date parsing

parse_date_files.py

+import json, os, sys
+from datetime import datetime
+
+"""
+Uso:
+python parse_date_files.py <nombre_del_crawler>
+
+Ej.
+python parse_date_files.py laJornadaBC2
+"""
+ 	
+this_file_path = os.path.dirname(os.path.realpath(__file__))
+json_file_path = this_file_path+'/crawlersNoticias/descarga_hacia_atras/'+sys.argv[1]
+destination_path = this_file_path+'/'+sys.argv[1]
+
+json_file = json.loads(open(json_file_path+'/noticias.json').read())
+date_set = set()
+
+for news in json_file:
+	if news['date'] is not None:
+		news_date = news['date'][:news['date'].rfind('T')]
+		
+		if not news_date in date_set:
+			date_set.add(news_date)
+			print(news_date)
+
+			try:
+				export_file = open(destination_path+'/'+news_date[:4]+'/'+news_date+'.json', 'a')
+			except:
+				os.makedirs(destination_path+'/'+news_date[:4])
+				export_file = open(destination_path+'/'+news_date[:4]+'/'+news_date+'.json', 'a')
+			
+			counter = 0
+			export_file.write('[\n')
+			
+			for line in json_file:
+				if line['date'] is not None:
+					line_date = line['date'][:line['date'].rfind('T')]
+
+					if line_date == news_date:
+						counter += 1
+						
+						if counter == 1:
+							export_file.write(json.dumps(line))
+						elif counter > 1:
+							export_file.write(','+'\n'+json.dumps(line))
+				
+			export_file.write('\n]')
+			export_file.close()
+