Multi-character register names not parsed correctly

[km-campbell]

Hi, firstly, thanks for the package! It has been a huge help to me.

I have found that when the names of qregs or cregs are multiple characters, only the last character of the name appears in the 'qreg_name' entry of elements in 'c_sect' with 'type' ASTType.QREG or ASTType.CREG. For example, a register called 'reg' would give return the value 'g' for the 'qreg_name' key.

This behaviour seems to stem from the line 233 to 235 of nuqasm2/qasmast.py (accessed 19/10/2023. The relevant lines are:

x = QTRegEx.REG_DECL.match(self.source)
self.qreg_name = x.group(1)
self.qreg_num = x.group(2)

One can replicate the issue without needing to run a specific file by modifying the above lines to read:

from nuqasm2.qasmast import QTRegEx

a_string = 'qreg reg[5]'
x = QTRegEx.REG_DECL.match(a_string)
print(x.group()) #gives 'qreg reg[5]'. This demonstrates that the matching is working as expected
qreg_name = x.group(1)
print(qreg_name) #gives 'g', rather than 'reg'

As I am new to regex, I personally fixed this using the 3rd-party pyparsing package by implementing the class:

import pyparsing as pp

class QasmParsingElement():
    l_sqr_brace = pp.Literal('[').suppress()
    r_sqr_brace = pp.Literal(']').suppress()
    idQasm = pp.Regex(r'[a-z][A-Za-z0-9]*') 
    nninteger = pp.Regex(r'[1-9]+[0-9]*|0')
    reg_index_slice = l_sqr_brace + nninteger + r_sqr_brace
    decl =  (pp.Keyword('qreg') + idQasm + reg_index_slice |
                 pp.Keyword('creg') + idQasm + reg_index_slice ) 

and then changing 233 to 235 of nuqasm2/qasmast.py to

parsed_line = QasmParsingElement.decl.parse_string(self.source)
self.qreg_name = parsed_line[1]
self.qreg_num = parsed_line[2]

However, obviously this does have the drawback of adding another dependency to the nuqasm2 package.

NOTE: The issue with cregs is identical.