Takes a document (string) or iterable of documents and returns a Pandas dataframe containing the number of occurrences of each unique word. Note that this is not efficient enough to replace Scikit's CountVectorizer class for a bag of words transformer.

get_word_counts.py

import numpy as np
import pandas as pd

def get_word_counts(document: str) -> pd.DataFrame:
    """
    Turns a document into a dataframe of word, counts
    
    Use preprocessing/lowercasing before this step for best results.
    
    If passing many documents, use document = '\n'.join(iterable_of_documents)
    """
    
    vocab, counts = np.unique(document.split(), return_counts=True)
    
    combined_df = pd.DataFrame({'vocab': vocab, 
                                'counts': counts})
    
    return combined_df.sort_values('counts', ascending=False).reset_index(drop=True)