Latex fractions to mathML in Python with regex

Question

I have some text:

\frac{A}{B}

i need to transform this text to form:

<mfrac>
 <mrow>
  A
 </mrow>
 <mrow>
  B
 </mrow>
</mfrac>

I have to use Python, and regex. A and B can be further fractions, so function must be recursive, for example text:

\frac{1+x}{1+\frac{1}{x}}

must change into

<mfrac>
 <mrow>
  1+x
 </mrow>
 <mrow>
  1+
  <mfrac>
   <mrow>
    1
   </mrow>
   <mrow>
    x
   </mrow>
   </mfrac>
 </mrow>
</mfrac>

please help with regex :)

Answer 1

If you need to match recursive pattern in default python re module, you can do like me for recursive comments I build recently for css preprocessor.

Generally use re just for splitting text to tokens and then use loops with nesting level variable to find all syntax. Here is my code:

COMMENTsRe = re.compile( r"""
                        //   |
                        \n   |
                        /\*  |
                        \*/  
                        """, re.X ) 

def rm_comments( cut ):
  nocomment = 0 # no inside comment
  c = 1 # c-like comments, but nested
  cpp = 2 # c++like comments

  mode = nocomment
  clevel = 0 # nesting level of c-like comments
  matchesidx = []

  # in pure RE we cannot find nestesd structuries
  # so we are just finding all boundires and parse it here
  matches = COMMENTsRe.finditer( str(cut) )
  start = 0
  for i in matches:
    m = i.group()
    if mode == cpp:
      if m == "\n":
        matchesidx.append( ( start, i.end()-1 ) ) # -1 because without \n
        mode = nocomment
    elif mode == c:
      if m == "/*":
        clevel += 1
      if m == "*/":
        clevel -= 1
      if clevel == 0:
        matchesidx.append( ( start, i.end() ) )
        mode = nocomment
    else:
      if m == "//":
        start = i.start()
        mode = cpp
      elif m == "/*":
        start = i.start()
        mode = c
        clevel += 1

  cut.rm_and_save( matchesidx )