#!/usr/bin/python

"""
Verifies that a text file is in the format required for a submission
to the TREC KBA 2012 task called "ccr-2012"

If the file is valid, this outputs nothing and exits with status 0.

If the file is invalid, this print diagnostics to stderr and exists
with an non-zero status code.
"""

import re
import sys
import json
from datetime import datetime
import argparse
import traceback
parser = argparse.ArgumentParser(description=__doc__)
parser.add_argument(dest="submission",   help="path to submission file to validate, can be gzip'ed or bzip2'ed")
args = parser.parse_args()

def log(mesg):
    sys.stderr.write("%s\n" % mesg)
    sys.stderr.flush()

if args.submission.endswith(".gz"):
    submission = gzip.GzipFile(args.submission)
elif args.submission.endswith(".bz2"):
    submission = bz2.BZ2File(args.submission)
else:
    submission = open(args.submission)

def get_comment(line):
    if line.startswith("#"):
        return line.lstrip("#")
    else:
        return None

try:
    first_line = submission.readline()
    should_be_filter_run = get_comment(first_line)
    filter_run = json.loads(should_be_filter_run)
except Exception, exc:
    log("failed to load a valid JSON string as filter_run instance from first line: length=%d\n%s" % (
            len(first_line), traceback.format_exc(exc)))
    sys.exit(1)

assert "task_id" in filter_run, "filter_run is missing task_id"
assert "topic_set_id" in filter_run, "filter_run is missing topic_step_id"
assert "corpus_id" in filter_run, "filter_run is missing corpus_id"
assert "team_id" in filter_run, "filter_run is missing team_id"
assert "system_id" in filter_run, "filter_run is missing system_id"
assert "run_type" in filter_run, "filter_run is missing run_type"

for line in submission:
    if get_comment(line):
        continue
    team_id, system_id, date_hour, stream_id, entity, relevance = line.strip().split()
    assert team_id == filter_run["team_id"]
    assert system_id == filter_run["system_id"]
    epoch_ticks, doc_id = stream_id.split("-")
    assert re.match("^[0-9a-f]{32}$", doc_id), "invalid doc_id=%r" % doc_id
    stream_id_date_hour = datetime.utcfromtimestamp(int(epoch_ticks)).strftime("%Y-%m-%d-%H")
    assert date_hour == stream_id_date_hour, "date_hour %s disagrees with stream_id-->%s" % (date_hour, stream_id_date_hour)
    relevance = int(relevance)
    assert 0 <= relevance <= 1000, "relevance not in range [0,1000]"
    assert len(entity) > 0, "blank entity"