Ġabra API v2.13.0 Updated 2020-07-04

This site is aimed at developers who want to access the Ġabra database programmatically, such as from a website or app. If you just want to search the database normally, you should do so from the Ġabra website.

General

Base URL: https://mlrs.research.um.edu.mt/resources/gabra-api
Everything should be UTF-8 encoded.
The API is sensitive to Unicode characters. This means that h will not match ħ.
Input arguments when searching are lower-cased, but results are not completely case-agnostic.
Searching in wordforms is implicitly a prefix search.
The order in the search results does not follow the Maltese alphabet correctly — the characters ċ, ġ, ħ, ż are sorted after z. Also, lemmas beginning with upper-case appear before lower-cased ones.
All fields are optional unless marked as required.

Operations

Search lexemes

This returns the lexeme entries (although it searches in wordforms too by default).

/lexemes/search?s=ktibt

Argument	Description	Example	Notes
`s`	Search query	`kiteb`
`l`	Search in lemmas?	`0` or `1` (default)	Mininum length 2
`wf`	Search in wordforms?	`0` or `1` (default)	Mininum length 3; prefix matching only
`g`	Search in gloss?	`0` or `1` (default)	Mininum length 3
`pos`	Part of speech	`NOUN` (see here)
`source`	Source	`Ellul2013`
`pending`	Include pending entries?	`0` (default) or `1`
`page`	Page number	`1` (default), `2`, ...

Paging

The results are paged. This works as follows:

A call will return up to query.page_size results (currently 20). The page size does not change.
To see how many results were actually returned, use results.length.
The total number of results from the database will be given in query.result_count. Use this to decide if you need to fetch further pages: query.page * query.page_size >= query.result_count
The first page of results is numbered 1. Anything below 1 in the page parameter will give you an error (HTTP 400).
There is no upper limit on the page number, but after the last page you will return back an empty results field.

Search lexeme glosses Since v2.9

Unlike /lexemes/search, this command uses a special text index on the glosses to return results which are sorted based on text-closeness. It also performs stemming and stop-word removal on your query. If only searching by English gloss, this command will return better results. For more flexibility and additional filtering, use the command above (with the g flag).

/lexemes/search_gloss?s=find

Argument	Description	Example	Notes
`s` (required)	Search query	`find`	Should be in English

Paging should work as above.

Load lexeme

Load a lexeme from its ID.

/lexemes/5200a366e36f237975000f26

Argument	Description	Example
`:id` (required)	Lexeme ID	`5200a366e36f237975000f26`

Load wordforms

Load the wordforms for a particular lexeme.

/lexemes/wordforms/5200a366e36f237975000f26

Argument	Description	Example
`:id` (required)	Lexeme ID	`5200a366e36f237975000f26`
`match`	Limit to surface forms matching substring	`kom`
`pending`	Include pending entries?	`0` (default) or `1`
`exclude_sources`	Exclude wordforms from given sources	`Camilleri2013,Apertium2014`
`sort`	Sort wordforms	`0` or `1` (default)

Load lexemes which are related to a given lexeme. Currently this only uses root information, but may be extended to include other relations. The results are sorted by part of speech and derived form, and will not include the original lexeme.

/lexemes/related/5200a366e36f237975000f26

Argument	Description	Example
`:id` (required)	Lexeme ID	`5200a366e36f237975000f26`

Search suggest Changed in v2.13

Find matching words which vary in spelling (diacritics, character case) of the search term, from lemmas or wordforms:

/lexemes/search_suggest?s=Hareg

/wordforms/search_suggest?s=ohorgu

Argument	Description	Example
`s` (required)	Search query	`Hareg`

Lemmatise

Similar to lexemes/search but also returns matching wordform info.

/lexemes/lemmatise?s=ktibniehom

Argument	Description	Example
`s` (required)	Surface form (regex)	`ktibnie?hom`

Search for root

Example:

/roots/search?s=k-t-b

/roots/search?c1=k&c3=b

Argument	Description	Example
`s` (required)	Search query (regex)	`k-t-b`, `s-r-v-j`, `^b-.-[jw]$`
`c1`,`c2`,`c3`,`c4`	Radicals at positions 1–4 (overrides `s`)	`għ`
`r`	Search in radicals?	`0` or `1` (default)
`l`	Search in lemma?	`0` or `1` (default)
`g`	Search in gloss?	`0` or `1` (default)
`t`	Root type	`strong`, `geminated`, `weak-initial`, `weak-medial`, `weak-final`, `irregular`
`page`	Page number	`1` (default), `2`, ...

All radicals should be in lower case. A dot . will also match the digraph għ.

The results are paged.

Load root Since v2.10.1

Load a root from its radicals and variant.

/roots/k-t-b

/roots/b-r-d/2

Argument	Description	Example
`:radicals` (required)	Root radicals	`k-t-b`, `s-r-v-j`
`:variant`	Root variant	`1`, `2`, ...

Load lexemes by root Since v2.5

Load lexemes with a given root. The results are sorted by part of speech and derived form.

/roots/lexemes/k-t-b

/roots/lexemes/b-r-d/2

Argument	Description	Example
`:radicals` (required)	Root radicals	`k-t-b`, `s-r-v-j`
`:variant`	Root variant	`1`, `2`, ...

List all sources Since v2.6

List all sources (from the sources collection).

/sources/

Load source Since v2.4

Load a source from its key.

/sources/Falzon2013

Argument	Description	Example
`:key` (required)	Source key	`Falzon2013`

Internationalisation (i18n) Since v2.3

You can get English & Maltese names for all the field names and values:

/i18n/all?lang=mlt

Argument	Description	Example
`lang`	Language	`eng` (default) or `mlt`

Counts Since v2.10.0

Get counts of all entities in database. These commands take no parameters.

/lexemes/count

/wordforms/count

/roots/count

Random Since v2.11.1

Get a random entry.

/lexemes/random