Is there any REST API available for Faceted Search?

[eguvenc]

Hi,

I'm currently using Wikibase 1.44 together with Faceted Search 1.0 and I was wondering if there is any REST API (or any other programmatic interface) available for Faceted Search.

I’d like to query the faceted search results (e.g., entities matching certain filters) directly via an API, instead of using the Special:FacetedSearch page.

Could you please let me know if:

• a REST or Action API endpoint exists for this extension, or
• if there are any plans to add such an API in the future?

Thank you for your work on this great extension!

[malberts]

There is no REST or Action API. We do not currently have plans to add it, but if anyone wishes to contribute, or sponsor development, then we can look at it.

(Duplicate of #275)

[JeroenDeDauw]

Can't you use the search keywords we add via whatever standard search APIs MediaWiki exposes?

[malberts]

Right, I misinterpreted this issue. #275 mentions it is possible to search using the Action API. That issue was about an API for retrieving facets, not search results.

[eguvenc]

Hi Jeroen,

Thanks for the clarification!

Just to make sure I understand correctly, do you mean that Faceted Search adds its keywords (e.g., properties or filters) to the standard MediaWiki search index, so that I can use them via action=query&list=search or the CirrusSearch API (e.g., srsearch= queries)?

If you have an example query or documentation reference for how to use these keywords programmatically, that would be really helpful.

For example, I tried the following request:

action=query&list=search&srsearch=haswbfacet:P5=Q2 haswbfacet:P4>=2020-01-01

Test URL

But the response is:

{
  "warnings": {
    "search": {
      "*": "haswbstatement keyword contains no valid statements"
    }
  },
  "query": {
    "searchinfo": { "totalhits": 0 },
    "search": []
  }
}

The query works fine in the Faceted Search UI, but fails through the API.
Is there a specific syntax or endpoint needed to run such facet-based queries programmatically ?

[JeroenDeDauw]

Query string is docuemnted at https://professional.wiki/en/extension/wikibase-faceted-search#QueryString. What you have looks correct.

Idk what the right endpoint is. Likley you have the correct one already. I suspect you run into some encoding issue, or something else specific to this API endpoint, unrelated to our extension. If you figure it out, posting here would be nice so I can update our docs with an example.

[eguvenc]

I'm currently working on a separate extension that queries the Elasticsearch index directly, as I couldn't find a readily available programmatic solution for fetching results that matches the Faceted Search UI filters via the standard MediaWiki API.

I have a quick question regarding the Faceted Search implementation itself:

I'm wondering if the Faceted Search Extension faces any performance issues when handling searches on string/text properties (like labels, descriptions, or string values) due to its use of the default analyzer, rather than an n-gram analyzer.

Given that I'm working with large datasets in Wikibase, and the Faceted Search extension seems to rely on the default Elasticsearch analyzers (not n-grams), will I run into significant performance issues when users perform partial text searches or try to use autocomplete on text fields (like labels and descriptions)?

Or is the extension primarily designed for high performance on structured/exact data queries ?

[JeroenDeDauw]

We did not specifically consider partial matching and I do not know if you should expect performance problems with those. We are not Elastic experts.

[colinrobinsonuib]

This is how I am doing searches:

/w/api.php?action=query&list=search&format=json&srsearch=Nelson%20haswbfacet:P1=Q804%20haswbfacet:P6=Q1180&srnamespace=120&srlimit=50

response:

{
  "batchcomplete": "",
  "query": {
    "searchinfo": {
      "totalhits": 2
    },
    "search": [
      {
        "ns": 120,
        "title": "Item:Q8815",
        "pageid": 8829,
        "size": 3962,
        "wordcount": 6,
        "snippet": "Curses\ncreative work by Graham \u003Cspan class=\"searchmatch\"\u003ENelson\u003C/span\u003E",
        "timestamp": "2025-02-21T17:35:02Z"
      },
      {
        "ns": 120,
        "title": "Item:Q10269",
        "pageid": 10283,
        "size": 3925,
        "wordcount": 9,
        "snippet": "The Inform Designer’s Manual\ncreative work by Graham \u003Cspan class=\"searchmatch\"\u003ENelson\u003C/span\u003E",
        "timestamp": "2025-02-12T07:19:26Z"
      }
    ]
  }
}

Works great