@remixmate/cli 0.9.71 → 0.9.73

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};

package/dist/handlers/shared.js

@@ -1,66 +1 @@
-/**
- * Shared helpers for TS skill handlers.
- *
- * These were previously copy-pasted across gen-image / gen-video /
- * gen-digital-human / gen-voice. Centralizing them keeps argument coercion and
- * image-input resolution identical across every atomic generator.
- *
- * Lives under handlers/ (not src/ root) so the compiled `dist/handlers/shared.js`
- * travels with the rest of the handler bundle when ab-agent copies `dist/`.
- */
-import { readFileSync, existsSync, statSync } from 'node:fs';
-import path from 'node:path';
-import { SkillError } from '../http.js';
-const MIME_BY_EXT = {
-    '.png': 'image/png',
-    '.jpg': 'image/jpeg',
-    '.jpeg': 'image/jpeg',
-    '.webp': 'image/webp',
-    '.gif': 'image/gif',
-};
-/** Truthy check tolerant of CLI string values (`--flag` → true, `--flag true` → 'true'). */
-export function isTrue(value) {
-    return value === true || value === 'true';
-}
-/**
- * Coerce a flag value into a finite number, or `undefined` when absent/empty.
- * Throws SkillError on a non-numeric value so callers fail fast with a clear msg.
- */
-export function toNumber(value, name) {
-    if (value === undefined || value === null || value === '')
-        return undefined;
-    const n = Number(value);
-    if (!Number.isFinite(n))
-        throw new SkillError(`❌ ${name} must be a number, received: ${value}`);
-    return n;
-}
-/** Coerce a repeatable flag (string | string[] | undefined) into a string array. */
-export function toStringArray(value) {
-    if (value === undefined || value === null)
-        return [];
-    if (Array.isArray(value))
-        return value.map(String);
-    return [String(value)];
-}
-/**
- * Resolve an image reference into a value the backend accepts:
- *   - http(s) URLs and data: URIs pass through unchanged
- *   - local file paths are read and base64-encoded into a `data:<mime>;base64,…` URI
- * Throws SkillError when a local path does not exist or is not a file.
- */
-export function resolveImageInput(pathOrUrl) {
-    const trimmed = pathOrUrl.trim();
-    if (!trimmed)
-        return '';
-    const lower = trimmed.toLowerCase();
-    if (lower.startsWith('http://') || lower.startsWith('https://') || lower.startsWith('data:')) {
-        return trimmed;
-    }
-    if (!existsSync(trimmed) || !statSync(trimmed).isFile()) {
-        throw new SkillError(`❌ image path does not exist or is not a file: ${trimmed}`);
-    }
-    const ext = path.extname(trimmed).toLowerCase();
-    const mime = MIME_BY_EXT[ext] ?? 'image/jpeg';
-    const b64 = readFileSync(trimmed).toString('base64');
-    return `data:${mime};base64,${b64}`;
-}
+var p=Object.defineProperty;var i=(t,r)=>p(t,"name",{value:r,configurable:!0});import{readFileSync as a,existsSync as g,statSync as f}from"node:fs";import u from"node:path";import{SkillError as n}from"../http.js";const c={".png":"image/png",".jpg":"image/jpeg",".jpeg":"image/jpeg",".webp":"image/webp",".gif":"image/gif"};function x(t){return t===!0||t==="true"}i(x,"isTrue");function S(t,r){if(t==null||t==="")return;const e=Number(t);if(!Number.isFinite(e))throw new n(`❌ ${r} must be a number, received: ${t}`);return e}i(S,"toNumber");function y(t){return t==null?[]:Array.isArray(t)?t.map(String):[String(t)]}i(y,"toStringArray");function j(t){const r=t.trim();if(!r)return"";const e=r.toLowerCase();if(e.startsWith("http://")||e.startsWith("https://")||e.startsWith("data:"))return r;if(!g(r)||!f(r).isFile())throw new n(`❌ image path does not exist or is not a file: ${r}`);const o=u.extname(r).toLowerCase(),s=c[o]??"image/jpeg",m=a(r).toString("base64");return`data:${s};base64,${m}`}i(j,"resolveImageInput");export{x as isTrue,j as resolveImageInput,S as toNumber,y as toStringArray};

package/dist/handlers/template-init.js

@@ -1,21 +1 @@
-/**
- * template-authoring skill → `remixmate template init`.
- *
- * Agent hosts call skills as `remixmate template-authoring --dir X [--force]`;
- * this handler is a thin adapter so both entry points share one implementation.
- * Purely local (auth: none): fetches the starter via npm, no ab-api calls.
- */
-import { EXIT, SkillError } from '../errors.js';
-import { runInit } from '../template/init.js';
-export async function templateInit(input, _ctx) {
-    const dir = typeof input.dir === 'string' ? input.dir.trim() : '';
-    if (!dir)
-        throw new SkillError('❌ 缺少 --dir：模板工程的目标目录', EXIT.USAGE);
-    await runInit({
-        dir,
-        force: input.force === true || input.force === 'true',
-        json: input.json_output === true || input.json === true,
-        skipInstall: input.skip_install === true,
-    });
-}
-export default templateInit;
+var i=Object.defineProperty;var e=(r,t)=>i(r,"name",{value:t,configurable:!0});import{EXIT as s,SkillError as f}from"../errors.js";import{runInit as l}from"../template/init.js";async function n(r,t){const o=typeof r.dir=="string"?r.dir.trim():"";if(!o)throw new f("❌ 缺少 --dir：模板工程的目标目录",s.USAGE);await l({dir:o,force:r.force===!0||r.force==="true",json:r.json_output===!0||r.json===!0,skipInstall:r.skip_install===!0})}e(n,"templateInit");var d=n;export{d as default,n as templateInit};

package/dist/http.js

@@ -1,155 +1,3 @@
-/**
- * Shared HTTP helpers for ab-api endpoints.
- *
- * Centralizes the auth contract (X-Priv-Token + x-invoke-skill / x-invoke-agent /
- * x-conversation-id), business-code error handling (result.code != 0), and the
- * env var resolution order that every Python skill used to duplicate.
- *
- * Token resolution order (high → low), shared with the CLI auth commands via
- * src/auth/resolve.ts:
- *   1. opts.privateToken              — caller override (maps to --token/--priv-token)
- *   2. process.env.PRIV_TOKEN         — process env (CI / agent-embedded hosts)
- *   3. Credential_Store[apiBaseUrl]   — written by `remixmate login`
- *   (none) → throws SkillError        — non-interactive failure
- *
- * Base URL resolution:
- *   1. opts.apiBaseUrl                  — caller override
- *   2. process.env.MM_API_BASE_URL      — process env (point at local/staging here)
- *   3. process.env.MM_BACKEND_API_URL   — ab-agent's convention (back-compat)
- *   4. https://api.remixmate.ai/api    — production default (zero-config)
- */
-import { traceHeaders, cleanDiagnosticText } from './diagnostics.js';
-import { recordBilling } from './billing.js';
-import { resolvePrivToken, NOT_AUTHENTICATED_HINT } from './auth/resolve.js';
-import { attemptAutoLogin } from './auth/auto-login.js';
-import { EXIT, SkillError } from './errors.js';
-const DEFAULT_API_BASE_URL = 'https://api.remixmate.ai/api';
-// Re-exported so the many `import { SkillError } from '../http.js'` call sites
-// stay valid; the definition moved to errors.ts to break an import cycle with
-// the auth gateway (see errors.ts).
-export { SkillError, EXIT } from './errors.js';
-/**
- * Resolve the ab-api base URL the same way for authenticated and device-flow calls.
- *
- * `MM_BACKEND_API_URL` is honored because several Python skills already fall back
- * to it (registry_loader, render_job_client, _vod_polling). Since runner.ts now
- * injects the resolved base URL into those child processes, leaving it out here
- * would let a prod default silently override an ab-agent-injected backend.
- */
-export function resolveApiBaseUrl(flag) {
-    const fromEnv = (process.env.MM_API_BASE_URL ?? '').trim() || (process.env.MM_BACKEND_API_URL ?? '').trim();
-    return ((flag ?? '').trim() || fromEnv || DEFAULT_API_BASE_URL).replace(/\/+$/, '');
-}
-export async function resolveHttpContext(skillName, opts = {}) {
-    const apiBaseUrl = resolveApiBaseUrl(opts.apiBaseUrl);
-    // An explicit --token flag still wins over the preflight result.
-    let resolved = await resolvePrivToken({
-        flag: opts.privateToken ?? opts.preflightToken,
-        apiBaseUrl,
-    });
-    // Only when flag/env/store are all empty do we try automatic browser auth.
-    // attemptAutoLogin self-gates (cloud env / headless / disabled → returns null
-    // immediately), so the env-injected path never reaches it (zero regression).
-    if (!resolved) {
-        resolved = await attemptAutoLogin(apiBaseUrl);
-    }
-    if (!resolved) {
-        throw new SkillError(NOT_AUTHENTICATED_HINT, EXIT.NOT_AUTHENTICATED);
-    }
-    return {
-        apiBaseUrl,
-        privateToken: resolved.privToken,
-        skillName,
-        agentName: process.env.AGENT_NAME || undefined,
-        conversationId: process.env.CONVERSATION_ID || undefined,
-    };
-}
-function buildHeaders(ctx, extra) {
-    const headers = {
-        'Content-Type': 'application/json',
-        'X-Priv-Token': ctx.privateToken,
-        'x-invoke-skill': ctx.skillName,
-    };
-    if (ctx.agentName)
-        headers['x-invoke-agent'] = ctx.agentName;
-    if (ctx.conversationId)
-        headers['x-conversation-id'] = ctx.conversationId;
-    return { ...headers, ...extra, ...traceHeaders() };
-}
-/**
- * POST JSON to ab-api and return the parsed business payload.
- * Throws SkillError on HTTP errors or business code != 0.
- */
-export async function mmPost(ctx, pathOrUrl, body, opts = {}) {
-    const url = pathOrUrl.startsWith('http') ? pathOrUrl : `${ctx.apiBaseUrl}${pathOrUrl}`;
-    const controller = new AbortController();
-    const timeout = setTimeout(() => controller.abort(), opts.timeoutMs ?? 120_000);
-    let resp;
-    try {
-        resp = await fetch(url, {
-            method: 'POST',
-            headers: buildHeaders(ctx),
-            body: JSON.stringify(body),
-            signal: controller.signal,
-        });
-    }
-    catch (err) {
-        if (err instanceof Error && err.name === 'AbortError') {
-            throw new SkillError(`❌ request timed out: ${url}`, EXIT.BACKEND_UNREACHABLE);
-        }
-        throw new SkillError(`❌ network error: ${err.message}`, EXIT.BACKEND_UNREACHABLE);
-    }
-    finally {
-        clearTimeout(timeout);
-    }
-    const text = await resp.text();
-    if (!resp.ok) {
-        // 401/403 is not a generic API failure — it means the credential is missing,
-        // expired or revoked. Surfacing it as such (with exit code 4) is what lets a
-        // host offer to re-authorize instead of reporting an opaque backend error.
-        if (resp.status === 401 || resp.status === 403) {
-            throw new SkillError(`${NOT_AUTHENTICATED_HINT}\n   （后端返回 HTTP ${resp.status}）`, EXIT.NOT_AUTHENTICATED);
-        }
-        if (resp.status >= 500) {
-            throw new SkillError(`❌ API request failed (HTTP ${resp.status}): [response omitted]`, EXIT.BACKEND_UNREACHABLE);
-        }
-        throw new SkillError(`❌ API request failed (HTTP ${resp.status}): [response omitted]`);
-    }
-    let parsed;
-    try {
-        parsed = JSON.parse(text);
-    }
-    catch {
-        throw new SkillError(`❌ failed to parse response, body is not JSON: [response omitted]`);
-    }
-    // Before the code check: a charged response is always code=0 today, but a
-    // partial-failure envelope that still billed must not lose its billing line.
-    recordBilling(parsed.billing);
-    if (parsed.code !== 0) {
-        // ab-api reports auth failures in the envelope (HTTP 200 + code=401), so the
-        // business-code path needs the same 401 → "re-authorize" mapping as above.
-        if (parsed.code === 401 || parsed.code === 403) {
-            throw new SkillError(`${NOT_AUTHENTICATED_HINT}\n   （后端返回 code=${parsed.code}${parsed.msg ? `: ${parsed.msg}` : ''}）`, EXIT.NOT_AUTHENTICATED);
-        }
-        throw new SkillError(`❌ API returned a business error: ${cleanDiagnosticText(parsed.msg ?? 'unknown error')} (code=${parsed.code})`);
-    }
-    return parsed.data ?? undefined;
-}
-/**
- * Generic polling helper. Invokes `probe()` immediately, then every
- * `intervalMs`, until it returns a non-null value or `timeoutMs` elapses.
- * Probing first (rather than sleeping first) avoids a wasted initial interval
- * when the backend job is already complete.
- */
-export async function pollUntil(probe, opts = {}) {
-    const intervalMs = opts.intervalMs ?? 5_000;
-    const timeoutMs = opts.timeoutMs ?? 120_000;
-    const start = Date.now();
-    while (Date.now() - start < timeoutMs) {
-        const result = await probe();
-        if (result !== null)
-            return result;
-        await new Promise((r) => setTimeout(r, intervalMs));
-    }
-    throw new SkillError(`❌ polling timed out (${Math.round(timeoutMs / 1000)}s)`);
-}
+var f=Object.defineProperty;var a=(e,t)=>f(e,"name",{value:t,configurable:!0});import{traceHeaders as A,cleanDiagnosticText as w}from"./diagnostics.js";import{recordBilling as E}from"./billing.js";import{resolvePrivToken as N,NOT_AUTHENTICATED_HINT as d}from"./auth/resolve.js";import{attemptAutoLogin as v}from"./auth/auto-login.js";import{EXIT as l,SkillError as n}from"./errors.js";const _="https://api.remixmate.ai/api";import{SkillError as S,EXIT as b}from"./errors.js";function h(e){const t=(process.env.MM_API_BASE_URL??"").trim()||(process.env.MM_BACKEND_API_URL??"").trim();return((e??"").trim()||t||_).replace(/\/+$/,"")}a(h,"resolveApiBaseUrl");async function P(e,t={}){const o=h(t.apiBaseUrl);let i=await N({flag:t.privateToken??t.preflightToken,apiBaseUrl:o});if(i||(i=await v(o)),!i)throw new n(d,l.NOT_AUTHENTICATED);return{apiBaseUrl:o,privateToken:i.privToken,skillName:e,agentName:process.env.AGENT_NAME||void 0,conversationId:process.env.CONVERSATION_ID||void 0}}a(P,"resolveHttpContext");function g(e,t){const o={"Content-Type":"application/json","X-Priv-Token":e.privateToken,"x-invoke-skill":e.skillName};return e.agentName&&(o["x-invoke-agent"]=e.agentName),e.conversationId&&(o["x-conversation-id"]=e.conversationId),{...o,...t,...A()}}a(g,"buildHeaders");async function U(e,t,o,i={}){const m=t.startsWith("http")?t:`${e.apiBaseUrl}${t}`,c=new AbortController,T=setTimeout(()=>c.abort(),i.timeoutMs??12e4);let s;try{s=await fetch(m,{method:"POST",headers:g(e),body:JSON.stringify(o),signal:c.signal})}catch(u){throw u instanceof Error&&u.name==="AbortError"?new n(`❌ request timed out: ${m}`,l.BACKEND_UNREACHABLE):new n(`❌ network error: ${u.message}`,l.BACKEND_UNREACHABLE)}finally{clearTimeout(T)}const p=await s.text();if(!s.ok)throw s.status===401||s.status===403?new n(`${d}
+   （后端返回 HTTP ${s.status}）`,l.NOT_AUTHENTICATED):s.status>=500?new n(`❌ API request failed (HTTP ${s.status}): [response omitted]`,l.BACKEND_UNREACHABLE):new n(`❌ API request failed (HTTP ${s.status}): [response omitted]`);let r;try{r=JSON.parse(p)}catch{throw new n("❌ failed to parse response, body is not JSON: [response omitted]")}if(E(r.billing),r.code!==0)throw r.code===401||r.code===403?new n(`${d}
+   （后端返回 code=${r.code}${r.msg?`: ${r.msg}`:""}）`,l.NOT_AUTHENTICATED):new n(`❌ API returned a business error: ${w(r.msg??"unknown error")} (code=${r.code})`);return r.data??void 0}a(U,"mmPost");async function D(e,t={}){const o=t.intervalMs??5e3,i=t.timeoutMs??12e4,m=Date.now();for(;Date.now()-m<i;){const c=await e();if(c!==null)return c;await new Promise(T=>setTimeout(T,o))}throw new n(`❌ polling timed out (${Math.round(i/1e3)}s)`)}a(D,"pollUntil");export{b as EXIT,S as SkillError,U as mmPost,D as pollUntil,h as resolveApiBaseUrl,P as resolveHttpContext};

package/dist/index.js

@@ -1,8 +1 @@
-/**
- * remixmate — public library exports.
- *
- * The CLI entrypoint lives in `cli.ts` (`bin: remixmate`). This file is the
- * library surface for type-sharing consumers; the published npm package
- * exposes it via `main: dist/index.js` / `types: dist/index.d.ts`.
- */
-export { loadSkills, findSkill, SKILLS_DIR } from './registry.js';
+import{loadSkills as i,findSkill as o,SKILLS_DIR as d}from"./registry.js";export{d as SKILLS_DIR,o as findSkill,i as loadSkills};