Support

Entity Extraction API Migration Guide

This guide will describe how to successfully use the new version if you were using the previous one.

dataTXT ➡ Entity Extraction API

dataTXT has been migrated under the Dandelion API umbrella, and it is now called Entity Extraction API. Please note the the Dandelion API now also offers services for text similarity (Text Similarity API) and for language identification (Language Detection API).

If you are looking for more information you can check out the Entity Extraction API reference or the getting started guide.

Did you know?
The old app_key and app_id are no longer valid and have been replaced by a single parameter called token. In order to get new credentials you must register on dandelion.

In your dandelion dashboard you will find the access token that you can use to access Entity Extraction API as well as all the other Dandelion API services. At the moment you can use the system for free with some limitation on the number of API calls you can make per day. If you want to know more check out our pricing page.

Changes in API

The endpoint for the requests is now:

https://api.dandelion.eu/datatxt/nex/v1

Some parameters both in the request and in the response were added, renamed, deprecated or removed. If a parameter is deprecated it can still be used but it might be removed without any notice.

Request parameter

app_id & app_key were removed. Use the token parameter to identify your calls
lang is now optional. if it is not provided dataTXT will infer the language (using dataTXT-LI)
dbpedia was removed. The "include"="types" parameter should now be used in order to get dbpedia types from entities. The use of "include" will no longer cause performance penalty. You can also get categories and images from dbpedia, setting the "include" parameter to "categories,image"
tweet was renamed to social.hashtag
rho was renamed to min_confidence. It is now a value from 0 to 1 representing how good a match is (0 is not to be trusted and 1 is perfect). As a rule of thumb, entities with a confidence of 0.6 or above are good
_callback was removed
long_text was deprecated
include_types was removed. You can get the same result setting the include parameter to "type"
include_categories was removed. You can get the same result setting the include parameter to "categories"
include_abstract was removed. You can get the same result setting the include parameter to "abstract"
include_image was removed. You can get the same result setting the include parameter to "image"

JSON Response

status, error were removed. They are now handled consistently with the rest of dandelion’s api. Documentation about it can be found here
text was removed
langConfidence was added. It is only included in the output if the language was inferred. It has a value from 0 to 1, describing how precise the inference was (0 is not to be trusted and 1 is perfect)
annotation.rho was renamed to confidence (semantic was altered as in the input parameter)
annotation.type parameter was renamed to types, and is only present if the parameter "include" is set to "types"
annotation.ref parameter was renamed to lod, and is only present if the parameter "include" is set to "lod"

More features are available in the new version, check them out in the full api documentation!

SpazioDati Via A. Olivetti 13, 38122, Trento (TN) -


Dandelion API built with by

site privacy | api privacy | tos | cookies

Contact Us

@dandelionapi

Need more info or a custom project? Write us: hello@dandelion.eu

About Us

We're a startup based in Italy, specialized in Semantics & Big Data.
Find out more about us at spaziodati.eu