[Question] substring search about saxpy

[nphoff]

Hey @kuzeko

Thanks for the question. There was a typo in the readme, which I've fixed which would have made this tough.

What I would do in this example is:

from saxpy import SAX
import math

query = [ math.sin(x)*2 for x in range(2, 5)]
target = [math.sin(x)*2 for x in range(0, 100)]

s = SAX(3, 7) # 3 letter words, 7 letter alphabet
(queryString, queryIndices) = s.to_letter_rep(query)
(targetStrings, targetIndices) = s.sliding_window(target, 33, 0.6)
comparisonScores = s.batch_compare(targetStrings, queryString)

From there, comparisonScores will be a list of comparison scores between subsections of the target and the query.

The lowest comparisonScores values will map to which sections were the most similar. The index of the lowest values can be used with targetIndices to find the position in the original dataset which were the lowest.

For instance when I ran it, comparisonScores[2] and comparisonScores[3] were both 0.0, and targetIndices[2], and targetIndices[3] were respectively (2,5) (exactly the same as the query, which is expected) and (3,6) which is pretty close. There will be other entries in comparisonScores which are 0.0 because the sine function is periodic.

Good luck!

[kuzeko]

Fantastic! Thanks a lot!