Skip to content

Latest commit

 

History

History
97 lines (84 loc) · 3.84 KB

README.md

File metadata and controls

97 lines (84 loc) · 3.84 KB
Raw

Ultimate Dictionary API

A Spring Boot app that exposes a dictionary API for definitions and translations for a wide variety of languages. It supports translations for each language pair.

API

The main API is /translation/{source_lang_code}/{target_lang_code}/{word}. I currently have an instance running on http://116.202.96.240:8080/ (I can't guarantee 100 percent uptime though, and it might change without warning).

It returns a JSON object in the form that contains 3 keys:

  • entries, which contains a list of all entry taken from Wiktionary, such as IPA, part of speech, definitions, examples, ...
  • posTranslations, which contains IPA + part of speech, a number of translations, but no detailed definitions or examples
  • translations, a list of unstructured translations taken from the translation boxes of all Wiktionaries

For example, a request to /translation/es/en/estimar will return

{
    "entries": [
        {
            "word": "estimar",
            "ipas": [
                "/estiˈmaɾ/",
                "[es.t̪iˈmaɾ]"
            ],
            "etymology": "Borrowed from Latin aestimāre.",
            "pos": "verb",
            "senses": [
                {
                    "glosses": [
                        "to esteem"
                    ]
                },
                {
                    "glosses": [
                        "to estimate"
                    ]
                },
                {
                    "glosses": [
                        "to think, to believe"
                    ]
                }
            ]
        }
    ],
    "translations": [
        "accuse",
        "appraise",
        "appreciate",
        "assay",
        "assess",
        "enjoy",
        "esteem",
        "estimate",
        "evaluate",
        "fancy",
        "gauge",
        "have a high regard for",
        "hold in (high) esteem",
        "honour",
        "judge",
        "like",
        "love",
        "prize",
        "rate",
        "respect",
        "tax",
        "think highly of",
        "think well of",
        "value"
    ]
}

(The below is not yet running on the VPS and so far only exists in the code:) There exists also a random word api under /random/{source_lang_code}/{target_lang_code}, which will return a random word in a specific source_lang and its associated data from a Wiktionary in a specific target_lang.

Another endpoint is the synonym endpoint under synonyms/{lang_code}/{word}.

Statistics

  • 18 620 000 dictionary entries with all sort of metadata
  • 7 282 354 translations
  • 6 283 different supported languages

Future work

I don't know how much time I will have, but there are a few things that I haven't implemented.

  • Lookup for inflected forms - my VPS currently does not have enough disk space. (Maybe one could also exclude languages with too many inflections, or I need to rework my deployment.)
  • Integrating word frequency data from wordfreq

Data source

The database is populated with extracted Wiktionary data from 6 Wiktionaries, powered by Wiktextract. The dictionary data is dual-licensed to the public under both the Creative Commons Attribution-ShareAlike 3.0 Unported License (CC-BY-SA) and the GNU Free Documentation License (GFDL).

Caveats

Data may be incorrect. This is mainly caused by some nonstandard formatting on Wiktionary that wiktextract doesn't parse correctly. Feel free to report (and fix) possible problems over there. If the problem does not lie with wiktextract, often the translation tables in some Wiktionary editions may be wrong. You can also open an issue here (especially if it is a big problem), I might be able to find the source (no promises).

Contributing

  • Feel free to report bugs, send pull requests and request new features