Skip to content
Search! & Match! API
Introduction
latest

Matching🔗

The Match feature auto-generates a Search query from a given URL or document, and is available when it is enabled for your environment.

Matching in the Textkernel Search UI🔗

The Match capabilities in the Search UI depend on your particular setup. In general the following features are available:

Matching from a document or URL🔗

Users can run a Match query by uploading a document, dropping a document, or entering a URL in the search box at the top of the screen. Note that in a Talent Search environment, the input document should always be a job description (e.g. public job URL). In a Job Search environment, the input should always be a person's profile (e.g. CV).

Match from search results🔗

Matching between Candidates and Jobs from the result listing. For each search result, there is a button that users can click to match that search result (people or job) against other people or jobs.

Pre-loading a Match query🔗

Opening the Search UI with a Match query by referencing to an already indexed document (document URI) or to a URL using the match URL Parameter. See Match Queries for the query language. Can be combined with other filters using the search URL Parameter (for example to rank candidates against a job: combination of Match query and filtering candidates on the job IDs they have applied to).

Match APIs🔗

Matching with an already indexed document🔗

You can match based on an already indexed document (candidate or job) by using the Search Service - method searchWithUrl. The query should contain a reference to the indexed document (document URI). See Match Queries for more information.

Because the document that is used for generating the match query has already been parsed and indexed, the results will be returned faster than when matching based on file or URL (see below).

Note: Matching with an already indexed document does not support pagination since it would require generating the query for each page. If you also need pagination you may flow one of the following approaches: 1. Use Search Service - method searchWithUrl to match from an indexed document. This returns not only the first page of results, but also the query, which you can then use for subsequent further pages, like with any other search request as well. So you continue after the searchWithURL endpoint with the normal search endpoint for the 2nd or any further page. 2. Generate query by calling Query Extraction Service, and then call Search Service - method search with the generated query.

Matching based on a file or URL🔗

When the document that you want to match is not already indexed, you can use one of the services below. This will first parse the document and then generate the match query.

Because of the additional parsing time, this method is typically a few seconds slower than when matching based on an already indexed document or when matching based with a pre-generated query.

Pre-generating the Match query🔗

You can use the Query Extraction Service for retrieving the query based on either a file or URL, and storing it for later use. This will give you the fastest response time compared to the other two methods below. But the disadvantage is that the Match query may become outdated if the document was changed in the meantime.