list comprehension of tuples

[bluefrog]

Hi

Is it possible to have a list comprehension so that I can end up with the following result:

file_word_list=[('f1',[('you',0.4),('jump',0.12),('the',0.48)]), ('f2',[('hi',0.167),('at',0.333),('you',0.5)])]

based on this:

file_word_list=[('f1',[('you',10),('jump',3),('the',12)]), ('f2',[('hi',2),('at',4),('you',6)])]

I have a function, which I think gives me what I want, but I was wondering if I could use a comprehension instead. I am unsure how to split the inner tuple though within a comprehension. Also, as you can see from the code, wc_lst is iterated over twice, once to obtain the total word count, and a second time to generate the output and to calculate the term frequency. Would it be possible to only iterate over once?


This is the function:

def calculate_term_frequency(f_wc):
    f,wc_lst = f_wc
    qty = 0
    wctn = []

    for wc in wc_lst:
        w,c = wc    # splitting the inner tuple
        qty = qty + c

    for wc in wc_lst:
        w,c = wc
        wctn.append( (w,c/qty))

    return (f,wctn)

[ichabod801]

Technically, you can:

file_word_freq = [(f, [(word, count / sum([wc[1] for wc in counts])) for word, count in counts]) for f, counts in file_word_list]

However, it's very confusing and rather inefficient. Maybe someone better than me with list comprehensions could fix the inefficiency, but I think it would still be a mess. What's wrong with using the function?

[wavic]

It may look impressive at the end but more important is to be readable.