Skip to content

Instantly share code, notes, and snippets.

@kidig
Created February 8, 2017 15:37
Show Gist options
  • Save kidig/4cc9a72b1117d8d6dbd35adf9a8a995a to your computer and use it in GitHub Desktop.
Save kidig/4cc9a72b1117d8d6dbd35adf9a8a995a to your computer and use it in GitHub Desktop.
PGN chess format parser
#
# Demonstration of the parsing module, implementing a pgn parser.#
# Originally by Alberto Santini http://www.albertosantini.it/chess/
#
import sys
from pyparsing import Combine, Forward, Group, Literal, oneOf, OneOrMore, Optional, Suppress, ZeroOrMore, Word
from pyparsing import ParseException
from pyparsing import alphanums, nums, quotedString, removeQuotes
unicodePrintables = ''.join(chr(c) for c in range(sys.maxunicode) if not chr(c).isspace())
unicodePrintablesNoBraces = unicodePrintables.replace('{', '').replace('}', '')
class Node:
def __init__(self, tokens):
self.tokens = tokens
def __repr__(self):
return "{}: ({})".format(self.__class__.__name__, self.tokens.asList())
class TagNode(Node): pass
class MoveNode(Node): pass
class CommentNode(Node): pass
#
# define pgn grammar
#
tag_value = Combine(quotedString)
tag_value.setParseAction(removeQuotes)
tag = Suppress("[") + Group(Word(alphanums) + tag_value) + Suppress("]")
tag.setParseAction(lambda toks: {'name': toks[0][0], 'value': toks[0][1]})
# comment_text = Word(unicodePrintablesNoBraces + " ")
# comment = Group(Suppress("{") + OneOrMore(comment_text) + Suppress("}"))
# comment.addParseAction(lambda toks: toks)
comment_text = Forward()
comment = Suppress("{") + Combine(comment_text, joinString=" ") + Suppress("}")
comment_text << OneOrMore(Word(unicodePrintablesNoBraces + " "))
dot = Literal(".")
piece = oneOf("K Q B N R")
file_coord = oneOf("a b c d e f g h")
rank_coord = oneOf("1 2 3 4 5 6 7 8")
capture = oneOf("x :")
promote = Literal("=")
castle_queenside = Literal("O-O-O") | Literal("0-0-0") | Literal("o-o-o")
castle_kingside = Literal("O-O") | Literal("0-0") | Literal("o-o")
move_number = Optional(comment) + Word(nums) + dot
m1 = file_coord + rank_coord # pawn move e.g. d4
m2 = file_coord + capture + file_coord + rank_coord # pawn capture move e.g. dxe5
m3 = file_coord + "8" + promote + piece # pawn promotion e.g. e8=Q
m4 = piece + file_coord + rank_coord # piece move e.g. Be6
m5 = piece + file_coord + file_coord + rank_coord # piece move e.g. Nbd2
m6 = piece + rank_coord + file_coord + rank_coord # piece move e.g. R4a7
m7 = piece + capture + file_coord + rank_coord # piece capture move e.g. Bxh7
m8 = castle_queenside | castle_kingside # castling e.g. o-o
check = oneOf("+ ++")
mate = Literal("#")
annotation = Word("!?", max=2)
nag = " $" + Word(nums)
decoration = check | mate | annotation | nag
variant = Forward()
half_move = Combine((m3 | m1 | m2 | m4 | m5 | m6 | m7 | m8) + Optional(decoration)) \
+ Optional(comment) + Optional(variant)
# move = Suppress(move_number) + half_move + Optional(half_move)
# variant << "(" + OneOrMore(move) + ")"
# grouping the plies (half-moves) for each move: useful to group annotations, variants...
# suggested by Paul McGuire :)
move = Group(Suppress(move_number) + half_move + Optional(half_move))
variant << Group(Suppress("(") + OneOrMore(move) + Suppress(")"))
game_terminator = oneOf("1-0 0-1 1/2-1/2 *")
tags = Group(ZeroOrMore(tag))
grammar = tags + Group(Optional(comment) + ZeroOrMore(move)) + Suppress(game_terminator)
# tag.addParseAction(TagNode)
tags.setParseAction(lambda toks: {'headers': toks[0].asList()})
def parse_pgn(text, bnf=grammar):
try:
return bnf.parseString(text)
except ParseException as err:
print(err.line)
print(" " * (err.column - 1) + "^")
print(err)
if __name__ == "__main__":
# input string
pgn = """
[Event "ICC 5 0 u"]
[Site "Internet Chess Club"]
[Date "2004.01.25"]
[Round "-"]
[White "guest920"]
[Black "IceBox"]
[Result "0-1"]
[ICCResult "White checkmated"]
[BlackElo "1498"]
[Opening "French defense"]
[ECO "C00"]
[NIC "FR.01"]
[Time "04:44:56"]
[TimeControl "300+0"]
{start йцу,
"12312"
Comment} 1. e4 e6 2. Nf3 d5 $2 3. exd5 (3. e5 g6 4. h4) exd5 4. Qe2+ Qe7 5. Qxe7+ Bxe7 6. d3 Nf6 7. Be3
Bg4 8. Nbd2 c5 9. h3 Be6 10. O-O-O Nc6 11. g4 Bd6 12. g5 Nd7 13. Rg1 d4 14.
g6 fxg6 15. Bg5 Rf8 16. a3 Bd5 17. Re1+ Nde5 18. Nxe5 Nxe5 19. Bf4 Rf5 20.
Bxe5 Rxe5 21. Rg5 Rxe1# {Black wins} 0-1
"""
# parse input string
tokens = parse_pgn(pgn, grammar)
print(tokens[1])
# print("\n\n".join([str(t) for t in tokens]))
# print(repr(tokens[0]))
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment