########################################################################################
# This script takes a CSV file as input with data at the period-location-specific level.
# From it, it outputs the same data in a json file that can be imported into MapVisualizer on my website.
#
# The input file should have the following format (one and only one row for each period-location):
#	Week,DEP,Var1,Var2,Var3
#	Week1,01,3,2,6
#	Week1,2A,2,5,4
#	Week2,01,4,8,2
#	Week2,2A,1,3,6
#
# The id's for period and for geographic entity are treated as strings, but the values are converted to float,
# thus they should be numeric.
#
# The output file looks as follows:
# {
# 	"periodIdField" :	"Week",
# 	"geoIdField" :		"DEP",
# 	"periodIds" :		["Week1", "Week2"],
# 	"geoIds" :			["01", "2A"],
# 	"VariableNames":	["Var1", "Var2", "Var3"],
# 	"values":			[[[3,2,6], [2,5,4]], [[4,8,2], [1,3,6]]]
# }
# 
# Example use:
#	python convertCsv2Json.py input.csv output.json
########################################################################################

import sys, json
from file_access import readDataFromFile

# Read input arguments
[inputFile, outputFile] = sys.argv[1:3]

# Get unique set of geoId's and unique set of periodId's and geoId's
[periodIds, geoIds] = readDataFromFile(inputFile, ',', True, [1, 2])
periodIds = sorted(list(set(periodIds)))
geoIds = sorted(list(set(geoIds)))

#### Read file line by line
fin = open(inputFile, 'r')

# Get variable names from headers
line = fin.readline()
headers = line.rstrip().split(',')
periodIdField = headers[0]
geoIdField = headers[1]
NumVars = len(headers) - 2
varNames = headers[2:]

values = [[ [] for gg in geoIds  ] for pp in periodIds]

# Read the other values
line = fin.readline()
while line:
    elements = line.rstrip().split(',')
    periodId = elements[0]
    geoId = elements[1]
    # Convert strings to numeric
    recValues = elements[2:]
    recValues = [float(vv) for vv in recValues]
    
    # 
    periodIdx	= periodIds.index(periodId)
    geoIdx		= geoIds.index(geoId)
    values[periodIdx][geoIdx] = recValues
    
    # Move to next line
    line = fin.readline()

fin.close()

# Put everything togeter
res = {}
res['periodIdField']	= periodIdField
res['geoIdField']		= geoIdField
res['periodIds']		= periodIds
res['geoIds']			= geoIds
res['values']			= values
res['VariableNames']	= varNames

with open(outputFile, "w") as myfile:
    json.dump(res, myfile)