Add get similar document method

[the-sinner]

Pull Request

Related issue

Fixes #642

What does this PR do?

• Add get similar document method

PR checklist

Please check if your PR fulfills the following requirements:

• Does this PR fix an existing issue, or have you listed the changes applied in the PR description (and why they are needed)?
• Have you read the contributing guidelines?
• Have you made sure that the title is accurate and descriptive of the changes?

Thank you so much for contributing to Meilisearch!

[curquiza]

Thank you very much for your PR @the-sinner 🙏 ❤️

[curquiza]

Why null? this parameter is mandatory

[the-sinner]

will fix this

[curquiza]

Suggested change

	public function testBasicSearchWithFilters(): void
	public function testBasicSearchWithSimilarDocuments(): void

[curquiza]

Can we check we have some expected fields in the answer (specific to similar documents query, like id) even if we don't check the value?

[the-sinner]

can you share the response format of this api, i could not find it.
i tried running the rc1 locally, however i was getting "Cannot find embedder with name default"

[curquiza]

I will ask my colleagues next week for having a basic tests with basic documents to make it test simply 😊

[dureuill]

Hello!

To use the similar endpoints, you need to:

1. enable the vectorStore experimental feature
2. declare at least one embedder in your configuration

For quick sanity tests, the simplest is probably to declare a userProvided embedder.

The settings could be:

"embedders": {
            "manual": {
                "source": "userProvided",
                "dimensions": 3,
            }
        }

Then declare a set of documents that contain the _vectors.manual field, for example:

[
        {
            "title": "Shazam!",
            "release_year": 2019,
            "id": "287947",
            // Three semantic properties:
            // 1. magic, anything that reminds you of magic
            // 2. authority, anything that inspires command
            // 3. horror, anything that inspires fear or dread
            "_vectors": { "manual": [0.8, 0.4, -0.5]},
        },
        {
            "title": "Captain Marvel",
            "release_year": 2019,
            "id": "299537",
            "_vectors": { "manual": [0.6, 0.8, -0.2] },
        },
        {
            "title": "Escape Room",
            "release_year": 2019,
            "id": "522681",
            "_vectors": { "manual": [0.1, 0.6, 0.8] },
        },
        {
            "title": "How to Train Your Dragon: The Hidden World",
            "release_year": 2019,
            "id": "166428",
            "_vectors": { "manual": [0.7, 0.7, -0.4] },
        },
        {
            "title": "All Quiet on the Western Front",
            "release_year": 1930,
            "id": "143",
            "_vectors": { "manual": [-0.5, 0.3, 0.85] },
        }
    ]

The example are from the the engine test file for the similar endpoints.

can you share the response format of this api, i could not find it.

You're right I added the response format to the public usage page, please feel free to ask for more information as needed ☀️

[curquiza]

@the-sinner do you think it would be relevant to create a SimilarDocumentsSearchResult file like https://github.com/meilisearch/meilisearch-php/blob/main/src/Search/SearchResult.php?

[curquiza]

@the-sinner can you update tests according to dureuill message? 😇

[the-sinner]

@curquiza i have added the test case shared by dureuill in the code.

[curquiza]

Thank you @the-sinner

There is an issue with the tests

[the-sinner]

found the issue,
I am using Union types, but they are only supported after 8.0. ref

[the-sinner]

@norkunas @curquiza
For removing the union type, there are two ways:

1. make id string and only accept string values
2. remove type from it and manually check the type while assigning the value

which one do you think would be better ?

In the reference link provided, the given example is using string ids, however the linked documentation allows both string and integer keys.

[norkunas]

Use phpdoc and that's all. People should respect the contract in the same way

[norkunas]

Suggested change

	* @param string[] $attributesToRetrieve an array of attribute names to retrieve
	* @param list<non-empty-string> $attributesToRetrieve an array of attribute names to retrieve

[norkunas]

Could this shape be more specific?

[norkunas]

Suggested change

	public function setOffset(?int $offset): SimilarDocumentsQuery
	/**
	* @param non-negative-int $offset
	*/
	public function setOffset(?int $offset): SimilarDocumentsQuery

[norkunas]

Suggested change

	public function setLimit(?int $limit): SimilarDocumentsQuery
	/**
	* @param positive-int $limit
	*/
	public function setLimit(?int $limit): SimilarDocumentsQuery

[norkunas]

Suggested change

	public function setEmbedder(string $embedder): SimilarDocumentsQuery
	/**
	* @param non-empty-string $embedder
	*/
	public function setEmbedder(string $embedder): SimilarDocumentsQuery

[norkunas]

I still don't see shape defined :)

[curquiza]

Test failing are expected and due to the new RC