Algorithm for parsing expressions in python?

Question

I have next algorithm for parsing expressions in Python:

def parse(strinput):
  for operator in ["+-", "*/"]:
    depth = 0
    for p in range(len(strinput) - 1, -1, -1):
      if strinput[p] == ')': depth += 1
      elif strinput[p] == '(': depth -= 1
      elif depth==0 and strinput[p] in operator:
        # strinput is a compound expression
        return (strinput[p], parse(strinput[:p]), parse(strinput[p+1:]))
  strinput = strinput.strip()
  if strinput[0] == '(':
    # strinput is a parenthesized expression?
    return parse(strinput[1:-1])
  # strinput is an atom!
  return strinput

(it can be found here: http://news.ycombinator.com/item?id=284842)

I have hard time understanding it, since I don't find Python docs very helpful for this situation. Can someone tell me what line: for operator in ["+-", "*/"]:means? I know it's structure like for each string variable which is operator in array of this 2 elements, but why is it written like this ["+-, */"]? How does Python separate it? In first iteration, operator is "+-"?

Any help would mean a lot. Thanks

It looks like a very unpythonic algorithm to me, if I can tell. — Elazar, Jun 02 '13 at 01:23
`for operator in ["+-", "*/"]:` it's a loop iterating through list which has two elements "+-" and "*/" — solusipse, Jun 02 '13 at 01:23
@Blender truth is, I don't really know. It looks too similar to something I wrote in an introduction-to-CS course. in C. Can't *you* improve it? For one, it takes advantage of the fact that an operator is single character (which breaks in python3). It just feels... awkward. — Elazar, Jun 02 '13 at 01:40
@Elazar: Python 2 has floor division as well. This snippet isn't supposed to parse Python expressions. — Blender, Jun 02 '13 at 01:45
@Blender: One better, though slightly more complicated, way is to use [top-down parsing](http://effbot.org/zone/simple-top-down-parsing.htm). — unutbu, Jun 02 '13 at 01:49
I think I know what bothers me: it uses both `+= 1` / `-= 1` *and* recursion. But recursion means you already have a stack. — Elazar, Jun 02 '13 at 01:51

score 4 · Accepted Answer · answered Jun 02 '13 at 01:23

4

You're correct; for operator in ["+-", "*/"]: means operator will be "+-" the first time through and "*/" the second time through the loop.

Notice how later it checks if strinput[p] in operator. Python treats a string as a list of characters, so this expression will only be true if strinput[p] is equal to "+" or "-" on the first time through and "*" or "/" the second time through.

(The reason they do this is for order of operations- "+" and "-" get equal but lower precedence to "*" and "/")

answered Jun 02 '13 at 01:23

ahuff44

1,100
1
9
9

Yea, thank you! :) I already got by time what it means by it and exactly got it on 6*3+2. It goes from right to left, first checking lower priority operators, because it can group by them, since higher priority operators need to be executed for first. :) – Tommz Jun 02 '13 at 01:27

Algorithm for parsing expressions in python?

1 Answers1