Tolerance

[mateonunez]

This PR is a proposal to resolve the problem of #2.

The changes don't implement a boolean to manage the typo errors. Instead of that, a new parameter has been added to the search method.

As the documentation shows: tolerance is the maximum distance between the searched term and a word inside a document.

I've added a few tests for this implementation.

Coverage

Without Tolerance implementation:

File	% Stmts	% Branch	% Funcs	% Lines	Uncovered Line #s
All files	85.6	65.38	89.47	85.82
...refix-tree	95.18	78.57	100	97.5
trie.ts	94.02	78.57	100	96.87	104,129

With Tolerance implementation:

File	% Stmts	% Branch	% Funcs	% Lines	Uncovered Line #s
All files	86.44	67.41	90	87.14
src	82.43	58.18	84.61	82.38
...shtein.ts	91.66	60	100	100	2-3
...refix-tree	95.55	82.35	100	97.7
trie.ts	94.59	82.35	100	97.18	115,140

Time elapsed

Without Tolerance implementation:

{
  elapsed: '135μs',
  hits: [
    {
      id: 'ECbOj5pixIOJc9KfQpwss',
      quote: 'Be yourself; everyone else is already taken.',
      author: 'Oscar Wilde',
      isFavorite: false,
      rating: 4
    },
    {
      id: 'jMfGCEX1zf-TX0Tc5-e-8',
      quote: 'So many books, so little time.',
      author: 'Frank Zappa',
      isFavorite: true,
      rating: 5
    }
  ],
  count: 2
}

With Tolerance implementation:

{
  elapsed: '147μs',
  hits: [
    {
      id: 'FFU-21lklMCkviNQgSPCc',
      quote: 'Be yourself; everyone else is already taken.',
      author: 'Oscar Wilde',
      isFavorite: false,
      rating: 4
    },
    {
      id: '4FAnVKAP2wygWEbnZOUhT',
      quote: 'So many books, so little time.',
      author: 'Frank Zappa',
      isFavorite: true,
      rating: 5
    }
  ],
  count: 2
}

[micheleriva]

Thank you @mateonunez, this is gold. I'll review the code tomorrow morning and let you know how we want to proceed.

I actually planned to introduce a tolerance-like property in one of the next releases of Lyra, but if we get it merged sooner, it's better.

[mateonunez]

Sounds great @micheleriva! Can't wait for the new release and the new features. 👌

[micheleriva]

@mateonunez this PR looks really good! I just left a few minor comments, it would be awesome if you could address them.

In a future release, as an improvement, we should aim at running the Levenshtein algorithm against the tree instead of words.

Let me give you an example:

• If you're searching for "hello", but you mistype and write "hllo" instead, the current algorithm will work just fine.
• If you're searching for "hello", but you mistype and write "yellow" (with a tolerance of 2), this won't work, as we discard words with y as a prefix.

This is something to consider for future releases, as for now, I am happy with the current PR

[mateonunez]

I got the point about the next Levenshtein algorithm implementation @micheleriva. It might be fun to work on it.

[micheleriva]

LGTM