propertysuggester/analyzer/impl/MainAnalyzer.py

Author: Dacry

propertysuggester/analyzer/RuleGenerator.py

@@ -1,18 +1,24 @@
+import ConfigParser
+import os
 from collections import defaultdict
 import itertools
 from propertysuggester.analyzer.impl.MainAnalyzer import ItemAnalyzer
 from propertysuggester.analyzer.impl.QualifierReferenceAnalyzer import QualifierAnalyzer, ReferenceAnalyzer
 from propertysuggester.analyzer.rule import Rule
 from propertysuggester.utils.datamodel import Entity
 
+config = ConfigParser.ConfigParser()
+config.read(os.path.join(os.path.abspath(os.path.dirname(__file__)), 'analyzer.ini'))
+classifying_pids = config.get("mainAnalyzer","classifying_properties").split(",")
+classifying_pids = map(int, classifying_pids)
 
 def compute_rules(entities, min_probability=0.01):
     """
     @type entities: collections.Iterable[Entity]
     @return: list[Rule]
     """
 
-    analyzers = [ItemAnalyzer(), QualifierAnalyzer(), ReferenceAnalyzer()]
+    analyzers = [ItemAnalyzer(classifying_pids), QualifierAnalyzer(), ReferenceAnalyzer()]
 
     for i, entity in enumerate(entities):
         if i % 100000 == 0 and i > 0:

propertysuggester/analyzer/analyzer.ini

@@ -0,0 +1,3 @@
+[mainAnalyzer]
+#List of classifying properties - at the moment limited to 'instance of' (31) and 'subclass of' (279)
+classifying_properties = 31,279

propertysuggester/analyzer/impl/MainAnalyzer.py

@@ -2,11 +2,10 @@
 from propertysuggester.analyzer.impl.Analyzer import Analyzer
 from propertysuggester.analyzer.rule import Rule
 
-classiying_property_ids = [31,279]
-
 class ItemAnalyzer(Analyzer):
-    def __init__(self):
+    def __init__(self, classiying_property_ids = [31,279]):
         Analyzer.__init__(self)
+        self.classiying_pids = classiying_property_ids
         self.tuple_occurrences = defaultdict(int)
         self.pair_occurrences = defaultdict(lambda: defaultdict(int))
 
@@ -17,7 +16,7 @@ def process(self, item):
 
     def _count_occurrences(self, distinct_ids, property_value_pairs):
         for pid1 in distinct_ids:
-            if pid1 in classiying_property_ids:
+            if pid1 in self.classiying_pids:
                 continue
             currentTuple = (pid1, None)
             self.tuple_occurrences[currentTuple] += 1
@@ -26,8 +25,7 @@ def _count_occurrences(self, distinct_ids, property_value_pairs):
                     self.pair_occurrences[currentTuple][pid2] += 1
 
         for pid1, value in property_value_pairs:
-
-            if pid1 in classiying_property_ids:
+            if pid1 in self.classiying_pids:
                 self.tuple_occurrences[pid1, int(value[1:])] += 1
                 for pid2 in distinct_ids:
                     if pid1 != pid2:

propertysuggester/test/analyzer/test_rule_generator.py

@@ -10,7 +10,8 @@
 
 test_data1 = [Entity('Q15', [Claim(Snak(31, 'wikibase-entityid', 'Q5107')),
                              Claim(Snak(373, 'string', 'Africa'))]),
-              Entity('Q16', [Claim(Snak(31, 'wikibase-entityid', 'Q384'))])]
+              Entity('Q16', [Claim(Snak(31, 'wikibase-entityid', 'Q5107'))]),
+              Entity('Q17', [Claim(Snak(31, 'wikibase-entityid', 'Q1337'))])]
 
 test_data2 = [Entity('Q15', [Claim(Snak(31, 'wikibase-entityid', 'Q5107')),
                              Claim(Snak(373, 'string', 'Africa')),