embedded_to_numpy.py

import numpy as np
import pandas as pd
from ast import literal_eval

# issue with reading csv file where if i take in the file using genfromtxt or loadtxt with dtype str, it
# considers the ',' inside the arrays as additional columns and gets an error
def embedded_to_numpy():
    # Loads the embedded data
    df = pd.read_csv('embedded.csv')
    all_arrays = []
    # Goes through all the ada_embeddings
    for arr in df['ada_embedding']:
        all_arrays.append(literal_eval(arr)) # Evaluate into a float array
    np_array = np.matrix(all_arrays) # Turn into np array
    return np_array

#test
a = embedded_to_numpy()
print(a)
print(a.shape)
print(a[0,0]) # -0.01962471939623356
print(type(a[0,0]))