User guide
The API can be accessed with an HTTP client. Send a POST request to the /extractor
endpoint with a text
parameter to annotate a given text. For example:
POST /extractor?text=This is a sample sentence. HTTP/1.1
Host: localhost
You can use Postman to query the demo API. If you need to analyse a large number of texts, you will need set up POLKE locally and send the texts to the API one by one with an HTTP client of your choice.
The API returns the list of English Grammar Profile (EGP) constructs and/or tokens found in the text in the JSON format.
{
"annotationList": [
{
"begin": 9,
"end": 21,
"constructID": 36
}
],
"message": ""
}
annotationList | All the EGP constructs found in the text. Each list item is one found construct. |
begin and end | The positions of the construct in the text string. |
constructID | The ID of the construct in the EGP mapping table. |
Parameters
text | The text to annotate. |
tokenize |
Return list of tokens.
annotate
Return list of annotations. Used only for returning just the list of tokens without the annotations. Set to
false
and tokenize=true
to return only the list of tokens.Mapping table
Each construct ID maps to an entry in the mapping table and contains the data from the EGP for the given construct: supercategory, subcategory, CEFR level, guideword, can-do statement, and learner example.
The mapping table can be downloaded here.