两个不同的结构化python列表之间的匹配

Question

I use python 2.7 and I have 2 lists, one is of this shape:

t1 = [('go', 'VB'), [('like', 'IN'), [('i', 'PR')]], [('to', 'TO')], [('there', 'RB')]]

the other is in a text file stored in this format:

t2 = [go:VB, [like:IN, [i:PR]], [to:TO], [there:RB]]

I would like to see if ( t1 ) is matching ( t2 ) or not.

A problem that I face is that the ones inside the text file without ( '' ) so they looks like variables.

Can you please help in finding a way of matching these two.

def match(t1, t2):
    #check here if the nested lists match or not.   
    return True

I tried to turn ( t1 ) to string and delete ' ( ' and ' ) ' by replacing them with an empty '' then replace ' , ' with ' : ' but it has given a lot of quotation marks and I thought this is not a great idea of fixing this issue.

Answer 1

This answer is not using eval() which is a really insecure thing.

1. use str to convert your t1 to string.
2. delete all whitespaces in t1 and t1 with help of replace
3. use re 's sub for conversion of t2 .
4. finally, compare strings

### 1
t1 = str(t1)

### 2
t1 = t1.replace(" ", "")
t2 = t2.replace(" ", "")

### 3
t2 = re.sub(r"(\w+):(\w+)", r"('\1','\2')", t2)

### 4
print(t1 == t2)

---

Edit

If you want to support tabs and newlines, you need to do this

### 2
t1 = "".join(t1.split())
t2 = "".join(t2.split())

Answer 2

A naive and simple approach - use regex substitution to transform the string from the file to a Python evaluable form, then evil eval it:

import re

s2 = '[go:VB, [like:IN, [i:PR]], [to:TO], [there:RB]]'

# 'go:VB' -> '("go", "VB")'
s2_pyth = re.sub(r'(\w+):(\w+)', r'("\1", "\2")', s2)
# '[("go", "VB"), [("like", "IN"), [("i", "PR")]], [("to", "TO")], [("there", "RB")]]'

l2 = eval(s2_pyth)
# [('go', 'VB'), [('like', 'IN'), [('i', 'PR')]], [('to', 'TO')], [('there', 'RB')]]

if l1 == l2:
    # or whatever more specific comparison

I think, using eval in this context (seems to be a harmless academic NLP task) is ok. If the tokens in your text file aren't strictly alphanumerical, you might need a smarter regex as r'\\w+' to match them, maybe sth. like r'[^\\[\\]]+' ...

Answer 3

Assuming that your structure is composed only of lists and tuples containing two strings, the following function should do what you whishes by generating your target string recursively :

def format_list(l):
  res = "["
  items = []
  for item in l:
    if isinstance(item,list):
      items.append(format_list(item))
    elif isinstance(item,tuple):
      items.append(item[0] + ':' + item[1])
  res += ", ".join(items) + "]"
  return res