koishi-plugin-bns-blacklist 1.0.13 → 1.0.15

package/lib/index.js

@@ -175,52 +175,124 @@ function bruteForceExtractStrings(buf) {
     return strings;
 }
 // ─── Row reconstruction ──────────────────────────────────
-// Heuristic: strings appear in row-major order in the protobuf.
-// We identify rows by looking for date patterns (YYYY/M/D) as row delimiters.
+// Group strings into rows. Each row starts with a player name (contains CJK).
+// Single-char strings between fields are protobuf metadata - skip them.
 function reconstructRows(strings, logger) {
-    // Filter out metadata strings (colors, fonts, hex values, etc.)
-    const dataStrings = strings.filter(s => {
-        if (/^[0-9A-Fa-f]{6}$/.test(s))
-            return false; // hex colors
-        if (/^\d+\.\d+\.\d+$/.test(s))
-            return false; // version numbers
-        if (s === 'Microsoft YaHei' || s === '工作表1' || s.startsWith('BB08J2'))
+    // First pass: merge adjacent non-metadata strings into fields
+    const merged = [];
+    let buf = '';
+    for (const s of strings) {
+        // Skip single-char protobuf metadata
+        if (s.length === 1 && !/[一-鿿]/.test(s)) {
+            if (buf) {
+                merged.push(buf.trim());
+                buf = '';
+            }
+        }
+        else {
+            buf += (buf ? '' : '') + s;
+        }
+    }
+    if (buf.trim())
+        merged.push(buf.trim());
+    // Filter out header/metadata lines
+    const data = merged.filter(s => {
+        if (s.startsWith('工作表') || s === '序号' || s === '事发ID')
+            return false;
+        if (s === '服务器' || s === '小号1' || s === '小号2')
+            return false;
+        if (s === '小号3' || s === '小号4' || s === '上榜原因')
+            return false;
+        if (s === '副本' || s === '举证截图' || s === '时间')
             return false;
-        if (s.length === 1 && /[\x00-\x1f]/.test(s))
+        if (s === '脚本混野队' || s.startsWith('——申诉'))
             return false;
-        return s.trim().length >= 2;
+        return s.length >= 2;
     });
     if (logger)
-        logger.info(`Data strings after filter: ${dataStrings.length}`);
-    // Try to group into rows using date pattern as row delimiter
+        logger.info(`Filtered to ${data.length} fields: ${data.slice(0, 30).join(' ||| ')}`);
+    // Group into rows: a row starts when we see a field that looks like a player name
+    // (not a pure number, not a pure date, contains CJK, not starting with digits)
     const rows = [];
-    let currentRow = [];
-    const datePattern = /^\d{4}\/\d{1,2}\/\d{1,2}/;
-    for (const s of dataStrings) {
-        currentRow.push(s);
-        if (datePattern.test(s)) {
-            // Date marks end of a row, start new one
-            if (currentRow.length >= 3) {
-                rows.push([...currentRow]);
-            }
-            currentRow = [];
+    let current = [];
+    for (const s of data) {
+        const isPlayerName = /[一-鿿]/.test(s) && !/^\d{1,2}$/.test(s) && !/^\d{4}\/\d{1,2}\/\d{1,2}/.test(s) && !/^\d+$/.test(s);
+        if (isPlayerName && current.length > 0) {
+            // New row starting - save previous if it has enough data
+            if (current.length >= 2)
+                rows.push([...current]);
+            current = [];
         }
+        current.push(s);
     }
-    // Don't forget the last row
-    if (currentRow.length >= 3) {
-        rows.push([...currentRow]);
-    }
-    // If date-based splitting didn't work, try fixed-width approach
-    if (rows.length < 2) {
-        // Assume each row has the same number of fields
-        // Find the most common field count by looking for patterns
-        const nonMetaStrings = dataStrings.filter(s => !/^\d+$/.test(s) && s.length > 1);
-        if (logger)
-            logger.info(`Non-meta strings: ${nonMetaStrings.join(' | ')}`);
-        return [];
+    if (current.length >= 2)
+        rows.push([...current]);
+    if (logger)
+        logger.info(`Grouped into ${rows.length} rows`);
+    // Normalize rows to our expected format
+    const result = [];
+    for (const row of rows) {
+        if (row.length < 2)
+            continue;
+        // Expected: 0=name, 1=server, [2-5]=alts, then description, dungeon, date
+        const mapped = [];
+        mapped[0] = row[0] || ''; // name (player name)
+        mapped[1] = ''; // class - not in this sheet, try to extract from name like "九天揽月丷（剑士）"
+        mapped[2] = row[1] || ''; // server
+        // violation is embedded in description, we'll infer it
+        mapped[3] = ''; // violation (inferred from description)
+        // combine alts (positions 2-5 in original)
+        const alts = [];
+        for (let i = 2; i < row.length && i < 6; i++) {
+            if (row[i] && !/^\d{4}\/\d/.test(row[i]) && !/^\d+$/.test(row[i]))
+                alts.push(row[i]);
+        }
+        mapped[4] = alts.join('、'); // alts
+        // description: typically the field before dungeon name
+        const dungeonIdx = row.findIndex((s, idx) => idx > 2 && isDungeonName(s));
+        if (dungeonIdx > 2) {
+            mapped[5] = row.slice(6, dungeonIdx).filter(s => !/^\d+$/.test(s)).join(' '); // description
+            mapped[7] = row[dungeonIdx]; // dungeon
+        }
+        else {
+            mapped[5] = ''; // description
+            mapped[7] = '';
+        }
+        mapped[6] = ''; // witness
+        mapped[8] = row[row.length - 1] || ''; // date (last field)
+        // Infer violation type from description
+        mapped[3] = inferViolation(mapped[5]);
+        // Extract class from name if present
+        const classMatch = mapped[0].match(/[（(]([^)）]+)[）)]/);
+        if (classMatch) {
+            mapped[1] = classMatch[1];
+            mapped[0] = mapped[0].replace(/[（(][^)）]+[）)]/, '').trim();
+        }
+        result.push(mapped);
     }
-    return rows;
+    return result;
+}
+function isDungeonName(s) {
+    const dungeons = ['铁拳', '狗头', '狼人P3', '冰库', '雪人', '36金', '狼人', '—'];
+    return dungeons.some(d => s.includes(d)) && s.length <= 10;
+}
+function inferViolation(desc) {
+    const d = desc || '';
+    if (/脚本/.test(d))
+        return '脚本';
+    if (/跳车/.test(d))
+        return '跳车';
+    if (/双开/.test(d))
+        return '双开';
+    if (/挂机/.test(d))
+        return '挂机';
+    if (/嘴臭|骂/.test(d))
+        return '嘴臭';
+    if (/打铁/.test(d))
+        return '打铁';
+    return '其他';
 }
+// ─── Row reconstruction ends ─────────────────────────────
 // ─── KV Format Parser ────────────────────────────────────
 function parseTencentKvFormat(text, logger) {
     const parts = text.split('\n').map(s => s.trim()).filter(s => s.length > 0);
@@ -378,20 +450,35 @@ ${itemsHtml}
 // ─── Scraping ────────────────────────────────────────────
 async function scrapeSheet(page, url, logger) {
     await page.goto(url, { waitUntil: 'networkidle2', timeout: 60000 });
-    await new Promise(r => setTimeout(r, 8000));
+    // Wait longer for the sheet data to fully load
+    await new Promise(r => setTimeout(r, 15000));
     if (logger)
         logger.info(`Page: "${await page.title()}"`);
-    // Fetch the dop-api data
-    const apiText = await page.evaluate(async () => {
-        const resp = await fetch(`https://docs.qq.com/dop-api/sheet/data?id=DZXRuRmtTS21mWU9V`, { credentials: 'include' });
-        return await resp.text();
-    });
+    // Fetch the dop-api data - retry a few times in case of incomplete load
+    let apiText = '';
+    for (let attempt = 0; attempt < 3; attempt++) {
+        apiText = await page.evaluate(async () => {
+            const resp = await fetch(`https://docs.qq.com/dop-api/sheet/data?id=DZXRuRmtTS21mWU9V`, { credentials: 'include' });
+            return await resp.text();
+        });
+        if (apiText.length > 5000)
+            break;
+        if (logger)
+            logger.info(`API response too short (${apiText.length}), retrying...`);
+        await new Promise(r => setTimeout(r, 3000));
+    }
+    if (logger)
+        logger.info(`API response: ${apiText.length} bytes`);
     const kvData = parseTencentKvFormat(apiText, logger);
     if (!kvData) {
         if (logger)
             logger.info('KV parse failed');
         return [];
     }
+    // Log all keys and sizes
+    const kvInfo = Object.entries(kvData).map(([k, v]) => `${k}(${v.length})`).join(', ');
+    if (logger)
+        logger.info(`KV keys: ${kvInfo}`);
     const chunkKey = Object.keys(kvData).find(k => k.startsWith('chunk_'));
     if (!chunkKey) {
         if (logger)

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "koishi-plugin-bns-blacklist",
-  "version": "1.0.13",
+  "version": "1.0.15",
   "description": "剑灵怀旧服副本黑名单查询 - 从腾讯文档实时拉取数据，输入角色名即可生成黑名单截图",
   "main": "lib/index.js",
   "typings": "lib/index.d.ts",