date_comparison.py

import os, sys
currentdir = os.path.dirname(os.path.abspath(__file__))
parentdir = os.path.dirname(currentdir)
sys.path.insert(0, parentdir)

import wikilib
import parsing_dates
import apiRequest
import parseToDB
import sys

#name = 'Battle_of_Kulikovo'
#name = 'Battle_of_Stalingrad'
infoboxType = sys.argv[1]
for line in sys.stdin:
	try:
		name = line.strip()

		page = apiRequest.getPage(name)
		infoboxData = parseToDB.infobox_parser(infoboxType)(wikilib.get_infobox(page))
		if 'date' not in infoboxData:
			print('Missing `date` field in infobox for article', name, file = fw)
			continue
		our_parse_date = infoboxData['date']
		date_from_wikidata = apiRequest.getWikidataDate(apiRequest.getWikidataProperties(apiRequest.getWikidataID(name)))

		message = ''
		if our_parse_date == ([None, None, None], [None, None, None]):
			message = 'Error - We can not parse this date'
		elif (date_from_wikidata is None) or (date_from_wikidata == ([None, None, None], [None, None, None])):
			message = 'Error - Wikidata has not this date. We can add our date'
		elif our_parse_date == date_from_wikidata:
			message = 'Ok'
		elif our_parse_date != date_from_wikidata:
			message = 'Error - Our date does not coincide with date from Wikidata'

		fw = sys.stdout
		#print('infobox ->', infoboxData, file = fw)
		print('Name:', name, file = fw)
		print('Infobox:', file = fw)
		for (key, value) in infoboxData.items():
			print(key, value, file = fw)
		print('Our date:', our_parse_date, file = fw)
		print('Date from Wikidata:', date_from_wikidata, file = fw)

		print('Result:', message, file = fw)
		print('-------------------------------------', file = fw)
	except Exception as e:
		print('Exception:', e, file = fw)