minor modifications

kornai · Jul 18, 2016 · 995baf14285d71b3462374361e3fe7d8b89c8e8e · 995baf1
1 parent 6a25bcc
commit 995baf14285d71b3462374361e3fe7d8b89c8e8e
Unified Split

Showing with 8 additions and 32 deletions.

+1 −1 dep_to_4lang.txt

+3 −3 src/fourlang/lemmatizer.py

+2 −24 src/fourlang/sim_feats.py

+2 −4 src/fourlang/similarity.py
diff --git a/dep_to_4lang.txt b/dep_to_4lang.txt
@@ -488,7 +488,7 @@ nmod:except_for	-,-
 nmod:far_from	-,-
 nmod:following	-,-
 nmod:for	-,-	FOR
-  
-nmod:from	-,-	from
+  
+nmod:from	-,-
 nmod:given	-,-
 nmod:if	-,-
 nmod:in	-,-	IN

diff --git a/src/fourlang/lemmatizer.py b/src/fourlang/lemmatizer.py
@@ -36,10 +36,10 @@ def _analyze(self, word):
        self.cache[word] = (stem, lemma, candidates)
    def _lemmatize_with_stopwords(self, word, uppercase):
-  
-        if not uppercase:
-  
-            return word
-  
-        elif word == 'have':
+  
+        if word == 'have':
            return 'HAS'
+  
+        elif not uppercase:
+  
+            return word
        elif word in self.stopwords:
            return word.upper()
        else:

diff --git a/src/fourlang/sim_feats.py b/src/fourlang/sim_feats.py
@@ -103,7 +103,6 @@ def zero_connected(self, name1, links1, links1_expand, name2, links2, links2_exp
        if val == -1:
            if name1 in links2_expand or name2 in links1_expand:
                val2 = 1
-  
-        #ret.update({ "0-connected_exp" : val2 })
        return ret
    def is_antonym(self, name1, nodes1, name2, nodes2):
@@ -124,6 +123,8 @@ def fullgraph(self, name1, name2):
        ####################
        if self.calc_path:
            length = 0
+  
+            if name1 not in self.UG.nodes() or name2 not in self.UG.nodes():
+  
+                return {"shortest_path" : length}
            if nx.has_path(self.UG, name1, name2):
                path = nx.shortest_path(self.UG, name1, name2)
                length = len(path)
@@ -177,29 +178,6 @@ def __init__(self, machine1, machine2, max_depth):
        name1 = machine1.printname()
        name2 = machine2.printname()
-  
-        # TODO: hack
-  
-        # G1_str = MachineGraph.create_from_machines([machine1], max_depth=max_depth, str_graph=True)
-  
-        # G2_str = MachineGraph.create_from_machines([machine2], max_depth=max_depth, str_graph=True)
-  
-        #
-  
-        # print name1
-  
-        # print G1.G.nodes()
-  
-        # print G1_str.G.nodes()
-  
-        # print G1.G.edges()
-  
-        # print G1_str.G.edges()
-  
-        # print name2
-  
-        # print G2.G.nodes()
-  
-        # print G2_str.G.nodes()
-  
-        # print G2.G.edges()
-  
-        # print G2_str.G.edges()
-  
-        #
-  
-        # if(name2 == 'intelligent'):
-  
-        #     G2_str.G = G2_str.G.to_undirected()
-  
-        #     print nx.shortest_path_length(G2_str.G, 'intelligent', 'intelligence')
-  
-        #     print G2_str.G.nodes(data=True)
-  
-        #     print G2_str.G.edges(data=True)
-  
-
-  
-        # TODO: end_hack
-  
-
        self.subgraph_dict = dict()
        # self.subgraph_dict.update(self._get_subgraph_N(G1.G, G2.G, name1, name2))
        # self.subgraph_dict.update(self._get_subgraph_N_X_N(G1.G, G2.G, name1, name2))

diff --git a/src/fourlang/similarity.py b/src/fourlang/similarity.py
@@ -98,9 +98,8 @@ def lemma_similarities(self, lemma1, lemma2):
    def word_similarities(self, word1, word2):
        if (word1, word2) in self.word_sim_cache:
            return self.word_sim_cache[(word1, word2)]
-  
-        # TODO: uppercase flag = ?
        lemma1, lemma2 = [self.lemmatizer.lemmatize(
-  
-            word, defined=self.defined_words, stem_first=True)
+  
+            word, defined=self.defined_words, stem_first=True, uppercase=True)
            for word in (word1, word2)]
        # self.log(u'lemmas: {0}, {1}'.format(lemma1, lemma2))
        if lemma1 is None or lemma2 is None:
@@ -330,10 +329,9 @@ def get_sims(self):
        logging.warning('lemmatizing words to determine machine-OOVs...')
        self.non_oov = set(
            (word for word in self.non_oov
-  
-                # TODO: uppercase flag = ?
                if self.sim_wrapper.lemmatizer.lemmatize(
                    word, defined=self.sim_wrapper.machine_wrapper.definitions,
-  
-                    stem_first=True) is not None))
+  
+                    stem_first=True, uppercase=True) is not None))
        logging.warning(
            'kept {0} words after discarding those not in machine sim'.format(