original: http://nlp.fi.muni.cz/trac/deb2/wiki/WordNetApi

[edit] API for Wordnet/Cornetto server

Replace servername with actual address of the server.

For more info about JSON, see http://www.json.org/

[edit] Web interface

Open servername/dictionary?action=listPreview in browser. Accepts the same queries as DEBVisDic.

[edit] Getting the list of available dictionaries for the user

Request: servername/doc?action=init

Response: JSON encoded hash with following structure

{
 "slovniky" => {
   client_package_id => {
     "id"    => client package id,
     "dict"  => dictionary code,
     "prava" => access rights r/w,
     "nazov" => package name,
     "dicslookup" => array of package names to allow lookups to,
     "eqtags" => array of eqtags+dictionaries,
     "dicsreload" => array of dictionaries to refresh,
     "key" => dictionary key tag
   }
 }
}

[edit] Querying dictionary

Request: servername/dictionary_code?action=queryList&word=search_query

search_query::

 search query, must be URL encoded UTF-8

Response: JSON encoded array of hashes with following structure

{
  "value" => entry ID,
  "label" => information about an entry, depends on dictionary, usually contains PoS, literals
}

Query types:

* word - search for entries containing whole word in literals
* word:number - search for entries containg whole word with sense number
* [*]word[*] - search for word substring - * can be put at the beginning/end of the word or at both of them
* xml_element=value - search for value of xml_element, for example SYNSET.POS=n
* xml_element=dictionary_code2:entry_id2:xml_element2 - search for xml_element with value from xml_element2 of entry_id2 in dictionary_code2

[edit] Display entry

Request: servername/dictionary_code?action=runQuery&query=entry_id&outtype=output_type[&arg=argument][&nojson=1][&default=1]

entry_id::
 ID of displayed entry
output_type::
 output type of entry, possible values:
* plain - plain form of XML entry from database
* xml - coloured, foldable XML representation for display in browser
* html - HTML representation of entry, use default XSLT template, it's possible to specify other template with optional arg argument
* tree - array representation of tree, arg specifies tree direction - ILR for normal tree, RILR for reversed tree
* editor - hash representation of XML for editor tab
arg::
 depends on output_type. For html optionally specifies other than default template. For tree specifies tree direction.
nojson::
 by default response is JSON encoded, by setting nojson=1 response is not encoded
default::
 optional argument for html output type - with default=1 HTML is displayed with default styles, otherwise style set in client is used

Response: by default JSON encoded, depends on output type

[edit] Display subtree of an entry

Request: servername/dictionary_code?action=subtree&query=entry_id&arg=argument

entry_id::
 ID of displayed entry
arg::
 specifies tree direction - ILR for normal tree, RILR for reversed tree

Response: JSON encoded array with tree structure (same as for basic tree in display) - only direct descendants of entry are returned

[edit] Display full subtree of an entry

Request: servername/dictionary_code?action=subtree&query=entry_id&arg=argument

entry_id::
 ID of displayed entry
arg::
 specifies tree direction - ILR for normal tree, RILR for reversed tree

Response: JSON encoded array with tree structure (same as for basic tree in display) - display all descendants of entry

[edit] Save entry

Request: servername/dictionary_code?action=save&id=entry_id&data=encoded data

entry_id::
 ID of saved entry
encoded data::
 JSON and URL UTF-8 encoded hash representation of an entry (same as returned for editor)

[edit] Delete entry

Request: servername/dictionary_code?action=delete&id=entry_id

entry_id::
 ID of entry