#!/usr/bin/env python

# Combines two map spreadsheets A0->B and A1->C to A->B, with B overwriting C

# Multi-safe (supports an input appearing multiple times). Note that if there is

# *any* non-empty mapping for an input in map_0, all mappings for that input in

# map_1 will be excluded.

import csv

import os.path

import sys

sys.path.append(os.path.dirname(__file__)+"/../lib")

import maps

import opts

import util

def main():

    ignore = opts.env_flag('ignore')

    header_num = util.cast(int, opts.get_env_var('header_num'))

        # selects which map's header to use as the output header

    all_ = opts.env_flag('all') # doesn't merge mappings that are in both maps

    try: _prog_name, map_1_path = sys.argv

    except ValueError:

        raise SystemExit('Usage: env [ignore=1] [header_num={0|1}] [all=1] '

            +sys.argv[0]+' <map_0 map_1 [| '+sys.argv[0]+' map_2]... >out_map')

    headers = [None]*2

    # Open map 0

    map_0_reader = csv.reader(sys.stdin)

    headers[0] = map_0_reader.next()

    # Open map 1

    stream = open(map_1_path, 'rb')

    map_1_reader = csv.reader(stream)

    headers[1] = map_1_reader.next()

    # Check col labels

    combinable = maps.combinable(*headers)

    if not combinable and not ignore:

        raise SystemExit('Map error: '

        'Map 0 column 0 label doesn\'t contain map 1 column 0 label')

    # Pass through map 0, storing which mappings it defines

    writer = csv.writer(sys.stdout)

    writer.writerow(maps.merge_headers(*headers, **dict(prefer=header_num)))

    mappings = set()

    for row in map_0_reader:

        if not all_ and row[0] != '' and row[1] != '':

            mappings.add(row[0])

            mappings.add(row[1])

        writer.writerow(row)

    if combinable:

        # Add mappings in map 1 that weren't already defined or mapped to

        for row in map_1_reader:

            if row[0] != '' and row[0] not in mappings: writer.writerow(row)

    stream.close()

main()