• Enhanced the text extraction parser of Wikipedia pages
• Enhanced the detection of category and disambiguation pages
• Converted Cython’s *.pyx code to * .py by adopting pure Python mode
• Added support of multi-step Wikipedia redirects
• Fixed an issue related to mmap (#79)

• Fixed a bug in MentionDB occurred when case_sensitive=True (#67)

Changelog:

• Pages that belong to Module and TimedText namespaces are now ignored while creating DumpDB
• Improved normalization rules of entity titles of Wikipedia links #36 #38
• Fixed Jieba tokenizer #27