#!/usr/bin/env python2

import sys, os, zipfile

# One package must be submitted for each run, for up to 3 runs.
# A submission package is a zip file containing two files:
# - cccs-<team-id>-<run-id>.tsv containing the output of the system for that run
# - cccs-<team-id>-<run-id>.README containing a description of the run
# The format for the system output file is one synopsis per line with the following tab-separated format:
# <conversation-id> \t <team-id>:<run-id> \t <synopsis-text> \n

if len(sys.argv) != 4:
    print 'USAGE: %s <team-id> <run-id> <submission-package>' % sys.argv[0]
    sys.exit(1)

team_id = sys.argv[1]
run_id = sys.argv[2]
submission_package = sys.argv[3]
run_filename = 'cccs-%s-%s.tsv' % (team_id, run_id)
description_filename = 'cccs-%s-%s.README' % (team_id, run_id)

def die(message):
    print message
    sys.exit(1)

if not os.path.exists(submission_package):
    die('ERROR: could not find submission package "%s"' % submission_package)

try:
    package_fp = zipfile.ZipFile(submission_package, 'r')
    files = package_fp.namelist()
    #if len(files) != 2:
    #    die('ERROR: zip package should contain exactly two files')
    if run_filename not in files:
        die('ERROR: zip package should contain a file named %s' % run_filename)
    if description_filename not in files:
        die('ERROR: zip package should contain a file named %s' % description_filename)
    description_fp = package_fp.open(description_filename)
    if len(description_fp.readlines()) < 1:
        die('ERROR: description file should contain some text')
    description_fp.close()
    run_fp = package_fp.open(run_filename)
    seen = {}
    for line_num, line in enumerate(run_fp):
        tokens = line.strip().split('\t')
        if len(tokens) != 3:
            die('ERROR: line %d does not contain exactly 3 tab-separated fields' % (line_num + 1))
        if len(tokens[0]) == 0:
            die('ERROR: empty conversation id at line %d' % (line_num + 1))
        if len(tokens[2]) == 0:
            print 'WARNING: empty synopsis at line %d' % (line_num + 1)
        if tokens[1] != '%s:%s' % (team_id, run_id):
            die('ERROR: invalid system identifier "%s" at line %d, should be "%s:%s"' % (tokens[1], line_num + 1, team_id, run_id))
        if tokens[0] in seen:
            die('ERROR: more than one synopsis for conversation id "%s", at line %d, previous was at line %d' % (tokens[0], line_num + 1, seen[tokens[0]]))
        else:
            seen[tokens[0]] = line_num + 1
    print 'INFO: %d synopses found in this submission' % len(seen)
    run_fp.close()
    package_fp.close()
except Exception as e:
    die('ERROR: exception while processing submission package "%s"' % str(e))

print 'INFO: all checks successful'