#Produces the social importance score for webpages in avro files
from avro import schema, datafile, io

OUTFILE_NAME = '/Users/vcedeno/Desktop/webpages/pagesocial.avro'

SCHEMA_STR = """{
    "namespace": "cs5604.tweet.social",
    "type": "record",
    "name": "PageSocial",
    "fields": [
        {   "name": "doc_id"   , "type": "string"   },
        {   "doc": "analysis"    , "name": "social_importance", "type": ["double", "null"]}
        ]
}"""
SCHEMA = schema.parse(SCHEMA_STR)

def makeObject(doc_id, social_importance):
    return {'doc_id': doc_id, 'social_importance': social_importance}

def write_avro_file():
	file = open("/Users/vcedeno/Desktop/webpages/webpages.txt",'r')
	rec_writer = io.DatumWriter(SCHEMA)
	df_writer = datafile.DataFileWriter(open(OUTFILE_NAME, 'wb'),rec_writer,writers_schema = SCHEMA,codec = 'deflate')
	for line in file:
		data=line.split(" ")
		id="shooting_B--webpage--"+data[0]
		df_writer.append(makeObject(id, float(data[1])))
	df_writer.close()
    
if __name__ == '__main__':
    # Write an AVRO file first
    write_avro_file()