#!/usr/bin/env python
# -*- coding: utf-8 -*-
#
# Copyright (C) 2012-2013 Tobias Klauser <tklauser@distanz.ch>
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License version 2 as
# published by the Free Software Foundation.

import os, sys
import getopt
import csv
import numpy as np

DEFAULT_NAME = 'control'
DEFAULT_MAXTIME = 100.0
DEFAULT_MAXY = 50.0

def usage():
    print("""usage: {} [OPTION...] CSV-FILE... STO-FILE

  -f        force overwrite of existing files
  -m NAMES  use names from comma-separated list NAMES to name muscles instead
            of CSV file name (# of names in list must match # of CSV files)
  -n        name of the activation data in sto file (default: {})
  -T N      scale time into range 0.0 to N
  -Y N      scale y values into range 0.0 to N
  -h        show this help and exit""".format(os.path.basename(sys.argv[0]), DEFAULT_NAME))

def main():
    try:
        opts, args = getopt.getopt(sys.argv[1:], "fm:n:hT:Y:")
    except getopt.GetoptError as err:
        print(str(err))
        usage()
        sys.exit(-1)

    if len(args) < 2:
        usage()
        sys.exit(-1)

    overwrite = False
    muscles = None
    name = DEFAULT_NAME
    maxtime = DEFAULT_MAXTIME
    maxy = DEFAULT_MAXY
    for o, a in opts:
        if o == '-f':
            overwrite = True
        elif o == '-m':
            muscles = [ x.strip() for x in a.split(',') ]
        elif o == '-n':
            name = a
        elif o == '-T':
            maxtime = float(a)
        elif o == '-Y':
            maxy = float(a)
        elif o == '-h':
            usage()
            sys.exit(0)
        else:
            assert False, "unhandled option"

    fs_csv = args[0:-1]
    for f_csv in fs_csv:
        if not os.path.exists(f_csv):
            print("Error: CSV file {} does not exist".format(f_csv))
            sys.exit(-1)

    if muscles and len(muscles) != len(fs_csv):
        print("Error: number of names ({}) provided does not match number of CSV files ({})".format(len(muscles), len(fs_csv)))
        sys.exit(-1)

    f_sto = args[-1]
    if not overwrite and os.path.exists(f_sto):
        print("Error: STO file {} already exists".format(f_sto))
        sys.exit(-1)

    # determine number of columns and rows
    tot_nRows = 0
    tot_nCols = 1   # initial time column (as we omit it when taking data from each CSV)
    for f_csv in fs_csv:
        fd = open(f_csv, 'r')
        dialect = csv.Sniffer().sniff(fd.read(1024))
        fd.seek(0)
        csv_reader = csv.reader(fd, dialect)

        nRows = nCols = 0
        for line in csv_reader:
            if nRows == 0:
                nCols = len(line)
                print("nCols: {}".format(nCols))
            nRows += 1

        if tot_nRows == 0:
            tot_nRows = nRows
        elif tot_nRows != nRows:
            print("Error: Number of rows in CSV files do not match")
            sys.exit(-1)

        # add up to the total number of colums but omit time column
        tot_nCols += nCols - 1

        fd.close()

    # don't count the header line
    tot_nRows -= 1

    print("Writing sto file {} with name {} ({} muscles)".format(f_sto, name, len(fs_csv)))
    if maxtime != DEFAULT_MAXTIME:
        print("Scaling to max. time {}".format(maxtime))
    if maxy != DEFAULT_MAXY:
        print("Scaling to max. y {}".format(maxy))

    fd_sto = open(f_sto, 'w')
    fd_sto.write(name + "\n")
    fd_sto.write("version=1\n")
    fd_sto.write("nRows=" + str(tot_nRows) + "\n")
    fd_sto.write("nColumns=" + str(tot_nCols) + "\n")
    fd_sto.write("endheader\n")

    fd_sto.write("time")
    i = 0
    for f_csv in fs_csv:
        if muscles:
            colname = muscles[i]
            i += 1
        else:
            colname = os.path.splitext(os.path.basename(f_csv))[0]
        print("[+] Adding muscle {}...".format(colname))
        fd_sto.write("\t" + colname)
    fd_sto.write("\n")

    fds_csv = [ open(f_csv, 'r') for f_csv in fs_csv ]
    crs = []
    for fd_csv in fds_csv:
        dialect = csv.Sniffer().sniff(fd_csv.read(1024))
        fd_csv.seek(0)
        crs.append(csv.reader(fd_csv, dialect))

    times = np.zeros((tot_nRows, len(fds_csv)))
    vals = np.zeros((tot_nRows, len(fds_csv)))
    for i, cr in enumerate(crs):
        cr.next()   # skip header line
        for j, row in enumerate(cr):
            t,y = map(float, row)
            if maxtime != DEFAULT_MAXTIME:
                # Assume default 0-100 time in all input files for now
                t = (t / 100.0) * maxtime
            if maxy != DEFAULT_MAXY:
                # Assume default 0-50 time in all input files for now
                y = (y / 50.0) * maxy
            times[j,i] = t
            vals[j,i] = y

    # XXX: For now assume same time spacing in all CSV files
    for i, row in enumerate(times):
        fd_sto.write(str(row[0]))
        for val in vals[i]:
            fd_sto.write("\t" + str(val))
        fd_sto.write("\n");

    for fd_csv in fds_csv:
        fd_csv.close()
    fd_sto.close()

if __name__ == '__main__':
    main()