Update PxCorpus.py

PxCorpus.py

@@ -36,6 +36,8 @@ of 55 participants (38% non-experts, 25% doctors, 36% medical practitioners), ma
 
 _URL = "https://zenodo.org/record/6524162/files/pxslu.zip?download=1"
 
+_VOCAB = StringIndex(vocab=open("./vocabulary_nachos_lowercased.txt","r").read().split("\n"))
+
 class StringIndex:
 
     def __init__(self, vocab):
@@ -137,8 +139,6 @@ class PxCorpus(datasets.GeneratorBasedBuilder):
 
     def getTokenTags(self, document):
 
-        VOCAB = StringIndex(vocab=open("./vocabulary_nachos_lowercased.txt","r").read().split("\n"))
-
         tokens = []
         ner_tags = []
         is_oov = []
@@ -151,7 +151,7 @@ class PxCorpus(datasets.GeneratorBasedBuilder):
             text, label = pair.split("\t")
             tokens.append(text)
             ner_tags.append(label)
-            is_oov.append(VOCAB.find(text))
+            is_oov.append(_VOCAB.find(text))
 
         return tokens, ner_tags, is_oov