infiplot-web/lib/types/index.ts

// ──────────────────────────────────────────────────────────────────────
//  Beat — one dialogue / narration moment within a Scene.
//  Multiple beats share the same background image; tapping or choosing
//  advances among them WITHOUT regenerating the image.
// ──────────────────────────────────────────────────────────────────────

export type Beat = {
  id: string;
  narration?: string;
  speaker?: string;
  line?: string;
  /** Free-form voice-acting direction for the line, sent to TTS only. Never displayed. */
  lineDelivery?: string;
  /**
   * Characters visible in this beat with their pose / expression for this moment.
   * Read by the Cinematographer when composing the scene's establishing shot —
   * the beat the entry beat lands in is the visual anchor for the image.
   */
  activeCharacters?: BeatActiveCharacter[];
  next: BeatNext;
};

export type BeatActiveCharacter = {
  name: string;
  /** Free-form 中文 description of pose / expression / what the character is doing. */
  pose?: string;
};

export type BeatNext =
  | { type: "continue"; nextBeatId: string }
  | { type: "choice"; choices: BeatChoice[] };

export type BeatChoice = {
  id: string;
  label: string;
  effect: BeatChoiceEffect;
};

export type BeatChoiceEffect =
  | { kind: "advance-beat"; targetBeatId: string }
  | { kind: "change-scene"; nextSceneSeed: string };

// ──────────────────────────────────────────────────────────────────────
//  Scene — one background image + a graph of beats.
//  The Director emits an entire Scene per call; the player navigates
//  through its beats locally with zero network until exiting.
// ──────────────────────────────────────────────────────────────────────

export type Scene = {
  id: string;
  scenePrompt: string;
  beats: Beat[];
  entryBeatId: string;
  /**
   * Stable English slug identifying the visual scene's location + time,
   * e.g. "classroom-dusk", "rooftop-night". When the next Scene shares this
   * key, the Painter slots the previous Scene's image into Runware's
   * `referenceImages` (alongside character portraits) so the same physical
   * space stays visually consistent across cuts.
   */
  sceneKey?: string;
  /**
   * Runware UUID of this Scene's generated image. Cheapest form to send back
   * to Runware's `referenceImages` in subsequent calls (UUID > URL > base64
   * in transport cost). Not shown to the client — `imageUrl` is what renders.
   */
  imageUuid?: string;
  /**
   * Public CDN URL of this Scene's generated image. Returned to the client for
   * `<img src>` rendering; the client also feeds it through a Canvas 2D click
   * annotator before posting to `/api/vision` (see
   * `VisionRequest.annotatedImageBase64`).
   *
   * For MOCK_IMAGE=true this is a `data:image/svg+xml;...` data URI, not a
   * Runware URL — the client renders both forms transparently.
   */
  imageUrl?: string;
};

export type SceneExit =
  | {
      kind: "choice";
      choiceId: string;
      label: string;
      nextSceneSeed: string;
    }
  | { kind: "freeform"; action: string };

export type SceneHistoryEntry = {
  scene: Scene;
  visitedBeatIds: string[];
  exit?: SceneExit;
};

// ──────────────────────────────────────────────────────────────────────
//  Characters & voices (TTS)
// ──────────────────────────────────────────────────────────────────────

export type CharacterVoice = {
  provider: "xiaomi";
  /** Xiaomi MiMo design output stored as reference audio for later clones. */
  referenceAudioBase64: string;
  mimeType: string;
};

export type Character = {
  name: string;
  /**
   * 中文 voice-acting direction card. Must begin with explicit gender, then
   * age / timbre / personality / speed / accent. Fed to Xiaomi MiMo's
   * voicedesign endpoint when the voice is first provisioned.
   */
  voiceDescription: string;
  /**
   * English appearance card — comma-separated visual attributes following
   * Runware/FLUX prompt-engineering convention. Fed to the Painter as a
   * character archetype anchor so the same face/outfit/style stays consistent
   * across every scene this character appears in.
   */
  visualDescription?: string;
  /**
   * Runware UUID for the base portrait. Generated by the CharacterDesigner
   * once, reused as a `referenceImages` entry on every subsequent scene the
   * character appears in. UUID is the cheapest reference form for Runware.
   */
  basePortraitUuid?: string;
  /**
   * Public CDN URL for the base portrait. Same image as `basePortraitUuid`;
   * kept around for the client (if it ever wants to render character cards)
   * and as a fallback reference form for `referenceImages` when UUID is absent.
   */
  basePortraitUrl?: string;
  /** Xiaomi MiMo voice reference audio. */
  voice?: CharacterVoice;
};

/** A single beat's synthesized audio, attached to the response. */
export type BeatAudio = {
  base64: string;
  mime: string;
};

// ──────────────────────────────────────────────────────────────────────
//  StoryState — the persistent "story bible" + evolving narrative memory.
//
//  Created once at session start by the Architect agent (rich opening
//  planning), then carried across every scene and incrementally updated by
//  the Writer. This is the single throughline that keeps tone, cast, and
//  stakes coherent across scene cuts — without it each Writer call would
//  re-derive the whole arc from a flat beat log and drift.
//
//  Split into STABLE fields (set by the Architect, rarely change) and
//  VOLATILE fields (rewritten each scene via StoryStatePatch).
// ──────────────────────────────────────────────────────────────────────

export type StoryState = {
  // ── Stable (Architect-authored; persists unless deliberately revised) ──
  /** One-line central dramatic question / 主线钩子. */
  logline: string;
  /** Genre + tone tags anchoring the 爽点 framework, e.g. "甜宠 / 校园 / 慢热治愈". */
  genreTags: string;
  /** Second-person protagonist card: who 你 are, the immediate situation, the
   *  core want, and a flaw/secret. The audience proxy — never rendered. */
  protagonist: string;
  /** Key supporting cast and their relationship/tension with 你 (one per line). */
  castNotes?: string;

  // ── Volatile (rewritten each scene by the Writer's StoryStatePatch) ──
  /** Rolling, compressed synopsis of what has happened so far (~3-5 句). */
  synopsis: string;
  /** Unresolved hooks / mysteries / questions still owed to the player. */
  openThreads?: string[];
  /** Current relationship/emotion state per character, e.g.
   *  "夏海：好感升温，刚向你告白了一半". */
  relationships?: string[];
  /** Where the story is heading next — the conflict/reversal/suspense the
   *  next scene should drive toward. Seeds the next scene's hook. */
  nextHook?: string;
};

/** The volatile subset the Writer rewrites after each scene. Stable fields
 *  (logline/genreTags/protagonist/castNotes) are preserved by the merge. */
export type StoryStatePatch = {
  synopsis?: string;
  openThreads?: string[];
  relationships?: string[];
  nextHook?: string;
};

// ──────────────────────────────────────────────────────────────────────
//  Session
// ──────────────────────────────────────────────────────────────────────

export type Session = {
  id: string;
  createdAt: number;
  worldSetting: string;
  styleGuide: string;
  history: SceneHistoryEntry[];
  /** Character registry — accumulates across scenes; voices + portraits persist for reuse. */
  characters: Character[];
  /**
   * Persistent story bible + evolving narrative memory. Set at session start
   * by the Architect, carried by the client across every /api/scene call, and
   * updated by the Writer each scene. Optional for back-compat with any
   * session payload created before this field existed.
   */
  storyState?: StoryState;
};

// ──────────────────────────────────────────────────────────────────────
//  Vision
// ──────────────────────────────────────────────────────────────────────

export type ClickIntent = {
  freeformAction: string;
  reasoning: string;
};

export type VisionClassify = "insert-beat" | "change-scene";

// ──────────────────────────────────────────────────────────────────────
//  Provider config
// ──────────────────────────────────────────────────────────────────────

export type ProviderConfig = {
  baseUrl: string;
  apiKey: string;
  model: string;
};

export type TtsConfig = {
  baseUrl: string;
  apiKey: string;
  /** Base model name; adapter derives "-voicedesign" / "-voiceclone" suffixes. */
  speechModel: string;
};

export type EngineConfig = {
  text: ProviderConfig;
  image: ProviderConfig;
  vision: ProviderConfig;
  /** Optional — when missing the game runs silently (no TTS). */
  tts?: TtsConfig;
  /** When true the renderer returns a placeholder PNG instead of calling the image API. */
  mockImage?: boolean;
};

// ──────────────────────────────────────────────────────────────────────
//  API contracts
// ──────────────────────────────────────────────────────────────────────

export type StartRequest = {
  worldSetting: string;
  styleGuide: string;
};

export type StartResponse = {
  sessionId: string;
  scene: Scene;
  /** Public CDN URL (or data URI in MOCK_IMAGE mode) for the rendered scene background. */
  imageUrl: string;
  /** Character registry with voice references + visual cards provisioned. */
  characters: Character[];
  /** Story bible created by the Architect + updated by the opening scene's
   *  Writer. The client persists this into the session for later /api/scene calls. */
  storyState: StoryState;
};

// /api/scene — generates the next Scene, given session whose latest
// history entry has `exit` set. Also used for prefetch speculation
// (frontend synthesizes a speculative exit).
export type SceneRequest = {
  session: Session;
};

export type SceneResponse = {
  scene: Scene;
  /** Public CDN URL (or data URI in MOCK_IMAGE mode) for the rendered scene background. */
  imageUrl: string;
  characters: Character[];
  /** Story bible after this scene's Writer applied its update. The client
   *  must persist this back into the session so the throughline survives the
   *  next scene cut. */
  storyState: StoryState;
};

// /api/beat-audio — lazily synthesize one beat's voice. Client fires this
// per beat after a scene loads; server has a per-call timeout so MiMo
// tail-latency cannot block the UI. A null audio response means "play silent."
export type BeatAudioRequest = {
  beat: {
    id: string;
    line: string;
    lineDelivery?: string;
  };
  voice: CharacterVoice;
};

export type BeatAudioResponse = {
  audio: BeatAudio | null;
};

// /api/vision — interprets a background click on the current image and
// classifies whether it should insert a beat (in-scene exploration) or
// trigger a scene change.
export type VisionRequest = {
  session: Session;
  /**
   * Raw PNG base64 (no `data:` prefix) of the scene image WITH the player's
   * click marker already drawn on it by the browser's Canvas 2D. The server
   * forwards this straight to the vision LLM as an OpenAI-compatible
   * image_url.
   *
   * Annotation lives client-side so the engine has no Node-native image
   * dependency (sharp doesn't run on Cloudflare Workers) and we save a
   * server-side image re-fetch per click.
   */
  annotatedImageBase64: string;
};

export type VisionResponse = {
  intent: ClickIntent;
  classify: VisionClassify;
};

// /api/insert-beat — generates a single transient beat in response to
// a freeform vision action. Does NOT regenerate the image.
export type InsertBeatRequest = {
  session: Session;
  freeformAction: string;
};

/** Partial beat fields produced by the insert-beat director. */
export type InsertBeatPartial = {
  narration?: string;
  speaker?: string;
  line?: string;
  lineDelivery?: string;
};

export type InsertBeatResponse = {
  partial: InsertBeatPartial;
  characters: Character[];
};