[nifconverter] Make TIKA converter a filter

[jnehring]

In #75 @fsasaki suggested to use the TIKA nifconverter without a pipeline. So with a single API call with informat=TIKAFile can convert e.g. PDF to Turtle and enrich it.

This will require us to write a filter that becomes active whenever informat=TIKAFile.

So this API request should be possible:

http://{{baseUrl}}/e-entity/freme-ner/documents?language=en&dataset=dbpedia&mode=all&nif-version=2.1&informat=TIKAFile&filename=test.pdf

In one API request the system converts the PDF to Turtle and runs FREME NER.

The filter can do a HTTP request to call the nif-converter controller to convert the document.

[bgrusdt]

This is finished now, see the code.

It does not work in combination with the url filter, unfortunately i couldn't find out why it does not work.

[jnehring]

Does not work for me.

request:

curl -X POST -H "Cache-Control: no-cache" -H "Postman-Token: 18ea10b6-fdf9-e597-5066-ed7fb86fefbe" "http://rv1443.1blu.de/current/e-entity/freme-ner/documents?language=en&dataset=dbpedia&mode=all&nif-version=2.1&informat=TIKAFile&filename=test.pptx"

response:

{
  "exception": "eu.freme.common.exception.BadRequestException",
  "path": "/e-entity/freme-ner/documents",
  "message": "parameter informat has invalid value \"TIKAFile\". Please use one of the registered serialization format values: rdf-xml, application/xml, application/ld+json, n-triples, application/x-turtle, application/json, turtle, n3, application/rdf+xml, text/xml, text/turtle, text/html, application/json+ld, xml, json, html, text, text/n3, application/x-openoffice, json-ld, application/n-triples, text/plain, application/x-xliff+xml",
  "error": "Bad Request",
  "status": 400,
  "timestamp": 1484665071834
}