Skip to content

Latest commit

 

History

History
82 lines (63 loc) · 4.79 KB

Graphene-Server.md

File metadata and controls

82 lines (63 loc) · 4.79 KB
Raw

Usage of the Graphene-Server

The Graphene-Server captures the same functionalities as the Graphene-Core API. The wrapped Core API functions can be executed by HTTP GET requests.

Resolving co-references

Coreference Resolution will resolve coreferences in the given input text. The Coreference Resolution service accepts the following endpoints with parameters:

Endpoint: /coreference/text

Returns the JSON-serialized versions of the CoreferenceContent class or a textual representation depending on the value of the Accept header in the POST request:

  • Accept: text/plain: returns a textual representation

  • Accept: application/json: returns JSON-serialized versions of CoreferenceContent

  • text: The input text (mandatory).

Example:

curl -X POST -H "Content-Type: application/json"  -d '{"text": "The text."}' -H "Accept: text/plain" "http://localhost:8080/coreference/text"
curl -X POST -H "Content-Type: application/json"  -d '{"text": "The text."}' -H "Accept: application/json" "http://localhost:8080/coreference/text"

Discourse Simplification

Discourse Simplification will recursively simplify the given input text and identify rhetorical relations, but will NOT extract subject-predicate-argument relations. The Discourse Simplification service accepts the following endpoints with parameters:

Endpoint: /discourseSimplification/text

Returns the JSON-serialized versions of the DiscourseSimplificationContent class or a textual representation depending on the value of the Accept header in the POST request:

  • Accept: text/plain: returns a textual representation specified by the format parameter

  • Accept: application/json: returns JSON-serialized versions of DiscourseSimplificationContent

  • text: The input text (mandatory).

  • doCoreference: Specifies whether coreference should be executed before Discourse-Simplification (true/false).

  • isolateSentences: Specifies whether the sentences from the input text should be processed individually (This will not extract relationships that occur between neighboured sentences). Set true, if you run Graphene over a collection of independent sentences and false for a full coherent text (true/false).

  • format: Specifies which textual representation should be returned:

    • DEFAULT: The default RDFNL format
    • DEFAULT_RESOLVED: The resolved default RDFNL format
    • FLAT: The flat RDFNL format
    • FLAT_RESSOLVED: The flat resolved RDFNL format

Examples:

curl -X POST -H "Content-Type: application/json"  -d '{"text": "The text.", "doCoreference": "true", "isolateSentences": "false", "format": "DEFAULT"}' -H "Accept: text/plain" "http://localhost:8080/discourseSimplification/text"
curl -X POST -H "Content-Type: application/json"  -d '{"text": "The text.", "doCoreference": "true", "isolateSentences": "false"}' -H "Accept: application/json" "http://localhost:8080/discourseSimplification/text"

Open Relation Extraction

Open Relation Extraction will recursively simplify the given input text and identify rhetorical relations, as well as subject-predicate-argument relations. The Open Relation Extraction service accepts the following endpoints with parameters:

Endpoint: /relationExtraction/text

Returns the JSON-serialized versions of the RelationExtractionContent class or a textual representation depending on the value of the Accept header in the POST request:

  • Accept: text/plain: returns a textual representation specified by the format parameter

  • Accept: application/json: returns JSON-serialized versions of RelationExtractionContent

  • text: The input text (mandatory).

  • doCoreference: Specifies whether coreference should be executed before Relation-Extraction (true/false).

  • isolateSentences: Specifies whether the sentences from the input text should be processed individually (This will not extract relationships that occur between neighboured sentences). Set true, if you run Graphene over a collection of independent sentences and false for a full coherent text (true/false).

  • format: Specifies which textual representation should be returned:

    • DEFAULT: The default RDFNL format
    • DEFAULT_RESOLVED: The resolved default RDFNL format
    • FLAT: The flat RDFNL format
    • FLAT_RESSOLVED: The flat resolved RDFNL format
    • RDF: The RDF N-Triples format

Examples:

curl -X POST -H "Content-Type: application/json"  -d '{"text": "The text.", "doCoreference": "true", "isolateSentences": "false", "format": "DEFAULT"}' -H "Accept: text/plain" "http://localhost:8080/relationExtraction/text"
curl -X POST -H "Content-Type: application/json"  -d '{"text": "The text.", "doCoreference": "true", "isolateSentences": "false"}' -H "Accept: application/json" "http://localhost:8080/relationExtraction/text"