UniTO/tesi/conv.py

import json
import re
from sys import argv

allsymbols = json.load(open('./unicode-latex.json'))
mysymbols = ['≡', '≼', '→', '⊀', '⋠', '≺', '∀', '∈', '₂', '₁', 'ₐ', 'ₘ', 'ₙ', 'ᵢ', 'ⁱ']

symbols = {s: allsymbols[s] for s in mysymbols}
mathsymbols = {s: '$'+allsymbols[s]+'$' for s in symbols}

def read_by_char(fname):
    # Yield character and True/False if inside mathmode block
    mathmode = False
    mathmode_begin = set(['\\begin{equation*}', '\\begin{equation}'])
    mathmode_end = set(['\\end{equation*}', '\\end{equation}'])
    cnt = 0
    with open(fname, 'r') as fp:
        for line in fp.readlines():
            cnt += 1
            words = [w.strip() for w in line.split(' ')]
            if mathmode_begin.intersection(words):
                assert mathmode == False
                mathmode = True
            elif mathmode_end.intersection(words):
                assert mathmode == True, f'Line: {words}, number: {cnt}'
                mathmode = False

            for ch in line:
                yield ch, mathmode

def convert(ch, mathmode):
    if not mathmode:
        return mathsymbols[ch] if ch in mathsymbols else ch
    else:
        return symbols[ch] if ch in symbols else ch

# convert symbols except the one requiring math mode modifiers
firstpass = [convert(*c) for c in read_by_char(argv[1])]

newfile = ''.join(firstpass)
with open(argv[2], 'w') as f:
    f.write(newfile)