salgo60/checkLevandemusikarv.py

## checkLevandemusikarv.py
# Check linkroot
# pip install sparqlwrapper
# https://rdflib.github.io/sparqlwrapper/
# see task https://github.com/salgo60/Litteraturbanken_wd_runes/issues/14
# gist https://gist.github.com/salgo60/51d56cc60dd382301cce33bd462f6363
from datetime import datetime
import urllib3
import sys
from SPARQLWrapper import SPARQLWrapper, JSON

endpoint_url = "https://query.wikidata.org/sparql"
#endpoint_url = "https://wcqs-beta.wmflabs.org/sparql"
http = urllib3.PoolManager()

# SPARQL
query = """SELECT ?item ?itemLabel ?LevandeMusik  WHERE {
?item wdt:P31 wd:Q2188189.
?item wdt:P4607 ?id
BIND(URI(CONCAT("http://www.swedishmusicalheritage.com/composers/",?id)) AS ?LevandeMusik)
SERVICE wikibase:label { bd:serviceParam wikibase:language "sv,en". }
}"""


def get_results(endpoint_url, query):
    user_agent = "user:salgo60/%s.%s" % (sys.version_info[0], sys.version_info[1])
    sparql = SPARQLWrapper(endpoint_url, agent=user_agent)
    sparql.setQuery(query)
    sparql.setReturnFormat(JSON)
    return sparql.query().convert()

''' Check if URLS is ok'''
def check(url):
    #print("\t" + url)
    r = http.request('GET', url)
    #print("\t" + url + "\t" + str(r.status))
    if r.status != 200:
        #print("Status: ",r.status, " \t", url)
        return False
    return True
start_time  = datetime.now()
print("Last run: ", start_time)
results = get_results(endpoint_url, query)

ok = 0
notok = 0
print("SPARQL result = ", str(len(results["results"]["bindings"])))
for result in results["results"]["bindings"]:
    #print(result)
    levURL = result["LevandeMusik"]["value"]
    #file = result["file"]["value"]
    if check(levURL):
        ok += 1
    else:
        print("Status not ok: ", levURL, " \t")

        notok += 1
print("OK: ",ok,"\t not ok",notok)
end = datetime.now()
print("Ended: ", end)
print('Time elapsed (hh:mm:ss.ms) {}'.format(datetime.now() - start_time))
	# Check linkroot
	# pip install sparqlwrapper
	# https://rdflib.github.io/sparqlwrapper/
	# see task https://github.com/salgo60/Litteraturbanken_wd_runes/issues/14
	# gist https://gist.github.com/salgo60/51d56cc60dd382301cce33bd462f6363
	from datetime import datetime
	import urllib3
	import sys
	from SPARQLWrapper import SPARQLWrapper, JSON

	endpoint_url = "https://query.wikidata.org/sparql"
	#endpoint_url = "https://wcqs-beta.wmflabs.org/sparql"
	http = urllib3.PoolManager()

	# SPARQL
	query = """SELECT ?item ?itemLabel ?LevandeMusik WHERE {
	?item wdt:P31 wd:Q2188189.
	?item wdt:P4607 ?id
	BIND(URI(CONCAT("http://www.swedishmusicalheritage.com/composers/",?id)) AS ?LevandeMusik)
	SERVICE wikibase:label { bd:serviceParam wikibase:language "sv,en". }
	}"""


	def get_results(endpoint_url, query):
	user_agent = "user:salgo60/%s.%s" % (sys.version_info[0], sys.version_info[1])
	sparql = SPARQLWrapper(endpoint_url, agent=user_agent)
	sparql.setQuery(query)
	sparql.setReturnFormat(JSON)
	return sparql.query().convert()

	''' Check if URLS is ok'''
	def check(url):
	#print("\t" + url)
	r = http.request('GET', url)
	#print("\t" + url + "\t" + str(r.status))
	if r.status != 200:
	#print("Status: ",r.status, " \t", url)
	return False
	return True
	start_time = datetime.now()
	print("Last run: ", start_time)
	results = get_results(endpoint_url, query)

	ok = 0
	notok = 0
	print("SPARQL result = ", str(len(results["results"]["bindings"])))
	for result in results["results"]["bindings"]:
	#print(result)
	levURL = result["LevandeMusik"]["value"]
	#file = result["file"]["value"]
	if check(levURL):
	ok += 1
	else:
	print("Status not ok: ", levURL, " \t")

	notok += 1
	print("OK: ",ok,"\t not ok",notok)
	end = datetime.now()
	print("Ended: ", end)
	print('Time elapsed (hh:mm:ss.ms) {}'.format(datetime.now() - start_time))