v6.7.8

What's Changed

• Handle encoding errors (#149)
• Bump supported Python version (#151)
• Remove numpy dependency (#156)
• Feature: distance comparer interface (#159)

Full Changelog: v6.7.7...v6.7.8

v6.7.8-rc1

What's Changed

• Handle encoding errors (#149)
• Bump supported Python version (#151)
• Remove numpy dependency (#156)
• Feature: distance comparer interface (#159)

Full Changelog: v6.7.7...v6.7.8-rc1

v6.7.7

• Remove support for Python 3.6
• Use compiled regex expression in create_dictionary() (#129)
• Configure module logger instead of modifying root logger (#132, #133)

v6.7.6

• Fix suggestion count in lookup_compound when ignore_words=True (#108)
• Log error message when loading dictionary fails (#109)

v6.7.5

• Fix replaced_words not being updated when best match is a combi (closes #103)
• Implement a way to change the edit distance comparer algorightm via distance_algorithm property. Available values are found in DistanceAlgorithm

v6.7.4

• Update editdistpy dependency version
• Remove unused code in fast distance comparer classes

v6.7.3

• Update editdistpy version
• Explicitly mention that ignore_term_with_digits=True has to be used with ignore_non_words=True

v6.7.2

• Fix typo of Dameruau to Damerau in various places. Can potentially break some setups that explicitly _distance_algorithm
• Implement fast distance comparers with editdistpy
• Set DamerauOsaFast as the default distance comparer

v6.7.1.post1

• Updated frequency_dictionary_en_82_765.txt dictionary with common contractions
• Added _below_threshold_words, _bigrams, _count_threshold, _max_dictionary_edit_distance, and _prefix_length when saving to pickle. (closes #93)
• Implemented to_bytes and from_bytes options to save and load pickle with bytes string
• Updated data_version to 3
• Removed Python 3.4 and Python 3.5 support
• Update package README

v6.7.0

• Removed numpy dependency
• word_segmentation now retains/preserves case.
• word_segmentation now keeps punctuation or apostrophe adjacent to previous word.
• word_segmentation now normalizes ligatures: "scientific" -> "scientific".
• word_segmentation now removes hyphens prior to word segmentation (untested).
• American English word forms added to dictionary in addition to British English e.g. favourable & favorable.