mulmocast 2.6.10 → 2.6.11-experimental.0

package/lib/agents/image_openai_agent.d.ts

@@ -1,5 +1,6 @@
 import { AgentFunction, AgentFunctionInfo } from "graphai";
 import type { AgentBufferResult, OpenAIImageAgentParams, OpenAIImageAgentInputs, OpenAIImageAgentConfig } from "../types/agent.js";
+export declare const buildDeprecatedModelMessage: (model: string) => string | null;
 export declare const imageOpenaiAgent: AgentFunction<OpenAIImageAgentParams, AgentBufferResult, OpenAIImageAgentInputs, OpenAIImageAgentConfig>;
 declare const imageOpenaiAgentInfo: AgentFunctionInfo;
 export default imageOpenaiAgentInfo;

package/lib/agents/image_openai_agent.js

@@ -3,8 +3,14 @@ import path from "path";
 import { GraphAILogger } from "graphai";
 import { toFile, AuthenticationError, RateLimitError, APIError } from "openai";
 import { createOpenAIClient } from "../utils/openai_client.js";
-import { provider2ImageAgent, gptImages } from "../types/provider2agent.js";
-import { apiKeyMissingError, agentGenerationError, openAIAgentGenerationError, agentIncorrectAPIKeyError, agentAPIRateLimitError, agentInvalidResponseError, imageAction, imageFileTarget, } from "../utils/error_cause.js";
+import { provider2ImageAgent, gptImages, deprecatedOpenAIImageModelHints } from "../types/provider2agent.js";
+import { apiKeyMissingError, agentGenerationError, openAIAgentGenerationError, agentIncorrectAPIKeyError, agentAPIRateLimitError, agentInvalidResponseError, imageAction, imageFileTarget, unsupportedModelTarget, } from "../utils/error_cause.js";
+const isDeprecatedOpenAIImageModel = (model) => model in deprecatedOpenAIImageModelHints;
+export const buildDeprecatedModelMessage = (model) => {
+    if (!isDeprecatedOpenAIImageModel(model))
+        return null;
+    return `OpenAI image model "${model}" is no longer available. ${deprecatedOpenAIImageModelHints[model]}`;
+};
 // https://platform.openai.com/docs/guides/image-generation
 export const imageOpenaiAgent = async ({ namedInputs, params, config, }) => {
     const { prompt, referenceImages } = namedInputs;
@@ -16,6 +22,12 @@ export const imageOpenaiAgent = async ({ namedInputs, params, config, }) => {
         });
     }
     const model = params.model ?? provider2ImageAgent["openai"].defaultModel;
+    const deprecatedMessage = buildDeprecatedModelMessage(model);
+    if (deprecatedMessage) {
+        throw new Error(deprecatedMessage, {
+            cause: agentGenerationError("imageOpenaiAgent", imageAction, unsupportedModelTarget),
+        });
+    }
     const openai = createOpenAIClient({ apiKey, baseURL, apiVersion });
     const size = (() => {
         if (gptImages.includes(model)) {
@@ -114,7 +126,7 @@ export const imageOpenaiAgent = async ({ namedInputs, params, config, }) => {
         }
         return { buffer: Buffer.from(image_base64, "base64") };
     }
-    // For dall-e-3
+    // URL response handling (legacy OpenAI image API response format)
     const res = await fetch(url);
     if (!res.ok) {
         throw new Error(`Failed to fetch ${url}: ${res.status} ${res.statusText}`, {

package/lib/methods/mulmo_presentation_style.js

@@ -84,7 +84,7 @@ export const MulmoPresentationStyleMethods = {
         const imageParams = { ...presentationStyle.imageParams, ...beat?.imageParams };
         const provider = MulmoPresentationStyleMethods.getText2ImageProvider(imageParams?.provider) ?? defaultProviders.text2image;
         const agentInfo = provider2ImageAgent[provider];
-        // The default text2image model is gpt-image-1 from OpenAI, and to use it you must have an OpenAI account and have verified your identity. If this is not possible, please specify dall-e-3 as the model.
+        // The default text2image model is gpt-image-1 from OpenAI.
         const defaultImageParams = {
             provider,
             model: agentInfo.defaultModel,
@@ -133,8 +133,7 @@ export const MulmoPresentationStyleMethods = {
         const imageAgentInfo = MulmoPresentationStyleMethods.getImageAgentInfo(presentationStyle);
         if (imageAgentInfo.imageParams.provider === "openai") {
             // NOTE: Here are the rate limits of OpenAI's text2image API (1token = 32x32 patch).
-            // dall-e-3: 7,500 RPM、15 images per minute (4 images for max resolution)
-            // gpt-image-1：3,000,000 TPM、150 images per minute
+            // gpt-image-1: 3,000,000 TPM, 150 images per minute
             if (imageAgentInfo.imageParams.model === provider2ImageAgent.openai.defaultModel) {
                 return 16;
             }

package/lib/types/provider2agent.d.ts

@@ -42,6 +42,11 @@ export declare const provider2TTSAgent: {
     };
 };
 export declare const gptImages: string[];
+export declare const deprecatedOpenAIImageModelHints: {
+    readonly "dall-e-2": "Use 'gpt-image-1' or another supported model.";
+    readonly "dall-e-3": "Use 'gpt-image-1' or another supported model.";
+};
+export type DeprecatedOpenAIImageModel = keyof typeof deprecatedOpenAIImageModelHints;
 export declare const provider2ImageAgent: {
     openai: {
         agentName: string;

package/lib/types/provider2agent.js

@@ -45,11 +45,16 @@ export const provider2TTSAgent = {
     },
 };
 export const gptImages = ["gpt-image-2", "gpt-image-1.5", "gpt-image-1", "gpt-image-1-mini"];
+const supportedOpenAIImageReplacementHint = "Use 'gpt-image-1' or another supported model.";
+export const deprecatedOpenAIImageModelHints = {
+    "dall-e-2": supportedOpenAIImageReplacementHint,
+    "dall-e-3": supportedOpenAIImageReplacementHint,
+};
 export const provider2ImageAgent = {
     openai: {
         agentName: "imageOpenaiAgent",
         defaultModel: "gpt-image-1",
-        models: ["dall-e-3", ...gptImages],
+        models: [...gptImages],
         keyName: "OPENAI_API_KEY",
         baseURLKeyName: "OPENAI_BASE_URL",
     },

package/lib/utils/ffmpeg_utils.d.ts

@@ -1,6 +1,7 @@
-import ffmpeg from "fluent-ffmpeg";
+import ffmpeg from "@modernized/fluent-ffmpeg";
+type FfmpegCommand = ReturnType<typeof ffmpeg>;
 export type FfmpegContext = {
-    command: ffmpeg.FfmpegCommand;
+    command: FfmpegCommand;
     inputCount: number;
     filterComplex: string[];
 };
@@ -29,3 +30,4 @@ export declare const extractImageFromMovie: (movieFile: string, imagePath: strin
 export declare const trimMusic: (inputFile: string, startTime: number, duration: number) => Promise<Buffer>;
 export declare const createSilentAudio: (filePath: string, durationSec: number) => Promise<void>;
 export declare const pcmToMp3: (rawPcm: Buffer, sampleRate?: number) => Promise<Buffer>;
+export {};

package/lib/utils/ffmpeg_utils.js

@@ -1,4 +1,4 @@
-import ffmpeg from "fluent-ffmpeg";
+import ffmpeg from "@modernized/fluent-ffmpeg";
 import { GraphAILogger } from "graphai";
 import { isFile } from "./file.js";
 import fs from "fs";

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "mulmocast",
-  "version": "2.6.10",
+  "version": "2.6.11-experimental.0",
   "description": "",
   "type": "module",
   "main": "lib/index.node.js",
@@ -88,7 +88,7 @@
   },
   "homepage": "https://github.com/receptron/mulmocast-cli#readme",
   "dependencies": {
-    "@google-cloud/text-to-speech": "^6.4.0",
+    "@google-cloud/text-to-speech": "^6.4.1",
     "@google/genai": "^1.50.1",
     "@graphai/anthropic_agent": "^2.0.12",
     "@graphai/browserless_agent": "^2.0.2",
@@ -102,41 +102,40 @@
     "@inquirer/input": "^5.0.12",
     "@inquirer/select": "^5.1.4",
     "@modelcontextprotocol/sdk": "^1.29.0",
+    "@modernized/fluent-ffmpeg": "^0.1.1",
     "@mozilla/readability": "^0.6.0",
     "@tavily/core": "^0.5.11",
     "archiver": "^7.0.1",
     "clipboardy": "^5.3.1",
     "dotenv": "^17.4.2",
-    "fluent-ffmpeg": "^2.1.3",
     "graphai": "^2.0.16",
-    "jsdom": "^29.1.0",
-    "marked": "^18.0.2",
+    "jsdom": "^29.1.1",
+    "marked": "^18.0.3",
     "mulmocast-vision": "^1.0.9",
     "ora": "^9.4.0",
-    "puppeteer": "^24.42.0",
+    "puppeteer": "^24.43.0",
     "replicate": "^1.4.0",
-    "yaml": "^2.8.3",
+    "yaml": "^2.8.4",
     "yargs": "^18.0.0",
-    "zod": "^4.3.6"
+    "zod": "^4.4.3"
   },
   "devDependencies": {
     "@eslint/js": "^10.0.1",
     "@receptron/test_utils": "^2.0.3",
     "@types/archiver": "^7.0.0",
-    "@types/fluent-ffmpeg": "^2.1.28",
     "@types/jsdom": "^28.0.1",
     "@types/yargs": "^17.0.35",
     "cross-env": "^10.1.0",
-    "eslint": "^10.2.1",
+    "eslint": "^10.3.0",
     "eslint-config-prettier": "^10.1.8",
     "eslint-plugin-import": "^2.32.0",
     "eslint-plugin-prettier": "^5.5.5",
     "eslint-plugin-sonarjs": "^4.0.3",
-    "globals": "^17.5.0",
+    "globals": "^17.6.0",
     "prettier": "^3.8.3",
     "tsx": "^4.21.0",
     "typescript": "6.0.3",
-    "typescript-eslint": "^8.59.1"
+    "typescript-eslint": "^8.59.2"
   },
   "engines": {
     "node": ">=22.0.0"

package/scripts/test/test_images.json

@@ -21,10 +21,10 @@
       }
     },
     {
-      "text": "Image with Dall-E 3",
+      "text": "Image with OpenAI gpt-image-1-mini",
       "imagePrompt": "Blue sky, a flock of birds",
       "imageParams": {
-        "model": "dall-e-3",
+        "model": "gpt-image-1-mini",
         "style": "anime-style",
         "provider": "openai"
       }

package/scripts/test/test_images_dalle_deprecated.json

@@ -0,0 +1,27 @@
+{
+  "$mulmocast": {
+    "version": "1.1"
+  },
+  "lang": "en",
+  "title": "Test Deprecated OpenAI Image Models",
+  "imageParams": {
+    "provider": "openai",
+    "style": "Photorealistic-style"
+  },
+  "beats": [
+    {
+      "text": "dall-e-2 is deprecated; mulmocast rejects this with a migration hint before calling OpenAI",
+      "imagePrompt": "Blue sky, a flock of birds",
+      "imageParams": {
+        "model": "dall-e-2"
+      }
+    },
+    {
+      "text": "dall-e-3 is deprecated; mulmocast rejects this with a migration hint before calling OpenAI",
+      "imagePrompt": "Blue sky, a flock of birds",
+      "imageParams": {
+        "model": "dall-e-3"
+      }
+    }
+  ]
+}

package/scripts/test/test_movie.json

@@ -12,7 +12,7 @@
   },
   "imageParams": {
     "provider": "openai",
-    "model": "dall-e-3",
+    "model": "gpt-image-1-mini",
     "style": "Photo realistic, cinematic style.",
     "images": {
       "optimus": {

package/scripts/test/test_movie_dalle_deprecated.json

@@ -0,0 +1,26 @@
+{
+  "$mulmocast": {
+    "version": "1.1"
+  },
+  "lang": "en",
+  "movieParams": {
+    "provider": "google"
+  },
+  "canvasSize": {
+    "width": 720,
+    "height": 1280
+  },
+  "imageParams": {
+    "provider": "openai",
+    "model": "dall-e-3",
+    "style": "Photo realistic, cinematic style."
+  },
+  "beats": [
+    {
+      "text": "dall-e-3 on movie flow is deprecated; mulmocast rejects upfront with a migration hint",
+      "imagePrompt": "A rocket on the launch pad at dusk.",
+      "moviePrompt": "Rocket lifts off.",
+      "duration": 5
+    }
+  ]
+}