Overview

Language Support

The languages are used in some concepts in the TextReveal API. Some languages have different implementation level.

Analyze

The TQL and Dataset routes accept both the language name and the language code for the language parameter.

For instance, to request documents in English you can use either english or eng as the language parameter.

You can also use the value all to search documents that match every language.

Workers

Some worker are available only for certain languages quality-score, ner-linking, embedder-indicators and entity-similarity. See the table below for the list of languages that support these workers.

The raw-matcher and concept workers are available on all our languages (see the table of languages in the Analyze section).

If a language is not present in the table, it means that only the basic workers are available for that language.

NameCodequality-scorener-linkingembedder-indicatorsentity-similarity
albaniansqi
arabicara
chinese (simplified)zho
czechcze
danishdan
dutchnld
englisheng
finnishfin
frenchfre
germandeu
hindihin
hungarianhun
italianita
japanesejpn
kazakhkaz
koreankor
kyrgyzkir
norwegiannor
polishpol
portuguesepor
romanianron
russianrus
serbo-croatianbos hrv srp
slovakslk
spanishspa
swedishswe
thaitha

If you use a language that support a worker along with a language that does not support it:

  • The qscore parameter (filtering feature) won't be available for both the /analyze/tql and /analyze/dataset routes. Only the annotation feature is supported using qscore: 0 or by removing the qscore parameter.
  • If the quality-score worker is selected, the Quality Score values will be returned only for the supported languages. For the other languages, null values will be returned.
  • If embedder-indicators worker is selected, indicators will be returned only for the supported languages, for the other languages null values will be returned.

Translation

The /batch/translate route accept the language code as the language parameter.