It’s best to visualize our grammar as a tree: There’s some trickiness about the last move because white might make one more move than black, but we’ll deal with that later. The game is a list of NUMBER., MOVE, MOVE, followed by an OUTCOME. The annotations block is a new line delimited list of the sequence. An “entry” in the file is composed of two blocks: an “annotations” block and a “game” block. Writing out the parse tree from the file above is straightforward. For this project, I use parsita, an open source python library for parsing written by one of my colleagues. The topic is pretty deep, but the gist is that parser-combinators allow us to write out a parse tree of a grammar and iteratively build complex parsers by combining simpler ones. In computer science, a parser-combinator is a higher order function that takes in multiple parsers and returns a single parser. We’re going to take a file of PGN entries and parse them. The last text declares the outcome, 1-0 for white wins, 0-1 for black, and 1/2-1/2 for a draw. After the annotations, the moves are listed by move number followed by white’s move, then black’s move ( in algebraic notation).
The annotations appear in brackets, starting with the tag followed by a quoted string.
For instance, this 2021 game between Magnus Carlsen and Jorden Van Foreest: 1. The important insight is that there are rules for how that text looks. The grammar of PGN includes annotations that are pairs of a tag and a string, the moves of the game, and the outcome. Lots of text data, PGN included, obeys a “grammar”. Photo by Sarah Pflug, used under CC0 The Grammar of PGN Files
0 kommentar(er)
