Matching arbitrary player names

Question

I'm using lark but can't figure out how I could match all players name as they can be complex with lark rules?

Example pattern is "Seat {number}: {player} ({chips} in chips)" and I want all the values from each line as well.

from lark import Lark, lexer
gram = r"""
start: seats+

seats: "Seat " seat_position player table_chips is_sitting_out? SPACE? NL
seat_position: NUMBER
is_sitting_out: " is sitting out"
table_chips: "(" chips " in chips"  (", " chips " bounty")? ")"
player: STRING_INNER
chips: "$"? (DECIMAL | NUMBERS)

NUMBERS: /[0-9]/+
NUMBER: /[0-9]/
DECIMAL: NUMBERS "." NUMBERS
SPACE: " "+
STRING_INNER: ": " /.*?/ " " 

CR : /\r/
LF : /\n/
NL: (CR? LF)+
"""
data = r"""Seat 1: Ruzzka(Rus) (1200 in chips)
Seat 1: Dladik Rzs38 (1200 in chips)
Seat 1: slum ^o_o^ (1200 in chips)
Seat 1: é=mc² (1200 in chips)
Seat 1: {’O_0`}/(nh) (1200 in chips)
Seat 1: °ÆND0c42Z4y° (1200 in chips)
Seat 1: $ salesovish (1200 in chips)
"""
parser = Lark(gram)
tree = parser.parse(data)
print(tree.pretty())

I guess the solution should be: player: / .*(?=\ $\$?[0-9]+\.?[0-9]* in chips$)/ — na a
– na a, Commented Nov 23, 2021 at 15:04

MegaIng · Accepted Answer · 2021-11-29 21:17:37Z

The problem is that there is apparently no real rule where names end, making it very hard to parse it, since lark is mostly non backtracking for high speed.

I would actually guess that just using a regex on each line directly will be easier unless you also need to parse a more complicated structure than what you showed here. But lark is able to deal with this kind of arbitrary content, example here, but at a large performance penalty.

Here a solution without lark:

import re

regex = re.compile(r"Seat\s*(?P<number>\d+)\s*:\s*(?P<player>[^\n]+?)\s+\((?P<chips>\d+) in chips\)")

seats = []
for line in data.splitlines():
    match = regex.match(line)
    if match is not None:
        values = match.groupdict()
        seats.append((values["number"], values["player"], values["chips"]))

print(seats)

It appears from your grammar that you actually need to extract a bit more information (e.g. is_sitting_out and bounty). For that you can slightly change the regex to this:

Seat\s*(?P<number>\d+)\s*:\s*(?P<player>[^\n]+?)\s+\((?P<chips>\d+) in chips\s*(?:,\s*(?P<bounty>\d+)\s*bounty)?\)(?P<is_sitting_out> is sitting out)?

You can check if a player is sitting out via values['is_sitting_out'] is not None, and values['bounty'] will be None if there is no bounty.

Collectives™ on Stack Overflow

Matching arbitrary player names

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related