Update from dev branch

[bact]

Comments on tokenizers and test cases

[pep8speaks]

Hello @bact! Thanks for opening this PR. We checked the lines you've touched for PEP 8 issues, and found:

• In the file tests/test_tokenize.py:

Line 169:80: E501 line too long (480 > 79 characters)
Line 170:80: E501 line too long (249 > 79 characters)
Line 171:80: E501 line too long (623 > 79 characters)
Line 172:80: E501 line too long (477 > 79 characters)
Line 173:80: E501 line too long (687 > 79 characters)
Line 174:80: E501 line too long (221 > 79 characters)
Line 175:80: E501 line too long (827 > 79 characters)
Line 176:80: E501 line too long (322 > 79 characters)
Line 177:80: E501 line too long (333 > 79 characters)
Line 178:80: E501 line too long (522 > 79 characters)
Line 179:80: E501 line too long (188 > 79 characters)
Line 180:80: E501 line too long (739 > 79 characters)
Line 181:80: E501 line too long (630 > 79 characters)
Line 182:80: E501 line too long (203 > 79 characters)
Line 183:80: E501 line too long (573 > 79 characters)
Line 184:80: E501 line too long (603 > 79 characters)
Line 185:80: E501 line too long (379 > 79 characters)
Line 186:80: E501 line too long (456 > 79 characters)