Can anyone explain this list comprehension?

Question

def unpack_dict(matrix, map_index_to_word):
    table = sorted(map_index_to_word, key=map_index_to_word.get)      
    data = matrix.data
    indices = matrix.indices
    indptr = matrix.indptr        
    num_doc = matrix.shape[0]    
    return [{k:v for k,v in zip([table[word_id] for word_id in 
    indices[indptr[i]:indptr[i+1]] ],
    data[indptr[i]:indptr[i+1]].tolist())} \
               for i in range(num_doc) ]

wiki['tf_idf'] = unpack_dict(tf_idf, map_index_to_word)

map_index_to_word is dictionary of word:index for few thousand words. tf_idf is TFIDF sparse vector DataFrame wiki is displayed in screenshot here

Answer 1

[{k: v for k, v in zip([table[word_id] for word_id in indices[indptr[i]:indptr[i + 1]]],data[indptr[i]:indptr[i + 1]].tolist())} for i in range(num_doc)]

is same as :

final_list = []
for i in range(num_doc):
    new_list = []
    for word_id in indices[indptr[i]:indptr[i + 1]]:
        new_list.append(table[word_id])

    new_dict = {}
    for k, v in zip(new_list, data[indptr[i]:indptr[i + 1]].tolist()):
        new_dict[k] = v
    final_list.append(new_dict)

Answer 2

This?

[{k:v for k,v in zip([table[word_id] for word_id in 
    indices[indptr[i]:indptr[i+1]] ],
    data[indptr[i]:indptr[i+1]].tolist())} \
               for i in range(num_doc) ]

The outer comprehension is

[... for i in range(num_doc) ]

Just a simple loop num_doc times.

Inside that is a dictionary comprehension.

{k:v for k,v in zip()}

The zip takes the k key from:

[table[word_id] for word_id in indices[indptr[i]:indptr[i+1]] ]

and v value from:

data[indptr[i]:indptr[i+1]].tolist()

So the i , outer variable creates the slicing range, indptr[i]:indptr[i+1] .

So it's making a list of dictionaries. The dictionary keys are from table[word_id] where word_id is found in a range of indices , and the value is the corresponding range of data .