#!/usr/bin/env python
#
# Create a header file for extracted Wikipedia text.

import argparse
from wikivector import wiki


def main(wiki_text_dir, header_file):
    df = wiki.dump_pages(wiki_text_dir)
    df.to_csv(header_file, index=False)


if __name__ == '__main__':
    parser = argparse.ArgumentParser()
    parser.add_argument('wiki_dir', help="Path to output from WikiExtractor")
    parser.add_argument('header_file', help="Path to output header file")
    args = parser.parse_args()

    main(args.wiki_dir, args.header_file)
