1.2.1

• Use bulk processing from spacy pipeline for improved performance #2392
• Modify TimeLogicAdapter to use SpaCy PhraseMatcher instead of NLTK NaiveBayesClassifier #2394