N-Gram Similarity

N-Gram Similarity

An experiment doing n-gram similarity.

Usage:

`$ python trigram.py '<left string>' '<right string>' <number of letters in a gram>`

Example:

```
$ python trigram.py 'hello' 'hallo' 3
Left to Right Similarity: 0.5
Right to Left Similarity: 0.5
```

trigram.py

import sys


def grammify(s, maxn=3):
    s = s.replace(' ', '')
    grams = []
    for n in range(1, len(s)):
        if n > maxn:
            break

        for i in range(len(s)):
            gram = s[i:i+n]
            if len(gram) == n:
                grams.append(gram)

    return grams


def main():
    s1 = sys.argv[1]
    s2 = sys.argv[2]
    maxn = int(sys.argv[3])

    left = grammify(s1, maxn=maxn)
    right = grammify(s2, maxn=maxn)
    common = set(left).intersection(set(right))
    # print("Left grams:", left)
    # print("Right grams:", right)
    # print("Common grams:", common)

    # how similar is s1 to s2
    right_score = len(common) / len(right)
    print("Left to Right Similarity:", right_score)

    # how similar is s2 to s1
    left_score = len(common) / len(left)
    print("Right to Left Similarity:", left_score)


if __name__ == "__main__":
    main()