Hacktoberfestbugdifficulty easygood first issueimpact LOWreach LOW
描述
Problem statement:
It seems to be a bug if the text is too short and number of words is lower than blocksize. In my case the values were: n_words (232.0) and blocksize (1024).
Log:
gensim\summarization\mz_entropy.py:127: RuntimeWarning: invalid value encountered in double_scalars
- __log_combinations(n_words, blocksize)
Dirty solution:
Override blocksize value from the default 1024 to something lower:
mz_keywords(text, blocksize=128)