def tfidf_counts(lyrics=clean_lyrics, n_samples=20, max_feats=MXF):
'''
Returns the nth highest weighted words in the dataset
'''
tfidf_weights = TfidfVectorizer(
max_features=max_feats).fit_transform(lyrics)
weights = np.asarray(tfidf_weights.mean(axis=0)).ravel().tolist()
weights_df = pd.DataFrame({
'word':
TfidfVectorizer(
max_features=max_feats).fit(lyrics).get_feature_names(),
'weight':
weights
})
sort_df = weights_df.sort_values(by='weight',
ascending=False).reset_index(drop=True)
sort_df.index = np.arange(1, len(weights_df) + 1)
print(
tabulate(sort_df.head(n_samples),
headers=weights_df.columns,
tablefmt='pipe'))
return sort_df.head(n_samples)
