a simple BibTeX parser based on finite state machine

bibparse.py

#!/usr/bin/env python
# -*- coding: utf-8 -*-
# CC0 dedicated to public domai by Akihiro Uchida
import sys

class BibParser(object):
    def __init__(self):
        self.parse = self.parse_main
        self.buffer = ''
        self.item = dict()
        self.items = []
        self.brace_count = 0
        return

    def feed(self, text):
        i = 0
        while i < len(text):
            (self.parse, i) = self.parse(text, i)
        return
                
    def parse_main(self, text, i):
        c = text[i]
        if c == '@':
            self.item = dict()
            return (self.parse_bibtype, i+1)
        return (self.parse_main, i+1)
        
    def parse_bibtype(self, text, i):
        c = text[i]
        if c == '{':
            self.item['type'] = self.buffer
            self.buffer = ''
            return (self.parse_key, i+1)
        elif c not in [' ', '\t', '\n']:
            self.buffer += c
        return (self.parse_bibtype, i+1)
            
    def parse_key(self, text, i):
        c = text[i]
        if c == ',':
            self.item['key'] = self.buffer
            self.buffer = ''
            return (self.parse_lhs, i+1)
        self.buffer += c
        return (self.parse_key, i+1)

    def parse_lhs(self, text, i):
        c = text[i]
        if c == '=':
            self.key = self.buffer.lower()
            self.buffer = ''
            return (self.parse_rhs, i+1)
        elif c not in [' ', '\t', '\n']:
            self.buffer += c
        return (self.parse_lhs, i+1)

    def parse_rhs(self, text, i):
        c = text[i]
        if c == '}':
            self.buffer = self.buffer.replace('\t', ' ')
            if self.key == 'author':
                self.item['author'] = author_list(self.buffer)
            else:
                self.item[self.key] = self.buffer
            self.finish_item()
            return (self.parse_main, i+1)
        elif c == ',':
            self.buffer = self.buffer.replace('\t', ' ')
            if self.key == 'author':
                self.item['author'] = author_list(self.buffer)
            else:
                self.item[self.key] = self.buffer
            self.buffer = ''
            return (self.parse_lhs, i+1)
        elif c == '{':
            return (self.parse_in_brace, i+1)
        elif c == '"':
            return (self.parse_in_dquote, i+1)
        elif c not in [' ', '\t', '\n']:
            self.buffer += c
        return (self.parse_rhs, i+1)

    def parse_in_brace(self, text, i):
        c = text[i]
        if self.brace_count == 0 and c == '}':
            return (self.parse_rhs, i+1)
        elif c == '{':
            self.brace_count += 1
        elif c == '}':
            self.brace_count -= 1
        else:
            self.buffer += c
        return (self.parse_in_brace, i+1)

    def parse_in_dquote(self, text, i):
        c = text[i]
        if c == '"':
            return (self.parse_rhs, i+1)
            self.buffer += c
        if c == '\\':
            return (self.parse_bs, i+1)
        return (self.parse_in_dquote, i+1)

    def parse_bs(self, text, i):
        c = text[i]
        self.buffer += c
        if c == ' ':
            return (self.parse_dquote, i+1)
        return (self.parse_bs, i+1)

    def finish_item(self):
        self.items.append(self.item)
        self.key = ''
        self.buffer = ''
        return
        
    def finish(self):
        self.finish_item()
        return self.items

def author_list(author_text):
    authors = []
    for author in author_text.split(' and '):
        if ',' in author:
            p = author.partition(',')
            authors.append('%s %s' % (p[2], p[0]))
        else:
            authors.append(author)
    return authors

def dump_rst(dic):
    separator = ""2 * '\n' + 2 * ' '
    fmt_str = '* '
    fmt_str += ', '.join(dic['author'])
    fmt_str += '**%(title)s**'
    fmt_str += separator

    reference_info = []
    if dic['type'] == 'article':
        reference_info = []
        reference_info.append('*%(journal)s*')
        if 'volume' in dic:
            reference_info.append('**%(volume)s**')
        if 'number' in dic:
            reference_info.append('%(number)s)')
        if 'pages' in dic:
            reference_info.append('pages. %(pages)s')
        if 'annote' in dic:
            reference_info.append('(%(annote)s)')
        fmt_str += ' '.join(reference_info) + ', ' + '*%(year)s*'

    if dic['type'] == 'inproceedings':
        reference_info = []
        if 'editor' in dic:
            reference_info.append('%(editor)s editors')
        reference_info.append('*%(booktitle)s*')
        if 'pages' in dic:
            reference_info.append('pages. %(pages)s, ')
        reference_info.append('%(year)s')
        if 'annote' in dic:
            reference_info.append('(%(annote)s)')
        fmt_str += 'In ' + ', '.join(reference_info)

    if dic['type'] == 'incollections':
        if 'editor' in dic:
            reference_info.append('Editors %(editor)s')
        reference_info.append('*%(booktitle)s*')
        if 'pages' in dic:
            reference_info.append('pages. %(pages)s, ')
        reference_info.append('%(publisher)s, %(year)s')
        reference_info.append('%(year)s')
        if 'annote' in dic:
            reference_info.append('(%(annote)s)' )
        fmt_str += 'In ' + ', '.join(reference_info)

    fmt_str += separator

    url_info = []
    if 'url' in dic:
        url_info.append('`URL <%(url)s>`_')
    if 'doi' in dic:
        url_info.append('`DOI: %(doi)s <http://dx.doi.org/%(doi)s>`_')
    if 'arxiv' in dic:
        url_info.append('`arXiv:%(arxiv)s <http://arxiv.org/abs/%(arxiv)s>`_')
    fmt_str += ', '.join(url_info)
    fmt_str += '\n'
    print fmt_str % dic
    return


if __name__ == '__main__':
    f = open(sys.argv[1], 'r')
    parser = BibParser()
    parser.feed(f.read())
    items = parser.finish()
    items.sort(key=lambda d: d["year"], reverse=True)
    for item in items:
        dump_rst(item)