Update codebleu.py

codebleu.py

@@ -15,10 +15,10 @@
 
 import evaluate
 import datasets
-from .bleu import *
-from .weighted_ngram_match import *
-from .syntax_match import *
-from .dataflow_match import *
+from .bleu import corpus_bleu
+from .weighted_ngram_match import corpus_weighted_ngram_match
+from .syntax_match import corpus_syntax_match
+from .dataflow_match import corpus_dataflow_match
 from tree_sitter import Language, Parser
 
 
@@ -131,7 +131,7 @@ class CodeBLEU(evaluate.Metric):
       tokenized_hyps = [x.split() for x in hypothesis]
       tokenized_refs = [[x.split() for x in reference] for reference in references]
 
-      ngram_match_score = bleu.corpus_bleu(tokenized_refs,tokenized_hyps)
+      ngram_match_score = corpus_bleu(tokenized_refs,tokenized_hyps)
 
       # calculate weighted ngram match
       # from os import listdir
@@ -145,13 +145,13 @@ class CodeBLEU(evaluate.Metric):
       tokenized_refs_with_weights = [[[reference_tokens, make_weights(reference_tokens, keywords)]\
                   for reference_tokens in reference] for reference in tokenized_refs]
 
-      weighted_ngram_match_score = weighted_ngram_match.corpus_bleu(tokenized_refs_with_weights,tokenized_hyps)
+      weighted_ngram_match_score = corpus_weighted_ngram_match(tokenized_refs_with_weights,tokenized_hyps)
 
       # calculate syntax match
-      syntax_match_score = syntax_match.corpus_syntax_match(references, hypothesis, language)
+      syntax_match_score = corpus_syntax_match(references, hypothesis, language)
 
       # calculate dataflow match
-      dataflow_match_score = dataflow_match.corpus_dataflow_match(references, hypothesis, language)
+      dataflow_match_score = corpus_dataflow_match(references, hypothesis, language)