LittleWhiteBox/modules/variables/state2/parser.js

import jsyaml from '../../../libs/js-yaml.mjs';

const STATE_TAG_RE = /<\s*state\b[^>]*>([\s\S]*?)<\s*\/\s*state\s*>/gi;

export function extractStateBlocks(text) {
    const s = String(text ?? '');
    if (!s || s.toLowerCase().indexOf('<state') === -1) return [];
    const out = [];
    STATE_TAG_RE.lastIndex = 0;
    let m;
    while ((m = STATE_TAG_RE.exec(s)) !== null) {
        const inner = String(m[1] ?? '');
        if (inner.trim()) out.push(inner);
    }
    return out;
}

export function computeStateSignature(text) {
    const s = String(text ?? '');
    if (!s || s.toLowerCase().indexOf('<state') === -1) return '';
    const chunks = [];
    STATE_TAG_RE.lastIndex = 0;
    let m;
    while ((m = STATE_TAG_RE.exec(s)) !== null) chunks.push(String(m[0] ?? '').trim());
    return chunks.length ? chunks.join('\n---\n') : '';
}

/**
 * 解析 <state> 块内容 -> ops[]
 * 单行支持运算符，多行只支持覆盖 set（YAML）
 */
export function parseStateBlock(content) {
    const results = [];
    const lines = String(content ?? '').split(/\r?\n/);

    let pendingPath = null;
    let pendingLines = [];

    const flushPending = () => {
        if (!pendingPath) return;
        // 没有任何缩进行：视为 set 空字符串
        if (!pendingLines.length) {
            results.push({ path: pendingPath, op: 'set', value: '' });
            pendingPath = null;
            pendingLines = [];
            return;
        }
        try {
            // 去除公共缩进
            const nonEmpty = pendingLines.filter(l => l.trim());
            const minIndent = nonEmpty.length
                ? Math.min(...nonEmpty.map(l => l.search(/\S/)))
                : 0;

            const yamlText = pendingLines
                .map(l => (l.trim() ? l.slice(minIndent) : ''))
                .join('\n');

            const obj = jsyaml.load(yamlText);
            results.push({ path: pendingPath, op: 'set', value: obj });
        } catch (e) {
            results.push({ path: pendingPath, op: 'set', value: null, warning: `YAML 解析失败: ${e.message}` });
        } finally {
            pendingPath = null;
            pendingLines = [];
        }
    };

    for (const raw of lines) {
        const trimmed = raw.trim();
        if (!trimmed || trimmed.startsWith('#')) continue;

        const indent = raw.search(/\S/);

        if (indent === 0) {
            flushPending();
            const colonIdx = findTopLevelColon(trimmed);
            if (colonIdx === -1) continue;

            const path = trimmed.slice(0, colonIdx).trim();
            const rhs = trimmed.slice(colonIdx + 1).trim();
            if (!path) continue;

            if (!rhs) {
                pendingPath = path;
                pendingLines = [];
            } else {
                results.push({ path, ...parseInlineValue(rhs) });
            }
        } else if (pendingPath) {
            pendingLines.push(raw);
        }
    }

    flushPending();
    return results;
}

function findTopLevelColon(line) {
    let inQuote = false;
    let q = '';
    let esc = false;
    for (let i = 0; i < line.length; i++) {
        const ch = line[i];
        if (esc) { esc = false; continue; }
        if (ch === '\\') { esc = true; continue; }
        if (!inQuote && (ch === '"' || ch === "'")) { inQuote = true; q = ch; continue; }
        if (inQuote && ch === q) { inQuote = false; q = ''; continue; }
        if (!inQuote && ch === ':') return i;
    }
    return -1;
}

function unescapeString(s) {
    return String(s ?? '')
        .replace(/\\n/g, '\n')
        .replace(/\\t/g, '\t')
        .replace(/\\r/g, '\r')
        .replace(/\\"/g, '"')
        .replace(/\\'/g, "'")
        .replace(/\\\\/g, '\\');
}

/**
 * 单行内联值解析
 */
export function parseInlineValue(raw) {
    const t = String(raw ?? '').trim();

    if (t === 'null') return { op: 'del' };

    // (负数) 用于强制 set -5，而不是 inc -5
    const parenNum = t.match(/^\((-?\d+(?:\.\d+)?)\)$/);
    if (parenNum) return { op: 'set', value: Number(parenNum[1]) };

    // +10 / -20
    if (/^\+\d/.test(t) || /^-\d/.test(t)) {
        const n = Number(t);
        if (Number.isFinite(n)) return { op: 'inc', delta: n };
    }

    // +"str" / +'str'
    const pushD = t.match(/^\+"((?:[^"\\]|\\.)*)"\s*$/);
    if (pushD) return { op: 'push', value: unescapeString(pushD[1]) };
    const pushS = t.match(/^\+'((?:[^'\\]|\\.)*)'\s*$/);
    if (pushS) return { op: 'push', value: unescapeString(pushS[1]) };

    // +[...]
    if (t.startsWith('+[')) {
        try {
            const arr = JSON.parse(t.slice(1));
            if (Array.isArray(arr)) return { op: 'push', value: arr };
            return { op: 'set', value: t, warning: '+[] 不是数组，作为字符串' };
        } catch {
            return { op: 'set', value: t, warning: '+[] JSON 解析失败，作为字符串' };
        }
    }

    // -"str" / -'str'
    const popD = t.match(/^-"((?:[^"\\]|\\.)*)"\s*$/);
    if (popD) return { op: 'pop', value: unescapeString(popD[1]) };
    const popS = t.match(/^-'((?:[^'\\]|\\.)*)'\s*$/);
    if (popS) return { op: 'pop', value: unescapeString(popS[1]) };

    // -[...]
    if (t.startsWith('-[')) {
        try {
            const arr = JSON.parse(t.slice(1));
            if (Array.isArray(arr)) return { op: 'pop', value: arr };
            return { op: 'set', value: t, warning: '-[] 不是数组，作为字符串' };
        } catch {
            return { op: 'set', value: t, warning: '-[] JSON 解析失败，作为字符串' };
        }
    }

    // 裸数字 set
    if (/^-?\d+(?:\.\d+)?$/.test(t)) return { op: 'set', value: Number(t) };

    // "str" / 'str'
    const strD = t.match(/^"((?:[^"\\]|\\.)*)"\s*$/);
    if (strD) return { op: 'set', value: unescapeString(strD[1]) };
    const strS = t.match(/^'((?:[^'\\]|\\.)*)'\s*$/);
    if (strS) return { op: 'set', value: unescapeString(strS[1]) };

    if (t === 'true') return { op: 'set', value: true };
    if (t === 'false') return { op: 'set', value: false };

    // JSON set
    if (t.startsWith('{') || t.startsWith('[')) {
        try { return { op: 'set', value: JSON.parse(t) }; }
        catch { return { op: 'set', value: t, warning: 'JSON 解析失败，作为字符串' }; }
    }

    // 兜底 set 原文本
    return { op: 'set', value: t };
}