"""File variables_swissmetro.py

Michel Bierlaire, EPFL
Mon Aug 04 2025, 10:05:55

"""

import pandas as pd
from biogeme.database import Database
from biogeme.expressions import Variable

df = pd.read_csv('swissmetro.dat', sep='\t')
database = Database('swissmetro', df)

CHOICE = Variable('CHOICE')
AGE = Variable('AGE')
MALE = Variable('MALE')
GA = Variable('GA')
FIRST = Variable('FIRST')
SM_CO = Variable('SM_CO')
TRAIN_CO = Variable('TRAIN_CO')
TRAIN_TT = Variable('TRAIN_TT')
SM_TT = Variable('SM_TT')
CAR_TT = Variable('CAR_TT')
CAR_CO = Variable('CAR_CO')
TRAIN_HE = Variable('TRAIN_HE')
SM_HE = Variable('SM_HE')
TRAIN_AV = Variable('TRAIN_AV')
SM_AV = Variable('SM_AV')
CAR_AV = Variable('CAR_AV')
PURPOSE = Variable('PURPOSE')

# Definition of new variables
SM_COST = SM_CO * (GA == 0)
TRAIN_COST = TRAIN_CO * (GA == 0)
TRAIN_TT_SCALED = TRAIN_TT / 60
TRAIN_COST_SCALED = TRAIN_COST / 100
SM_TT_SCALED = SM_TT / 60
SM_COST_SCALED = SM_COST / 100
CAR_TT_SCALED = CAR_TT / 60
CAR_COST_SCALED = CAR_CO / 100
TRAIN_HE_SCALED = TRAIN_HE / 60
SM_HE_SCALED = SM_HE / 60


# Removing some observations
# exclude = CHOICE == 0
exclude = ((PURPOSE != 1) * (PURPOSE != 3) + (CHOICE == 0)) > 0
database.remove(exclude)