@remixmate/cli 0.9.72 → 0.9.74

package/dist/handlers/gen-post.js

@@ -1,192 +1,15 @@
-/**
- * gen-post handler — POST ab-api /publish/caption/generate.
- *
- * Generates the text of a social post — per-platform title / description / tags —
- * for a finished video (the video itself is not touched or uploaded): the same endpoint and prompts the web studio's 「发布文案」
- * panel uses. Prompt wording, compliance rules and the platform roster all live
- * in ab-api (publish_prompts.go + the user's saved prompt overrides); this
- * handler only collects context and forwards it.
- *
- * Behavior:
- *   - --platform is repeatable and/or comma-separated; `all` (the default) expands
- *     to the three platforms the web panel shows. Each platform is one backend
- *     call (one LLM run, billed separately), issued concurrently.
- *   - --render-id links the result to a studio take so it shows up in the web
- *     panel; without it the backend generates without persisting anything.
- *   - --instruction is the user's "how should it read" request. With a draft
- *     (--title / --description / --tag) it rewrites that caption — single platform
- *     only, since the draft belongs to one platform. Without a draft it is an extra
- *     requirement applied to a fresh generation on every requested platform.
- */
-import { EXIT, mmPost, resolveHttpContext, SkillError } from '../http.js';
-import { emitProgress } from '../progress.js';
-import { isTrue, toNumber, toStringArray } from './shared.js';
-/** Platforms the web panel enables by default; `all` expands to these. */
-const DEFAULT_PLATFORMS = ['douyin', 'xiaohongshu', 'tiktok'];
-/** Friendly aliases → backend platform codes. Unknown values pass through for ab-api to validate. */
-const PLATFORM_ALIASES = {
-    dy: 'douyin',
-    抖音: 'douyin',
-    xhs: 'xiaohongshu',
-    rednote: 'xiaohongshu',
-    小红书: 'xiaohongshu',
-    tt: 'tiktok',
-};
-/** Display names for human-readable output only. */
-const PLATFORM_LABELS = {
-    douyin: '抖音 Douyin',
-    xiaohongshu: '小红书 Xiaohongshu',
-    tiktok: 'TikTok',
-};
-/** Backend waits on a non-streaming LLM call (80s server-side cap); leave headroom. */
-const GENERATE_TIMEOUT_MS = 90_000;
-function resolvePlatforms(value) {
-    const raw = toStringArray(value)
-        .flatMap((v) => v.split(/[,，\s]+/))
-        .map((v) => v.trim())
-        .filter(Boolean);
-    if (raw.length === 0)
-        return [...DEFAULT_PLATFORMS];
-    const out = [];
-    for (const item of raw) {
-        const key = item.toLowerCase();
-        // 「全部」: the web param card writes option labels into the message, and the agent
-        // may forward the label verbatim.
-        const codes = key === 'all' || item === '全部' ? [...DEFAULT_PLATFORMS] : [PLATFORM_ALIASES[item] ?? PLATFORM_ALIASES[key] ?? key];
-        for (const code of codes)
-            if (!out.includes(code))
-                out.push(code);
-    }
-    return out;
-}
-function optString(value) {
-    if (typeof value !== 'string')
-        return undefined;
-    const trimmed = value.trim();
-    return trimmed || undefined;
-}
-/** Tags may arrive as `--tag a --tag b`, `--tag "a,b"`, or with a leading `#`. */
-function normalizeTags(value) {
-    return toStringArray(value)
-        .flatMap((v) => v.split(/[,，]+/))
-        .map((v) => v.trim().replace(/^#+/, ''))
-        .filter(Boolean);
-}
-async function generateOne(ctx, platform, base) {
-    emitProgress({ phase: 'gen-post:request', platform });
-    try {
-        const caption = await mmPost(ctx, '/publish/caption/generate', { ...base, platform }, { timeoutMs: GENERATE_TIMEOUT_MS });
-        if (!caption || (!caption.title && !caption.description)) {
-            return { platform, error: 'backend returned an empty caption' };
-        }
-        emitProgress({ phase: 'gen-post:done', platform });
-        return { platform, caption: { ...caption, platform, tags: caption.tags ?? [] } };
-    }
-    catch (err) {
-        // Auth failures abort the whole run: every other platform would fail the same way.
-        if (err instanceof SkillError && err.exitCode === EXIT.NOT_AUTHENTICATED)
-            throw err;
-        return { platform, error: err instanceof Error ? err.message : String(err) };
-    }
-}
-function printHuman(results) {
-    for (const r of results) {
-        const label = PLATFORM_LABELS[r.platform] ?? r.platform;
-        process.stdout.write(`\n━━━━━━━━ ${label} ━━━━━━━━\n`);
-        if (!r.caption) {
-            process.stdout.write(`❌ ${r.error ?? 'generation failed'}\n`);
-            continue;
-        }
-        const { title, description, tags } = r.caption;
-        process.stdout.write(`标题 / Title:\n${title}\n\n`);
-        process.stdout.write(`描述 / Description:\n${description}\n\n`);
-        process.stdout.write(`标签 / Tags:\n${tags.map((t) => `#${t}`).join(' ')}\n`);
-    }
-    const ok = results.filter((r) => r.caption).length;
-    process.stdout.write(`\n🎉 Generated publish copy for ${ok}/${results.length} platform(s)\n`);
-}
-export async function genPost(input, ctxIn) {
-    const topic = optString(input.topic);
-    const hook = optString(input.hook);
-    const instruction = optString(input.instruction);
-    const renderId = optString(input.render_id);
-    const platforms = resolvePlatforms(input.platform);
-    const current = {
-        title: optString(input.title) ?? '',
-        description: optString(input.description) ?? '',
-        tags: normalizeTags(input.tag),
-    };
-    const hasDraft = !!(current.title || current.description || current.tags.length);
-    if (instruction && instruction.length > 500) {
-        throw new SkillError('❌ --instruction must be at most 500 characters');
-    }
-    // With a draft, --instruction rewrites that one platform's caption. Without one it is
-    // an extra requirement for a fresh generation, which needs the video's context.
-    const rewriting = !!instruction && hasDraft;
-    if (hasDraft && !instruction) {
-        throw new SkillError('❌ --title / --description / --tag are the draft to rewrite; pass --instruction with them');
-    }
-    if (rewriting && platforms.length !== 1) {
-        throw new SkillError('❌ rewriting a caption (--instruction with --title / --description / --tag) needs exactly one --platform');
-    }
-    if (!rewriting && !topic && !hook) {
-        throw new SkillError('❌ Please provide --topic (what the video is about); --hook is optional extra context');
-    }
-    const durationSec = toNumber(input.duration, 'duration');
-    if (durationSec !== undefined && durationSec < 0) {
-        throw new SkillError('❌ --duration must be a non-negative number of seconds');
-    }
-    const base = {};
-    if (renderId)
-        base.renderId = renderId;
-    const videoProjectId = optString(input.video_project_id);
-    if (videoProjectId)
-        base.videoProjectId = videoProjectId;
-    if (topic)
-        base.topic = topic;
-    if (hook)
-        base.hook = hook;
-    const sourceUrl = optString(input.source_url);
-    if (sourceUrl)
-        base.sourceUrl = sourceUrl;
-    if (durationSec !== undefined)
-        base.durationSec = durationSec;
-    const aspectRatio = optString(input.aspect_ratio);
-    if (aspectRatio)
-        base.aspectRatio = aspectRatio;
-    const language = optString(input.language);
-    if (language)
-        base.language = language;
-    const promptOverride = optString(input.prompt);
-    if (promptOverride)
-        base.promptOverride = promptOverride;
-    if (instruction)
-        base.instruction = instruction;
-    if (rewriting)
-        base.current = current;
-    const ctx = await resolveHttpContext(ctxIn.skillName, {
-        apiBaseUrl: input.api_base_url,
-        privateToken: input.priv_token,
-        preflightToken: ctxIn.auth?.token,
-    });
-    const results = await Promise.all(platforms.map((p) => generateOne(ctx, p, base)));
-    const succeeded = results.filter((r) => r.caption);
-    if (isTrue(input.json_output)) {
-        process.stdout.write(JSON.stringify({
-            captions: succeeded.map((r) => ({
-                platform: r.platform,
-                title: r.caption.title,
-                description: r.caption.description,
-                tags: r.caption.tags,
-            })),
-            errors: results.filter((r) => !r.caption).map((r) => ({ platform: r.platform, error: r.error })),
-            persisted: !!renderId,
-        }) + '\n');
-    }
-    else {
-        printHuman(results);
-    }
-    if (succeeded.length === 0) {
-        throw new SkillError(`❌ publish copy generation failed for every platform: ${results.map((r) => `${r.platform}: ${r.error}`).join('; ')}`);
-    }
-}
+var P=Object.defineProperty;var d=(t,e)=>P(t,"name",{value:e,configurable:!0});import{EXIT as $,mmPost as O,resolveHttpContext as L,SkillError as f}from"../http.js";import{emitProgress as v}from"../progress.js";import{isTrue as M,toNumber as I,toStringArray as x}from"./shared.js";const E=["douyin","xiaohongshu","tiktok"],S={dy:"douyin",抖音:"douyin",xhs:"xiaohongshu",rednote:"xiaohongshu",小红书:"xiaohongshu",tt:"tiktok"},j={douyin:"抖音 Douyin",xiaohongshu:"小红书 Xiaohongshu",tiktok:"TikTok"},N=9e4;function R(t){const e=x(t).flatMap(o=>o.split(/[,，\s]+/)).map(o=>o.trim()).filter(Boolean);if(e.length===0)return[...E];const r=[];for(const o of e){const i=o.toLowerCase(),c=i==="all"||o==="全部"?[...E]:[S[o]??S[i]??i];for(const p of c)r.includes(p)||r.push(p)}return r}d(R,"resolvePlatforms");function a(t){return typeof t!="string"?void 0:t.trim()||void 0}d(a,"optString");function U(t){return x(t).flatMap(e=>e.split(/[,，]+/)).map(e=>e.trim().replace(/^#+/,"")).filter(Boolean)}d(U,"normalizeTags");async function D(t,e,r){v({phase:"gen-post:request",platform:e});try{const o=await O(t,"/publish/caption/generate",{...r,platform:e},{timeoutMs:N});return!o||!o.title&&!o.description?{platform:e,error:"backend returned an empty caption"}:(v({phase:"gen-post:done",platform:e}),{platform:e,caption:{...o,platform:e,tags:o.tags??[]}})}catch(o){if(o instanceof f&&o.exitCode===$.NOT_AUTHENTICATED)throw o;return{platform:e,error:o instanceof Error?o.message:String(o)}}}d(D,"generateOne");function B(t){for(const r of t){const o=j[r.platform]??r.platform;if(process.stdout.write(`
+━━━━━━━━ ${o} ━━━━━━━━
+`),!r.caption){process.stdout.write(`❌ ${r.error??"generation failed"}
+`);continue}const{title:i,description:c,tags:p}=r.caption;process.stdout.write(`标题 / Title:
+${i}
+
+`),process.stdout.write(`描述 / Description:
+${c}
+
+`),process.stdout.write(`标签 / Tags:
+${p.map(l=>`#${l}`).join(" ")}
+`)}const e=t.filter(r=>r.caption).length;process.stdout.write(`
+🎉 Generated publish copy for ${e}/${t.length} platform(s)
+`)}d(B,"printHuman");async function X(t,e){const r=a(t.topic),o=a(t.hook),i=a(t.instruction),c=a(t.render_id),p=R(t.platform),l={title:a(t.title)??"",description:a(t.description)??"",tags:U(t.tag)},m=!!(l.title||l.description||l.tags.length);if(i&&i.length>500)throw new f("❌ --instruction must be at most 500 characters");const h=!!i&&m;if(m&&!i)throw new f("❌ --title / --description / --tag are the draft to rewrite; pass --instruction with them");if(h&&p.length!==1)throw new f("❌ rewriting a caption (--instruction with --title / --description / --tag) needs exactly one --platform");if(!h&&!r&&!o)throw new f("❌ Please provide --topic (what the video is about); --hook is optional extra context");const u=I(t.duration,"duration");if(u!==void 0&&u<0)throw new f("❌ --duration must be a non-negative number of seconds");const s={};c&&(s.renderId=c);const w=a(t.video_project_id);w&&(s.videoProjectId=w),r&&(s.topic=r),o&&(s.hook=o);const k=a(t.source_url);k&&(s.sourceUrl=k),u!==void 0&&(s.durationSec=u);const T=a(t.aspect_ratio);T&&(s.aspectRatio=T);const y=a(t.language);y&&(s.language=y);const _=a(t.prompt);_&&(s.promptOverride=_),i&&(s.instruction=i),h&&(s.current=l);const A=await L(e.skillName,{apiBaseUrl:t.api_base_url,privateToken:t.priv_token,preflightToken:e.auth?.token}),g=await Promise.all(p.map(n=>D(A,n,s))),b=g.filter(n=>n.caption);if(M(t.json_output)?process.stdout.write(JSON.stringify({captions:b.map(n=>({platform:n.platform,title:n.caption.title,description:n.caption.description,tags:n.caption.tags})),errors:g.filter(n=>!n.caption).map(n=>({platform:n.platform,error:n.error})),persisted:!!c})+`
+`):B(g),b.length===0)throw new f(`❌ publish copy generation failed for every platform: ${g.map(n=>`${n.platform}: ${n.error}`).join("; ")}`)}d(X,"genPost");export{X as genPost};

package/dist/handlers/gen-video.js

@@ -1,185 +1,7 @@
-/**
- * gen-video handler — POST ab-api /model/genVideo, then poll /model/getVideoStatus.
- *
- * Thin forwarder: the model roster, friendly aliases, defaults and per-model
- * constraints (aspect ratios / resolutions / durations / reference-image caps)
- * come from ab-api's capabilities catalog (see capabilities.ts), not from this
- * package. Pre-flight validation here is purely data-driven from that catalog —
- * the CLI carries no model matrix — and ab-api re-validates authoritatively.
- *
- * --model accepts a friendly alias or a full id; ab-api also resolves aliases
- * server-side, so an unknown value still passes through to the backend.
- *
- * Local image paths for first/last/reference frames are read and base64-encoded
- * into a data: URI; https / data: URIs pass through unchanged. Reference videos
- * and audio are the exception — they must already be public URLs, see
- * `requirePublicUrls`.
- */
-import { mmPost, pollUntil, resolveHttpContext, SkillError } from '../http.js';
-import { getCapabilities, matchModel, defaultModel } from '../capabilities.js';
-import { emitProgress } from '../progress.js';
-import { isTrue, resolveImageInput, toNumber, toStringArray } from './shared.js';
-/** Generic, catalog-driven validation. No model names live here. */
-function validateAgainstConstraints(descriptor, duration, ratio, resolution, refs) {
-    const c = descriptor.constraints;
-    const label = descriptor.label || descriptor.id;
-    if (c.durations && c.durations.length > 0) {
-        if (!c.durations.includes(duration)) {
-            throw new SkillError(`❌ ${label} duration must be one of: ${c.durations.join(', ')} s; got: ${duration}`);
-        }
-    }
-    else if (c.durationMin !== undefined && c.durationMax !== undefined) {
-        if (duration < c.durationMin || duration > c.durationMax) {
-            throw new SkillError(`❌ ${label} duration must be ${c.durationMin}–${c.durationMax} s, got: ${duration}`);
-        }
-    }
-    if (c.aspectRatios && c.aspectRatios.length > 0 && !c.aspectRatios.includes(ratio)) {
-        throw new SkillError(`❌ ${label} aspect ratio must be one of: ${c.aspectRatios.join(', ')}; got: ${ratio}`);
-    }
-    if (c.resolutions && c.resolutions.length > 0 && !c.resolutions.includes(resolution)) {
-        throw new SkillError(`❌ ${label} resolution must be one of: ${c.resolutions.join(', ')}; got: ${resolution}`);
-    }
-    if (c.refImageMax !== undefined && refs.images > c.refImageMax) {
-        if (c.refImageMax === 0) {
-            throw new SkillError(`❌ ${label} does not support a reference-image array; use --first-frame / --last-frame instead`);
-        }
-        throw new SkillError(`❌ ${label} accepts at most ${c.refImageMax} reference images; got: ${refs.images}`);
-    }
-    if (c.refVideoMax !== undefined && refs.videos > c.refVideoMax) {
-        throw new SkillError(`❌ ${label} accepts at most ${c.refVideoMax} reference videos; got: ${refs.videos}`);
-    }
-    if (c.refAudioMax !== undefined && refs.audios > c.refAudioMax) {
-        throw new SkillError(`❌ ${label} accepts at most ${c.refAudioMax} reference audio clips; got: ${refs.audios}`);
-    }
-}
-/**
- * Reference videos / audio must be public https URLs.
- *
- * Unlike images, they are NOT inlined as data URIs: the caps are 50 MB per video
- * and 15 MB per clip, and base64 inflates that by a third before it ever reaches
- * the gateway. Failing here with the offending path beats a truncated upload or
- * an opaque 413 three hops downstream.
- */
-function requirePublicUrls(values, flag) {
-    return values.map((value) => {
-        const url = value.trim();
-        if (!/^https?:\/\//i.test(url)) {
-            throw new SkillError(`❌ ${flag} needs a public http(s) URL (local files are not uploaded), received: ${value}`);
-        }
-        return url;
-    });
-}
-async function pollVideoStatus(ctx, taskId) {
-    return pollUntil(async () => {
-        const data = await mmPost(ctx, '/model/getVideoStatus', { taskId }, { timeoutMs: 30_000 });
-        const status = data.status ?? 'unknown';
-        emitProgress({ phase: 'gen-video:poll', taskId, status });
-        if (status === 'succeeded' || status === 'completed') {
-            return data.videoUrl ?? '';
-        }
-        if (status === 'failed') {
-            const err = data.errorMsg ?? data.error ?? data.message ?? 'unknown error';
-            throw new SkillError(`❌ video generation failed: ${err}`);
-        }
-        return null;
-    }, { intervalMs: 10_000, timeoutMs: 300_000 });
-}
-export async function genVideo(input, ctxIn) {
-    const apiBaseUrl = input.api_base_url;
-    const caps = await getCapabilities(apiBaseUrl);
-    // Resolve the model against the catalog: explicit alias/id, env override, or default.
-    const requested = input.model ?? process.env.MM_VIDEO_MODEL;
-    const descriptor = requested ? matchModel(caps.video, requested) : defaultModel(caps.video);
-    const model = descriptor?.id ?? requested ?? defaultModel(caps.video)?.id ?? '';
-    const firstFrame = input.first_frame ? resolveImageInput(String(input.first_frame)) : '';
-    const lastFrame = input.last_frame ? resolveImageInput(String(input.last_frame)) : '';
-    const references = toStringArray(input.reference).map(resolveImageInput).filter(Boolean);
-    const referenceVideos = requirePublicUrls(toStringArray(input.reference_video).filter(Boolean), '--reference-video');
-    const referenceAudios = requirePublicUrls(toStringArray(input.reference_audio).filter(Boolean), '--reference-audio');
-    const prompt = (input.prompt ?? '').trim();
-    const hasVisual = Boolean(firstFrame || lastFrame || references.length > 0 || referenceVideos.length > 0);
-    if (!prompt && !hasVisual) {
-        throw new SkillError('❌ Provide at least one of: a non-empty --prompt, or --first-frame / --last-frame / --reference / --reference-video');
-    }
-    /**
-     * Two shapes the backend rejects outright; catching them here saves a round
-     * trip and names the flag that caused it.
-     *
-     * - first/last frame pins exact frames, omni-modal reference lets the model
-     *   re-cut — the API has no meaning for both at once;
-     * - audio-only (or text + audio) input is not a supported modality mix.
-     */
-    if ((firstFrame || lastFrame) && (referenceVideos.length > 0 || referenceAudios.length > 0)) {
-        throw new SkillError('❌ --reference-video / --reference-audio cannot be combined with --first-frame / --last-frame; use --reference for images instead');
-    }
-    if (referenceAudios.length > 0 && references.length === 0 && referenceVideos.length === 0) {
-        throw new SkillError('❌ --reference-audio needs at least one --reference image or --reference-video alongside it');
-    }
-    const ratio = input.ratio ?? '16:9';
-    const resolution = input.resolution ?? '720p';
-    const duration = toNumber(input.duration, 'duration') ?? descriptor?.constraints.durationDefault ?? 5;
-    const seed = toNumber(input.seed, 'seed');
-    // Pre-flight validation only when we recognize the model (have its constraints).
-    if (descriptor) {
-        validateAgainstConstraints(descriptor, duration, ratio, resolution, {
-            images: references.length,
-            videos: referenceVideos.length,
-            audios: referenceAudios.length,
-        });
-    }
-    const ctx = await resolveHttpContext(ctxIn.skillName, {
-        apiBaseUrl,
-        privateToken: input.priv_token,
-        preflightToken: ctxIn.auth?.token,
-    });
-    const payload = {
-        model,
-        duration,
-        ratio,
-        resolution,
-        generateAudio: isTrue(input.generate_audio),
-        cameraFixed: isTrue(input.camera_fixed),
-    };
-    if (prompt)
-        payload.prompt = prompt;
-    if (firstFrame)
-        payload.firstFrameImage = firstFrame;
-    if (lastFrame)
-        payload.lastFrameImage = lastFrame;
-    if (references.length > 0)
-        payload.referenceImages = references;
-    if (referenceVideos.length > 0)
-        payload.referenceVideos = referenceVideos;
-    if (referenceAudios.length > 0)
-        payload.referenceAudios = referenceAudios;
-    // Web search only fires on text-only input; sending it alongside attachments
-    // would be a knob the model silently ignores.
-    if (isTrue(input.web_search) && !hasVisual)
-        payload.webSearch = true;
-    if (isTrue(input.return_last_frame))
-        payload.returnLastFrame = true;
-    if (input.negative_prompt && String(input.negative_prompt).trim()) {
-        payload.negativePrompt = String(input.negative_prompt).trim();
-    }
-    if (seed !== undefined)
-        payload.seed = seed;
-    if (input.person_generation)
-        payload.personGeneration = String(input.person_generation);
-    emitProgress({ phase: 'gen-video:request', model, ratio, resolution, duration });
-    const data = await mmPost(ctx, '/model/genVideo', payload, { timeoutMs: 120_000 });
-    const taskId = data.taskId;
-    if (!taskId) {
-        throw new SkillError(`❌ could not parse the task id; API returned: ${JSON.stringify(data)}`);
-    }
-    emitProgress({ phase: 'gen-video:async', taskId });
-    const videoUrl = await pollVideoStatus(ctx, taskId);
-    if (!videoUrl) {
-        throw new SkillError(`❌ video generated but no URL was returned, task id: ${taskId}`);
-    }
-    if (isTrue(input.json_output)) {
-        process.stdout.write(JSON.stringify({ url: videoUrl }) + '\n');
-        return;
-    }
-    process.stdout.write(`\n🔗 Video URL (renders directly, no download needed):\n${videoUrl}\n`);
-    process.stdout.write(`\n🎉 Video generation done!\n`);
-}
+var P=Object.defineProperty;var m=(e,o)=>P(e,"name",{value:o,configurable:!0});import{mmPost as A,pollUntil as B,resolveHttpContext as L,SkillError as r}from"../http.js";import{getCapabilities as j,matchModel as N,defaultModel as U}from"../capabilities.js";import{emitProgress as x}from"../progress.js";import{isTrue as u,resolveImageInput as b,toNumber as R,toStringArray as V}from"./shared.js";function O(e,o,n,i,a){const t=e.constraints,d=e.label||e.id;if(t.durations&&t.durations.length>0){if(!t.durations.includes(o))throw new r(`❌ ${d} duration must be one of: ${t.durations.join(", ")} s; got: ${o}`)}else if(t.durationMin!==void 0&&t.durationMax!==void 0&&(o<t.durationMin||o>t.durationMax))throw new r(`❌ ${d} duration must be ${t.durationMin}–${t.durationMax} s, got: ${o}`);if(t.aspectRatios&&t.aspectRatios.length>0&&!t.aspectRatios.includes(n))throw new r(`❌ ${d} aspect ratio must be one of: ${t.aspectRatios.join(", ")}; got: ${n}`);if(t.resolutions&&t.resolutions.length>0&&!t.resolutions.includes(i))throw new r(`❌ ${d} resolution must be one of: ${t.resolutions.join(", ")}; got: ${i}`);if(t.refImageMax!==void 0&&a.images>t.refImageMax)throw t.refImageMax===0?new r(`❌ ${d} does not support a reference-image array; use --first-frame / --last-frame instead`):new r(`❌ ${d} accepts at most ${t.refImageMax} reference images; got: ${a.images}`);if(t.refVideoMax!==void 0&&a.videos>t.refVideoMax)throw new r(`❌ ${d} accepts at most ${t.refVideoMax} reference videos; got: ${a.videos}`);if(t.refAudioMax!==void 0&&a.audios>t.refAudioMax)throw new r(`❌ ${d} accepts at most ${t.refAudioMax} reference audio clips; got: ${a.audios}`)}m(O,"validateAgainstConstraints");function F(e,o){return e.map(n=>{const i=n.trim();if(!/^https?:\/\//i.test(i))throw new r(`❌ ${o} needs a public http(s) URL (local files are not uploaded), received: ${n}`);return i})}m(F,"requirePublicUrls");async function q(e,o){return B(async()=>{const n=await A(e,"/model/getVideoStatus",{taskId:o},{timeoutMs:3e4}),i=n.status??"unknown";if(x({phase:"gen-video:poll",taskId:o,status:i}),i==="succeeded"||i==="completed")return n.videoUrl??"";if(i==="failed"){const a=n.errorMsg??n.error??n.message??"unknown error";throw new r(`❌ video generation failed: ${a}`)}return null},{intervalMs:1e4,timeoutMs:3e5})}m(q,"pollVideoStatus");async function G(e,o){const n=e.api_base_url,i=await j(n),a=e.model??process.env.MM_VIDEO_MODEL,t=a?N(i.video,a):U(i.video),d=t?.id??a??U(i.video)?.id??"",g=e.first_frame?b(String(e.first_frame)):"",w=e.last_frame?b(String(e.last_frame)):"",c=V(e.reference).map(b).filter(Boolean),f=F(V(e.reference_video).filter(Boolean),"--reference-video"),l=F(V(e.reference_audio).filter(Boolean),"--reference-audio"),p=(e.prompt??"").trim(),S=!!(g||w||c.length>0||f.length>0);if(!p&&!S)throw new r("❌ Provide at least one of: a non-empty --prompt, or --first-frame / --last-frame / --reference / --reference-video");if((g||w)&&(f.length>0||l.length>0))throw new r("❌ --reference-video / --reference-audio cannot be combined with --first-frame / --last-frame; use --reference for images instead");if(l.length>0&&c.length===0&&f.length===0)throw new r("❌ --reference-audio needs at least one --reference image or --reference-video alongside it");const v=e.ratio??"16:9",$=e.resolution??"720p",M=R(e.duration,"duration")??t?.constraints.durationDefault??5,k=R(e.seed,"seed");t&&O(t,M,v,$,{images:c.length,videos:f.length,audios:l.length});const y=await L(o.skillName,{apiBaseUrl:n,privateToken:e.priv_token,preflightToken:o.auth?.token}),s={model:d,duration:M,ratio:v,resolution:$,generateAudio:u(e.generate_audio),cameraFixed:u(e.camera_fixed)};p&&(s.prompt=p),g&&(s.firstFrameImage=g),w&&(s.lastFrameImage=w),c.length>0&&(s.referenceImages=c),f.length>0&&(s.referenceVideos=f),l.length>0&&(s.referenceAudios=l),u(e.web_search)&&!S&&(s.webSearch=!0),u(e.return_last_frame)&&(s.returnLastFrame=!0),e.negative_prompt&&String(e.negative_prompt).trim()&&(s.negativePrompt=String(e.negative_prompt).trim()),k!==void 0&&(s.seed=k),e.person_generation&&(s.personGeneration=String(e.person_generation)),x({phase:"gen-video:request",model:d,ratio:v,resolution:$,duration:M});const I=await A(y,"/model/genVideo",s,{timeoutMs:12e4}),h=I.taskId;if(!h)throw new r(`❌ could not parse the task id; API returned: ${JSON.stringify(I)}`);x({phase:"gen-video:async",taskId:h});const _=await q(y,h);if(!_)throw new r(`❌ video generated but no URL was returned, task id: ${h}`);if(u(e.json_output)){process.stdout.write(JSON.stringify({url:_})+`
+`);return}process.stdout.write(`
+🔗 Video URL (renders directly, no download needed):
+${_}
+`),process.stdout.write(`
+🎉 Video generation done!
+`)}m(G,"genVideo");export{G as genVideo};

package/dist/handlers/gen-voice.js

@@ -1,133 +1,19 @@
-/**
- * gen-voice handler — Minimax TTS via ab-api unified /tool/tts (provider=minimax).
- *
- * Operations:
- *   --list-voices true                 → GET available voice IDs from /voice/page (live)
- *   --list-voices true --local true    → print the built-in language-tagged
- *                                        fallback catalog (offline, no API call).
- *                                        **CLI-only escape hatch for offline dev.**
- *   default                            → synthesize text into an MP3 URL
- *
- * `local` is deliberately absent from `skills/gen-voice/skill.json`'s
- * `parameters`, which is what ab-agent turns into the model-facing tool schema.
- * A model asked to check a voice id would reach for `local: true`, get a
- * two-entry hardcoded list back, and conclude that a perfectly valid voice does
- * not exist — which is exactly what happened on the hosted web agent
- * ("本地回退列表中没有该音色"). The remote catalog is the only authority on
- * which voices exist; the local list is a developer convenience for working
- * without a token, and it is only reachable by typing the flag yourself.
- *
- * The previous Python script also offered `--download` to write the MP3 to
- * disk. That's intentionally dropped here: the agent flow never used it (the
- * audio URL is persisted in cloud) and removing local file I/O keeps the
- * handler stateless.
- */
-import { mmPost, resolveHttpContext, SkillError } from '../http.js';
-import { getCapabilities } from '../capabilities.js';
-import { emitProgress } from '../progress.js';
-import { isTrue, toNumber } from './shared.js';
-/**
- * Fallback voice catalog printed when `--list-voices --local` is passed.
- * Mirrors the language-keyed defaults that gen_script.py uses, kept in sync
- * by hand because the two callers live on different runtimes.
- *
- * This is **not** a list of available voices — it is the two ids the resolver
- * falls back to when nothing else is specified, and one of them is still a
- * placeholder. Never treat it as evidence about what the service offers.
- *
- * Promote to a JSON config once this grows beyond a handful of entries.
- */
-const LOCAL_VOICE_CATALOG = [
-    { id: 'Chinese (Mandarin)_Male_Announcer', language: 'zh', name: 'Mandarin Male Announcer (legacy default)' },
-    // TODO: replace once gen-voice service confirms an English default voice id.
-    { id: 'english-male-friendly-01', language: 'en', name: 'Friendly Male (English, TBD)' },
-];
-async function listVoicesRemote(ctx) {
-    const data = await mmPost(ctx, '/voice/page', { current: 1, pageSize: 100 }, { timeoutMs: 30_000 });
-    const voices = data.list ?? [];
-    if (voices.length === 0) {
-        process.stdout.write('⚠️  No voices available\n');
-        return;
-    }
-    process.stdout.write(`\n${'Voice ID'.padEnd(30)} Name\n`);
-    process.stdout.write(`${'-'.repeat(30)} ${'-'.repeat(20)}\n`);
-    for (const v of voices) {
-        process.stdout.write(`${(v.uniqId ?? '').padEnd(30)} ${v.name ?? ''}\n`);
-    }
-    process.stdout.write(`\n${voices.length} voice(s) available\n`);
-    process.stdout.write(`💡 Use --voice-id <id> to select a voice for synthesis\n`);
-}
-function listVoicesLocal() {
-    process.stderr.write('⚠️  --local prints a hardcoded 2-entry fallback list, NOT the available voices. ' +
-        'Drop --local to query the live catalog.\n');
-    for (const v of LOCAL_VOICE_CATALOG) {
-        process.stdout.write(`${v.id}\t${v.language}\t${v.name}\n`);
-    }
-}
-export async function genVoice(input, ctxIn) {
-    // ── --list-voices: catalog mode (no synthesis) ──────────────────────────
-    if (isTrue(input.list_voices)) {
-        if (isTrue(input.local))
-            return listVoicesLocal();
-        const ctx = await resolveHttpContext(ctxIn.skillName, {
-            apiBaseUrl: input.api_base_url,
-            privateToken: input.priv_token,
-            preflightToken: ctxIn.auth?.token,
-        });
-        return listVoicesRemote(ctx);
-    }
-    // ── synthesize ──────────────────────────────────────────────────────────
-    const text = input.text;
-    if (!text || typeof text !== 'string' || !text.trim()) {
-        throw new SkillError('❌ Please provide --text to specify the text to synthesize');
-    }
-    const apiBaseUrl = input.api_base_url;
-    const voice = (await getCapabilities(apiBaseUrl)).voice;
-    const voiceId = input.voice_id ?? voice.defaultVoiceId;
-    const speed = toNumber(input.speed, 'speed') ?? 1.0;
-    if (speed < voice.speedMin || speed > voice.speedMax) {
-        throw new SkillError(`❌ speed must be in the range ${voice.speedMin}~${voice.speedMax}`);
-    }
-    // Language hint for mixed-script narration. Like `local`, this is deliberately
-    // absent from skill.json's `parameters`: the service already applies a sane
-    // default for every caller, so handing the model a 40-value enum only invites
-    // it to guess. Reachable from the DSL (global/scene `narration.languageBoost`,
-    // forwarded by render_video) and by typing the flag yourself. Omitted =
-    // the service default.
-    const languageBoost = input.language_boost?.trim() || undefined;
-    const ctx = await resolveHttpContext(ctxIn.skillName, {
-        apiBaseUrl,
-        privateToken: input.priv_token,
-        preflightToken: ctxIn.auth?.token,
-    });
-    emitProgress({ phase: 'gen-voice:request', voiceId, speed });
-    const data = await mmPost(ctx, '/tool/tts', {
-        provider: voice.provider,
-        text,
-        voiceSetting: { voiceId, speed },
-        outputFormat: 'url',
-        ...(languageBoost ? { languageBoost } : {}),
-    }, { timeoutMs: 120_000 });
-    const audioUrl = data.audio;
-    if (!audioUrl) {
-        throw new SkillError(`❌ API response did not contain an audio URL: ${JSON.stringify(data)}`);
-    }
-    if (isTrue(input.json_output)) {
-        process.stdout.write(JSON.stringify({
-            url: audioUrl,
-            audio_length_ms: data.extraInfo?.audioLength ?? null,
-            subtitles: (data.subtitles ?? []).map((s) => ({
-                text: s.text ?? '',
-                timeBegin: s.timeBegin ?? 0,
-                timeEnd: s.timeEnd ?? 0,
-            })),
-        }) + '\n');
-        return;
-    }
-    const lengthMs = data.extraInfo?.audioLength;
-    if (typeof lengthMs === 'number') {
-        process.stdout.write(`\n🕐 Audio length: ${(lengthMs / 1000).toFixed(1)} s\n`);
-    }
-    process.stdout.write(`\n🔗 Audio URL (plays directly, no download needed):\n${audioUrl}\n`);
-    process.stdout.write(`\n🎉 TTS synthesis done!\n`);
-}
+var w=Object.defineProperty;var l=(e,o)=>w(e,"name",{value:o,configurable:!0});import{mmPost as m,resolveHttpContext as v,SkillError as c}from"../http.js";import{getCapabilities as _}from"../capabilities.js";import{emitProgress as y}from"../progress.js";import{isTrue as u,toNumber as $}from"./shared.js";const b=[{id:"Chinese (Mandarin)_Male_Announcer",language:"zh",name:"Mandarin Male Announcer (legacy default)"},{id:"english-male-friendly-01",language:"en",name:"Friendly Male (English, TBD)"}];async function k(e){const t=(await m(e,"/voice/page",{current:1,pageSize:100},{timeoutMs:3e4})).list??[];if(t.length===0){process.stdout.write(`⚠️  No voices available
+`);return}process.stdout.write(`
+${"Voice ID".padEnd(30)} Name
+`),process.stdout.write(`${"-".repeat(30)} ${"-".repeat(20)}
+`);for(const s of t)process.stdout.write(`${(s.uniqId??"").padEnd(30)} ${s.name??""}
+`);process.stdout.write(`
+${t.length} voice(s) available
+`),process.stdout.write(`💡 Use --voice-id <id> to select a voice for synthesis
+`)}l(k,"listVoicesRemote");function x(){process.stderr.write(`⚠️  --local prints a hardcoded 2-entry fallback list, NOT the available voices. Drop --local to query the live catalog.
+`);for(const e of b)process.stdout.write(`${e.id}	${e.language}	${e.name}
+`)}l(x,"listVoicesLocal");async function E(e,o){if(u(e.list_voices)){if(u(e.local))return x();const a=await v(o.skillName,{apiBaseUrl:e.api_base_url,privateToken:e.priv_token,preflightToken:o.auth?.token});return k(a)}const t=e.text;if(!t||typeof t!="string"||!t.trim())throw new c("❌ Please provide --text to specify the text to synthesize");const s=e.api_base_url,i=(await _(s)).voice,p=e.voice_id??i.defaultVoiceId,r=$(e.speed,"speed")??1;if(r<i.speedMin||r>i.speedMax)throw new c(`❌ speed must be in the range ${i.speedMin}~${i.speedMax}`);const g=e.language_boost?.trim()||void 0,h=await v(o.skillName,{apiBaseUrl:s,privateToken:e.priv_token,preflightToken:o.auth?.token});y({phase:"gen-voice:request",voiceId:p,speed:r});const n=await m(h,"/tool/tts",{provider:i.provider,text:t,voiceSetting:{voiceId:p,speed:r},outputFormat:"url",...g?{languageBoost:g}:{}},{timeoutMs:12e4}),d=n.audio;if(!d)throw new c(`❌ API response did not contain an audio URL: ${JSON.stringify(n)}`);if(u(e.json_output)){process.stdout.write(JSON.stringify({url:d,audio_length_ms:n.extraInfo?.audioLength??null,subtitles:(n.subtitles??[]).map(a=>({text:a.text??"",timeBegin:a.timeBegin??0,timeEnd:a.timeEnd??0}))})+`
+`);return}const f=n.extraInfo?.audioLength;typeof f=="number"&&process.stdout.write(`
+🕐 Audio length: ${(f/1e3).toFixed(1)} s
+`),process.stdout.write(`
+🔗 Audio URL (plays directly, no download needed):
+${d}
+`),process.stdout.write(`
+🎉 TTS synthesis done!
+`)}l(E,"genVoice");export{E as genVoice};

package/dist/handlers/index.js

@@ -1,20 +1 @@
-/**
- * Handler registry — maps `entry.handler` strings to their implementation.
- *
- * Only `entry.type` in ('http' | 'builtin') uses this registry; python
- * skills go through spawn() in runner.ts and bypass handlers entirely.
- */
-import { genImage } from './gen-image.js';
-import { genVoice } from './gen-voice.js';
-import { genVideo } from './gen-video.js';
-import { genDigitalHuman } from './gen-digital-human.js';
-import { templateInit } from './template-init.js';
-import { genPost } from './gen-post.js';
-export const HANDLERS = {
-    'gen-image': genImage,
-    'gen-voice': genVoice,
-    'gen-video': genVideo,
-    'gen-digital-human': genDigitalHuman,
-    'template-init': templateInit,
-    'gen-post': genPost,
-};
+import{genImage as e}from"./gen-image.js";import{genVoice as o}from"./gen-voice.js";import{genVideo as i}from"./gen-video.js";import{genDigitalHuman as m}from"./gen-digital-human.js";import{templateInit as t}from"./template-init.js";import{genPost as n}from"./gen-post.js";const c={"gen-image":e,"gen-voice":o,"gen-video":i,"gen-digital-human":m,"template-init":t,"gen-post":n};export{c as HANDLERS};