#!/usr/bin/env python

"""
Solve an example problem with lp_solve.
"""

from math import acos, sin, cos
import os
import site
import sys

# Basically, add '../src' to our path.
# Needed for the imports that follow.
site.addsitedir(os.path.dirname(os.path.abspath(sys.argv[0])) + '/../../src')

import LinearProgramming


class Facility(object):

    def __init__(self, fac_id, fac_capacity, fac_latitude, fac_longitude):
        self.id = fac_id
        self.capacity = fac_capacity
        self.latitude = fac_latitude
        self.longitude = fac_longitude
        
    def distance(self, other_facility):
        return 3963.0 * acos(sin(self.latitude/57.2958) * sin(other_facility.latitude/57.2958) + cos(self.latitude/57.2958) * cos(other_facility.latitude/57.2958) *  cos(other_facility.longitude/57.2958 - self.longitude/57.2958))


class Consumer(Facility):
    def __init__(self, fac_id, fac_capacity, fac_latitude, fac_longitude):
        super(Consumer, self).__init__(fac_id,
                                       fac_capacity,
                                       fac_latitude,
                                       fac_longitude)

class Producer(Facility):
    def __init__(self, fac_id, fac_capacity, fac_latitude, fac_longitude):
        super(Producer, self).__init__(fac_id,
                                       fac_capacity,
                                       fac_latitude,
                                       fac_longitude)

f = open('midatl.csv', 'r')
f.readline() # Skip the header

producers = []
consumers = []

for line in f:
    row = line.split(',')
    fac_id = int(row[0])
    fac_type = row[1]
    fac_capacity = float(row[2])
    fac_latitude = float(row[3])
    fac_longitude = float(row[4])
    
    if fac_type == '"D"':
        c = Consumer(fac_id, fac_capacity, fac_latitude, fac_longitude)
        consumers.append(c)
    else:
        p = Producer(fac_id, fac_capacity, fac_latitude, fac_longitude)
        producers.append(p)



lp = LinearProgramming.LinearProgram()

# Loop through the consumer/producer pairs, calculating costs
# (distances) and adding them to the array of objective function
# coefficients.
lp.objective_coefficients = []

for c_idx in range(0, len(consumers)):
    for p_idx in range(0, len(producers)):
        distance = consumers[c_idx].distance(producers[p_idx])
        lp.objective_coefficients.append(distance)


num_cols = len(lp.objective_coefficients)

lp.constraint_matrix = []
lp.inequalities = []
lp.rhs = []

for c_idx in range(0, len(consumers)):
    demand = consumers[c_idx].capacity
    lp.rhs.append(demand)
    
    lp.inequalities.append(LinearProgramming.GE)

    demand_row = [0]*num_cols
    offset = c_idx * len(producers)
    for idx in range(0, len(producers)):
        demand_row[offset+idx] = 1

    lp.constraint_matrix.append(demand_row)

for p_idx in range(0, len(producers)):
    supply = producers[p_idx].capacity
    lp.rhs.append(supply)

    lp.inequalities.append(LinearProgramming.LE)
    
    supply_row = [0]*num_cols
    
    for idx in range(0,len(consumers)):
        offset = p_idx + len(producers)*idx
        supply_row[offset] = 1
    lp.constraint_matrix.append(supply_row)

lp.type = LinearProgramming.MINIMIZE

[v,x,duals] = lp.solve()

print 'Optimal objective function value: ', v
print 'Optimal solution vector: ', x