@sns-parse/core 0.3.0-alpha.1 → 0.4.0-alpha.2

package/lib/utils/concurrency.js

@@ -0,0 +1,29 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.ConcurrencyLimiter = void 0;
+class ConcurrencyLimiter {
+    constructor(max) {
+        this.max = max;
+        this.running = 0;
+        this.queue = [];
+    }
+    async acquire() {
+        if (this.running < this.max) {
+            this.running++;
+            return;
+        }
+        return new Promise(resolve => {
+            this.queue.push(() => {
+                this.running++;
+                resolve();
+            });
+        });
+    }
+    release() {
+        this.running--;
+        const next = this.queue.shift();
+        if (next)
+            next();
+    }
+}
+exports.ConcurrencyLimiter = ConcurrencyLimiter;

package/lib/utils/field-mapping.d.ts

@@ -0,0 +1,2 @@
+export declare function getNestedValue(obj: any, path: string): any;
+export declare function parseFieldMapping(mappingStr: string): Record<string, string> | undefined;

package/lib/utils/field-mapping.js

@@ -0,0 +1,29 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.getNestedValue = getNestedValue;
+exports.parseFieldMapping = parseFieldMapping;
+function getNestedValue(obj, path) {
+    if (!path)
+        return obj;
+    const keys = path.split('.');
+    let current = obj;
+    for (const key of keys) {
+        if (current === null || current === undefined)
+            return undefined;
+        current = current[key];
+    }
+    return current;
+}
+function parseFieldMapping(mappingStr) {
+    if (!mappingStr || mappingStr.trim() === '{}' || mappingStr.trim() === '')
+        return undefined;
+    try {
+        const obj = JSON.parse(mappingStr);
+        if (typeof obj === 'object' && !Array.isArray(obj))
+            return obj;
+        return undefined;
+    }
+    catch {
+        return undefined;
+    }
+}

package/lib/utils/format.d.ts

@@ -0,0 +1,4 @@
+import type { ParsedData } from '../types';
+export declare function formatDuration(seconds: number): string;
+export declare function formatPublishTime(ms: number): string;
+export declare function generateFormattedText(p: ParsedData, format: string, index?: number, total?: number): string;

package/lib/utils/format.js

@@ -0,0 +1,79 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.formatDuration = formatDuration;
+exports.formatPublishTime = formatPublishTime;
+exports.generateFormattedText = generateFormattedText;
+const translate_1 = require("../engine/translate");
+const formatVarRegex = /\$\{([^}]+)\}/g;
+function formatDuration(seconds) {
+    if (!seconds || seconds <= 0)
+        return '';
+    const h = Math.floor(seconds / 3600);
+    const m = Math.floor((seconds % 3600) / 60);
+    const s = Math.floor(seconds % 60);
+    if (h > 0)
+        return `${h}:${m.toString().padStart(2, '0')}:${s.toString().padStart(2, '0')}`;
+    return `${m.toString().padStart(2, '0')}:${s.toString().padStart(2, '0')}`;
+}
+function formatPublishTime(ms) {
+    if (!ms)
+        return '';
+    const d = new Date(ms);
+    const y = d.getFullYear();
+    const mo = (d.getMonth() + 1).toString().padStart(2, '0');
+    const day = d.getDate().toString().padStart(2, '0');
+    const H = d.getHours().toString().padStart(2, '0');
+    const i = d.getMinutes().toString().padStart(2, '0');
+    return `${y}年${mo}月${day}日 ${H}:${i}`;
+}
+function generateFormattedText(p, format, index, total) {
+    const imageCount = p.images.length || p.live_photo.length;
+    const vars = {
+        '标题': p.title,
+        '作者': p.author,
+        '简介': p.desc,
+        '视频时长': p.duration > 0 ? formatDuration(p.duration) : '',
+        '点赞数': String(p.like),
+        '收藏数': String(p.collect),
+        '转发数': String(p.share),
+        '播放数': String(p.play),
+        '评论数': String(p.comment),
+        '发布时间': p.publishTime ? formatPublishTime(p.publishTime) : '',
+        '图片数量': String(imageCount),
+        '作者ID': p.uid,
+        '音乐标题': p.music.title || '',
+        '音乐作者': p.music.author || '',
+        '翻译': p.translation || '',
+        // 原文语言与提供方仅在有译文时展示（同一行变量全空则整行隐藏）
+        '原文语言': p.translation ? (0, translate_1.langName)(p.lang) : '',
+        '翻译提供方': p.translationProvider || '',
+    };
+    const lines = format.split('\n');
+    const resultLines = [];
+    for (const line of lines) {
+        const varMatches = line.match(formatVarRegex);
+        if (varMatches && varMatches.length > 0) {
+            let allEmptyOrZero = true;
+            for (const match of varMatches) {
+                const varName = match.slice(2, -1);
+                const val = vars[varName];
+                if (val && val !== '0') {
+                    allEmptyOrZero = false;
+                    break;
+                }
+            }
+            if (allEmptyOrZero)
+                continue;
+        }
+        let newLine = line;
+        for (const [key, val] of Object.entries(vars)) {
+            newLine = newLine.replace(new RegExp(`\\$\\{${key}\\}`, 'g'), val);
+        }
+        resultLines.push(newLine);
+    }
+    let text = resultLines.join('\n').trim();
+    if (index !== undefined && total !== undefined && total > 1) {
+        text = `【${index}/${total}】\n${text}`;
+    }
+    return text;
+}

package/lib/utils/gif.d.ts

@@ -0,0 +1,10 @@
+import type { ParserRuntime } from '../engine/runtime';
+import type { GifOptions } from '../extensions';
+export type { GifOptions };
+export declare function resolveFfmpeg(): string;
+/** 生成 ffmpeg 滤镜串（导出供测试） */
+export declare function gifFilter(width: number, fps: number): string;
+/** 裁剪实际转换时长：动图时长与上限取小；未知时长用上限 */
+export declare function gifDuration(durationSec: number, maxDurationSec: number): number;
+/** 下载视频并转 GIF；任何失败返回 null */
+export declare function mp4ToGif(rt: ParserRuntime, url: string, durationSec: number, opts: GifOptions): Promise<Buffer | null>;

package/lib/utils/gif.js

@@ -0,0 +1,104 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.resolveFfmpeg = resolveFfmpeg;
+exports.gifFilter = gifFilter;
+exports.gifDuration = gifDuration;
+exports.mp4ToGif = mp4ToGif;
+/**
+ * 视频转 GIF（推文动图用）。
+ *
+ * ffmpeg 管道两遍法（palettegen + paletteuse），不落盘。
+ * ffmpeg 不可用 / 转换失败 / 下载失败 → 返回 null，调用方回退发送原视频。
+ */
+const child_process_1 = require("child_process");
+const logger_1 = require("../logger");
+let ffmpegMissingWarned = false;
+/**
+ * ffmpeg 路径解析：内置静态二进制（optionalDependency ffmpeg-static，随插件自动安装）
+ * → 系统 PATH。解析结果记忆化，来源记入 debug 日志。
+ */
+let resolvedFfmpeg = null;
+function resolveFfmpeg() {
+    if (resolvedFfmpeg !== null)
+        return resolvedFfmpeg;
+    try {
+        const p = require('ffmpeg-static');
+        resolvedFfmpeg = typeof p === 'string' && p ? p : 'ffmpeg';
+        (0, logger_1.debugLog)(`ffmpeg 解析：内置静态二进制 ${resolvedFfmpeg}`);
+    }
+    catch {
+        resolvedFfmpeg = 'ffmpeg';
+        (0, logger_1.debugLog)('ffmpeg 解析：未安装 ffmpeg-static（--no-optional 安装？），回退系统 PATH');
+    }
+    return resolvedFfmpeg;
+}
+/** 生成 ffmpeg 滤镜串（导出供测试） */
+function gifFilter(width, fps) {
+    const w = Math.max(120, Math.min(1080, Math.round(width)));
+    const f = Math.max(5, Math.min(30, Math.round(fps)));
+    return `fps=${f},scale=${w}:-2:flags=lanczos,split[s0][s1];[s0]palettegen=max_colors=128[p];[s1][p]paletteuse=dither=bayer`;
+}
+/** 裁剪实际转换时长：动图时长与上限取小；未知时长用上限 */
+function gifDuration(durationSec, maxDurationSec) {
+    const max = Math.max(1, Math.min(60, maxDurationSec));
+    if (!durationSec || durationSec <= 0)
+        return max;
+    return Math.max(1, Math.min(max, Math.round(durationSec)));
+}
+function runFfmpeg(input, vf, durationSec) {
+    return new Promise((resolve) => {
+        const child = (0, child_process_1.spawn)(resolveFfmpeg(), [
+            '-hide_banner', '-loglevel', 'error',
+            '-i', 'pipe:0',
+            '-t', String(durationSec),
+            '-vf', vf,
+            '-f', 'gif', 'pipe:1',
+        ]);
+        const out = [];
+        let err = '';
+        child.stdout.on('data', (d) => out.push(d));
+        child.stderr.on('data', (d) => { err += d.toString(); });
+        child.on('error', (e) => {
+            if (!ffmpegMissingWarned) {
+                ffmpegMissingWarned = true;
+                logger_1.logger.warn(`GIF 转换需要 ffmpeg，但二进制不可用（动图将回退为视频发送）：${e.message}。请安装 optionalDependency ffmpeg-static，或在 PATH 中提供 ffmpeg。`);
+            }
+            resolve(null);
+        });
+        child.on('close', (code) => {
+            if (code === 0 && out.length) {
+                const gif = Buffer.concat(out);
+                logger_1.logger.info(`动图已转 GIF 发送：${Math.round(gif.length / 1024)}KB（源视频 ${Math.round(input.length / 1024)}KB）`);
+                resolve(gif);
+            }
+            else {
+                logger_1.logger.info(`动图转 GIF 失败，回退发送原视频（ffmpeg 退出码 ${code}${err ? '：' + err.slice(0, 120) : ''}）`);
+                resolve(null);
+            }
+        });
+        child.stdin.on('error', () => { });
+        child.stdin.end(input);
+    });
+}
+/** 下载视频并转 GIF；任何失败返回 null */
+async function mp4ToGif(rt, url, durationSec, opts) {
+    try {
+        // X 媒体 CDN (video.twimg.com) 需要浏览器级 UA + Referer 才不会 403（与 vault 下载一致）
+        const res = await rt.http.get(url, {
+            responseType: 'arraybuffer',
+            timeout: Math.min(rt.config.timeout || 60000, 120000),
+            headers: {
+                'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/124.0.0.0 Safari/537.36',
+                'Referer': 'https://twitter.com/',
+            },
+        });
+        const input = Buffer.from(res.data);
+        if (!input.length)
+            return null;
+        return await runFfmpeg(input, gifFilter(opts.maxWidth, opts.fps), gifDuration(durationSec, opts.maxDurationSec));
+    }
+    catch (e) {
+        logger_1.logger.info(`动图 GIF 转换跳过（源视频下载失败，回退发送原视频）：${e?.message || e}`);
+        return null;
+    }
+}

package/lib/utils/image-size.d.ts

@@ -0,0 +1,10 @@
+/**
+ * 图片尺寸探测（无依赖）：PNG / GIF / JPEG / WebP 头部解析。
+ * 仅供同源切图识别用：只读宽高，不解码像素。
+ */
+export interface ImageSize {
+    width: number;
+    height: number;
+}
+/** 探测图片缓冲区尺寸；无法识别返回 null */
+export declare function probeImageSize(buf: Buffer): ImageSize | null;

package/lib/utils/image-size.js

@@ -0,0 +1,69 @@
+"use strict";
+/**
+ * 图片尺寸探测（无依赖）：PNG / GIF / JPEG / WebP 头部解析。
+ * 仅供同源切图识别用：只读宽高，不解码像素。
+ */
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.probeImageSize = probeImageSize;
+/** 探测图片缓冲区尺寸；无法识别返回 null */
+function probeImageSize(buf) {
+    if (!buf || buf.length < 12)
+        return null;
+    // PNG: 89 50 4E 47 0D 0A 1A 0A，IHDR 宽高在 16..24
+    if (buf[0] === 0x89 && buf[1] === 0x50 && buf[2] === 0x4e && buf[3] === 0x47) {
+        if (buf.length < 24)
+            return null;
+        return { width: buf.readUInt32BE(16), height: buf.readUInt32BE(20) };
+    }
+    // GIF: GIF8
+    if (buf.toString('ascii', 0, 4) === 'GIF8') {
+        if (buf.length < 10)
+            return null;
+        return { width: buf.readUInt16LE(6), height: buf.readUInt16LE(8) };
+    }
+    // JPEG: FF D8，扫描 SOF0/1/2/...（C0-CF，除 C4/C8/CC）段
+    if (buf[0] === 0xff && buf[1] === 0xd8) {
+        let off = 2;
+        while (off + 9 < buf.length) {
+            if (buf[off] !== 0xff) {
+                off++;
+                continue;
+            }
+            const marker = buf[off + 1];
+            if (marker >= 0xc0 && marker <= 0xcf && marker !== 0xc4 && marker !== 0xc8 && marker !== 0xcc) {
+                const height = buf.readUInt16BE(off + 5);
+                const width = buf.readUInt16BE(off + 7);
+                if (width && height)
+                    return { width, height };
+                return null;
+            }
+            if (marker === 0xd8 || (marker >= 0xd0 && marker <= 0xd9)) {
+                off += 2;
+                continue;
+            }
+            const segLen = buf.readUInt16BE(off + 2);
+            if (segLen < 2)
+                return null;
+            off += 2 + segLen;
+        }
+        return null;
+    }
+    // WebP: RIFF....WEBP
+    if (buf.toString('ascii', 0, 4) === 'RIFF' && buf.toString('ascii', 8, 12) === 'WEBP') {
+        const fourcc = buf.toString('ascii', 12, 16);
+        if (fourcc === 'VP8 ' && buf.length >= 30) {
+            return { width: buf.readUInt16LE(26) & 0x3fff, height: buf.readUInt16LE(28) & 0x3fff };
+        }
+        if (fourcc === 'VP8L' && buf.length >= 25) {
+            const b = buf.readUInt32LE(21);
+            return { width: (b & 0x3fff) + 1, height: ((b >> 14) & 0x3fff) + 1 };
+        }
+        if (fourcc === 'VP8X' && buf.length >= 30) {
+            const w = 1 + (buf[24] | (buf[25] << 8) | (buf[26] << 16));
+            const h = 1 + (buf[27] | (buf[28] << 8) | (buf[29] << 16));
+            return { width: w, height: h };
+        }
+        return null;
+    }
+    return null;
+}

package/lib/utils/merge.d.ts

@@ -0,0 +1,61 @@
+import type { ParserRuntime } from '../engine/runtime';
+import { type ImageSize } from './image-size';
+export type MergeLayout = {
+    kind: 'grid';
+    cols: number;
+    rows: number;
+} | {
+    kind: 'v';
+} | {
+    kind: 'h';
+};
+/** 仅凭数量与尺寸给出全部可行布局（内容验证前） */
+export declare function candidateLayouts(n: number, sizes: ImageSize[]): MergeLayout[];
+export interface SeamVerdict {
+    /** false = 接缝不可验证（边界带平坦，无纹理可判）或不连续 */
+    ok: boolean;
+    score: number;
+    /** 证据明细（供 debug 日志） */
+    dir: 'V' | 'H';
+    from: number;
+    to: number;
+    /** 边界带高频纹理能量（两侧最大值；低于 TEXTURE_MIN 即不可验证） */
+    texture: number;
+    /** 平滑后接缝差 */
+    trendMAD: number;
+    /** 片内自然步进基线 */
+    baseline: number;
+    /** 结论文案 */
+    reason: string;
+}
+/** 内容验证：所有接缝（趋势+纹理）均通过才成立；返回最大接缝比值与逐缝证据 */
+export declare function verifyLayout(layout: MergeLayout, grays: Buffer[]): {
+    pass: boolean;
+    score: number;
+    seams: SeamVerdict[];
+};
+/** 内容识别：候选布局逐一经接缝连续性验证，取通过者中比值最小的一个 */
+export declare function detectMergeLayout(sizes: ImageSize[], grays: Buffer[]): MergeLayout | null;
+/** 布局名（日志用） */
+export declare function layoutName(layout: MergeLayout): string;
+/**
+ * 候选布局逐一验证并产出证据链；返回通过者中证据最强（worst 接缝比值最小）的一个。
+ * 每次调用都会把逐候选、逐接缝的判定依据写入 debug 日志（含尺寸与灰度行数）。
+ */
+export declare function pickMergeLayout(sizes: ImageSize[], grays: Buffer[]): {
+    layout: MergeLayout;
+    score: number;
+    seams: SeamVerdict[];
+} | null;
+/** xstack 均匀网格布局串（各片已统一缩放为 w×h，行优先） */
+export declare function xstackLayout(cols: number, rows: number, w: number, h: number): string;
+/** 生成 ffmpeg filter_complex（输出标签恒为 [out]，导出供测试） */
+export declare function buildMergeFilter(n: number, layout: MergeLayout, sizes: ImageSize[]): string;
+/**
+ * 下载 → 灰度解码 → 内容识别接缝 → 合并。
+ * 不构成母图切片（接缝不连续）或任一环节失败 → null。
+ */
+export declare function mergeImages(rt: ParserRuntime, urls: string[]): Promise<{
+    buffer: Buffer;
+    layout: MergeLayout;
+} | null>;