Collecting stats about paper titles per year in a .bib file

bibtex_collect_stats.py

#!/usr/bin/python3
# -*- coding: utf-8 -*-

import argparse
import bibtexparser
from collections import Counter
import matplotlib.pyplot as plt
import seaborn as sns
import sys

keywords = ["neural", "deep learning", "lstm", "rnn", "long short", "cnn"]

def main(args):
    sys.stderr.write("### Loading BibTeX file... ")
    sys.stderr.flush()
    bib = bibtexparser.load(args.bibfile)
    sys.stderr.write("done!\n")

    per_year_match = Counter()
    per_year_total = Counter()
    skipped = Counter()

    for entry in bib.entries:
        if 'author' not in entry:
            # probably a full proceedings entry
            skipped['no author'] += 1
            continue
        if 'title' not in entry or 'year' not in entry:
            # probably a parse error...
            skipped['no title/year'] += 1
            continue
        year = entry['year']
        title = entry['title']
        if "Proceedings" in title:
            if 'chapter' in entry:
                title = entry['chapter']
            else:
                # probably a full proceedings entry
                skipped['proceedings in title'] += 1
                continue
        per_year_total[year] += 1
        title = title.lower()
        if any(kw in title for kw in keywords):
            per_year_match[year] += 1

    sys.stderr.write("### Skipped {} items:\n".format(sum(skipped.values())))
    for reason, count in skipped.items():
        sys.stderr.write("  {0:4d} = {1}\n".format(count, reason))

    years = sorted(per_year_total.keys())
    matches = []
    for year in years:
        match = per_year_match[year]
        total = per_year_total[year]
        percentage = match / total * 100
        print("{0}:  {1:3n}/{2:4n} = {3:2.2f}%".format(year, match, total, percentage))
        matches.append(match)

    if args.plot is not False:
        if args.fromyear:
            idx = years.index(args.fromyear)
            years, matches = years[idx:], matches[idx:]
        fig, ax = plt.subplots(figsize=(16,9))
        sns.barplot(years, matches, palette=sns.xkcd_palette(("deep sea blue",)), ax=ax)
        if args.plot is None:
            plt.show()
        else:
            fig.savefig(args.plot, dpi=700)


if __name__ == "__main__":
    description = ""
    epilog = ""
    parser = argparse.ArgumentParser(description=description, epilog=epilog)
    parser.add_argument('bibfile',
                        metavar='BIBFILE',
                        type=argparse.FileType('r', encoding="UTF-8"),
                        default=sys.stdin,
                        help='BibTeX file to analyze (default: STDIN)')
    parser.add_argument('-p', '--plot',
                        nargs='?',
                        type=str,
                        default=False,
                        help='Plot the output (default: %(default)b)')
    parser.add_argument('-f', '--fromyear',
                        metavar='YEAR',
                        type=str,
                        default=None,
                        help='Plot should start from year (default: %(default)s)')

    args = parser.parse_args()
    main(args)