CODEWITHSUNDEEP
pythonpython-2.7
P A N
P A N

Reputation: 5922

"TypeError: Bool is not iterable" in any() iteration with boolean criteria

In Python 2.7, I want to check the similarity between one string, and strings in a list, until it finds a match.

from difflib import SequenceMatcher

def similar(a, b):
    return SequenceMatcher(None, a, b).ratio()

correctList = ["thanks", "believe", "definitely"]

myString = "thansk"

for correctWord in correctList:
    ratio = similar(correctWord, myString)
    if ratio > 0.9:
        myString = correctWord
        break

print myString
>>> "thanks"

I would like to simplify the for iteration into fewer lines, to something like:

if similar(myString, any([correctWord for correctWord in correctList])) > 0.9:
    myString = correctWord

I'm not entirely sure about the correct logic here, but in either case variants of this syntax throws the error:

TypeError: ("'bool' object is not iterable", u'occurred at index 0')

What would be the proper way to achieve this?

Upvotes: 0

Views: 691

Answers (2)

obataku
obataku

Reputation: 29646

any should only take a list of boolean expressions, so we need to find a way to evaluating similar between myString and every element of correctList first. We can use map here alongside a predicate lambda s: similar(myString, s) > 0.9:

any(map(lambda s: similar(s, myString) > 0.9, correctList))

This evaluates to True if there is at least one element of correctList 'similar enough' to myString.

... but you'll notice that we want to determine which elements of correctList are similar to myString, so perhaps we should really be using filter:

candidates = filter(lambda s: similar(s, myString) > 0.9, correctList)

You could just take the first result, in which case the next expression would work, but it wouldn't necessarily be the most similar element of correctList.

We can, however, use map, filter, and max to accomplish that. Consider:

pairs = map(lambda s: (s, similar(s, myString)), correctList)

returns a list of pairs, each consisting of an element of correctList and its 'degree of similarity' with myString. We can then filter out the candidates with similarity below 0.9:

pairs = filter(lambda (s, d): d > 0.9, pairs)

and lastly we choose the candidate with maximum similarity from those remaining (using operator.itemgetter as our key function) or myString if none exists:

myString = (max(pairs, key = itemgetter(1)) or [myString])[0]

Of course, we could also use max without pre-filtering and then disregard the answer if its similarity is insufficient:

pairs = map(lambda s: (s, similar(s, myString)), correctList)
candidate = max(pairs, key = itemgetter(1))
myString = candidate[0] if candidate[1] > 0.9 else myString

Upvotes: 2

lurker
lurker

Reputation: 58294

You are stopping as soon as you find the first string with similarity > 0.9, starting with "thansk" as a candidate. So I think this is equivalent:

myString = "thansk"

myString = next((w for w in correctList if similar(w, myString) > 0.9), myString)

Upvotes: 2

Related Questions