pycrea 0.1.3

A Python package to parse and manage CREA word vectors. Learn more about this project on GitHub: https://github.com/WiredBrains-Lab/

CREA-Vectors

This is a package to access CREA vectors via pycrea. Various data collection methods are also included. Future updates will allow users to add their collected vectors to the main set.

How to run the code

1. pip install pycrea in your terminal
2. In python import the library
   • from crea import CREA
   • c = CREA() # Automatically loads main dictionary or
   • c = CREA('path_to_file.json') # Load custom json file
     • c_cols = CREA(col_names = True) loads feature names

   Usage

   • all_vectors = c.get_all_vectors()
   • vector = c.get_vector('word1')
   • vectors = c.get_vectors(['word1', 'word2'])
   • vectors_and_cols = c_cols.select_cols(['word1', 'word2'], [feature_1, feature_2])
   • similarity = c.cosine_similarity('word1', 'word2')
   • n_similar = c.top_n_similar('word1', 5) # defaults to 5

PsychoPy Data Collection

• Results are saved as a JSON file named 'results.json'
• Currently it is overwritten each time
• JSON file has the structure {'target_word': {'Rating Category 1': rating_1, ... 'Rating Category n': rating_n}}
• CREA.get_raw('results.json') gets the new raw data from results.json and adds a new row to raw results.csv
• CREA.calculate_averages('results.csv') calculates the average of results.csv and returns a DataFrame of the averages