specific-ai-cli 0.0.0-stage → 0.1.1

package/dist/platforms/jd/web/util.js

@@ -0,0 +1,175 @@
+import * as n from '../../../core/normalize.js';
+import { pyStr } from '../../../core/py.js';
+import * as rand from '../../../core/rand.js';
+import { DEFAULT_AREA } from './profile.js';
+/** 纯算工具（上游 utils/jd_util.py、utils/jd_cookie.py、utils/trace_headers.py、utils/aks.py 的编码部分）。 */
+export { sha256Hex } from '../../../core/hash.js';
+/** 商品页 URL 里的 SKU：PC、移动（item.m.jd.com/product/）、全球购（item.jd.hk、npcitem.jd.hk）。 */
+export const ITEM_URL_SKU = /(?:item(?:\.m)?\.jd\.(?:com|hk)|npcitem\.jd\.hk)\/(?:product\/)?(\d{4,})\.html/;
+/** 去掉搜索结果里的高亮标签。 */
+export function stripTags(text) {
+    if (!text)
+        return '';
+    return String(text).replace(/<[^>]+>/g, '');
+}
+/** 仿浏览器的 jQuery jsonp 回调名（jQuery2905208）。 */
+export function randomJqueryCallback() {
+    return `jQuery${rand.randint(1000000, 9999999)}`;
+}
+/** 咚咚 WS 的 `_wid_`：8-4-4-4-12 的随机十六进制。 */
+export function generateWid() {
+    const seg = () => rand.randint(0x10000, 0x1ffff).toString(16).slice(1);
+    return `${seg()}${seg()}-${seg()}-${seg()}-${seg()}-${seg()}${seg()}${seg()}`;
+}
+/** 咚咚会话 ID（getSessionId）。 */
+export function sessionId(pin, app, venderId) {
+    return `${pin.toLowerCase()}:${app}:${venderId}`;
+}
+/** 从 `ipLoc-djd` 取收货地区；full 时保留 `.addressId` 后缀（area_of）。 */
+export function areaOf(cookies, full = false) {
+    const raw = (cookies.get('ipLoc-djd') ?? '').trim();
+    let area = raw ? raw.replaceAll('-', '_') : DEFAULT_AREA;
+    if (!full)
+        area = area.split('.')[0];
+    return area;
+}
+/** 搜索页用 `__jda` 的第二段作 uuid，缺省为 -1（search_uuid_of）。 */
+export function searchUuidOf(cookies) {
+    const fields = String(cookies.get('__jda') ?? '').split('.');
+    if (fields.length > 1 && fields[1] !== '' && fields[1] !== '-')
+        return fields[1];
+    return '-1';
+}
+/** JS `encodeURIComponent`（jQuery.param 用的编码，aks.encode_component）。 */
+export function encodeComponent(value) {
+    const s = value == null ? '' : pyStr(value);
+    let out = '';
+    for (const byte of Buffer.from(s, 'utf8')) {
+        const ch = String.fromCharCode(byte);
+        out += byte < 0x80 && /[A-Za-z0-9\-_.!~*'()]/.test(ch) ? ch : '%' + byte.toString(16).toUpperCase().padStart(2, '0');
+    }
+    return out;
+}
+/** 按 jQuery 的方式序列化有序键值对（aks.encode_query_pairs）。 */
+export function encodeQueryPairs(pairs) {
+    return pairs.map(([k, v]) => `${encodeComponent(k)}=${encodeComponent(v)}`).join('&');
+}
+// ---------------------------------------------------------------- 埋点 cookie 冷启动（utils/jd_cookie.py）
+/** 各站点的埋点 siteId。 */
+export const SITE_IDS = {
+    www: '122270672',
+    search: '143920055',
+    item: '122270672',
+    passport: '95931165',
+    chat: '23334881',
+};
+/** `__jdu` = 13 位毫秒时间戳 + 10 位随机数字。 */
+export function generateJdu() {
+    return `${rand.now()}${rand.randint(10 ** 9, 10 ** 10 - 1)}`;
+}
+/** `__jdu` / `__jdc` / `__jda` / `__jdb` / `__jdv` 一组埋点 cookie。 */
+export function trackingCookies(site = 'search', jdu, source = 'direct', visitCount = 1, pv = 1) {
+    const siteId = SITE_IDS[site] ?? SITE_IDS.www;
+    const u = jdu || generateJdu();
+    const ms = rand.now();
+    const sec = Math.floor(ms / 1000);
+    return [
+        ['__jdu', u],
+        ['__jdc', siteId],
+        ['__jda', `${siteId}.${u}.${sec}.${sec}.${sec}.${visitCount}`],
+        ['__jdb', `${siteId}.${pv}.${u}|${visitCount}.${sec}`],
+        ['__jdv', `${siteId}|${source}|-|direct|-|${ms}`],
+    ];
+}
+/** 给空会话灌冷启动 cookie（bootstrap）：埋点 cookie + 收货地区。 */
+export function bootstrapCookies(site = 'search', area) {
+    const a = area || DEFAULT_AREA.replaceAll('_', '-');
+    return [...trackingCookies(site), ['areaId', a.split('-')[0]], ['ipLoc-djd', a]];
+}
+// ---------------------------------------------------------------- 登录页 SGM / JDAS 链路头（utils/trace_headers.py）
+/** 取 `secrets.randbits(k)`：与对拍框架的 getrandbits 一致，每 32 位一个随机数。 */
+function randbits(k) {
+    let out = 0n;
+    let shift = 0n;
+    while (k > 0) {
+        const bits = Math.min(k, 32);
+        out |= BigInt(Math.floor(rand.random() * 2 ** bits)) << shift;
+        shift += BigInt(bits);
+        k -= bits;
+    }
+    return out;
+}
+/** RFC 9562 UUIDv7。 */
+export function uuid7() {
+    const ts = BigInt(rand.now()) & ((1n << 48n) - 1n);
+    const randA = randbits(12);
+    const randB = randbits(62);
+    let v = (ts << 80n) | (0x7n << 76n) | (randA << 64n);
+    v |= (2n << 62n) | randB;
+    const h = v.toString(16).padStart(32, '0');
+    return `${h.slice(0, 8)}-${h.slice(8, 12)}-${h.slice(12, 16)}-${h.slice(16, 20)}-${h.slice(20)}`;
+}
+/** sgm-web 3.3.0 的数字 key。 */
+function sgmKey() {
+    let out = '';
+    for (const ch of '1000-1000-4000-8000-1000000') {
+        if (!'018'.includes(ch)) {
+            out += ch;
+            continue;
+        }
+        const digit = Number(ch);
+        const shift = Math.trunc(digit / 4);
+        const part = Number(randbits(8)) & (15 >> shift);
+        out += String(digit ^ part);
+    }
+    const value = out.replaceAll('-', '').slice(0, 18);
+    return BigInt(value || '0').toString();
+}
+/** 登录页的 `#uuid` 取作 JDAS-Page-Id。 */
+export function pageIdFromHtml(html) {
+    const inputs = html.match(/<input\b[^>]*>/gi) ?? [];
+    for (const tag of inputs) {
+        const attrs = parseAttrs(tag);
+        if (attrs.id === 'uuid' && attrs.value)
+            return attrs.value.trim();
+    }
+    throw new Error('登录页 HTML 缺少 #uuid，无法生成 JDAS-Page-Id');
+}
+/** 一个 input 标签的属性（小写名 → 值，实体已解码）。 */
+export function parseAttrs(tag) {
+    const out = {};
+    const re = /([^\s"'<>/=]+)(?:\s*=\s*(?:"([^"]*)"|'([^']*)'|([^\s"'=<>`]+)))?/g;
+    const body = tag.replace(/^<\s*[a-zA-Z0-9]+/, '').replace(/\/?>$/, '');
+    let m;
+    while ((m = re.exec(body))) {
+        const name = m[1].toLowerCase();
+        if (name in out)
+            continue;
+        out[name] = n.unescapeHtml(m[2] ?? m[3] ?? m[4] ?? '');
+    }
+    return out;
+}
+/** 同一登录页内的链路上下文：Page-Id 取自页面，Session-Id 页面内复用（LoginTraceContext）。 */
+export class TraceContext {
+    pageId;
+    sessionId;
+    constructor(pageId, sessionId = (() => {
+        const g = uuid7();
+        return g.slice(0, -4) + '0001';
+    })()) {
+        this.pageId = pageId;
+        this.sessionId = sessionId;
+    }
+    static fromHtml(html) {
+        return new TraceContext(pageIdFromHtml(html));
+    }
+    nextHeaders() {
+        const key = sgmKey();
+        return {
+            'sgm-context': `${key};${key}`,
+            'jdas-trace-id': uuid7(),
+            'jdas-page-id': this.pageId,
+            'jdas-session-id': this.sessionId,
+        };
+    }
+}

package/dist/platforms/jd/web/webm.js

@@ -0,0 +1,133 @@
+import * as realFs from 'node:fs';
+import { createRequire } from 'node:module';
+import { join } from 'node:path';
+import vm from 'node:vm';
+import { SpecificAIError } from '../../../core/errors.js';
+import { staticFile } from '../../../core/paths.js';
+import { DEFAULT_NOW, DEFAULT_SEED, isDeterministic, now } from '../../../core/rand.js';
+import { createContext } from '../../../core/vm.js';
+import { PROFILE } from './profile.js';
+/**
+ * 京东 PC WebM 指纹（utils/webm.py + static/webm/env/run.js）：原版 jdwebm.js 跑在 jsdom 里，产出 wsgw_getinfo 的正文。
+ * run.js 本是命令行脚本（读 stdin、写 stdout），这里放进 vm context，stdin / stdout 换成内存。
+ */
+const hostRequire = createRequire(import.meta.url);
+const RESULT_PREFIX = '__WEBM_RESULT__';
+/** jsdom window realm 的确定性（对拍时与 scripts/golden/jd/preload.cjs 相同）。 */
+const REALM_JS = `(() => {
+  let a = __specific_ai_seed >>> 0
+  Math.random = () => { a = (a + 0x6d2b79f5) >>> 0; let t = Math.imul(a ^ (a >>> 15), a | 1); t = (t + Math.imul(t ^ (t >>> 7), t | 61)) ^ t; return ((t ^ (t >>> 14)) >>> 0) / 4294967296 }
+  const D = Date
+  globalThis.Date = class extends D { constructor(...x) { super(...(x.length ? x : [__specific_ai_now])) } static now() { return __specific_ai_now } }
+  try { const fill = (arr) => { const span = 2 ** (8 * arr.BYTES_PER_ELEMENT); for (let i = 0; i < arr.length; i++) arr[i] = Math.floor(Math.random() * span); return arr }; Object.defineProperty(crypto, 'getRandomValues', { value: fill, configurable: true, writable: true }) } catch (e) {}
+  try { const t0 = __specific_ai_now; Object.defineProperty(performance, 'now', { value: () => 0, configurable: true, writable: true }); Object.defineProperty(performance, 'timeOrigin', { value: t0, configurable: true }) } catch (e) {}
+})()`;
+/** 给脚本的 jsdom：确定性模式下，每个新 window 的 realm 固定随机数与时钟。 */
+export function deterministicJsdom() {
+    const m = hostRequire('jsdom');
+    if (!isDeterministic())
+        return m;
+    const Base = m.JSDOM;
+    class JSDOM extends Base {
+        constructor(...args) {
+            super(...args);
+            try {
+                const ctx = this.getInternalVMContext();
+                ctx.__specific_ai_seed = DEFAULT_SEED;
+                ctx.__specific_ai_now = now() || DEFAULT_NOW;
+                vm.runInContext(REALM_JS, ctx);
+            }
+            catch { }
+        }
+    }
+    return Object.assign(Object.create(m), { JSDOM });
+}
+/** 在 vm 里跑一个命令行脚本：stdin 给定，stdout 收集，脚本调用 process.exit 或写出结果行后结束。 */
+export async function runCliScript(file, input, options) {
+    const dir = join(file, '..');
+    let out = '';
+    let resolveDone;
+    const done = new Promise((r) => (resolveDone = r));
+    const onChunk = (chunk) => {
+        out += String(chunk);
+        let i;
+        while ((i = out.indexOf('\n')) >= 0) {
+            const line = out.slice(0, i);
+            out = out.slice(i + 1);
+            if (line.startsWith(options.resultPrefix))
+                resolveDone(line.slice(options.resultPrefix.length));
+            else
+                options.log?.(line);
+        }
+        return true;
+    };
+    const fakeStdout = { write: onChunk, isTTY: false };
+    const fs = {
+        ...realFs,
+        readFileSync: (p, ...rest) => (p === 0 ? input : realFs.readFileSync(p, ...rest)),
+    };
+    const modules = { fs, ...options.modules };
+    const fakeProcess = new Proxy(process, {
+        get: (t, k) => (k === 'stdout' ? fakeStdout : k === 'exit' ? () => resolveDone('') : k === 'env' ? {} : Reflect.get(t, k)),
+        set: (_t, k) => k === 'exitCode',
+    });
+    const moduleShim = {
+        createRequire: () => (name) => (name in modules ? modules[name] : hostRequire(name)),
+    };
+    const req = (name) => (name === 'module' ? moduleShim : name in modules ? modules[name] : hostRequire(name));
+    const quiet = (...a) => options.log?.(a.map(String).join(' '));
+    const context = createContext({
+        require: req,
+        process: fakeProcess,
+        __dirname: dir,
+        __filename: file,
+        module: { exports: {} },
+        exports: {},
+        console: { log: quiet, info: quiet, warn: quiet, error: quiet, debug: quiet },
+        Buffer,
+    });
+    context.global = context;
+    vm.runInContext(realFs.readFileSync(file, 'utf8'), context, { filename: file });
+    let timer;
+    const timeout = new Promise((_, reject) => {
+        timer = setTimeout(() => reject(new SpecificAIError('ERROR', `${file} 运行超时`)), options.timeout);
+    });
+    try {
+        return await Promise.race([done, timeout]);
+    }
+    finally {
+        clearTimeout(timer);
+    }
+}
+/** 跑原版 jdwebm.js，返回完整的 wsgw_getinfo 正文，同时把它写的 cookie 和 localStorage 交回会话（build_search_payload）。 */
+export async function buildSearchPayload(jd, pageUrl, configData = '', localStorage) {
+    const input = {
+        pageUrl: String(pageUrl),
+        userAgent: PROFILE.ua,
+        cookies: Object.fromEntries([...jd.cookies].filter(([, v]) => v != null && v !== '')),
+        localStorage: localStorage ?? jd.localStorageFor(pageUrl),
+        configData: String(configData ?? ''),
+    };
+    const marker = await runCliScript(staticFile('jd', 'webm/env/run.js'), JSON.stringify(input), {
+        modules: { jsdom: deterministicJsdom() },
+        resultPrefix: RESULT_PREFIX,
+        timeout: 30_000,
+        log: (line) => jd.ctx.log.debug(`[webm] ${line.slice(0, 300)}`),
+    });
+    if (!marker)
+        throw new SpecificAIError('ERROR', '纯程序 WebM 指纹进程异常：无结果标记');
+    let result;
+    try {
+        result = JSON.parse(marker);
+    }
+    catch {
+        throw new SpecificAIError('ERROR', '纯程序 WebM 指纹返回格式异常');
+    }
+    if (!result.ok)
+        throw new SpecificAIError('ERROR', `纯程序 WebM 指纹生成失败：${String(result.errorMessage || result.error || '').slice(0, 240)}`);
+    jd.absorbScript(pageUrl, result);
+    const payload = result.payload;
+    if (!payload || typeof payload !== 'object' || typeof payload.body !== 'object' || !payload.body)
+        throw new SpecificAIError('ERROR', '纯程序 WebM 指纹缺少上报正文');
+    return payload;
+}

package/dist/platforms/kuaishou/index.js

@@ -0,0 +1,98 @@
+import { filter, PRODUCT } from '../../core/options.js';
+import { definePlatform, handlers } from '../../core/registry.js';
+/** 懒加载 web 端的 handler：只有执行到这条命令时才加载实现。 */
+const { h, impl } = handlers(() => import('./web/commands.js'));
+export default definePlatform({
+    id: 'kuaishou',
+    name: '快手',
+    aliases: ['ks'],
+    item: '作品',
+    endpoints: {
+        web: {
+            login: { methods: ['qrcode', 'sms', 'cookie'], default: 'qrcode' },
+            // 直播页的退出登录：passport logout + userLogout
+            logout: h('serverLogout'),
+            commands: {
+                'auth login': impl('full', 'authLogin'),
+                'auth status': impl('full', 'authStatus'),
+                'user get': impl('full', 'userGet'),
+                'user search': impl('full', 'userSearch'),
+                'user items': impl('full', 'userItems'),
+                'user likes': impl('full', 'userLikes', { note: '只支持 me' }),
+                'user collects': impl('full', 'userCollects'),
+                'user followers': impl('full', 'userFollowers', { note: '只支持 me' }),
+                'user following': impl('full', 'userFollowing', { note: '只支持 me' }),
+                'user follow': 'none',
+                'user unfollow': 'none',
+                'item get': impl('full', 'itemGet'),
+                'item search': impl('partial', 'itemSearch', { note: '没有筛选（--sort / --type / --time），只按关键词搜' }),
+                'item related': impl('full', 'itemRelated'),
+                'item list': impl('full', 'itemList'),
+                'item media': impl('full', 'itemMedia'),
+                'item download': impl('full', 'itemDownload'),
+                'item like': 'none',
+                'item unlike': 'none',
+                'item collect': 'none',
+                'item uncollect': 'none',
+                'item publish': impl('partial', 'itemPublish', {
+                    note: '不支持 --cover、--mention、--poi、--category、--price；--image 1～31 张（每张不超过 15MB），与 --video 二选一',
+                    supports: ['title', 'text', 'image', 'video', 'tag', 'topic', 'visibility', 'schedule'],
+                }),
+                'item delete': 'none',
+                'product get': 'none',
+                'comment list': impl('partial', 'commentList', { note: '--product 规划中', options: { product: PRODUCT } }),
+                'comment replies': impl('partial', 'commentReplies', { note: '上游的二级评论接口只核对过一组抓包参数，其余评论的结果没有经过上游核对' }),
+                'comment add': 'none',
+                'comment delete': 'none',
+                'comment like': 'none',
+                'comment unlike': 'none',
+                // recommend 是 new-reco 推荐流（上游 get_feed_hot）；快手没有单独的热门流
+                'feed list': impl('partial', 'feedList', {
+                    note: 'following 只有首屏',
+                    options: { kind: filter.kind('recommend', 'following') },
+                }),
+                'live get': impl('full', 'liveGet'),
+                'live list': impl('full', 'liveList'),
+                'live search': 'none',
+                'live categories': impl('full', 'liveCategories'),
+                'live listen': impl('full', 'liveListen'),
+                'live history': 'none',
+                // 上游没有发直播弹幕、直播间点赞的方法（README：“直播间主动发送弹幕尚未接入”）
+                'live send': 'none',
+                'live like': 'none',
+                'live rank': 'none',
+                'live gifts': impl('full', 'liveGifts'),
+                'live products': 'none',
+                'live media': 'none',
+                'live replays': impl('full', 'liveReplays'),
+                'live start': 'none',
+                'live stop': 'none',
+                'keyword suggest': 'none',
+                'keyword hot': 'none',
+                'notice list': 'none',
+                'notice count': impl('full', 'noticeCount'),
+                'msg list': 'none',
+                'msg history': 'none',
+                'msg send': 'none',
+                'msg listen': 'none',
+                'msg read': 'none',
+                'msg revoke': 'none',
+                'msg delete': 'none',
+                'media upload': impl('full', 'mediaUpload'),
+                'folder list': 'none',
+                'folder items': 'none',
+                'folder create': 'none',
+                'folder update': 'none',
+                'folder delete': 'none',
+                'series list': 'none',
+                'series items': 'none',
+                // 上游 get_history_list 已下线（服务端 404，上游直接拒绝发包）
+                'history list': 'none',
+                'topic search': 'none',
+                'poi search': 'none',
+            },
+        },
+        app: 'planned',
+        pc: 'planned',
+    },
+});