mulmocast 0.0.9 → 0.0.11

package/lib/agents/movie_google_agent.d.ts

@@ -0,0 +1,24 @@
+import type { AgentFunction, AgentFunctionInfo } from "graphai";
+export type MovieGoogleConfig = {
+    projectId?: string;
+    token?: string;
+};
+export declare const getAspectRatio: (canvasSize: {
+    width: number;
+    height: number;
+}) => string;
+export declare const movieGoogleAgent: AgentFunction<{
+    model: string;
+    canvasSize: {
+        width: number;
+        height: number;
+    };
+    duration?: number;
+}, {
+    buffer: Buffer;
+}, {
+    prompt: string;
+    imagePath?: string;
+}, MovieGoogleConfig>;
+declare const movieGoogleAgentInfo: AgentFunctionInfo;
+export default movieGoogleAgentInfo;

package/lib/agents/movie_google_agent.js

@@ -0,0 +1,122 @@
+import { readFileSync } from "fs";
+import { GraphAILogger, sleep } from "graphai";
+async function generateMovie(projectId, model, token, prompt, imagePath, aspectRatio, duration) {
+    const GOOGLE_IMAGEN_ENDPOINT = `https://us-central1-aiplatform.googleapis.com/v1/projects/${projectId}/locations/us-central1/publishers/google/models/${model}`;
+    const payload = {
+        instances: [
+            {
+                prompt: prompt,
+                image: undefined,
+            },
+        ],
+        parameters: {
+            sampleCount: 1,
+            aspectRatio: aspectRatio,
+            safetySetting: "block_only_high",
+            personGeneration: "allow_all",
+            durationSeconds: duration,
+        },
+    };
+    if (imagePath) {
+        const buffer = readFileSync(imagePath);
+        const bytesBase64Encoded = buffer.toString("base64");
+        payload.instances[0].image = {
+            bytesBase64Encoded,
+            mimeType: "image/png",
+        };
+    }
+    // Make the API call using fetch
+    const response = await fetch(`${GOOGLE_IMAGEN_ENDPOINT}:predictLongRunning`, {
+        method: "POST",
+        headers: {
+            Authorization: `Bearer ${token}`,
+            "Content-Type": "application/json",
+        },
+        body: JSON.stringify(payload),
+    });
+    if (!response.ok) {
+        GraphAILogger.info("create project on google cloud console and setup the project. More details see readme.");
+        throw new Error(`Error: ${response.status} - ${response.statusText}`);
+    }
+    const initialResponse = await response.json();
+    const fetchBody = {
+        operationName: initialResponse.name,
+    };
+    const completeResponse = await (async () => {
+        while (true) {
+            GraphAILogger.info("...waiting for movie generation...");
+            await sleep(3000);
+            const response = await fetch(`${GOOGLE_IMAGEN_ENDPOINT}:fetchPredictOperation`, {
+                method: "POST",
+                headers: {
+                    Authorization: `Bearer ${token}`,
+                    "Content-Type": "application/json",
+                },
+                body: JSON.stringify(fetchBody),
+            });
+            if (!response.ok) {
+                throw new Error(`Error: ${response.status} - ${response.statusText}`);
+            }
+            const responseData = await response.json();
+            if (responseData.done) {
+                if (responseData.error) {
+                    GraphAILogger.info("Prompt: ", prompt);
+                    throw new Error(`Error: ${responseData.error.message}`);
+                }
+                if (!responseData.response.videos) {
+                    throw new Error(`No video: ${JSON.stringify(responseData, null, 2)}`);
+                }
+                return responseData.response;
+            }
+        }
+    })();
+    const encodedMovie = completeResponse.videos[0].bytesBase64Encoded;
+    if (encodedMovie) {
+        return Buffer.from(encodedMovie, "base64");
+    }
+    return undefined;
+}
+export const getAspectRatio = (canvasSize) => {
+    if (canvasSize.width > canvasSize.height) {
+        return "16:9";
+    }
+    else if (canvasSize.width < canvasSize.height) {
+        return "9:16";
+    }
+    else {
+        return "1:1";
+    }
+};
+export const movieGoogleAgent = async ({ namedInputs, params, config }) => {
+    const { prompt, imagePath } = namedInputs;
+    const aspectRatio = getAspectRatio(params.canvasSize);
+    const model = params.model ?? "veo-2.0-generate-001"; // "veo-3.0-generate-preview";
+    const duration = params.duration ?? 8;
+    const projectId = config?.projectId;
+    const token = config?.token;
+    try {
+        const buffer = await generateMovie(projectId, model, token, prompt, imagePath, aspectRatio, duration);
+        if (buffer) {
+            return { buffer };
+        }
+        throw new Error("ERROR: geneateImage returned undefined");
+    }
+    catch (error) {
+        GraphAILogger.info("Failed to generate movie:", error);
+        throw error;
+    }
+};
+const movieGoogleAgentInfo = {
+    name: "movieGoogleAgent",
+    agent: movieGoogleAgent,
+    mock: movieGoogleAgent,
+    samples: [],
+    description: "Google Movie agent",
+    category: ["movie"],
+    author: "Receptron Team",
+    repository: "https://github.com/receptron/mulmocast-cli/",
+    // source: "https://github.com/receptron/mulmocast-cli/blob/main/src/agents/image_google_agent.ts",
+    license: "MIT",
+    environmentVariables: [],
+};
+export default movieGoogleAgentInfo;

package/lib/cli/bin.js

@@ -36,6 +36,18 @@ export const main = async () => {
         .demandCommand()
         .strict()
         .help()
+        .showHelpOnFail(false)
+        .fail((msg, err, y) => {
+        // if yargs detect error, show help and exit
+        if (msg) {
+            y.showHelp();
+            GraphAILogger.info("\\n" + msg);
+            process.exit(1);
+        }
+        if (err) {
+            throw err;
+        }
+    })
         .alias("help", "h");
     await cli.parseAsync();
 };

package/lib/index.d.ts

@@ -0,0 +1,5 @@
+export * from "./actions/index.js";
+export * from "./cli/helpers.js";
+export * from "./utils/file.js";
+export * from "./methods/index.js";
+export * from "./types/index.js";

package/lib/index.js

@@ -0,0 +1,5 @@
+export * from "./actions/index.js";
+export * from "./cli/helpers.js";
+export * from "./utils/file.js";
+export * from "./methods/index.js";
+export * from "./types/index.js";

package/lib/methods/mulmo_script.d.ts

@@ -2,7 +2,6 @@ import "dotenv/config";
 import { MulmoCanvasDimension, MulmoScript, MulmoBeat, SpeechOptions, Text2SpeechProvider, Text2ImageAgentInfo, BeatMediaType } from "../types/index.js";
 export declare const MulmoScriptMethods: {
     getCanvasSize(script: MulmoScript): MulmoCanvasDimension;
-    getAspectRatio(script: MulmoScript): string;
     getSpeechProvider(script: MulmoScript): Text2SpeechProvider;
     getTextSlideStyle(script: MulmoScript, beat: MulmoBeat): string;
     getSpeechOptions(script: MulmoScript, beat: MulmoBeat): SpeechOptions | undefined;

package/lib/methods/mulmo_script.js

@@ -18,11 +18,6 @@ export const MulmoScriptMethods = {
     getCanvasSize(script) {
         return mulmoCanvasDimensionSchema.parse(script.canvasSize);
     },
-    getAspectRatio(script) {
-        // Google's text2image specific parameter
-        const size = this.getCanvasSize(script);
-        return size.width > size.height ? "16:9" : "9:16";
-    },
     getSpeechProvider(script) {
         return text2SpeechProviderSchema.parse(script.speechParams?.provider);
     },

package/lib/methods/mulmo_studio.d.ts

@@ -1,6 +1,6 @@
 import { MulmoStudio } from "../types/index.js";
 type SessionType = "audio" | "image" | "video" | "multiLingual" | "caption" | "pdf";
-type BeatSessionType = "audio" | "image" | "multiLingual" | "caption";
+type BeatSessionType = "audio" | "image" | "multiLingual" | "caption" | "movie";
 export declare const MulmoStudioMethods: {
     setSessionState(studio: MulmoStudio, sessionType: SessionType, value: boolean): void;
     setBeatSessionState(studio: MulmoStudio, sessionType: BeatSessionType, index: number, value: boolean): void;

package/lib/tools/create_mulmo_script_from_url.js

@@ -4,7 +4,7 @@ import { openAIAgent } from "@graphai/openai_agent";
 import { anthropicAgent } from "@graphai/anthropic_agent";
 import { geminiAgent } from "@graphai/gemini_agent";
 import { groqAgent } from "@graphai/groq_agent";
-import vanillaAgents from "@graphai/vanilla";
+import * as agents from "@graphai/vanilla";
 import { fileWriteAgent } from "@graphai/vanilla_node_agents";
 import { browserlessAgent } from "@graphai/browserless_agent";
 import validateSchemaAgent from "../agents/validate_schema_agent.js";
@@ -14,7 +14,7 @@ import { mulmoScriptSchema, urlsSchema } from "../types/schema.js";
 import { cliLoadingPlugin } from "../utils/plugins.js";
 import { graphDataScriptFromUrlPrompt } from "../utils/prompt.js";
 import { llmPair } from "../utils/utils.js";
-// const { default: __, ...vanillaAgents } = agents;
+const vanillaAgents = agents.default ?? agents;
 const graphData = {
     version: 0.5,
     // Execute sequentially because the free version of browserless API doesn't support concurrent execution.

package/lib/tools/create_mulmo_script_interactively.js

@@ -6,7 +6,7 @@ import { openAIAgent } from "@graphai/openai_agent";
 import { anthropicAgent } from "@graphai/anthropic_agent";
 import { geminiAgent } from "@graphai/gemini_agent";
 import { groqAgent } from "@graphai/groq_agent";
-import vanillaAgents from "@graphai/vanilla";
+import * as agents from "@graphai/vanilla";
 import { fileWriteAgent } from "@graphai/vanilla_node_agents";
 import { readTemplatePrompt, mkdir } from "../utils/file.js";
 import { browserlessCacheGenerator } from "../utils/filters.js";
@@ -16,7 +16,7 @@ import validateSchemaAgent from "../agents/validate_schema_agent.js";
 import { llmPair } from "../utils/utils.js";
 import { interactiveClarificationPrompt, prefixPrompt } from "../utils/prompt.js";
 // import { cliLoadingPlugin } from "../utils/plugins.js";
-// const { default: __, ...vanillaAgents } = agents;
+const vanillaAgents = agents.default ?? agents;
 const agentHeader = "\x1b[34m● \x1b[0m\x1b[1mAgent\x1b[0m:\x1b[0m";
 const graphDataForScraping = {
     version: 0.5,

package/lib/tools/story_to_script.js

@@ -5,14 +5,14 @@ import { openAIAgent } from "@graphai/openai_agent";
 import { anthropicAgent } from "@graphai/anthropic_agent";
 import { geminiAgent } from "@graphai/gemini_agent";
 import { groqAgent } from "@graphai/groq_agent";
-import vanillaAgents from "@graphai/vanilla";
+import * as agents from "@graphai/vanilla";
 import { graphDataScriptGeneratePrompt, sceneToBeatsPrompt, storyToScriptInfoPrompt, storyToScriptPrompt } from "../utils/prompt.js";
 import { fileWriteAgent } from "@graphai/vanilla_node_agents";
 import validateSchemaAgent from "../agents/validate_schema_agent.js";
 import { llmPair } from "../utils/utils.js";
 import { storyToScriptGenerateMode } from "../utils/const.js";
 import { cliLoadingPlugin } from "../utils/plugins.js";
-// const { default: __, ...vanillaAgents } = agents;
+const vanillaAgents = agents.default ?? agents;
 const createValidatedScriptGraphData = ({ systemPrompt, prompt, schema, llmAgent, llmModel, maxTokens, }) => {
     return {
         loop: {

package/lib/types/index.d.ts

@@ -1,2 +1,3 @@
 export * from "./type.js";
 export * from "./schema.js";
+export * from "./cli_types.js";

package/lib/types/index.js

@@ -1,2 +1,3 @@
 export * from "./type.js";
 export * from "./schema.js";
+export * from "./cli_types.js";