also use amended version of python_2_unicode_compatible for matching annotator

sciencebeam_gym/preprocess/annotation/matching_annotator.py

@@ -5,10 +5,12 @@ import csv
 from builtins import str as text
 from itertools import tee, islice
 
-from future.utils import python_2_unicode_compatible
-
 from six.moves import zip_longest
 
+from sciencebeam_gym.utils.compat import (
+  python_2_unicode_compatible
+)
+
 from sciencebeam_gym.utils.csv import (
   csv_delimiter_by_filename,
   write_csv_row