Ontology API Reference

Last update: July 2021

Introduction

The Ontology API provides insights into the relations of objects from our Textkernel's knowledge graph, in particular professions and skills. It works in combination with Textkernel’s Professions API and Skills API, which provide functionalities around extraction, normalization and autocompletion for profession and skill data. The endpoints of the Ontology API operate on the level of IDs; they do not return the descriptions of professions and skills. To translate IDs into descriptions, and vice versa, the Professions API and Skills API can be used.

Language support

The Ontology API is language agnostic. The language support is therefore limited to those languages that are supported by both the Skills API and Professions APIs. These are currently:

• English
• French
• German
• Dutch
• Spanish
• Italian
• Portuguese
• Polish
• Japanese

Endpoint Overview

The Ontology API offers the following endpoints:

• /professions/suggest_skills - Given a profession, suggest associated skills
• /professions/compare_skills - Given two professions, compare the associated skills
• /skills/suggest_professions - Given a set of skills, suggest associated professions
• /skills/compare_to_profession - Compare a given skill set to the inferred skill set for a given profession.
• /skills/suggest_skills - Given a set of skills, suggest more similar skills
• /skills/similarity_score - Given two sets of skills, calculate the similarity score between the skill sets

How to map professions to skills

The pipeline for mapping professions to skills (or vice versa) generally consists of three steps, though this may depend on the details of the integration.

Step 1: obtain the Profession ID

There are three options:

• If the starting point is a CV or job ad, the profession ID is in the output of the Extract API.
• If the starting point is a job title or vacancy title, the profession ID can be retrieved from the normalization endpoint of the Professions API.
• If the starting point is user-typed job title, the profession ID can be retrieved using the autocomplete endpoint of the Professions API.

Step 2: Map the profession to skill IDs

In order to retrieve the most relevant skills for a given profession, make a call to the suggest_skills endpoint.

How to map skills to professions or to other skills:

The pipeline for the other direction - skills to professions - is conceptually very similar.

Step 1: Obtain the Skill ID(s)

• If the starting point is a CV or job ad, the skills IDs are in the output of the Extract API.
• If the starting point is any other type of document, the Skills API returns IDs of the extracted skills.
• If the starting point is a skill description in raw text, it can be normalized to an ID value with the /skills/extract (threshold=0)
• If the starting point is an input field with autocomplete, the profession ID can be retrieved using the autocomplete endpoint of the Skills API.

Step 2: Map the skill ID(s) to professions in a given language

• In order to retrieve the most relevant professions for a given skill set, make a call to /skills/suggest_professions.
• In order to retrieve the most relevant skills for a given skill set, make a call to /skills/suggest_skills.