Language Services

Natural Language Processing Centre provides several APIs for different language services.

How to use the API

You can call the API via GET requests. See examples in the service description.

Parameters

Required parameters: call=service name, text=text

Optional parameters: output=output (text/JSON applicable only for some services), lang=language (applicable only for some services)

Service-specific parameters: see service description

Returns

If output is set to json, returns always a JSON object. See example calls to check the object structure. If output set to text, outputs a text. The default format is text for most services.

Rate limits

You can call the API once a second maximum from one IP address.

The daily limit is set to 500 calls.

Available languages

language	service name
cs	Morphological analysis of words
	Intensional logical analysis
	Get string with diacritics added
	Inflection of words
	Get topics of a text
	Extract (sub)phrases
	Get politeness service
	Get vocative for a name
	Shallow ontology for Czech words
	Word forms generator
	Get location service
	Declension of noun phrases
	Get lemmata and tags for tokens
en	Hello Service
	Get lemmata and tags for tokens

Available services

service name	description	parameters	returns	example call	text length limit
sholva Shallow ontology for Czech words	This service returns negative or positive membership of the word in the following classes: person, person-individual, event, substance.	text type string desc word	concepts type list desc The output format is a list of positive or negative memberships to some (or all) of the classes.	service.py?call=sholva&lang=cs&output=json&text=šampaňské	1000
topics Get topics of a text	This service returns a list of topics found in a short text. The topic candidates are noun phrases ordered by importance.	text type string desc text	response type list desc list of noun phrases ordered by score	service.py?call=topics&lang=cs&output=json&text=Jednoho letního večera na návsi pod starou lípou hostinský Antonín Kučera vyvalil soudeček s pípou, nebylo to posvícení, nebyla to neděle, v naší obci mezi kopci plnily se korbele. Byl to ten slavný den, kdy k nám byl zaveden elektrický proud, Byl to ten slavný den, kdy k nám byl zaveden elektrický proud, střídavý, střídavý, silný elektrický proud, střídavý, střídavý, zkrátka elektrický proud.	1000
phrases Extract (sub)phrases	This service processes an input sentence and returns all phrases identified in the text (and their tags).	text type string desc text	phrases type list desc The list of phrases, i.e. lists of phrase text and tag.	service.py?call=phrases&lang=cs&output=json&text=Konečně levný mobil - nástupce kompaktního modelu má být velkým lákadlem	1000
vocative Get vocative for a name	This service returns vocative case for a name plus gender if possible.	text type string desc a person name (first names, surnames)	gender type string desc gender of the name (m, f, or nothing) last type string desc last name name type string desc the name in vocative case first type string desc first name	service.py?call=vocative&lang=cs&output=json&text=Jan Petr Obr	200
gen Word forms generator	For the input word (lemma, base form) returns all inflected forms with tags (tagset reference is here or here)	text type string desc Input word (lemma, base form)	forms type string desc Input word (just to be sure it works) and pairs wordform:tag	service.py?call=gen&lang=cs&output=json&text=slovo	40
majka Morphological analysis of words	This service returns all lemmata and morphological tags for each of the input words (comma separated).	text type string desc list of words (comma separated)	Word type string desc one input word Annot type list desc lemma and tag annotations	service.py?call=majka&lang=cs&output=json&text=cestou,jedem	1000
declension Declension of noun phrases	This service returns a the input phrase in a different case and/or number.	text type string desc the noun phrase you want to inflect c1 type string desc input case: c1-nominative, c2-genitive, c3-dative, c4-accusative, c6-locative, c7-instrumental n type string desc output number: nS-singular, nP-plural, empty-do not change number c2 type string desc output case: c1-nominative, c2-genitive, c3-dative, c4-accusative, c6-locative, c7-instrumental	response type string desc the input phrase in output case and number	service.py?call=declension&lang=cs&output=json&text=velký růžový slon&c1=c1&c2=c4&n=nP	500
diacritics Get string with diacritics added	This service returns the phrase with the most probable diacritics added.	text type string desc text without diacritics	text type string desc text with diacritics	service.py?call=diacritics&lang=cs&output=json&text=dama v treti rade	1000
polite Get politeness service	This service returns politeness level of texts.	text type string desc text	politeness type string desc politeness level: one of polite, rude, veryrude rudewords type list desc list of rude words identified in the text	service.py?call=polite&lang=cs&output=json&text=Tento text obsahuje nepěkné slovo kretén.	1000
get_location Get location service	This service returns a list of locations incl. nominatives contained in the text.	text type string desc text in which are location names	tokens type list desc list of tokens places_nom type list desc list of place names converted to nominative case places type list desc list of place names	service.py?call=get_location&lang=cs&output=json&text=V létě pojedeme do New Yorku, Paříže a Loučné nad Desnou.	1000
logic Intensional logical analysis	This service processes an input sentence and returns logical construction in Transparent Intensional Logic of the first possible syntactic tree.	text type string desc text	construction type list desc The list of constructions	service.py?call=logic&lang=cs&output=json&text=Ministr dopravy ČR navštívil Ostravsko.	1000
tagger Get lemmata and tags for tokens	This service splits the text into sentences, split the sentences into tokens, and determines lemma and tag for each token.	lang type string desc language code text type string desc text	vertical type list desc The vertical format is a list of lists, each item in the list is either a structural token (1-item list) such as sentence start or end, or a 3-item list [word, lemma, tag].	service.py?call=tagger&lang=cs&output=json&text=Na okně seděla kočka, byl krásný letní den. Na okně seděla kočka a koukala se ven.	100000
inflection Inflection of words	This service returns all input words in the specified form (comma separated list of tag-values).	text type string desc list of lemmata (comma separated) tag type string desc list of tag-values (comma separated)	Lemma type string desc one input lemma Annot type list desc word and tag annotations	service.py?call=inflection&lang=cs&output=json&text=cesta,jed&tag=c7,nS	1000
hello Hello Service	This service returns a greeting. It's practically useless since it is only an example.	name type string desc the person you want to greet or 'World' if not provided	response type string desc a greeting service type string desc service name	service.py?call=hello&lang=en&output=json&text=Dolly	40

Licence

All services are provided under the NLP Centre Web Service Licence.

How to add a service

1. Find a unique name <service_name> for the call.
2. Create <service_name.json> with the description in the languageservices/json directory. See hello.json as an example.
3. Add limit to text length (number of characters).
4. Test it.