search a string in a long text

[uriva]

hi, I'm trying to use fuse.js to fuzzy search within a large string.

const text ='...very large text';

const query = 'some query';

const result = new Fuse([text], {
  threshold: 0.1,
  includeMatches: true,
  includeScore: true,
  isCaseSensitive: false,
  ignoreLocation: true,
}).search(query);

console.log(
  result[0].matches[0].indices.map(([start, end]) => start, end, text.slice(start, end + 1))
);

this does match but I don't know which match is best. most are are useless. it's not the longest one, and it's apparently not even the longest one that is shorter than the query.

I think this might require some changes in the source code, to get scoring per match or get them sorted.

Is there any way to achieve this?

Thanks

[offsky]

I realize this question is nearly 2 years old. I am also trying to use this library to search large blocks of text and am having the same trouble with the returned matches mostly being useless for highlighting purposes. This screenshot illustrates it.

I have searched for "roots in various cultures" in the source text. Fuse found the displayed document and gave it a 37% match even though the exact string is found in the text. Fuse also returned 160 matches which I have highlighted. The exact match is highlighted as expected but so are a ton of other useless substrings.

I haven't dug into the code to understand this yet. If we could get a score for each individual match, then we could selectively highlight only the good ones. Or maybe this library isn't built for this use case. I dont know.

[uriva]

fyi I ended up using npm:levenshtein-search and other complementary approaches (tf-idf)