Latexの参考文献の順番を直すpythonプログラム

Latexのコードを入力として、参考文献の出てくる順番通りにbibitemを並べかえたものを出力します。

以下のリンクのDaveWさんの回答を基にしました。

sctex.py

import argparse

#https://tex.stackexchange.com/questions/17354/sort-thebibliography-by-citation-order
#DaveW Mar 27, 2017 at 7:43

parser = argparse.ArgumentParser(description='latex citation sorter')
parser.add_argument('filein')

args = parser.parse_args()

with open(args.filein) as f:
    string = f.read()

start = '\\begin{document}'
end = '\\begin{thebibliography}'
si = string.find(start) + len(start)
ei = string.rfind(end)

body = string[si:ei]

start = '\\begin{thebibliography}'
end = '\\end{thebibliography}'
si = string.find(start) + len(start)
ei = string.rfind(end)

bib = string[si:ei]

authors = []

for c in range(0,len(body)-10):
    if body[c:c+6] == '\\cite{':
        author = ''
        c+=6
        while(body[c]!='}'):
            if(body[c]==' '):
                c+=1
            elif(body[c]==','):
                c+=1
                if author in authors:
                    author = ''
                else:
                    authors.append(author)
                    author=''
            else:
                author += (body[c])
                c+=1
        if author in authors:
            pass
        else:
            authors.append(author)
            
labels = []
refs = []

for c in range(0,len(bib)-7):
    ref=''
    if bib[c:c+9] == '\\bibitem{':
        c+=9
        label=''
        while(bib[c]!='}'):
            label += bib[c]
            c+=1
        labels.append(label)
        c+=1
        while(bib[c:c+9] != '\\bibitem{'):
            if c == len(bib)-1:
                break
            elif (bib[c]) == '\n':
                c+=1
            else:
                ref += (bib[c])
                c+=1
        refs.append(ref)
        
labref = dict(zip(labels,refs))

for author in authors:
    print('\\bibitem{' + author + '}\t' + labref[author])

print('orphans:')

if len(authors) < len(labels):
    orphans = list(set(labels)-set(authors))
    for orphan in orphans:
        print('\033[91m' + '\\bibitem{' + orphan + '}\t' + labref[orphan] + '\033[0m')