ekrizdis367_obsidian-caissa/scripts/fetch-wcc-games.mjs

/**
 * One-shot build script. Reads scripts/wcc-curated.json, downloads each
 * match's PGN from PGN Mentor, extracts the curated game rounds, and emits
 * src/chess/wcc-data.generated.ts with everything inlined.
 *
 * Run from the plugin root: `node scripts/fetch-wcc-games.mjs`
 *
 * The generated file is committed to the repo so end users never make any
 * network calls. Only re-run this script when curating new matches/games.
 */

import { Chess } from "chess.js";
import fs from "node:fs";
import path from "node:path";
import { fileURLToPath } from "node:url";

const __dirname = path.dirname(fileURLToPath(import.meta.url));
const ROOT = path.resolve(__dirname, "..");
const CURATED_PATH = path.join(ROOT, "scripts", "wcc-curated.json");
const OUT_PATH = path.join(ROOT, "src", "chess", "wcc-data.generated.ts");

const USER_AGENT = "obsidian-caissa-build-script";

async function fetchText(url) {
	const res = await fetch(url, {
		headers: { "User-Agent": USER_AGENT, Accept: "application/x-chess-pgn,*/*" },
	});
	if (!res.ok) {
		throw new Error(`HTTP ${res.status} ${res.statusText} for ${url}`);
	}
	return await res.text();
}

/**
 * Split a multi-game PGN file on blank lines between games. Returns an array
 * of single-game PGN strings (each still containing its [Tag "value"] block
 * and the move text).
 */
function splitPgnIntoGames(text) {
	const lines = text.split(/\r?\n/);
	const games = [];
	let current = [];
	let inMoves = false;
	for (const line of lines) {
		const trimmed = line.trim();
		if (trimmed.startsWith("[")) {
			if (inMoves && current.length > 0) {
				games.push(current.join("\n").trim());
				current = [];
			}
			inMoves = false;
			current.push(line);
		} else if (trimmed === "") {
			if (current.length > 0) current.push(line);
		} else {
			inMoves = true;
			current.push(line);
		}
	}
	if (current.length > 0) games.push(current.join("\n").trim());
	return games.filter((g) => g.length > 0);
}

/** Pull a single tag value out of a single-game PGN string. */
function getHeader(pgn, tag) {
	const re = new RegExp(`\\[\\s*${tag}\\s+"((?:[^"\\\\]|\\\\.)*)"\\s*\\]`, "i");
	const m = pgn.match(re);
	if (!m) return undefined;
	return m[1].replace(/\\"/g, '"').replace(/\\\\/g, "\\");
}

/** Verify the PGN is parseable and at least one move was played. */
function validatePgn(pgn) {
	const game = new Chess();
	game.loadPgn(pgn, { strict: false });
	const history = game.history();
	return history.length;
}

function jsString(s) {
	return JSON.stringify(s);
}

async function main() {
	const curated = JSON.parse(fs.readFileSync(CURATED_PATH, "utf8"));
	const matches = curated.matches;

	const collected = [];
	let totalSizeBytes = 0;

	for (const match of matches) {
		console.log(`\n=== ${match.matchLabel} ===`);
		let pgnText;
		try {
			pgnText = await fetchText(match.pgnUrl);
		} catch (e) {
			console.warn(`  FETCH FAILED: ${e.message}`);
			continue;
		}
		const games = splitPgnIntoGames(pgnText);
		console.log(`  Downloaded ${games.length} games (${pgnText.length} bytes)`);

		for (const targetRound of match.games) {
			const found = games.find((g) => {
				const round = getHeader(g, "Round");
				if (!round) return false;
				const num = parseInt(round, 10);
				return !isNaN(num) && num === targetRound;
			});
			if (!found) {
				console.warn(`  MISS  round ${targetRound}`);
				continue;
			}
			let plyCount;
			try {
				plyCount = validatePgn(found);
			} catch (e) {
				console.warn(`  INVALID round ${targetRound}: ${e.message}`);
				continue;
			}
			if (plyCount === 0) {
				console.warn(`  EMPTY round ${targetRound} (no moves recorded — forfeit?)`);
				continue;
			}
			const slug = `${match.matchSlug}-${String(targetRound).padStart(2, "0")}`;
			const white = getHeader(found, "White") ?? "?";
			const black = getHeader(found, "Black") ?? "?";
			const result = getHeader(found, "Result") ?? "*";
			console.log(`  OK    round ${targetRound}: ${white} vs ${black} (${result}, ${plyCount} ply)`);
			totalSizeBytes += found.length;
			collected.push({
				id: slug,
				matchSlug: match.matchSlug,
				matchLabel: match.matchLabel,
				matchYear: match.matchYear,
				gameNumber: targetRound,
				white,
				black,
				result,
				pgn: found,
			});
		}
	}

	console.log(`\nCollected ${collected.length} games (${totalSizeBytes} bytes of raw PGN).`);

	const matchesOrdered = matches
		.map((m) => ({
			matchSlug: m.matchSlug,
			matchLabel: m.matchLabel,
			matchYear: m.matchYear,
		}))
		.filter((m) => collected.some((g) => g.matchSlug === m.matchSlug));

	const out = [];
	out.push("// AUTO-GENERATED by scripts/fetch-wcc-games.mjs — do not edit by hand.");
	out.push("// Re-run that script to refresh.");
	out.push("");
	out.push("export interface WccGameData {");
	out.push("\tid: string;");
	out.push("\tmatchSlug: string;");
	out.push("\tmatchLabel: string;");
	out.push("\tmatchYear: number;");
	out.push("\tgameNumber: number;");
	out.push("\twhite: string;");
	out.push("\tblack: string;");
	out.push("\tresult: string;");
	out.push("\tpgn: string;");
	out.push("}");
	out.push("");
	out.push("export interface WccMatchData {");
	out.push("\tmatchSlug: string;");
	out.push("\tmatchLabel: string;");
	out.push("\tmatchYear: number;");
	out.push("}");
	out.push("");
	out.push("export const WCC_MATCHES: WccMatchData[] = [");
	for (const m of matchesOrdered) {
		out.push("\t{");
		out.push(`\t\tmatchSlug: ${jsString(m.matchSlug)},`);
		out.push(`\t\tmatchLabel: ${jsString(m.matchLabel)},`);
		out.push(`\t\tmatchYear: ${m.matchYear},`);
		out.push("\t},");
	}
	out.push("];");
	out.push("");
	out.push("export const WCC_GAMES: WccGameData[] = [");
	for (const g of collected) {
		out.push("\t{");
		out.push(`\t\tid: ${jsString(g.id)},`);
		out.push(`\t\tmatchSlug: ${jsString(g.matchSlug)},`);
		out.push(`\t\tmatchLabel: ${jsString(g.matchLabel)},`);
		out.push(`\t\tmatchYear: ${g.matchYear},`);
		out.push(`\t\tgameNumber: ${g.gameNumber},`);
		out.push(`\t\twhite: ${jsString(g.white)},`);
		out.push(`\t\tblack: ${jsString(g.black)},`);
		out.push(`\t\tresult: ${jsString(g.result)},`);
		out.push(`\t\tpgn: ${jsString(g.pgn)},`);
		out.push("\t},");
	}
	out.push("];");
	out.push("");

	fs.writeFileSync(OUT_PATH, out.join("\n"), "utf8");
	console.log(`\nWrote ${OUT_PATH}`);
}

main().catch((e) => {
	console.error(e);
	process.exit(1);
});