Jeff Schnurr

The sentiment analysis is trained on movie reviews in English, so you won't get expected results with Spanish input. There are adapters available for some langues (i.e. [French](https://github.com/sloria/textblob-fr)), but I'm...

TextBlob is training the classifier on each iteration of the loop. Try creating an instance of Blobber once (which invokes training), and then run it against each string of text...

Here are my `timeit` results for training sets of various sizes, using `nltk.corpus.movie_reviews` data and the `nltk.classify.NaiveBayesClassifier.train(data)` method: 2000: 2.27268099785 4000: 3.63234615326 10000: 7.66381907463 20000: 14.5906889439 40000: 28.350331068 It seems...

Interesting - the Google Translate API returns 500 Internal Server Error on strings less than 3 characters long, which is presumably the reason for this limitation in the code: ```...

These are lazy-loaded already. The TextBlob constructor configures the tokenizers/taggers etc, but they don't get called until you invoke a method or access a property that requires it. Can you...

PR #325 should resolve it.

The [documented](http://textblob.readthedocs.io/en/dev/install.html) method is to set the `NLTK_DATA` environment variable, though [this](https://stackoverflow.com/questions/44857382/change-nltk-download-path-directory-from-default-ntlk-data) suggests that the setting is not honoured by the NLTK downloader anyway, though it's used to find the...

Verify the file is accessible by the user executing the script. For example, for Apache: `sudo -u apache cat /nltk_data/tokenizers/punkt/english.pickle`

`NotTranslated` means that the Google API simply returned the same string unchanged; it doesn't think there is any work to do. Are the strings in question the same in both...

botkit-middleware-dialogflow doesn't appear in the stacktrace; this looks like a different problem. Did you check the log?