====== API ======
Selected applications can be queried via an [[https://en.wikipedia.org/wiki/API|API]]. API access requires user authentication at all times, which holds also for applications where this would not be necessary for their use via the standard user web interface. For the authentication, personal access tokens should be used.
===== Getting your personal access token =====
- Log in to the CNC web portal.
- Go to your [[https://korpus.cz/account|user account]], "Personal access tokens" section. Click on "New", select the requested validity and confirm by "OK".
- Copy the token, it will be displayed only once.
It is possible to create any number of personal access tokens. The tokens can be deleted. After a token expires, it is necessary to create a new one.
===== Use =====
Token is sent in a HTTP request to ''https://korpus.cz/login'' using POST in a ''personal_access_token'' parameter.
A session cookie will be returned that should be used in every other request. It may be useful to save the session cookie and use it repeatedly within the next few hours.
Individual API usage manuals:
* [[https://github.com/czcorpus/kontext/wiki/HTTP-API|KonText]]
* [[https://wiki.korpus.cz/doku.php/manualy:treq:api|Treq]]
The APIs can be queried at the following URLs:
* KonText: ''https://korpus.cz/kontext-api/v0.17''
* Treq: ''https://treq.korpus.cz/api/v1''
Every application also imposes its own **limits**. After exceeding them, there is a risk of blocking access to the API.
* **KonText**: max. 12 requests per minute and max. 5000 requests per day
* **Treq**: max. 30 requests per minute and max. 5000 requests per day
===== Usage examples =====
==== 1) curl and Treq API ====
#!/usr/bin/env bash
# Log in
curl --cookie cookies.txt --cookie-jar cookies.txt -X POST -F 'personal_access_token=0a1b2c3d4e5f6-abc012...' 'https://korpus.cz/login'
# Query Treq
curl --cookie cookies.txt --cookie-jar cookies.txt 'https://treq.korpus.cz/api/v1?from=cs&to=en&multiword=false®ex=true&lemma=true&ci=true&pkgs[0]=SYNDICATE&pkgs[1]=CORE&query=pravda&order=perc&asc=false'
==== 2) Python and KonText API ====
#!/usr/bin/env python3
import pickle, requests
personal_access_token = '0a1b2c3d4e5f6-abc012...'
cookies_file = 'cookies.pickle'
with requests.Session() as s:
# Load cookies
try
with open(cookies_file, 'rb') as f:
s.cookies.update(pickle.load(f))
except FileNotFoundError:
pass
# Log in
r = s.post('https://korpus.cz/login', data={'personal_access_token': personal_access_token})
# Creating a concordance query
request_body = {
"type": "concQueryArgs",
"maincorp": "syn2015",
"usesubcorp": None,
"viewmode": "kwic",
"pagesize": 40,
"attrs": ["word","tag"],
"attr_vmode": "visible-kwic",
"base_viewattr": "word",
"ctxattrs": [],
"structs": ["text","p","g"],
"refs": [],
"fromp": 0,
"shuffle": 0,
"queries": [
{
"qtype": "advanced",
"corpname": "syn2015",
"query": "[word=\"celou\"] [lemma=\"pravda\"]",
"pcq_pos_neg": "pos",
"include_empty": False,
"default_attr": "word"
}
],
"text_types": {},
"context":
{
"fc_lemword_wsize": [-5, 5],
"fc_lemword": "",
"fc_lemword_type": "all",
"fc_pos_wsize": [-5, 5],
"fc_pos": [],
"fc_pos_type": "all"
},
"async": False
}
r = s.post('https://korpus.cz/kontext-api/v0.17/query_submit', params={'format': 'json'}, json=request_body)
response_json = r.json()
print(response_json)
# Displaying a concordance
conc_persistence_op_id = response_json['conc_persistence_op_id']
r = s.get('https://korpus.cz/kontext-api/v0.17/view', params={'format': 'json', 'q': '~'+conc_persistence_op_id})
print(r.json())
# Store cookies
with open(cookies_file, 'wb') as f:
pickle.dump(s.cookies, f)