xjtei.md

The XJ-TEI format

This document describe the JSON schema that is used to describe entries from the nb+nn dictionary. The format is inspired by TEI Dictionary XML.

The dictionary is a collection of entries that have the following structure:

{
    "id": 99999,
    "lang": "nn",
    "lemmas": ["foo"],
    "pos": "v",
    "pos2": "v1",
    "forms": [["fooe", "fooer", "fooet", "fooet"]],
    "etym": {
        "langs": [
            {
                "lang": "latin",
                "intro": "...",
                "text": "..."
            }
        ],
        "cits": [
            {
                "quote": "..."
            }
        ]
    },
    "senses": [
        {
            "n": "1",
            "defs": [
                {
                    "def": "..."
                }
            ],
            "cits": [
                {
                    "quote": "...",
                    "usg": "..."
                },
            ],
            "sub-defs": [
                {
                    "def": "..."
                }
            ],
            "sub-cits": [
                {
                    "quote": "...",
                    "usg": "..."
                },
            ],
        }
    ],
    "links": [
        {
            "rel": "compare",
            "target": 99901,
            "text": "..."
        },
        {
            "rel": "related",
            "target": 99904,
            "text": "...",
            "intro": "..."
        }
    ]
}

Description of the elements of the entry object:

• id numeric identification of this entry.

• lang is the ISO 639-1 code for the language the described word belongs to. The descriptions follow the language. The code is nn for Nynorsk and nb for Bokmål.

• lemmas The base form of the word. Note that the same lemma can be used for different entries and that an entry can be indexed by multiple lemmas (when there are multiple ways to write the same word). Ref wikipedia. The field is an array of at least one element. The common case is that there is only one lemma for each entry.

• pos is the word class the word belongs to. It's a string like 'v=verb', 'n=noun', 'a=adjective', 'av=adverb',...

• pos2 the expanded word class code. Given this code it's possible to algorithmically derive the forms from the lemma. It's a string like 'n1', 'n2', 'v1', 'v2', 'm1'

• forms lists how this word is written in its different forms. The interpretation of these lists depend on the pos. For instance for nouns the inner array consist of 4 inflections (singular/plural × indefinitive/definitive form).

• etym describes where where the word came from. Ref wikipedia

• etym.langs describe the origins of the word from other languages

• etym.langs[].lang language name (not ISO 631-1 this time)

• etym.cits describes the origins of the word from literature

• senses describes the meaning of the word and examples of use. These descriptions are grouped together.

• senses[].n This is the label for this group of sense statements. Display can be suppressed when there is only one sense.

• senses[].defs Array containing definitions. Each definition is an object with the attribute def containing the text of the definition.

• senses[].cits Array containing examples of use. Each example of use is an object with the attributes quote and usg.

• senses[].sub-defs Array containing subordinate definitions in the same form as for defs.

• senses[].sub-cits Array containing subordinate examples of use in the same form as for cits.

• links is a list of references to other relevant entries. The rel attribute encodes how this entry relates.

The following attributes repeats at various locations in the structure:

• intro is a short unescaped text string that makes sense to inline in front of a link or the main text

• text is a short HTML fragment (escaped). It's text with <span>...</span> elements. The span elements can have the class attribute set to one of the following values: "wordform",...

• def is a text string (unescaped)

• quote is a text string (unescaped)

• usg is a text string (unescaped)