From df8253dbf6cb09093018333f99dd9c19ba0ff02b Mon Sep 17 00:00:00 2001
From: Daniel Smith <rdnlsmith@gmail.com>
Date: Sat, 28 Nov 2020 21:24:40 -0500
Subject: Initial commit

`Tokenizer` can _just barely_ parse a basic, well-formed move list.

Initially, I wanted to provide the movetext as a `Stream` rather than a
string, the idea being that it could be processed as it was being read
from a file without having to read the entire file into memory first. I
had difficulties with the stream being unreadable in
`Tokenizer.ParseMoves()`, so I switched to a string in order to get the
actual parsing logic down first.

Because of the `yield return` strategy, the debug console output
includes all of the expected halfmoves multiple times in various orders.
After running a test, generally the full, in-order list seems to exist
at the bottom of the output.
---
 DotnetPgn/Tokenizer.cs | 75 ++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 75 insertions(+)
 create mode 100644 DotnetPgn/Tokenizer.cs

(limited to 'DotnetPgn/Tokenizer.cs')

diff --git a/DotnetPgn/Tokenizer.cs b/DotnetPgn/Tokenizer.cs
new file mode 100644
index 0000000..9f779cb
--- /dev/null
+++ b/DotnetPgn/Tokenizer.cs
@@ -0,0 +1,75 @@
+using System;
+using System.Collections.Generic;
+using System.IO;
+using System.Text;
+using System.Text.RegularExpressions;
+using DotnetPgn.Models;
+
+namespace DotnetPgn
+{
+    public static class Tokenizer
+    {
+        private static readonly Regex s_halfMoveRegex =
+            new(@"([KQRBNP]?)([a-h]?[1-8]?)(x?)([a-h][1-8])([+#]?)");
+
+        public static IEnumerable<HalfMove> ParseMoves(string moveText)
+        {
+            StringBuilder currToken = new();
+            char[] moveTextChars = moveText.ToCharArray();
+            int moveNumber = 1;
+            Player currPlayer = Player.White;
+
+            for (int i = 0; i < moveTextChars.Length; i++)
+            {
+                char nextChar = moveTextChars[i];
+
+                if (nextChar is ' ' or '\t' or '\n' or '\r')
+                {
+                    // We should have either a move number or a complete halfmove.
+                    string token = currToken.ToString();
+                    currToken.Clear();
+                    Match match = s_halfMoveRegex.Match(token);
+
+                    if (!match.Success)
+                    {
+                        Console.WriteLine($"Unrecognized token: `{token}`");
+                        Console.WriteLine();
+                        continue;
+                    }
+
+                    Console.WriteLine($"Recognized token: {token}");
+                    Console.WriteLine($"Matching groups: ({match.Groups[1].Value})" +
+                        $"({match.Groups[2].Value})({match.Groups[3].Value})({match.Groups[4].Value})" +
+                        $"({match.Groups[5].Value})");
+
+                    var move = new HalfMove
+                    {
+                        MoveNumber = moveNumber,
+                        Player = currPlayer,
+                        Piece = PieceParser.ParsePiece(match.Groups[1].Value),
+                        TargetSquare = new Square(match.Groups[4].Value[0], Convert.ToInt32(Char.GetNumericValue(match.Groups[4].Value[1]))),
+                    };
+
+                    Console.WriteLine($"Halfmove: {move.MoveNumber}. {move.Player} {move.Piece} from {move.SourceSquare} to {move.TargetSquare}");
+                    Console.WriteLine();
+
+                    if (currPlayer == Player.White)
+                    {
+                        currPlayer = Player.Black;
+                    }
+                    else
+                    {
+                        currPlayer = Player.White;
+                        moveNumber++;
+                    }
+
+                    yield return move;
+                }
+                else
+                {
+                    currToken.Append(nextChar);
+                }
+            }
+        }
+    }
+}
\ No newline at end of file
-- 
cgit v1.2.3