Fix encoding bug

drugprot.py

@@ -180,7 +180,7 @@ class DrugProtDataset(datasets.GeneratorBasedBuilder):
         """ """
         # Note: The split "test-background" does not contain any relations
         document_to_entities = collections.defaultdict(list)
-        for line in entities_file.read_text().splitlines():
+        for line in entities_file.read_text(encoding='utf-8').splitlines():
             columns = line.split("\t")
             document_id = columns[0]
 
@@ -196,7 +196,7 @@ class DrugProtDataset(datasets.GeneratorBasedBuilder):
         document_to_relations = collections.defaultdict(list)
 
         if relations_file is not None:
-            for line in relations_file.read_text().splitlines():
+            for line in relations_file.read_text(encoding='utf-8').splitlines():
                 columns = line.split("\t")
                 document_id = columns[0]
 
@@ -212,7 +212,7 @@ class DrugProtDataset(datasets.GeneratorBasedBuilder):
                 )
 
         document_to_source = {}
-        for line in abstracts_file.read_text().splitlines():
+        for line in abstracts_file.read_text(encoding='utf-8').splitlines():
             document_id, title, abstract = line.split("\t")
 
             document_to_source[document_id] = {