Following this product discussion: meilisearch/product#491 (comment)
Related to this spec: meilisearch/specifications#189

Goal

Allow using the wildcard * at the end of the index name (in searchRules field) when creating tenant tokens.

How does it work?

To create a tenant token, you need to provide a searchRules field containing one or several index names.

{
    "searchRules": {
        "medical_records": {
            "filter": "user_id = 1"
        },
        "medical_records_staging": {
            "filter": "user_id = 1"
        },     
    }
}

Also, the * is already allowing as following, meaning "all the indexes":

{
    "searchRules": {
        "*": {
            "filter": "user_id = 1"
        }
    }
}

But the users need an intermediate usage of the * character: allowing to use the * character at the end of the index name.

Instead of passing:

{
    "searchRules": {
        "medical_records": {
            "filter": "user_id = 1"
        },
        "medical_records_staging": {
            "filter": "user_id = 1"
        },     
    }
}

The users could pass:

{
    "searchRules": {
        "medical_records*": {
            "filter": "user_id = 1"
        }
}

And this tenant will be available for all indexes matching medical_records*, so medical_records and medical_records_staging will match, but not medical_record (without s)

Errors

• if no indexes are matched, no error are thrown during the tenant token creation. However this token will be useless until a matching index is created

Todo

• Implementation the changes by opening a PR in this repository -> open to external contributions. You can be inspired by this PR
• Update the spec regarding this -> API Keys - Introduce * wildcard char at the last position to express access to multiple indexes starting with the same string specifications#189

Impacted teams

Pinging @meilisearch/docs-team, but also @meilisearch/integration-team (I'm not sure for you)