Python3

[der-tim]

I changed some minor functions (print, range) in the notebook to make it run in a python 3 kernel. Renamed the notebook, so that python 2 and 3 files can co-exist (not sure if that is best practice / your preferred way of doing it, thought this might create the least problems). environment3.yaml contains my conda environment, which has way more than the necessary dependencies, but lists all versions I use.

As I did not have the original yelp data set, I created a mock one to test. However in my case, the cell containing

# this took about an hour on our 16GB Digital Ocean instance - go grab lunch!
yelp_model = gensim.models.word2vec.Word2Vec(['test'])

did not run, as line 431 in gensim\models\word2vec.py (self.build_vocab(sentences, trim_rule=trim_rule)) seem to set self.vocab to an empy dict (or not set it at all) or as trim_rule might not be set. (This is likely a problem with my mock data, and as all cells following this one should run in python 3, I did not look into it further.)

[AlJohri]

👍