N-grams are not used?

[zeitrechnen]

Before I installed the German n-gram data set the API responded to a check request with source = +fallback

"language": {
    "name": "German (Germany)",
    "code": "de-DE",
    "detectedLanguage": {
      "name": "German (Germany)",
      "code": "de-DE",
      "confidence": 0.99999934,
      "source": "+fallback"
    }
  }

Then I installed the data set and checked the .env file: Data set is installed and path is correct. After restarting the app I made the same request again and expected "ngram" as the source. At least this is the result when I make a request here: https://languagetool.org/http-api/swagger-ui/#!/default/post_check

Does this mean the n-gram data set is not used in my installation?

[girish]

@zeitrechnen not sure what source means above but you can check with the example strings at https://dev.languagetool.org/finding-errors-using-n-gram-data .

With n-grams:

Bitte prüfen Sie, ob ‚Meditation‘ (spirituelle Übung) hier das richtige Wort ist anstelle von ‚Mediation‘ (Verfahren zur Konfliktlösung).

Without n-grams:

You will see an error "ngram index dir /app/data/ngrams/de not found for German" and also no warning as above

[zeitrechnen]

@girish Many thanks for the link and the example. Then n-gram is used correctly on my server!