#!/usr/bin/python

import csv
from itertools import islice
import optparse
from rdflib.sparql.sparqlGraph import SPARQLGraph as Graph
from rdflib.sparql import GraphPattern
from rdflib import URIRef, Literal, BNode, Namespace
from rdflib import RDF, RDFS

from gasuse import DC, DATE, GAS, DOLLAR, MILE, GALLON
from gasuse import dateFromAmerican

def newURI(store, base):
    """find an unused uri with a name like base1 base2, etc"""
    suffix = ""
    while 1:
        uri = URIRef("%s%s" % (base, suffix))
        if not list(store.predicate_objects(uri)):
            break
        if suffix == "":
            suffix = 1
        else:
            suffix = suffix + 1
    return URIRef("%s%s" % (base, suffix))

def addOrGetStation(store, shortName, shortAddress):
    results = store.query("?station", GraphPattern([
        ("?station", GAS['shortName'], Literal(shortName)),
        ("?station", GAS['shortAddress'], Literal(shortAddress))]))

    if len(results) > 1:
        raise ValueError("found multiple stations with same name and address")
    if len(results) == 1:
        return results[0]
        
    station = newURI(store, GAS['station/' + shortName.replace(' ','')])
    for p,o in [
        (RDFS.Class, GAS['station']),
        (GAS['shortName'], Literal(shortName)),
        (GAS['shortAddress'], Literal(shortAddress)),
        (RDFS.label, Literal("%s at %s" % (shortName, shortAddress)))]:
        store.add((station, p, o))
    return station

def addGasPrice(store, date, dollars, station):
    gasPrice = BNode()
    for p, o in [ 
        (DC['date'], date),
        (GAS['pricePerGallon'], Literal(dollars, datatype=DOLLAR)),
        (GAS['gasStation'], station),
        ]:
        store.add((gasPrice, p, o))
    return gasPrice

parser = optparse.OptionParser(usage="""%prog input.csv
First row will be treated as headers.""")
opts, args = parser.parse_args()
if len(args)<1:
    parser.error("You need a file")

store = Graph()
dialect = csv.Sniffer().sniff("\n".join(list(islice(open(args[0]), 10))))

reader = csv.reader(open(args[0]), dialect)
headers = reader.next()
for row in reader:
    record = dict(zip(headers, [unicode(f, "latin") for f in row]))

    if record["gallons (+adj)"] == "":
        continue

    date = dateFromAmerican(record["date"])
    station = addOrGetStation(store, record["station"], record["addr"])
    gasPrice = addGasPrice(store, date, record["price gal"], station)
    
    fillUp = BNode()
    for p,o in [
        (RDFS.Class, GAS['fillUp']),
        (DC['date'], date),
        (GAS['notes'], Literal(record["notes"])),
        (GAS['gallons'], Literal(record["gallons (+adj)"], datatype=GALLON)),
        (GAS['fee'], Literal(record["fee"], datatype=DOLLAR)),
        (GAS['tripMeter'], Literal(record["miles (+adjs)"], datatype=MILE)),
        (GAS['odometer'],
         Literal(record["mileage"] or record["miles real"], datatype=MILE)),
        (GAS['gasStation'], station),
        (GAS['gasPrice'], gasPrice),
        (GAS['car'], GAS['car/drewHonda']),
        ]:
        if isinstance(o, basestring) and o == "":
            continue
        store.add((fillUp, p, o))

store.save("gas.rdf", format="pretty-xml")