@langchain/classic 1.0.17 → 1.0.18

package/dist/experimental/autogpt/prompt.d.ts

@@ -1,7 +1,7 @@
 import { ObjectTool } from "./schema.js";
-import { BaseChatPromptTemplate, SerializedBasePromptTemplate } from "@langchain/core/prompts";
-import * as _langchain_core_messages3 from "@langchain/core/messages";
+import * as _langchain_core_messages0 from "@langchain/core/messages";
 import { BaseMessage } from "@langchain/core/messages";
+import { BaseChatPromptTemplate, SerializedBasePromptTemplate } from "@langchain/core/prompts";
 import { PartialValues } from "@langchain/core/utils/types";
 import { VectorStoreRetrieverInterface } from "@langchain/core/vectorstores";
 
@@ -54,7 +54,7 @@ declare class AutoGPTPrompt extends BaseChatPromptTemplate implements AutoGPTPro
     memory: VectorStoreRetrieverInterface;
     messages: BaseMessage[];
     user_input: string;
-  }): Promise<BaseMessage<_langchain_core_messages3.MessageStructure<_langchain_core_messages3.MessageToolSet>, _langchain_core_messages3.MessageType>[]>;
+  }): Promise<BaseMessage<_langchain_core_messages0.MessageStructure<_langchain_core_messages0.MessageToolSet>, _langchain_core_messages0.MessageType>[]>;
   /**
    * This method is not implemented in the AutoGPTPrompt class and will
    * throw an error if called.

package/dist/experimental/autogpt/prompt.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"prompt.d.ts","names":["VectorStoreRetrieverInterface","BaseChatPromptTemplate","SerializedBasePromptTemplate","BaseMessage","PartialValues","ObjectTool","AutoGPTPromptInput","Promise","AutoGPTPrompt","goals","memory","previousMessages","user_input","_langchain_core_messages3","MessageToolSet","MessageStructure","MessageType"],"sources":["../../../src/experimental/autogpt/prompt.d.ts"],"sourcesContent":["import type { VectorStoreRetrieverInterface } from \"@langchain/core/vectorstores\";\nimport { BaseChatPromptTemplate, SerializedBasePromptTemplate } from \"@langchain/core/prompts\";\nimport { BaseMessage } from \"@langchain/core/messages\";\nimport { PartialValues } from \"@langchain/core/utils/types\";\nimport { ObjectTool } from \"./schema.js\";\n/**\n * Interface for the input parameters of the AutoGPTPrompt class.\n */\nexport interface AutoGPTPromptInput {\n    aiName: string;\n    aiRole: string;\n    tools: ObjectTool[];\n    tokenCounter: (text: string) => Promise<number>;\n    sendTokenLimit?: number;\n}\n/**\n * Class used to generate prompts for the AutoGPT model. It takes into\n * account the AI's name, role, tools, token counter, and send token\n * limit. The class also handles the formatting of messages and the\n * construction of the full prompt.\n */\nexport declare class AutoGPTPrompt extends BaseChatPromptTemplate implements AutoGPTPromptInput {\n    aiName: string;\n    aiRole: string;\n    tools: ObjectTool[];\n    tokenCounter: (text: string) => Promise<number>;\n    sendTokenLimit: number;\n    constructor(fields: AutoGPTPromptInput);\n    _getPromptType(): \"autogpt\";\n    /**\n     * Constructs the full prompt based on the provided goals.\n     * @param goals An array of goals.\n     * @returns The full prompt as a string.\n     */\n    constructFullPrompt(goals: string[]): string;\n    /**\n     * Formats the messages based on the provided parameters.\n     * @param goals An array of goals.\n     * @param memory A VectorStoreRetriever instance.\n     * @param messages An array of previous messages.\n     * @param user_input The user's input.\n     * @returns An array of formatted messages.\n     */\n    formatMessages({ goals, memory, messages: previousMessages, user_input }: {\n        goals: string[];\n        memory: VectorStoreRetrieverInterface;\n        messages: BaseMessage[];\n        user_input: string;\n    }): Promise<BaseMessage<import(\"@langchain/core/messages\").MessageStructure<import(\"@langchain/core/messages\").MessageToolSet>, import(\"@langchain/core/messages\").MessageType>[]>;\n    /**\n     * This method is not implemented in the AutoGPTPrompt class and will\n     * throw an error if called.\n     * @param _values Partial values.\n     * @returns Throws an error.\n     */\n    partial(_values: PartialValues): Promise<BaseChatPromptTemplate>;\n    serialize(): SerializedBasePromptTemplate;\n}\n//# sourceMappingURL=prompt.d.ts.map"],"mappings":";;;;;;;;;;;AAQiBM,UAAAA,kBAAAA,CAAkB;EAadE,MAAAA,EAAAA,MAAAA;EAGVH,MAAAA,EAAAA,MAAAA;EACyBE,KAAAA,EAdzBF,UAcyBE,EAAAA;EAEZD,YAAAA,EAAAA,CAAAA,IAAAA,EAAAA,MAAAA,EAAAA,GAfYC,OAeZD,CAAAA,MAAAA,CAAAA;EAgBHG,cAAAA,CAAAA,EAAAA,MAAAA;;;;;;;;AAK0DI,cA3B1DL,aAAAA,SAAsBP,sBAAAA,YAAkCK,kBA2B0FU,CAAAA;EAAvJb,MAAAA,EAAAA,MAAAA;EAARI,MAAAA,EAAAA,MAAAA;EAOaH,KAAAA,EA/BVC,UA+BUD,EAAAA;EAAwBH,YAAAA,EAAAA,CAAAA,IAAAA,EAAAA,MAAAA,EAAAA,GA9BTM,OA8BSN,CAAAA,MAAAA,CAAAA;EAARM,cAAAA,EAAAA,MAAAA;EACpBL,WAAAA,CAAAA,MAAAA,EA7BOI,kBA6BPJ;EAnC0BD,cAAAA,CAAAA,CAAAA,EAAAA,SAAAA;EAAkCK;AAAkB;;;;;;;;;;;;;;;;cAsBjDK;;;;YAE9BX;cACEG;;MAEVI,QAAQJ,YAAiHU,yBAAAA,CAAlEE,iBAFlCF,yBAAAA,CAEsFC,cAAAA,GAApCD,yBAAAA,CAAwFG,WAAAA;;;;;;;mBAOlJZ,gBAAgBG,QAAQN;eAC5BC"}
+{"version":3,"file":"prompt.d.ts","names":[],"sources":["../../../src/experimental/autogpt/prompt.ts"],"mappings":";;;;;;;;;;;UAiBiB,kBAAA;EACf,MAAA;EACA,MAAA;EACA,KAAA,EAAO,UAAA;EACP,YAAA,GAAe,IAAA,aAAiB,OAAA;EAChC,cAAA;AAAA;;;;;;;cASW,aAAA,SACH,sBAAA,YACG,kBAAA;EAEX,MAAA;EAEA,MAAA;EAEA,KAAA,EAAO,UAAA;EAEP,YAAA,GAAe,IAAA,aAAiB,OAAA;EAEhC,cAAA;EAEA,WAAA,CAAY,MAAA,EAAQ,kBAAA;EASpB,cAAA,CAAA;EAmCE;;;;;EA1BF,mBAAA,CAAoB,KAAA;EAkCnB;;;;;;;;EAVK,cAAA,CAAA;IACJ,KAAA;IACA,MAAA;IACA,QAAA,EAAU,gBAAA;IACV;EAAA;IAEA,KAAA;IACA,MAAA,EAAQ,6BAAA;IACR,QAAA,EAAU,WAAA;IACV,UAAA;EAAA,IACD,OAAA,CAAA,WAAA,CAAA,yBAAA,CAAA,gBAAA,CAFsB,yBAAA,CAEtB,cAAA,GAAA,yBAAA,CAAA,WAAA;EA9DD;;;;;;EAwIM,OAAA,CAAQ,OAAA,EAAS,aAAA,GAAgB,OAAA,CAAQ,sBAAA;EAI/C,SAAA,CAAA,GAAa,4BAAA;AAAA"}

package/dist/experimental/autogpt/prompt.js

@@ -1,6 +1,6 @@
 import { getPrompt } from "./prompt_generator.js";
-import { BaseChatPromptTemplate } from "@langchain/core/prompts";
 import { HumanMessage, SystemMessage } from "@langchain/core/messages";
+import { BaseChatPromptTemplate } from "@langchain/core/prompts";
 
 //#region src/experimental/autogpt/prompt.ts
 /**
@@ -37,12 +37,11 @@ var AutoGPTPrompt = class extends BaseChatPromptTemplate {
 	* @returns The full prompt as a string.
 	*/
 	constructFullPrompt(goals) {
-		const promptStart = `Your decisions must always be made independently
+		let fullPrompt = `You are ${this.aiName}, ${this.aiRole}\nYour decisions must always be made independently
             without seeking user assistance. Play to your strengths
             as an LLM and pursue simple strategies with no legal complications.
             If you have completed all your tasks,
-            make sure to use the "finish" command.`;
-		let fullPrompt = `You are ${this.aiName}, ${this.aiRole}\n${promptStart}\n\nGOALS:\n\n`;
+            make sure to use the "finish" command.\n\nGOALS:\n\n`;
 		goals.forEach((goal, index) => {
 			fullPrompt += `${index + 1}. ${goal}\n`;
 		});
@@ -62,33 +61,29 @@ var AutoGPTPrompt = class extends BaseChatPromptTemplate {
 		const timePrompt = new SystemMessage(`The current time and date is ${(/* @__PURE__ */ new Date()).toLocaleString("en-US")}`);
 		if (typeof basePrompt.content !== "string" || typeof timePrompt.content !== "string") throw new Error("Non-string message content is not supported.");
 		const usedTokens = await this.tokenCounter(basePrompt.content) + await this.tokenCounter(timePrompt.content);
-		const relevantDocs = await memory.invoke(JSON.stringify(previousMessages.slice(-10)));
-		const relevantMemory = relevantDocs.map((d) => d.pageContent);
+		const relevantMemory = (await memory.invoke(JSON.stringify(previousMessages.slice(-10)))).map((d) => d.pageContent);
 		let relevantMemoryTokens = await relevantMemory.reduce(async (acc, doc) => await acc + await this.tokenCounter(doc), Promise.resolve(0));
 		while (usedTokens + relevantMemoryTokens > 2500) {
 			relevantMemory.pop();
 			relevantMemoryTokens = await relevantMemory.reduce(async (acc, doc) => await acc + await this.tokenCounter(doc), Promise.resolve(0));
 		}
-		const contentFormat = `This reminds you of these events from your past:\n${relevantMemory.join("\n")}\n\n`;
-		const memoryMessage = new SystemMessage(contentFormat);
+		const memoryMessage = new SystemMessage(`This reminds you of these events from your past:\n${relevantMemory.join("\n")}\n\n`);
 		if (typeof memoryMessage.content !== "string") throw new Error("Non-string message content is not supported.");
 		const usedTokensWithMemory = usedTokens + await this.tokenCounter(memoryMessage.content);
 		const historicalMessages = [];
 		for (const message of previousMessages.slice(-10).reverse()) {
 			if (typeof message.content !== "string") throw new Error("Non-string message content is not supported.");
-			const messageTokens = await this.tokenCounter(message.content);
-			if (usedTokensWithMemory + messageTokens > this.sendTokenLimit - 1e3) break;
+			if (usedTokensWithMemory + await this.tokenCounter(message.content) > this.sendTokenLimit - 1e3) break;
 			historicalMessages.unshift(message);
 		}
 		const inputMessage = new HumanMessage(user_input);
-		const messages = [
+		return [
 			basePrompt,
 			timePrompt,
 			memoryMessage,
 			...historicalMessages,
 			inputMessage
 		];
-		return messages;
 	}
 	/**
 	* This method is not implemented in the AutoGPTPrompt class and will

package/dist/experimental/autogpt/prompt.js.map

@@ -1 +1 @@
-{"version":3,"file":"prompt.js","names":["fields: AutoGPTPromptInput","goals: string[]","d: { pageContent: string }","acc: Promise<number>","doc: string","historicalMessages: BaseMessage[]","messages: BaseMessage[]","_values: PartialValues"],"sources":["../../../src/experimental/autogpt/prompt.ts"],"sourcesContent":["import type { VectorStoreRetrieverInterface } from \"@langchain/core/vectorstores\";\nimport {\n  BaseChatPromptTemplate,\n  SerializedBasePromptTemplate,\n} from \"@langchain/core/prompts\";\nimport {\n  BaseMessage,\n  HumanMessage,\n  SystemMessage,\n} from \"@langchain/core/messages\";\nimport { PartialValues } from \"@langchain/core/utils/types\";\nimport { getPrompt } from \"./prompt_generator.js\";\nimport { ObjectTool } from \"./schema.js\";\n\n/**\n * Interface for the input parameters of the AutoGPTPrompt class.\n */\nexport interface AutoGPTPromptInput {\n  aiName: string;\n  aiRole: string;\n  tools: ObjectTool[];\n  tokenCounter: (text: string) => Promise<number>;\n  sendTokenLimit?: number;\n}\n\n/**\n * Class used to generate prompts for the AutoGPT model. It takes into\n * account the AI's name, role, tools, token counter, and send token\n * limit. The class also handles the formatting of messages and the\n * construction of the full prompt.\n */\nexport class AutoGPTPrompt\n  extends BaseChatPromptTemplate\n  implements AutoGPTPromptInput\n{\n  aiName: string;\n\n  aiRole: string;\n\n  tools: ObjectTool[];\n\n  tokenCounter: (text: string) => Promise<number>;\n\n  sendTokenLimit: number;\n\n  constructor(fields: AutoGPTPromptInput) {\n    super({ inputVariables: [\"goals\", \"memory\", \"messages\", \"user_input\"] });\n    this.aiName = fields.aiName;\n    this.aiRole = fields.aiRole;\n    this.tools = fields.tools;\n    this.tokenCounter = fields.tokenCounter;\n    this.sendTokenLimit = fields.sendTokenLimit || 4196;\n  }\n\n  _getPromptType() {\n    return \"autogpt\" as const;\n  }\n\n  /**\n   * Constructs the full prompt based on the provided goals.\n   * @param goals An array of goals.\n   * @returns The full prompt as a string.\n   */\n  constructFullPrompt(goals: string[]): string {\n    const promptStart = `Your decisions must always be made independently\n            without seeking user assistance. Play to your strengths\n            as an LLM and pursue simple strategies with no legal complications.\n            If you have completed all your tasks,\n            make sure to use the \"finish\" command.`;\n\n    let fullPrompt = `You are ${this.aiName}, ${this.aiRole}\\n${promptStart}\\n\\nGOALS:\\n\\n`;\n    goals.forEach((goal, index) => {\n      fullPrompt += `${index + 1}. ${goal}\\n`;\n    });\n\n    fullPrompt += `\\n\\n${getPrompt(this.tools)}`;\n    return fullPrompt;\n  }\n\n  /**\n   * Formats the messages based on the provided parameters.\n   * @param goals An array of goals.\n   * @param memory A VectorStoreRetriever instance.\n   * @param messages An array of previous messages.\n   * @param user_input The user's input.\n   * @returns An array of formatted messages.\n   */\n  async formatMessages({\n    goals,\n    memory,\n    messages: previousMessages,\n    user_input,\n  }: {\n    goals: string[];\n    memory: VectorStoreRetrieverInterface;\n    messages: BaseMessage[];\n    user_input: string;\n  }) {\n    const basePrompt = new SystemMessage(this.constructFullPrompt(goals));\n    const timePrompt = new SystemMessage(\n      `The current time and date is ${new Date().toLocaleString(\"en-US\")}`\n    );\n    if (\n      typeof basePrompt.content !== \"string\" ||\n      typeof timePrompt.content !== \"string\"\n    ) {\n      throw new Error(\"Non-string message content is not supported.\");\n    }\n    const usedTokens =\n      (await this.tokenCounter(basePrompt.content)) +\n      (await this.tokenCounter(timePrompt.content));\n    const relevantDocs = await memory.invoke(\n      JSON.stringify(previousMessages.slice(-10))\n    );\n    const relevantMemory = relevantDocs.map(\n      (d: { pageContent: string }) => d.pageContent\n    );\n    let relevantMemoryTokens = await relevantMemory.reduce(\n      async (acc: Promise<number>, doc: string) =>\n        (await acc) + (await this.tokenCounter(doc)),\n      Promise.resolve(0)\n    );\n\n    while (usedTokens + relevantMemoryTokens > 2500) {\n      relevantMemory.pop();\n      relevantMemoryTokens = await relevantMemory.reduce(\n        async (acc: Promise<number>, doc: string) =>\n          (await acc) + (await this.tokenCounter(doc)),\n        Promise.resolve(0)\n      );\n    }\n\n    const contentFormat = `This reminds you of these events from your past:\\n${relevantMemory.join(\n      \"\\n\"\n    )}\\n\\n`;\n    const memoryMessage = new SystemMessage(contentFormat);\n    if (typeof memoryMessage.content !== \"string\") {\n      throw new Error(\"Non-string message content is not supported.\");\n    }\n    const usedTokensWithMemory =\n      usedTokens + (await this.tokenCounter(memoryMessage.content));\n    const historicalMessages: BaseMessage[] = [];\n\n    for (const message of previousMessages.slice(-10).reverse()) {\n      if (typeof message.content !== \"string\") {\n        throw new Error(\"Non-string message content is not supported.\");\n      }\n      const messageTokens = await this.tokenCounter(message.content);\n      if (usedTokensWithMemory + messageTokens > this.sendTokenLimit - 1000) {\n        break;\n      }\n      historicalMessages.unshift(message);\n    }\n\n    const inputMessage = new HumanMessage(user_input);\n    const messages: BaseMessage[] = [\n      basePrompt,\n      timePrompt,\n      memoryMessage,\n      ...historicalMessages,\n      inputMessage,\n    ];\n    return messages;\n  }\n\n  /**\n   * This method is not implemented in the AutoGPTPrompt class and will\n   * throw an error if called.\n   * @param _values Partial values.\n   * @returns Throws an error.\n   */\n  async partial(_values: PartialValues): Promise<BaseChatPromptTemplate> {\n    throw new Error(\"Method not implemented.\");\n  }\n\n  serialize(): SerializedBasePromptTemplate {\n    throw new Error(\"Method not implemented.\");\n  }\n}\n"],"mappings":";;;;;;;;;;;AA+BA,IAAa,gBAAb,cACU,uBAEV;CACE;CAEA;CAEA;CAEA;CAEA;CAEA,YAAYA,QAA4B;EACtC,MAAM,EAAE,gBAAgB;GAAC;GAAS;GAAU;GAAY;EAAa,EAAE,EAAC;EACxE,KAAK,SAAS,OAAO;EACrB,KAAK,SAAS,OAAO;EACrB,KAAK,QAAQ,OAAO;EACpB,KAAK,eAAe,OAAO;EAC3B,KAAK,iBAAiB,OAAO,kBAAkB;CAChD;CAED,iBAAiB;AACf,SAAO;CACR;;;;;;CAOD,oBAAoBC,OAAyB;EAC3C,MAAM,cAAc,CAAC;;;;kDAIyB,CAAC;EAE/C,IAAI,aAAa,CAAC,QAAQ,EAAE,KAAK,OAAO,EAAE,EAAE,KAAK,OAAO,EAAE,EAAE,YAAY,cAAc,CAAC;EACvF,MAAM,QAAQ,CAAC,MAAM,UAAU;GAC7B,cAAc,GAAG,QAAQ,EAAE,EAAE,EAAE,KAAK,EAAE,CAAC;EACxC,EAAC;EAEF,cAAc,CAAC,IAAI,EAAE,UAAU,KAAK,MAAM,EAAE;AAC5C,SAAO;CACR;;;;;;;;;CAUD,MAAM,eAAe,EACnB,OACA,QACA,UAAU,kBACV,YAMD,EAAE;EACD,MAAM,aAAa,IAAI,cAAc,KAAK,oBAAoB,MAAM;EACpE,MAAM,aAAa,IAAI,cACrB,CAAC,6BAA6B,mBAAE,IAAI,QAAO,eAAe,QAAQ,EAAE;AAEtE,MACE,OAAO,WAAW,YAAY,YAC9B,OAAO,WAAW,YAAY,SAE9B,OAAM,IAAI,MAAM;EAElB,MAAM,aACH,MAAM,KAAK,aAAa,WAAW,QAAQ,GAC3C,MAAM,KAAK,aAAa,WAAW,QAAQ;EAC9C,MAAM,eAAe,MAAM,OAAO,OAChC,KAAK,UAAU,iBAAiB,MAAM,IAAI,CAAC,CAC5C;EACD,MAAM,iBAAiB,aAAa,IAClC,CAACC,MAA+B,EAAE,YACnC;EACD,IAAI,uBAAuB,MAAM,eAAe,OAC9C,OAAOC,KAAsBC,QAC1B,MAAM,MAAQ,MAAM,KAAK,aAAa,IAAI,EAC7C,QAAQ,QAAQ,EAAE,CACnB;AAED,SAAO,aAAa,uBAAuB,MAAM;GAC/C,eAAe,KAAK;GACpB,uBAAuB,MAAM,eAAe,OAC1C,OAAOD,KAAsBC,QAC1B,MAAM,MAAQ,MAAM,KAAK,aAAa,IAAI,EAC7C,QAAQ,QAAQ,EAAE,CACnB;EACF;EAED,MAAM,gBAAgB,CAAC,kDAAkD,EAAE,eAAe,KACxF,KACD,CAAC,IAAI,CAAC;EACP,MAAM,gBAAgB,IAAI,cAAc;AACxC,MAAI,OAAO,cAAc,YAAY,SACnC,OAAM,IAAI,MAAM;EAElB,MAAM,uBACJ,aAAc,MAAM,KAAK,aAAa,cAAc,QAAQ;EAC9D,MAAMC,qBAAoC,CAAE;AAE5C,OAAK,MAAM,WAAW,iBAAiB,MAAM,IAAI,CAAC,SAAS,EAAE;AAC3D,OAAI,OAAO,QAAQ,YAAY,SAC7B,OAAM,IAAI,MAAM;GAElB,MAAM,gBAAgB,MAAM,KAAK,aAAa,QAAQ,QAAQ;AAC9D,OAAI,uBAAuB,gBAAgB,KAAK,iBAAiB,IAC/D;GAEF,mBAAmB,QAAQ,QAAQ;EACpC;EAED,MAAM,eAAe,IAAI,aAAa;EACtC,MAAMC,WAA0B;GAC9B;GACA;GACA;GACA,GAAG;GACH;EACD;AACD,SAAO;CACR;;;;;;;CAQD,MAAM,QAAQC,SAAyD;AACrE,QAAM,IAAI,MAAM;CACjB;CAED,YAA0C;AACxC,QAAM,IAAI,MAAM;CACjB;AACF"}
+{"version":3,"file":"prompt.js","names":[],"sources":["../../../src/experimental/autogpt/prompt.ts"],"sourcesContent":["import type { VectorStoreRetrieverInterface } from \"@langchain/core/vectorstores\";\nimport {\n  BaseChatPromptTemplate,\n  SerializedBasePromptTemplate,\n} from \"@langchain/core/prompts\";\nimport {\n  BaseMessage,\n  HumanMessage,\n  SystemMessage,\n} from \"@langchain/core/messages\";\nimport { PartialValues } from \"@langchain/core/utils/types\";\nimport { getPrompt } from \"./prompt_generator.js\";\nimport { ObjectTool } from \"./schema.js\";\n\n/**\n * Interface for the input parameters of the AutoGPTPrompt class.\n */\nexport interface AutoGPTPromptInput {\n  aiName: string;\n  aiRole: string;\n  tools: ObjectTool[];\n  tokenCounter: (text: string) => Promise<number>;\n  sendTokenLimit?: number;\n}\n\n/**\n * Class used to generate prompts for the AutoGPT model. It takes into\n * account the AI's name, role, tools, token counter, and send token\n * limit. The class also handles the formatting of messages and the\n * construction of the full prompt.\n */\nexport class AutoGPTPrompt\n  extends BaseChatPromptTemplate\n  implements AutoGPTPromptInput\n{\n  aiName: string;\n\n  aiRole: string;\n\n  tools: ObjectTool[];\n\n  tokenCounter: (text: string) => Promise<number>;\n\n  sendTokenLimit: number;\n\n  constructor(fields: AutoGPTPromptInput) {\n    super({ inputVariables: [\"goals\", \"memory\", \"messages\", \"user_input\"] });\n    this.aiName = fields.aiName;\n    this.aiRole = fields.aiRole;\n    this.tools = fields.tools;\n    this.tokenCounter = fields.tokenCounter;\n    this.sendTokenLimit = fields.sendTokenLimit || 4196;\n  }\n\n  _getPromptType() {\n    return \"autogpt\" as const;\n  }\n\n  /**\n   * Constructs the full prompt based on the provided goals.\n   * @param goals An array of goals.\n   * @returns The full prompt as a string.\n   */\n  constructFullPrompt(goals: string[]): string {\n    const promptStart = `Your decisions must always be made independently\n            without seeking user assistance. Play to your strengths\n            as an LLM and pursue simple strategies with no legal complications.\n            If you have completed all your tasks,\n            make sure to use the \"finish\" command.`;\n\n    let fullPrompt = `You are ${this.aiName}, ${this.aiRole}\\n${promptStart}\\n\\nGOALS:\\n\\n`;\n    goals.forEach((goal, index) => {\n      fullPrompt += `${index + 1}. ${goal}\\n`;\n    });\n\n    fullPrompt += `\\n\\n${getPrompt(this.tools)}`;\n    return fullPrompt;\n  }\n\n  /**\n   * Formats the messages based on the provided parameters.\n   * @param goals An array of goals.\n   * @param memory A VectorStoreRetriever instance.\n   * @param messages An array of previous messages.\n   * @param user_input The user's input.\n   * @returns An array of formatted messages.\n   */\n  async formatMessages({\n    goals,\n    memory,\n    messages: previousMessages,\n    user_input,\n  }: {\n    goals: string[];\n    memory: VectorStoreRetrieverInterface;\n    messages: BaseMessage[];\n    user_input: string;\n  }) {\n    const basePrompt = new SystemMessage(this.constructFullPrompt(goals));\n    const timePrompt = new SystemMessage(\n      `The current time and date is ${new Date().toLocaleString(\"en-US\")}`\n    );\n    if (\n      typeof basePrompt.content !== \"string\" ||\n      typeof timePrompt.content !== \"string\"\n    ) {\n      throw new Error(\"Non-string message content is not supported.\");\n    }\n    const usedTokens =\n      (await this.tokenCounter(basePrompt.content)) +\n      (await this.tokenCounter(timePrompt.content));\n    const relevantDocs = await memory.invoke(\n      JSON.stringify(previousMessages.slice(-10))\n    );\n    const relevantMemory = relevantDocs.map(\n      (d: { pageContent: string }) => d.pageContent\n    );\n    let relevantMemoryTokens = await relevantMemory.reduce(\n      async (acc: Promise<number>, doc: string) =>\n        (await acc) + (await this.tokenCounter(doc)),\n      Promise.resolve(0)\n    );\n\n    while (usedTokens + relevantMemoryTokens > 2500) {\n      relevantMemory.pop();\n      relevantMemoryTokens = await relevantMemory.reduce(\n        async (acc: Promise<number>, doc: string) =>\n          (await acc) + (await this.tokenCounter(doc)),\n        Promise.resolve(0)\n      );\n    }\n\n    const contentFormat = `This reminds you of these events from your past:\\n${relevantMemory.join(\n      \"\\n\"\n    )}\\n\\n`;\n    const memoryMessage = new SystemMessage(contentFormat);\n    if (typeof memoryMessage.content !== \"string\") {\n      throw new Error(\"Non-string message content is not supported.\");\n    }\n    const usedTokensWithMemory =\n      usedTokens + (await this.tokenCounter(memoryMessage.content));\n    const historicalMessages: BaseMessage[] = [];\n\n    for (const message of previousMessages.slice(-10).reverse()) {\n      if (typeof message.content !== \"string\") {\n        throw new Error(\"Non-string message content is not supported.\");\n      }\n      const messageTokens = await this.tokenCounter(message.content);\n      if (usedTokensWithMemory + messageTokens > this.sendTokenLimit - 1000) {\n        break;\n      }\n      historicalMessages.unshift(message);\n    }\n\n    const inputMessage = new HumanMessage(user_input);\n    const messages: BaseMessage[] = [\n      basePrompt,\n      timePrompt,\n      memoryMessage,\n      ...historicalMessages,\n      inputMessage,\n    ];\n    return messages;\n  }\n\n  /**\n   * This method is not implemented in the AutoGPTPrompt class and will\n   * throw an error if called.\n   * @param _values Partial values.\n   * @returns Throws an error.\n   */\n  async partial(_values: PartialValues): Promise<BaseChatPromptTemplate> {\n    throw new Error(\"Method not implemented.\");\n  }\n\n  serialize(): SerializedBasePromptTemplate {\n    throw new Error(\"Method not implemented.\");\n  }\n}\n"],"mappings":";;;;;;;;;;;AA+BA,IAAa,gBAAb,cACU,uBAEV;CACE;CAEA;CAEA;CAEA;CAEA;CAEA,YAAY,QAA4B;AACtC,QAAM,EAAE,gBAAgB;GAAC;GAAS;GAAU;GAAY;GAAa,EAAE,CAAC;AACxE,OAAK,SAAS,OAAO;AACrB,OAAK,SAAS,OAAO;AACrB,OAAK,QAAQ,OAAO;AACpB,OAAK,eAAe,OAAO;AAC3B,OAAK,iBAAiB,OAAO,kBAAkB;;CAGjD,iBAAiB;AACf,SAAO;;;;;;;CAQT,oBAAoB,OAAyB;EAO3C,IAAI,aAAa,WAAW,KAAK,OAAO,IAAI,KAAK,OAAO;;;;;AACxD,QAAM,SAAS,MAAM,UAAU;AAC7B,iBAAc,GAAG,QAAQ,EAAE,IAAI,KAAK;IACpC;AAEF,gBAAc,OAAO,UAAU,KAAK,MAAM;AAC1C,SAAO;;;;;;;;;;CAWT,MAAM,eAAe,EACnB,OACA,QACA,UAAU,kBACV,cAMC;EACD,MAAM,aAAa,IAAI,cAAc,KAAK,oBAAoB,MAAM,CAAC;EACrE,MAAM,aAAa,IAAI,cACrB,iDAAgC,IAAI,MAAM,EAAC,eAAe,QAAQ,GACnE;AACD,MACE,OAAO,WAAW,YAAY,YAC9B,OAAO,WAAW,YAAY,SAE9B,OAAM,IAAI,MAAM,+CAA+C;EAEjE,MAAM,aACH,MAAM,KAAK,aAAa,WAAW,QAAQ,GAC3C,MAAM,KAAK,aAAa,WAAW,QAAQ;EAI9C,MAAM,kBAHe,MAAM,OAAO,OAChC,KAAK,UAAU,iBAAiB,MAAM,IAAI,CAAC,CAC5C,EACmC,KACjC,MAA+B,EAAE,YACnC;EACD,IAAI,uBAAuB,MAAM,eAAe,OAC9C,OAAO,KAAsB,QAC1B,MAAM,MAAQ,MAAM,KAAK,aAAa,IAAI,EAC7C,QAAQ,QAAQ,EAAE,CACnB;AAED,SAAO,aAAa,uBAAuB,MAAM;AAC/C,kBAAe,KAAK;AACpB,0BAAuB,MAAM,eAAe,OAC1C,OAAO,KAAsB,QAC1B,MAAM,MAAQ,MAAM,KAAK,aAAa,IAAI,EAC7C,QAAQ,QAAQ,EAAE,CACnB;;EAMH,MAAM,gBAAgB,IAAI,cAHJ,qDAAqD,eAAe,KACxF,KACD,CAAC,MACoD;AACtD,MAAI,OAAO,cAAc,YAAY,SACnC,OAAM,IAAI,MAAM,+CAA+C;EAEjE,MAAM,uBACJ,aAAc,MAAM,KAAK,aAAa,cAAc,QAAQ;EAC9D,MAAM,qBAAoC,EAAE;AAE5C,OAAK,MAAM,WAAW,iBAAiB,MAAM,IAAI,CAAC,SAAS,EAAE;AAC3D,OAAI,OAAO,QAAQ,YAAY,SAC7B,OAAM,IAAI,MAAM,+CAA+C;AAGjE,OAAI,uBADkB,MAAM,KAAK,aAAa,QAAQ,QAAQ,GACnB,KAAK,iBAAiB,IAC/D;AAEF,sBAAmB,QAAQ,QAAQ;;EAGrC,MAAM,eAAe,IAAI,aAAa,WAAW;AAQjD,SAPgC;GAC9B;GACA;GACA;GACA,GAAG;GACH;GACD;;;;;;;;CAUH,MAAM,QAAQ,SAAyD;AACrE,QAAM,IAAI,MAAM,0BAA0B;;CAG5C,YAA0C;AACxC,QAAM,IAAI,MAAM,0BAA0B"}

package/dist/experimental/autogpt/prompt_generator.cjs

@@ -1,7 +1,7 @@
-const require_rolldown_runtime = require('../../_virtual/rolldown_runtime.cjs');
+const require_runtime = require('../../_virtual/_rolldown/runtime.cjs');
 const require_schema = require('./schema.cjs');
-const __langchain_core_utils_types = require_rolldown_runtime.__toESM(require("@langchain/core/utils/types"));
-const __langchain_core_utils_json_schema = require_rolldown_runtime.__toESM(require("@langchain/core/utils/json_schema"));
+let _langchain_core_utils_types = require("@langchain/core/utils/types");
+let _langchain_core_utils_json_schema = require("@langchain/core/utils/json_schema");
 
 //#region src/experimental/autogpt/prompt_generator.ts
 /**
@@ -51,7 +51,7 @@ var PromptGenerator = class {
 	}
 	_generate_command_string(tool) {
 		let output = `"${tool.name}": ${tool.description}`;
-		const jsonSchema = (0, __langchain_core_utils_types.isInteropZodSchema)(tool.schema) ? (0, __langchain_core_utils_json_schema.toJsonSchema)(tool.schema) : tool.schema;
+		const jsonSchema = (0, _langchain_core_utils_types.isInteropZodSchema)(tool.schema) ? (0, _langchain_core_utils_json_schema.toJsonSchema)(tool.schema) : tool.schema;
 		output += `, args json schema: ${JSON.stringify(jsonSchema?.properties)}`;
 		return output;
 	}
@@ -74,9 +74,7 @@ var PromptGenerator = class {
 	_generate_numbered_list(items, item_type = "list") {
 		if (item_type === "command") {
 			const command_strings = items.map((item, i) => `${i + 1}. ${this._generate_command_string(item)}`);
-			const finish_description = "use this to signal that you have finished all your objectives";
-			const finish_args = "\"response\": \"final response to let people know you have finished your objectives\"";
-			const finish_string = `${items.length + 1}. ${require_schema.FINISH_NAME}: ${finish_description}, args: ${finish_args}`;
+			const finish_string = `${items.length + 1}. ${require_schema.FINISH_NAME}: use this to signal that you have finished all your objectives, args: "response": "final response to let people know you have finished your objectives"`;
 			return command_strings.concat([finish_string]).join("\n");
 		}
 		return items.map((item, i) => `${i + 1}. ${item}`).join("\n");
@@ -88,8 +86,7 @@ var PromptGenerator = class {
 	*/
 	generate_prompt_string() {
 		const formatted_response_format = JSON.stringify(this.response_format, null, 4);
-		const prompt_string = `Constraints:\n${this._generate_numbered_list(this.constraints)}\n\nCommands:\n${this._generate_numbered_list(this.commands, "command")}\n\nResources:\n${this._generate_numbered_list(this.resources)}\n\nPerformance Evaluation:\n${this._generate_numbered_list(this.performance_evaluation)}\n\nYou should only respond in JSON format as described below \nResponse Format: \n${formatted_response_format} \nEnsure the response can be parsed by Python json.loads`;
-		return prompt_string;
+		return `Constraints:\n${this._generate_numbered_list(this.constraints)}\n\nCommands:\n${this._generate_numbered_list(this.commands, "command")}\n\nResources:\n${this._generate_numbered_list(this.resources)}\n\nPerformance Evaluation:\n${this._generate_numbered_list(this.performance_evaluation)}\n\nYou should only respond in JSON format as described below \nResponse Format: \n${formatted_response_format} \nEnsure the response can be parsed by Python json.loads`;
 	}
 };
 /**
@@ -110,8 +107,7 @@ function getPrompt(tools) {
 	prompt_generator.add_performance_evaluation("Constructively self-criticize your big-picture behavior constantly.");
 	prompt_generator.add_performance_evaluation("Reflect on past decisions and strategies to refine your approach.");
 	prompt_generator.add_performance_evaluation("Every command has a cost, so be smart and efficient. Aim to complete tasks in the least number of steps.");
-	const prompt_string = prompt_generator.generate_prompt_string();
-	return prompt_string;
+	return prompt_generator.generate_prompt_string();
 }
 
 //#endregion

package/dist/experimental/autogpt/prompt_generator.cjs.map

@@ -1 +1 @@
-{"version":3,"file":"prompt_generator.cjs","names":["constraint: string","tool: ObjectTool","resource: string","evaluation: string","items: any[]","FINISH_NAME","tools: ObjectTool[]"],"sources":["../../../src/experimental/autogpt/prompt_generator.ts"],"sourcesContent":["import { isInteropZodSchema } from \"@langchain/core/utils/types\";\nimport {\n  type JsonSchema7Type,\n  toJsonSchema,\n} from \"@langchain/core/utils/json_schema\";\nimport { ObjectTool, FINISH_NAME } from \"./schema.js\";\n\n/**\n * Class that generates prompts for generative agents. It maintains a list\n * of constraints, commands, resources, and performance evaluations.\n */\nexport class PromptGenerator {\n  constraints: string[];\n\n  commands: ObjectTool[];\n\n  resources: string[];\n\n  performance_evaluation: string[];\n\n  response_format: object;\n\n  constructor() {\n    this.constraints = [];\n    this.commands = [];\n    this.resources = [];\n    this.performance_evaluation = [];\n    this.response_format = {\n      thoughts: {\n        text: \"thought\",\n        reasoning: \"reasoning\",\n        plan: \"- short bulleted\\n- list that conveys\\n- long-term plan\",\n        criticism: \"constructive self-criticism\",\n        speak: \"thoughts summary to say to user\",\n      },\n      command: { name: \"command name\", args: { \"arg name\": \"value\" } },\n    };\n  }\n\n  /**\n   * Adds a constraint to the list of constraints.\n   * @param constraint The constraint to add.\n   * @returns void\n   */\n  add_constraint(constraint: string): void {\n    this.constraints.push(constraint);\n  }\n\n  /**\n   * Adds a tool to the list of commands.\n   * @param tool The tool to add.\n   * @returns void\n   */\n  add_tool(tool: ObjectTool): void {\n    this.commands.push(tool);\n  }\n\n  _generate_command_string(tool: ObjectTool): string {\n    let output = `\"${tool.name}\": ${tool.description}`;\n    const jsonSchema = (\n      isInteropZodSchema(tool.schema) ? toJsonSchema(tool.schema) : tool.schema\n    ) as { properties?: Record<string, JsonSchema7Type> } | undefined;\n    output += `, args json schema: ${JSON.stringify(jsonSchema?.properties)}`;\n    return output;\n  }\n\n  /**\n   * Adds a resource to the list of resources.\n   * @param resource The resource to add.\n   * @returns void\n   */\n  add_resource(resource: string): void {\n    this.resources.push(resource);\n  }\n\n  /**\n   * Adds a performance evaluation to the list of performance evaluations.\n   * @param evaluation The performance evaluation to add.\n   * @returns void\n   */\n  add_performance_evaluation(evaluation: string): void {\n    this.performance_evaluation.push(evaluation);\n  }\n\n  // eslint-disable-next-line @typescript-eslint/no-explicit-any\n  _generate_numbered_list(items: any[], item_type = \"list\"): string {\n    if (item_type === \"command\") {\n      const command_strings = items.map(\n        (item, i) => `${i + 1}. ${this._generate_command_string(item)}`\n      );\n      const finish_description =\n        \"use this to signal that you have finished all your objectives\";\n      const finish_args =\n        '\"response\": \"final response to let people know you have finished your objectives\"';\n      const finish_string = `${\n        items.length + 1\n      }. ${FINISH_NAME}: ${finish_description}, args: ${finish_args}`;\n      return command_strings.concat([finish_string]).join(\"\\n\");\n    }\n\n    return items.map((item, i) => `${i + 1}. ${item}`).join(\"\\n\");\n  }\n\n  /**\n   * Generates a prompt string that includes the constraints, commands,\n   * resources, performance evaluations, and response format.\n   * @returns A string representing the prompt.\n   */\n  generate_prompt_string(): string {\n    const formatted_response_format = JSON.stringify(\n      this.response_format,\n      null,\n      4\n    );\n    const prompt_string =\n      `Constraints:\\n${this._generate_numbered_list(this.constraints)}\\n\\n` +\n      `Commands:\\n${this._generate_numbered_list(\n        this.commands,\n        \"command\"\n      )}\\n\\n` +\n      `Resources:\\n${this._generate_numbered_list(this.resources)}\\n\\n` +\n      `Performance Evaluation:\\n${this._generate_numbered_list(\n        this.performance_evaluation\n      )}\\n\\n` +\n      `You should only respond in JSON format as described below ` +\n      `\\nResponse Format: \\n${formatted_response_format} ` +\n      `\\nEnsure the response can be parsed by Python json.loads`;\n\n    return prompt_string;\n  }\n}\n\n/**\n * Function that generates a prompt string for a given list of tools.\n */\nexport function getPrompt(tools: ObjectTool[]): string {\n  const prompt_generator = new PromptGenerator();\n\n  prompt_generator.add_constraint(\n    \"~4000 word limit for short term memory. \" +\n      \"Your short term memory is short, \" +\n      \"so immediately save important information to files.\"\n  );\n  prompt_generator.add_constraint(\n    \"If you are unsure how you previously did something \" +\n      \"or want to recall past events, \" +\n      \"thinking about similar events will help you remember.\"\n  );\n  prompt_generator.add_constraint(\"No user assistance\");\n  prompt_generator.add_constraint(\n    'Exclusively use the commands listed in double quotes e.g. \"command name\"'\n  );\n\n  for (const tool of tools) {\n    prompt_generator.add_tool(tool);\n  }\n\n  prompt_generator.add_resource(\n    \"Internet access for searches and information gathering.\"\n  );\n  prompt_generator.add_resource(\"Long Term memory management.\");\n  prompt_generator.add_resource(\n    \"GPT-3.5 powered Agents for delegation of simple tasks.\"\n  );\n  prompt_generator.add_resource(\"File output.\");\n\n  prompt_generator.add_performance_evaluation(\n    \"Continuously review and analyze your actions \" +\n      \"to ensure you are performing to the best of your abilities.\"\n  );\n  prompt_generator.add_performance_evaluation(\n    \"Constructively self-criticize your big-picture behavior constantly.\"\n  );\n  prompt_generator.add_performance_evaluation(\n    \"Reflect on past decisions and strategies to refine your approach.\"\n  );\n  prompt_generator.add_performance_evaluation(\n    \"Every command has a cost, so be smart and efficient. \" +\n      \"Aim to complete tasks in the least number of steps.\"\n  );\n\n  const prompt_string = prompt_generator.generate_prompt_string();\n\n  return prompt_string;\n}\n"],"mappings":";;;;;;;;;;AAWA,IAAa,kBAAb,MAA6B;CAC3B;CAEA;CAEA;CAEA;CAEA;CAEA,cAAc;EACZ,KAAK,cAAc,CAAE;EACrB,KAAK,WAAW,CAAE;EAClB,KAAK,YAAY,CAAE;EACnB,KAAK,yBAAyB,CAAE;EAChC,KAAK,kBAAkB;GACrB,UAAU;IACR,MAAM;IACN,WAAW;IACX,MAAM;IACN,WAAW;IACX,OAAO;GACR;GACD,SAAS;IAAE,MAAM;IAAgB,MAAM,EAAE,YAAY,QAAS;GAAE;EACjE;CACF;;;;;;CAOD,eAAeA,YAA0B;EACvC,KAAK,YAAY,KAAK,WAAW;CAClC;;;;;;CAOD,SAASC,MAAwB;EAC/B,KAAK,SAAS,KAAK,KAAK;CACzB;CAED,yBAAyBA,MAA0B;EACjD,IAAI,SAAS,CAAC,CAAC,EAAE,KAAK,KAAK,GAAG,EAAE,KAAK,aAAa;EAClD,MAAM,kEACe,KAAK,OAAO,wDAAgB,KAAK,OAAO,GAAG,KAAK;EAErE,UAAU,CAAC,oBAAoB,EAAE,KAAK,UAAU,YAAY,WAAW,EAAE;AACzE,SAAO;CACR;;;;;;CAOD,aAAaC,UAAwB;EACnC,KAAK,UAAU,KAAK,SAAS;CAC9B;;;;;;CAOD,2BAA2BC,YAA0B;EACnD,KAAK,uBAAuB,KAAK,WAAW;CAC7C;CAGD,wBAAwBC,OAAc,YAAY,QAAgB;AAChE,MAAI,cAAc,WAAW;GAC3B,MAAM,kBAAkB,MAAM,IAC5B,CAAC,MAAM,MAAM,GAAG,IAAI,EAAE,EAAE,EAAE,KAAK,yBAAyB,KAAK,EAAE,CAChE;GACD,MAAM,qBACJ;GACF,MAAM,cACJ;GACF,MAAM,gBAAgB,GACpB,MAAM,SAAS,EAChB,EAAE,EAAEC,2BAAY,EAAE,EAAE,mBAAmB,QAAQ,EAAE,aAAa;AAC/D,UAAO,gBAAgB,OAAO,CAAC,aAAc,EAAC,CAAC,KAAK,KAAK;EAC1D;AAED,SAAO,MAAM,IAAI,CAAC,MAAM,MAAM,GAAG,IAAI,EAAE,EAAE,EAAE,MAAM,CAAC,CAAC,KAAK,KAAK;CAC9D;;;;;;CAOD,yBAAiC;EAC/B,MAAM,4BAA4B,KAAK,UACrC,KAAK,iBACL,MACA,EACD;EACD,MAAM,gBACJ,CAAC,cAAc,EAAE,KAAK,wBAAwB,KAAK,YAAY,CAAC,eAAI,EACtD,KAAK,wBACjB,KAAK,UACL,UACD,CAAC,gBAAI,EACS,KAAK,wBAAwB,KAAK,UAAU,CAAC,6BAAI,EACpC,KAAK,wBAC/B,KAAK,uBACN,CAAC,mFAAI,EAEkB,0BAA0B,yDAAC,CACO;AAE5D,SAAO;CACR;AACF;;;;AAKD,SAAgB,UAAUC,OAA6B;CACrD,MAAM,mBAAmB,IAAI;CAE7B,iBAAiB,eACf,+HAGD;CACD,iBAAiB,eACf,0IAGD;CACD,iBAAiB,eAAe,qBAAqB;CACrD,iBAAiB,eACf,6EACD;AAED,MAAK,MAAM,QAAQ,OACjB,iBAAiB,SAAS,KAAK;CAGjC,iBAAiB,aACf,0DACD;CACD,iBAAiB,aAAa,+BAA+B;CAC7D,iBAAiB,aACf,yDACD;CACD,iBAAiB,aAAa,eAAe;CAE7C,iBAAiB,2BACf,2GAED;CACD,iBAAiB,2BACf,sEACD;CACD,iBAAiB,2BACf,oEACD;CACD,iBAAiB,2BACf,2GAED;CAED,MAAM,gBAAgB,iBAAiB,wBAAwB;AAE/D,QAAO;AACR"}
+{"version":3,"file":"prompt_generator.cjs","names":["FINISH_NAME"],"sources":["../../../src/experimental/autogpt/prompt_generator.ts"],"sourcesContent":["import { isInteropZodSchema } from \"@langchain/core/utils/types\";\nimport {\n  type JsonSchema7Type,\n  toJsonSchema,\n} from \"@langchain/core/utils/json_schema\";\nimport { ObjectTool, FINISH_NAME } from \"./schema.js\";\n\n/**\n * Class that generates prompts for generative agents. It maintains a list\n * of constraints, commands, resources, and performance evaluations.\n */\nexport class PromptGenerator {\n  constraints: string[];\n\n  commands: ObjectTool[];\n\n  resources: string[];\n\n  performance_evaluation: string[];\n\n  response_format: object;\n\n  constructor() {\n    this.constraints = [];\n    this.commands = [];\n    this.resources = [];\n    this.performance_evaluation = [];\n    this.response_format = {\n      thoughts: {\n        text: \"thought\",\n        reasoning: \"reasoning\",\n        plan: \"- short bulleted\\n- list that conveys\\n- long-term plan\",\n        criticism: \"constructive self-criticism\",\n        speak: \"thoughts summary to say to user\",\n      },\n      command: { name: \"command name\", args: { \"arg name\": \"value\" } },\n    };\n  }\n\n  /**\n   * Adds a constraint to the list of constraints.\n   * @param constraint The constraint to add.\n   * @returns void\n   */\n  add_constraint(constraint: string): void {\n    this.constraints.push(constraint);\n  }\n\n  /**\n   * Adds a tool to the list of commands.\n   * @param tool The tool to add.\n   * @returns void\n   */\n  add_tool(tool: ObjectTool): void {\n    this.commands.push(tool);\n  }\n\n  _generate_command_string(tool: ObjectTool): string {\n    let output = `\"${tool.name}\": ${tool.description}`;\n    const jsonSchema = (\n      isInteropZodSchema(tool.schema) ? toJsonSchema(tool.schema) : tool.schema\n    ) as { properties?: Record<string, JsonSchema7Type> } | undefined;\n    output += `, args json schema: ${JSON.stringify(jsonSchema?.properties)}`;\n    return output;\n  }\n\n  /**\n   * Adds a resource to the list of resources.\n   * @param resource The resource to add.\n   * @returns void\n   */\n  add_resource(resource: string): void {\n    this.resources.push(resource);\n  }\n\n  /**\n   * Adds a performance evaluation to the list of performance evaluations.\n   * @param evaluation The performance evaluation to add.\n   * @returns void\n   */\n  add_performance_evaluation(evaluation: string): void {\n    this.performance_evaluation.push(evaluation);\n  }\n\n  // eslint-disable-next-line @typescript-eslint/no-explicit-any\n  _generate_numbered_list(items: any[], item_type = \"list\"): string {\n    if (item_type === \"command\") {\n      const command_strings = items.map(\n        (item, i) => `${i + 1}. ${this._generate_command_string(item)}`\n      );\n      const finish_description =\n        \"use this to signal that you have finished all your objectives\";\n      const finish_args =\n        '\"response\": \"final response to let people know you have finished your objectives\"';\n      const finish_string = `${\n        items.length + 1\n      }. ${FINISH_NAME}: ${finish_description}, args: ${finish_args}`;\n      return command_strings.concat([finish_string]).join(\"\\n\");\n    }\n\n    return items.map((item, i) => `${i + 1}. ${item}`).join(\"\\n\");\n  }\n\n  /**\n   * Generates a prompt string that includes the constraints, commands,\n   * resources, performance evaluations, and response format.\n   * @returns A string representing the prompt.\n   */\n  generate_prompt_string(): string {\n    const formatted_response_format = JSON.stringify(\n      this.response_format,\n      null,\n      4\n    );\n    const prompt_string =\n      `Constraints:\\n${this._generate_numbered_list(this.constraints)}\\n\\n` +\n      `Commands:\\n${this._generate_numbered_list(\n        this.commands,\n        \"command\"\n      )}\\n\\n` +\n      `Resources:\\n${this._generate_numbered_list(this.resources)}\\n\\n` +\n      `Performance Evaluation:\\n${this._generate_numbered_list(\n        this.performance_evaluation\n      )}\\n\\n` +\n      `You should only respond in JSON format as described below ` +\n      `\\nResponse Format: \\n${formatted_response_format} ` +\n      `\\nEnsure the response can be parsed by Python json.loads`;\n\n    return prompt_string;\n  }\n}\n\n/**\n * Function that generates a prompt string for a given list of tools.\n */\nexport function getPrompt(tools: ObjectTool[]): string {\n  const prompt_generator = new PromptGenerator();\n\n  prompt_generator.add_constraint(\n    \"~4000 word limit for short term memory. \" +\n      \"Your short term memory is short, \" +\n      \"so immediately save important information to files.\"\n  );\n  prompt_generator.add_constraint(\n    \"If you are unsure how you previously did something \" +\n      \"or want to recall past events, \" +\n      \"thinking about similar events will help you remember.\"\n  );\n  prompt_generator.add_constraint(\"No user assistance\");\n  prompt_generator.add_constraint(\n    'Exclusively use the commands listed in double quotes e.g. \"command name\"'\n  );\n\n  for (const tool of tools) {\n    prompt_generator.add_tool(tool);\n  }\n\n  prompt_generator.add_resource(\n    \"Internet access for searches and information gathering.\"\n  );\n  prompt_generator.add_resource(\"Long Term memory management.\");\n  prompt_generator.add_resource(\n    \"GPT-3.5 powered Agents for delegation of simple tasks.\"\n  );\n  prompt_generator.add_resource(\"File output.\");\n\n  prompt_generator.add_performance_evaluation(\n    \"Continuously review and analyze your actions \" +\n      \"to ensure you are performing to the best of your abilities.\"\n  );\n  prompt_generator.add_performance_evaluation(\n    \"Constructively self-criticize your big-picture behavior constantly.\"\n  );\n  prompt_generator.add_performance_evaluation(\n    \"Reflect on past decisions and strategies to refine your approach.\"\n  );\n  prompt_generator.add_performance_evaluation(\n    \"Every command has a cost, so be smart and efficient. \" +\n      \"Aim to complete tasks in the least number of steps.\"\n  );\n\n  const prompt_string = prompt_generator.generate_prompt_string();\n\n  return prompt_string;\n}\n"],"mappings":";;;;;;;;;;AAWA,IAAa,kBAAb,MAA6B;CAC3B;CAEA;CAEA;CAEA;CAEA;CAEA,cAAc;AACZ,OAAK,cAAc,EAAE;AACrB,OAAK,WAAW,EAAE;AAClB,OAAK,YAAY,EAAE;AACnB,OAAK,yBAAyB,EAAE;AAChC,OAAK,kBAAkB;GACrB,UAAU;IACR,MAAM;IACN,WAAW;IACX,MAAM;IACN,WAAW;IACX,OAAO;IACR;GACD,SAAS;IAAE,MAAM;IAAgB,MAAM,EAAE,YAAY,SAAS;IAAE;GACjE;;;;;;;CAQH,eAAe,YAA0B;AACvC,OAAK,YAAY,KAAK,WAAW;;;;;;;CAQnC,SAAS,MAAwB;AAC/B,OAAK,SAAS,KAAK,KAAK;;CAG1B,yBAAyB,MAA0B;EACjD,IAAI,SAAS,IAAI,KAAK,KAAK,KAAK,KAAK;EACrC,MAAM,iEACe,KAAK,OAAO,uDAAgB,KAAK,OAAO,GAAG,KAAK;AAErE,YAAU,uBAAuB,KAAK,UAAU,YAAY,WAAW;AACvE,SAAO;;;;;;;CAQT,aAAa,UAAwB;AACnC,OAAK,UAAU,KAAK,SAAS;;;;;;;CAQ/B,2BAA2B,YAA0B;AACnD,OAAK,uBAAuB,KAAK,WAAW;;CAI9C,wBAAwB,OAAc,YAAY,QAAgB;AAChE,MAAI,cAAc,WAAW;GAC3B,MAAM,kBAAkB,MAAM,KAC3B,MAAM,MAAM,GAAG,IAAI,EAAE,IAAI,KAAK,yBAAyB,KAAK,GAC9D;GAKD,MAAM,gBAAgB,GACpB,MAAM,SAAS,EAChB,IAAIA,2BAAY;AACjB,UAAO,gBAAgB,OAAO,CAAC,cAAc,CAAC,CAAC,KAAK,KAAK;;AAG3D,SAAO,MAAM,KAAK,MAAM,MAAM,GAAG,IAAI,EAAE,IAAI,OAAO,CAAC,KAAK,KAAK;;;;;;;CAQ/D,yBAAiC;EAC/B,MAAM,4BAA4B,KAAK,UACrC,KAAK,iBACL,MACA,EACD;AAeD,SAbE,iBAAiB,KAAK,wBAAwB,KAAK,YAAY,CAAC,iBAClD,KAAK,wBACjB,KAAK,UACL,UACD,CAAC,kBACa,KAAK,wBAAwB,KAAK,UAAU,CAAC,+BAChC,KAAK,wBAC/B,KAAK,uBACN,CAAC,qFAEsB,0BAA0B;;;;;;AAUxD,SAAgB,UAAU,OAA6B;CACrD,MAAM,mBAAmB,IAAI,iBAAiB;AAE9C,kBAAiB,eACf,+HAGD;AACD,kBAAiB,eACf,0IAGD;AACD,kBAAiB,eAAe,qBAAqB;AACrD,kBAAiB,eACf,6EACD;AAED,MAAK,MAAM,QAAQ,MACjB,kBAAiB,SAAS,KAAK;AAGjC,kBAAiB,aACf,0DACD;AACD,kBAAiB,aAAa,+BAA+B;AAC7D,kBAAiB,aACf,yDACD;AACD,kBAAiB,aAAa,eAAe;AAE7C,kBAAiB,2BACf,2GAED;AACD,kBAAiB,2BACf,sEACD;AACD,kBAAiB,2BACf,oEACD;AACD,kBAAiB,2BACf,2GAED;AAID,QAFsB,iBAAiB,wBAAwB"}

package/dist/experimental/autogpt/prompt_generator.js

@@ -73,9 +73,7 @@ var PromptGenerator = class {
 	_generate_numbered_list(items, item_type = "list") {
 		if (item_type === "command") {
 			const command_strings = items.map((item, i) => `${i + 1}. ${this._generate_command_string(item)}`);
-			const finish_description = "use this to signal that you have finished all your objectives";
-			const finish_args = "\"response\": \"final response to let people know you have finished your objectives\"";
-			const finish_string = `${items.length + 1}. ${FINISH_NAME}: ${finish_description}, args: ${finish_args}`;
+			const finish_string = `${items.length + 1}. ${FINISH_NAME}: use this to signal that you have finished all your objectives, args: "response": "final response to let people know you have finished your objectives"`;
 			return command_strings.concat([finish_string]).join("\n");
 		}
 		return items.map((item, i) => `${i + 1}. ${item}`).join("\n");
@@ -87,8 +85,7 @@ var PromptGenerator = class {
 	*/
 	generate_prompt_string() {
 		const formatted_response_format = JSON.stringify(this.response_format, null, 4);
-		const prompt_string = `Constraints:\n${this._generate_numbered_list(this.constraints)}\n\nCommands:\n${this._generate_numbered_list(this.commands, "command")}\n\nResources:\n${this._generate_numbered_list(this.resources)}\n\nPerformance Evaluation:\n${this._generate_numbered_list(this.performance_evaluation)}\n\nYou should only respond in JSON format as described below \nResponse Format: \n${formatted_response_format} \nEnsure the response can be parsed by Python json.loads`;
-		return prompt_string;
+		return `Constraints:\n${this._generate_numbered_list(this.constraints)}\n\nCommands:\n${this._generate_numbered_list(this.commands, "command")}\n\nResources:\n${this._generate_numbered_list(this.resources)}\n\nPerformance Evaluation:\n${this._generate_numbered_list(this.performance_evaluation)}\n\nYou should only respond in JSON format as described below \nResponse Format: \n${formatted_response_format} \nEnsure the response can be parsed by Python json.loads`;
 	}
 };
 /**
@@ -109,8 +106,7 @@ function getPrompt(tools) {
 	prompt_generator.add_performance_evaluation("Constructively self-criticize your big-picture behavior constantly.");
 	prompt_generator.add_performance_evaluation("Reflect on past decisions and strategies to refine your approach.");
 	prompt_generator.add_performance_evaluation("Every command has a cost, so be smart and efficient. Aim to complete tasks in the least number of steps.");
-	const prompt_string = prompt_generator.generate_prompt_string();
-	return prompt_string;
+	return prompt_generator.generate_prompt_string();
 }
 
 //#endregion

package/dist/experimental/autogpt/prompt_generator.js.map

@@ -1 +1 @@
-{"version":3,"file":"prompt_generator.js","names":["constraint: string","tool: ObjectTool","resource: string","evaluation: string","items: any[]","tools: ObjectTool[]"],"sources":["../../../src/experimental/autogpt/prompt_generator.ts"],"sourcesContent":["import { isInteropZodSchema } from \"@langchain/core/utils/types\";\nimport {\n  type JsonSchema7Type,\n  toJsonSchema,\n} from \"@langchain/core/utils/json_schema\";\nimport { ObjectTool, FINISH_NAME } from \"./schema.js\";\n\n/**\n * Class that generates prompts for generative agents. It maintains a list\n * of constraints, commands, resources, and performance evaluations.\n */\nexport class PromptGenerator {\n  constraints: string[];\n\n  commands: ObjectTool[];\n\n  resources: string[];\n\n  performance_evaluation: string[];\n\n  response_format: object;\n\n  constructor() {\n    this.constraints = [];\n    this.commands = [];\n    this.resources = [];\n    this.performance_evaluation = [];\n    this.response_format = {\n      thoughts: {\n        text: \"thought\",\n        reasoning: \"reasoning\",\n        plan: \"- short bulleted\\n- list that conveys\\n- long-term plan\",\n        criticism: \"constructive self-criticism\",\n        speak: \"thoughts summary to say to user\",\n      },\n      command: { name: \"command name\", args: { \"arg name\": \"value\" } },\n    };\n  }\n\n  /**\n   * Adds a constraint to the list of constraints.\n   * @param constraint The constraint to add.\n   * @returns void\n   */\n  add_constraint(constraint: string): void {\n    this.constraints.push(constraint);\n  }\n\n  /**\n   * Adds a tool to the list of commands.\n   * @param tool The tool to add.\n   * @returns void\n   */\n  add_tool(tool: ObjectTool): void {\n    this.commands.push(tool);\n  }\n\n  _generate_command_string(tool: ObjectTool): string {\n    let output = `\"${tool.name}\": ${tool.description}`;\n    const jsonSchema = (\n      isInteropZodSchema(tool.schema) ? toJsonSchema(tool.schema) : tool.schema\n    ) as { properties?: Record<string, JsonSchema7Type> } | undefined;\n    output += `, args json schema: ${JSON.stringify(jsonSchema?.properties)}`;\n    return output;\n  }\n\n  /**\n   * Adds a resource to the list of resources.\n   * @param resource The resource to add.\n   * @returns void\n   */\n  add_resource(resource: string): void {\n    this.resources.push(resource);\n  }\n\n  /**\n   * Adds a performance evaluation to the list of performance evaluations.\n   * @param evaluation The performance evaluation to add.\n   * @returns void\n   */\n  add_performance_evaluation(evaluation: string): void {\n    this.performance_evaluation.push(evaluation);\n  }\n\n  // eslint-disable-next-line @typescript-eslint/no-explicit-any\n  _generate_numbered_list(items: any[], item_type = \"list\"): string {\n    if (item_type === \"command\") {\n      const command_strings = items.map(\n        (item, i) => `${i + 1}. ${this._generate_command_string(item)}`\n      );\n      const finish_description =\n        \"use this to signal that you have finished all your objectives\";\n      const finish_args =\n        '\"response\": \"final response to let people know you have finished your objectives\"';\n      const finish_string = `${\n        items.length + 1\n      }. ${FINISH_NAME}: ${finish_description}, args: ${finish_args}`;\n      return command_strings.concat([finish_string]).join(\"\\n\");\n    }\n\n    return items.map((item, i) => `${i + 1}. ${item}`).join(\"\\n\");\n  }\n\n  /**\n   * Generates a prompt string that includes the constraints, commands,\n   * resources, performance evaluations, and response format.\n   * @returns A string representing the prompt.\n   */\n  generate_prompt_string(): string {\n    const formatted_response_format = JSON.stringify(\n      this.response_format,\n      null,\n      4\n    );\n    const prompt_string =\n      `Constraints:\\n${this._generate_numbered_list(this.constraints)}\\n\\n` +\n      `Commands:\\n${this._generate_numbered_list(\n        this.commands,\n        \"command\"\n      )}\\n\\n` +\n      `Resources:\\n${this._generate_numbered_list(this.resources)}\\n\\n` +\n      `Performance Evaluation:\\n${this._generate_numbered_list(\n        this.performance_evaluation\n      )}\\n\\n` +\n      `You should only respond in JSON format as described below ` +\n      `\\nResponse Format: \\n${formatted_response_format} ` +\n      `\\nEnsure the response can be parsed by Python json.loads`;\n\n    return prompt_string;\n  }\n}\n\n/**\n * Function that generates a prompt string for a given list of tools.\n */\nexport function getPrompt(tools: ObjectTool[]): string {\n  const prompt_generator = new PromptGenerator();\n\n  prompt_generator.add_constraint(\n    \"~4000 word limit for short term memory. \" +\n      \"Your short term memory is short, \" +\n      \"so immediately save important information to files.\"\n  );\n  prompt_generator.add_constraint(\n    \"If you are unsure how you previously did something \" +\n      \"or want to recall past events, \" +\n      \"thinking about similar events will help you remember.\"\n  );\n  prompt_generator.add_constraint(\"No user assistance\");\n  prompt_generator.add_constraint(\n    'Exclusively use the commands listed in double quotes e.g. \"command name\"'\n  );\n\n  for (const tool of tools) {\n    prompt_generator.add_tool(tool);\n  }\n\n  prompt_generator.add_resource(\n    \"Internet access for searches and information gathering.\"\n  );\n  prompt_generator.add_resource(\"Long Term memory management.\");\n  prompt_generator.add_resource(\n    \"GPT-3.5 powered Agents for delegation of simple tasks.\"\n  );\n  prompt_generator.add_resource(\"File output.\");\n\n  prompt_generator.add_performance_evaluation(\n    \"Continuously review and analyze your actions \" +\n      \"to ensure you are performing to the best of your abilities.\"\n  );\n  prompt_generator.add_performance_evaluation(\n    \"Constructively self-criticize your big-picture behavior constantly.\"\n  );\n  prompt_generator.add_performance_evaluation(\n    \"Reflect on past decisions and strategies to refine your approach.\"\n  );\n  prompt_generator.add_performance_evaluation(\n    \"Every command has a cost, so be smart and efficient. \" +\n      \"Aim to complete tasks in the least number of steps.\"\n  );\n\n  const prompt_string = prompt_generator.generate_prompt_string();\n\n  return prompt_string;\n}\n"],"mappings":";;;;;;;;;AAWA,IAAa,kBAAb,MAA6B;CAC3B;CAEA;CAEA;CAEA;CAEA;CAEA,cAAc;EACZ,KAAK,cAAc,CAAE;EACrB,KAAK,WAAW,CAAE;EAClB,KAAK,YAAY,CAAE;EACnB,KAAK,yBAAyB,CAAE;EAChC,KAAK,kBAAkB;GACrB,UAAU;IACR,MAAM;IACN,WAAW;IACX,MAAM;IACN,WAAW;IACX,OAAO;GACR;GACD,SAAS;IAAE,MAAM;IAAgB,MAAM,EAAE,YAAY,QAAS;GAAE;EACjE;CACF;;;;;;CAOD,eAAeA,YAA0B;EACvC,KAAK,YAAY,KAAK,WAAW;CAClC;;;;;;CAOD,SAASC,MAAwB;EAC/B,KAAK,SAAS,KAAK,KAAK;CACzB;CAED,yBAAyBA,MAA0B;EACjD,IAAI,SAAS,CAAC,CAAC,EAAE,KAAK,KAAK,GAAG,EAAE,KAAK,aAAa;EAClD,MAAM,aACJ,mBAAmB,KAAK,OAAO,GAAG,aAAa,KAAK,OAAO,GAAG,KAAK;EAErE,UAAU,CAAC,oBAAoB,EAAE,KAAK,UAAU,YAAY,WAAW,EAAE;AACzE,SAAO;CACR;;;;;;CAOD,aAAaC,UAAwB;EACnC,KAAK,UAAU,KAAK,SAAS;CAC9B;;;;;;CAOD,2BAA2BC,YAA0B;EACnD,KAAK,uBAAuB,KAAK,WAAW;CAC7C;CAGD,wBAAwBC,OAAc,YAAY,QAAgB;AAChE,MAAI,cAAc,WAAW;GAC3B,MAAM,kBAAkB,MAAM,IAC5B,CAAC,MAAM,MAAM,GAAG,IAAI,EAAE,EAAE,EAAE,KAAK,yBAAyB,KAAK,EAAE,CAChE;GACD,MAAM,qBACJ;GACF,MAAM,cACJ;GACF,MAAM,gBAAgB,GACpB,MAAM,SAAS,EAChB,EAAE,EAAE,YAAY,EAAE,EAAE,mBAAmB,QAAQ,EAAE,aAAa;AAC/D,UAAO,gBAAgB,OAAO,CAAC,aAAc,EAAC,CAAC,KAAK,KAAK;EAC1D;AAED,SAAO,MAAM,IAAI,CAAC,MAAM,MAAM,GAAG,IAAI,EAAE,EAAE,EAAE,MAAM,CAAC,CAAC,KAAK,KAAK;CAC9D;;;;;;CAOD,yBAAiC;EAC/B,MAAM,4BAA4B,KAAK,UACrC,KAAK,iBACL,MACA,EACD;EACD,MAAM,gBACJ,CAAC,cAAc,EAAE,KAAK,wBAAwB,KAAK,YAAY,CAAC,eAAI,EACtD,KAAK,wBACjB,KAAK,UACL,UACD,CAAC,gBAAI,EACS,KAAK,wBAAwB,KAAK,UAAU,CAAC,6BAAI,EACpC,KAAK,wBAC/B,KAAK,uBACN,CAAC,mFAAI,EAEkB,0BAA0B,yDAAC,CACO;AAE5D,SAAO;CACR;AACF;;;;AAKD,SAAgB,UAAUC,OAA6B;CACrD,MAAM,mBAAmB,IAAI;CAE7B,iBAAiB,eACf,+HAGD;CACD,iBAAiB,eACf,0IAGD;CACD,iBAAiB,eAAe,qBAAqB;CACrD,iBAAiB,eACf,6EACD;AAED,MAAK,MAAM,QAAQ,OACjB,iBAAiB,SAAS,KAAK;CAGjC,iBAAiB,aACf,0DACD;CACD,iBAAiB,aAAa,+BAA+B;CAC7D,iBAAiB,aACf,yDACD;CACD,iBAAiB,aAAa,eAAe;CAE7C,iBAAiB,2BACf,2GAED;CACD,iBAAiB,2BACf,sEACD;CACD,iBAAiB,2BACf,oEACD;CACD,iBAAiB,2BACf,2GAED;CAED,MAAM,gBAAgB,iBAAiB,wBAAwB;AAE/D,QAAO;AACR"}
+{"version":3,"file":"prompt_generator.js","names":[],"sources":["../../../src/experimental/autogpt/prompt_generator.ts"],"sourcesContent":["import { isInteropZodSchema } from \"@langchain/core/utils/types\";\nimport {\n  type JsonSchema7Type,\n  toJsonSchema,\n} from \"@langchain/core/utils/json_schema\";\nimport { ObjectTool, FINISH_NAME } from \"./schema.js\";\n\n/**\n * Class that generates prompts for generative agents. It maintains a list\n * of constraints, commands, resources, and performance evaluations.\n */\nexport class PromptGenerator {\n  constraints: string[];\n\n  commands: ObjectTool[];\n\n  resources: string[];\n\n  performance_evaluation: string[];\n\n  response_format: object;\n\n  constructor() {\n    this.constraints = [];\n    this.commands = [];\n    this.resources = [];\n    this.performance_evaluation = [];\n    this.response_format = {\n      thoughts: {\n        text: \"thought\",\n        reasoning: \"reasoning\",\n        plan: \"- short bulleted\\n- list that conveys\\n- long-term plan\",\n        criticism: \"constructive self-criticism\",\n        speak: \"thoughts summary to say to user\",\n      },\n      command: { name: \"command name\", args: { \"arg name\": \"value\" } },\n    };\n  }\n\n  /**\n   * Adds a constraint to the list of constraints.\n   * @param constraint The constraint to add.\n   * @returns void\n   */\n  add_constraint(constraint: string): void {\n    this.constraints.push(constraint);\n  }\n\n  /**\n   * Adds a tool to the list of commands.\n   * @param tool The tool to add.\n   * @returns void\n   */\n  add_tool(tool: ObjectTool): void {\n    this.commands.push(tool);\n  }\n\n  _generate_command_string(tool: ObjectTool): string {\n    let output = `\"${tool.name}\": ${tool.description}`;\n    const jsonSchema = (\n      isInteropZodSchema(tool.schema) ? toJsonSchema(tool.schema) : tool.schema\n    ) as { properties?: Record<string, JsonSchema7Type> } | undefined;\n    output += `, args json schema: ${JSON.stringify(jsonSchema?.properties)}`;\n    return output;\n  }\n\n  /**\n   * Adds a resource to the list of resources.\n   * @param resource The resource to add.\n   * @returns void\n   */\n  add_resource(resource: string): void {\n    this.resources.push(resource);\n  }\n\n  /**\n   * Adds a performance evaluation to the list of performance evaluations.\n   * @param evaluation The performance evaluation to add.\n   * @returns void\n   */\n  add_performance_evaluation(evaluation: string): void {\n    this.performance_evaluation.push(evaluation);\n  }\n\n  // eslint-disable-next-line @typescript-eslint/no-explicit-any\n  _generate_numbered_list(items: any[], item_type = \"list\"): string {\n    if (item_type === \"command\") {\n      const command_strings = items.map(\n        (item, i) => `${i + 1}. ${this._generate_command_string(item)}`\n      );\n      const finish_description =\n        \"use this to signal that you have finished all your objectives\";\n      const finish_args =\n        '\"response\": \"final response to let people know you have finished your objectives\"';\n      const finish_string = `${\n        items.length + 1\n      }. ${FINISH_NAME}: ${finish_description}, args: ${finish_args}`;\n      return command_strings.concat([finish_string]).join(\"\\n\");\n    }\n\n    return items.map((item, i) => `${i + 1}. ${item}`).join(\"\\n\");\n  }\n\n  /**\n   * Generates a prompt string that includes the constraints, commands,\n   * resources, performance evaluations, and response format.\n   * @returns A string representing the prompt.\n   */\n  generate_prompt_string(): string {\n    const formatted_response_format = JSON.stringify(\n      this.response_format,\n      null,\n      4\n    );\n    const prompt_string =\n      `Constraints:\\n${this._generate_numbered_list(this.constraints)}\\n\\n` +\n      `Commands:\\n${this._generate_numbered_list(\n        this.commands,\n        \"command\"\n      )}\\n\\n` +\n      `Resources:\\n${this._generate_numbered_list(this.resources)}\\n\\n` +\n      `Performance Evaluation:\\n${this._generate_numbered_list(\n        this.performance_evaluation\n      )}\\n\\n` +\n      `You should only respond in JSON format as described below ` +\n      `\\nResponse Format: \\n${formatted_response_format} ` +\n      `\\nEnsure the response can be parsed by Python json.loads`;\n\n    return prompt_string;\n  }\n}\n\n/**\n * Function that generates a prompt string for a given list of tools.\n */\nexport function getPrompt(tools: ObjectTool[]): string {\n  const prompt_generator = new PromptGenerator();\n\n  prompt_generator.add_constraint(\n    \"~4000 word limit for short term memory. \" +\n      \"Your short term memory is short, \" +\n      \"so immediately save important information to files.\"\n  );\n  prompt_generator.add_constraint(\n    \"If you are unsure how you previously did something \" +\n      \"or want to recall past events, \" +\n      \"thinking about similar events will help you remember.\"\n  );\n  prompt_generator.add_constraint(\"No user assistance\");\n  prompt_generator.add_constraint(\n    'Exclusively use the commands listed in double quotes e.g. \"command name\"'\n  );\n\n  for (const tool of tools) {\n    prompt_generator.add_tool(tool);\n  }\n\n  prompt_generator.add_resource(\n    \"Internet access for searches and information gathering.\"\n  );\n  prompt_generator.add_resource(\"Long Term memory management.\");\n  prompt_generator.add_resource(\n    \"GPT-3.5 powered Agents for delegation of simple tasks.\"\n  );\n  prompt_generator.add_resource(\"File output.\");\n\n  prompt_generator.add_performance_evaluation(\n    \"Continuously review and analyze your actions \" +\n      \"to ensure you are performing to the best of your abilities.\"\n  );\n  prompt_generator.add_performance_evaluation(\n    \"Constructively self-criticize your big-picture behavior constantly.\"\n  );\n  prompt_generator.add_performance_evaluation(\n    \"Reflect on past decisions and strategies to refine your approach.\"\n  );\n  prompt_generator.add_performance_evaluation(\n    \"Every command has a cost, so be smart and efficient. \" +\n      \"Aim to complete tasks in the least number of steps.\"\n  );\n\n  const prompt_string = prompt_generator.generate_prompt_string();\n\n  return prompt_string;\n}\n"],"mappings":";;;;;;;;;AAWA,IAAa,kBAAb,MAA6B;CAC3B;CAEA;CAEA;CAEA;CAEA;CAEA,cAAc;AACZ,OAAK,cAAc,EAAE;AACrB,OAAK,WAAW,EAAE;AAClB,OAAK,YAAY,EAAE;AACnB,OAAK,yBAAyB,EAAE;AAChC,OAAK,kBAAkB;GACrB,UAAU;IACR,MAAM;IACN,WAAW;IACX,MAAM;IACN,WAAW;IACX,OAAO;IACR;GACD,SAAS;IAAE,MAAM;IAAgB,MAAM,EAAE,YAAY,SAAS;IAAE;GACjE;;;;;;;CAQH,eAAe,YAA0B;AACvC,OAAK,YAAY,KAAK,WAAW;;;;;;;CAQnC,SAAS,MAAwB;AAC/B,OAAK,SAAS,KAAK,KAAK;;CAG1B,yBAAyB,MAA0B;EACjD,IAAI,SAAS,IAAI,KAAK,KAAK,KAAK,KAAK;EACrC,MAAM,aACJ,mBAAmB,KAAK,OAAO,GAAG,aAAa,KAAK,OAAO,GAAG,KAAK;AAErE,YAAU,uBAAuB,KAAK,UAAU,YAAY,WAAW;AACvE,SAAO;;;;;;;CAQT,aAAa,UAAwB;AACnC,OAAK,UAAU,KAAK,SAAS;;;;;;;CAQ/B,2BAA2B,YAA0B;AACnD,OAAK,uBAAuB,KAAK,WAAW;;CAI9C,wBAAwB,OAAc,YAAY,QAAgB;AAChE,MAAI,cAAc,WAAW;GAC3B,MAAM,kBAAkB,MAAM,KAC3B,MAAM,MAAM,GAAG,IAAI,EAAE,IAAI,KAAK,yBAAyB,KAAK,GAC9D;GAKD,MAAM,gBAAgB,GACpB,MAAM,SAAS,EAChB,IAAI,YAAY;AACjB,UAAO,gBAAgB,OAAO,CAAC,cAAc,CAAC,CAAC,KAAK,KAAK;;AAG3D,SAAO,MAAM,KAAK,MAAM,MAAM,GAAG,IAAI,EAAE,IAAI,OAAO,CAAC,KAAK,KAAK;;;;;;;CAQ/D,yBAAiC;EAC/B,MAAM,4BAA4B,KAAK,UACrC,KAAK,iBACL,MACA,EACD;AAeD,SAbE,iBAAiB,KAAK,wBAAwB,KAAK,YAAY,CAAC,iBAClD,KAAK,wBACjB,KAAK,UACL,UACD,CAAC,kBACa,KAAK,wBAAwB,KAAK,UAAU,CAAC,+BAChC,KAAK,wBAC/B,KAAK,uBACN,CAAC,qFAEsB,0BAA0B;;;;;;AAUxD,SAAgB,UAAU,OAA6B;CACrD,MAAM,mBAAmB,IAAI,iBAAiB;AAE9C,kBAAiB,eACf,+HAGD;AACD,kBAAiB,eACf,0IAGD;AACD,kBAAiB,eAAe,qBAAqB;AACrD,kBAAiB,eACf,6EACD;AAED,MAAK,MAAM,QAAQ,MACjB,kBAAiB,SAAS,KAAK;AAGjC,kBAAiB,aACf,0DACD;AACD,kBAAiB,aAAa,+BAA+B;AAC7D,kBAAiB,aACf,yDACD;AACD,kBAAiB,aAAa,eAAe;AAE7C,kBAAiB,2BACf,2GAED;AACD,kBAAiB,2BACf,sEACD;AACD,kBAAiB,2BACf,oEACD;AACD,kBAAiB,2BACf,2GAED;AAID,QAFsB,iBAAiB,wBAAwB"}

package/dist/experimental/autogpt/schema.d.cts

@@ -1,7 +1,6 @@
 import { StructuredTool, ToolInterface } from "@langchain/core/tools";
 
 //#region src/experimental/autogpt/schema.d.ts
-
 /**
  * Type alias for StructuredTool. It is part of the tools module in
  * LangChain, which includes a variety of tools used for different

package/dist/experimental/autogpt/schema.d.cts.map

@@ -1 +1 @@
-{"version":3,"file":"schema.d.cts","names":["StructuredTool","ToolInterface","ObjectTool","FINISH_NAME","AutoGPTAction","Record"],"sources":["../../../src/experimental/autogpt/schema.d.ts"],"sourcesContent":["import { StructuredTool, ToolInterface } from \"@langchain/core/tools\";\n/**\n * Type alias for StructuredTool. It is part of the tools module in\n * LangChain, which includes a variety of tools used for different\n * purposes.\n */\nexport type ObjectTool = StructuredTool | ToolInterface;\nexport declare const FINISH_NAME = \"finish\";\n/**\n * Interface that describes an action that can be performed by the AutoGPT\n * model in LangChain. It has a `name` property, which is a string that\n * represents the name of the action, and an `args` property, which is an\n * object that contains the arguments for the action.\n */\nexport interface AutoGPTAction {\n    name: string;\n    args: Record<string, any>;\n}\n//# sourceMappingURL=schema.d.ts.map"],"mappings":";;;;;;AAMA;AAQA;;KARYE,UAAAA,GAAaF,iBAAiBC;;;;;;;UAQzBG,aAAAA;;QAEPC"}
+{"version":3,"file":"schema.d.cts","names":[],"sources":["../../../src/experimental/autogpt/schema.ts"],"mappings":";;;;;AAOA;;;KAAY,UAAA,GAAa,cAAA,GAAiB,aAAA;AAU1C;;;;;;AAAA,UAAiB,aAAA;EACf,IAAA;EAEA,IAAA,EAAM,MAAA;AAAA"}

package/dist/experimental/autogpt/schema.d.ts

@@ -1,7 +1,6 @@
 import { StructuredTool, ToolInterface } from "@langchain/core/tools";
 
 //#region src/experimental/autogpt/schema.d.ts
-
 /**
  * Type alias for StructuredTool. It is part of the tools module in
  * LangChain, which includes a variety of tools used for different

package/dist/experimental/autogpt/schema.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"schema.d.ts","names":["StructuredTool","ToolInterface","ObjectTool","FINISH_NAME","AutoGPTAction","Record"],"sources":["../../../src/experimental/autogpt/schema.d.ts"],"sourcesContent":["import { StructuredTool, ToolInterface } from \"@langchain/core/tools\";\n/**\n * Type alias for StructuredTool. It is part of the tools module in\n * LangChain, which includes a variety of tools used for different\n * purposes.\n */\nexport type ObjectTool = StructuredTool | ToolInterface;\nexport declare const FINISH_NAME = \"finish\";\n/**\n * Interface that describes an action that can be performed by the AutoGPT\n * model in LangChain. It has a `name` property, which is a string that\n * represents the name of the action, and an `args` property, which is an\n * object that contains the arguments for the action.\n */\nexport interface AutoGPTAction {\n    name: string;\n    args: Record<string, any>;\n}\n//# sourceMappingURL=schema.d.ts.map"],"mappings":";;;;;;AAMA;AAQA;;KARYE,UAAAA,GAAaF,iBAAiBC;;;;;;;UAQzBG,aAAAA;;QAEPC"}
+{"version":3,"file":"schema.d.ts","names":[],"sources":["../../../src/experimental/autogpt/schema.ts"],"mappings":";;;;;AAOA;;;KAAY,UAAA,GAAa,cAAA,GAAiB,aAAA;AAU1C;;;;;;AAAA,UAAiB,aAAA;EACf,IAAA;EAEA,IAAA,EAAM,MAAA;AAAA"}

package/dist/experimental/babyagi/agent.cjs

@@ -1,9 +1,9 @@
-const require_rolldown_runtime = require('../../_virtual/rolldown_runtime.cjs');
+const require_runtime = require('../../_virtual/_rolldown/runtime.cjs');
 const require_base = require('../../chains/base.cjs');
 const require_task_creation = require('./task_creation.cjs');
 const require_task_execution = require('./task_execution.cjs');
 const require_task_prioritization = require('./task_prioritization.cjs');
-const __langchain_core_documents = require_rolldown_runtime.__toESM(require("@langchain/core/documents"));
+let _langchain_core_documents = require("@langchain/core/documents");
 
 //#region src/experimental/babyagi/agent.ts
 /**
@@ -98,16 +98,14 @@ var BabyAGI = class BabyAGI extends require_base.BaseChain {
 	* @returns Promise resolving to an array of tasks without taskID.
 	*/
 	async getNextTasks(result, task_description, objective, runManager) {
-		const taskNames = this.taskList.map((t) => t.taskName);
-		const incomplete_tasks = taskNames.join(", ");
+		const incomplete_tasks = this.taskList.map((t) => t.taskName).join(", ");
 		const { [this.creationChain.outputKeys[0]]: text } = await this.creationChain.call({
 			result,
 			task_description,
 			incomplete_tasks,
 			objective
 		}, runManager?.getChild());
-		const newTasks = text.split("\n");
-		return newTasks.filter((taskName) => taskName.trim()).map((taskName) => ({ taskName }));
+		return text.split("\n").filter((taskName) => taskName.trim()).map((taskName) => ({ taskName }));
 	}
 	/**
 	* Prioritizes the tasks based on the current task ID and the objective.
@@ -181,7 +179,7 @@ var BabyAGI = class BabyAGI extends require_base.BaseChain {
 			const result = await this.executeTask(objective, task.taskName, runManager);
 			const thisTaskID = parseInt(task.taskID, 10);
 			this.printTaskResult(result);
-			await this.vectorstore.addDocuments([new __langchain_core_documents.Document({
+			await this.vectorstore.addDocuments([new _langchain_core_documents.Document({
 				pageContent: result,
 				metadata: { task: task.taskName }
 			})]);
@@ -210,20 +208,18 @@ var BabyAGI = class BabyAGI extends require_base.BaseChain {
 	* @param rest Optional additional parameters.
 	* @returns A new instance of BabyAGI.
 	*/
-	static fromLLM({ llm, vectorstore, executionChain, verbose, callbacks,...rest }) {
-		const creationChain = require_task_creation.TaskCreationChain.fromLLM({
-			llm,
-			verbose,
-			callbacks
-		});
-		const prioritizationChain = require_task_prioritization.TaskPrioritizationChain.fromLLM({
-			llm,
-			verbose,
-			callbacks
-		});
+	static fromLLM({ llm, vectorstore, executionChain, verbose, callbacks, ...rest }) {
 		return new BabyAGI({
-			creationChain,
-			prioritizationChain,
+			creationChain: require_task_creation.TaskCreationChain.fromLLM({
+				llm,
+				verbose,
+				callbacks
+			}),
+			prioritizationChain: require_task_prioritization.TaskPrioritizationChain.fromLLM({
+				llm,
+				verbose,
+				callbacks
+			}),
 			executionChain: executionChain || require_task_execution.TaskExecutionChain.fromLLM({
 				llm,
 				verbose,

package/dist/experimental/babyagi/agent.cjs.map

@@ -1 +1 @@
-{"version":3,"file":"agent.cjs","names":["BaseChain","task: Task","result: string","task_description: string","objective: string","runManager?: CallbackManagerForChainRun","thisTaskID: number","query: string","task: string","Document","TaskCreationChain","TaskPrioritizationChain","TaskExecutionChain"],"sources":["../../../src/experimental/babyagi/agent.ts"],"sourcesContent":["import type { BaseLanguageModelInterface } from \"@langchain/core/language_models/base\";\nimport type { VectorStoreInterface } from \"@langchain/core/vectorstores\";\nimport { Document } from \"@langchain/core/documents\";\nimport { ChainValues } from \"@langchain/core/utils/types\";\nimport { CallbackManagerForChainRun } from \"@langchain/core/callbacks/manager\";\nimport { BaseChain, ChainInputs } from \"../../chains/base.js\";\nimport { SerializedBaseChain } from \"../../chains/serde.js\";\nimport { Optional } from \"../../types/type-utils.js\";\nimport { TaskCreationChain } from \"./task_creation.js\";\nimport { TaskExecutionChain } from \"./task_execution.js\";\nimport { TaskPrioritizationChain } from \"./task_prioritization.js\";\n\n/**\n * Interface defining the structure of a task. A task has a `taskID` and a\n * `taskName`.\n */\nexport interface Task {\n  taskID: string;\n  taskName: string;\n}\n\n/**\n * Interface defining the structure of the inputs for the `BabyAGI` class.\n * It extends the `ChainInputs` interface, omitting the 'memory' and\n * 'callbackManager' properties, and adds properties specific to\n * `BabyAGI`.\n */\nexport interface BabyAGIInputs extends Omit<\n  ChainInputs,\n  \"memory\" | \"callbackManager\"\n> {\n  creationChain: BaseChain;\n  prioritizationChain: BaseChain;\n  executionChain: BaseChain;\n  vectorstore: VectorStoreInterface;\n  maxIterations?: number;\n}\n\n/**\n * Class responsible for managing tasks, including their creation,\n * prioritization, and execution. It uses three chains for these\n * operations: `creationChain`, `prioritizationChain`, and\n * `executionChain`.\n * @example\n * ```typescript\n * const babyAGI = BabyAGI.fromLLM({\n *   llm: new OpenAI({ temperature: 0 }),\n *   vectorstore: new MemoryVectorStore(new OpenAIEmbeddings()),\n *   maxIterations: 3,\n * });\n *\n * const result = await babyAGI.call({\n *   objective: \"Write a weather report for SF today\",\n * });\n * ```\n */\nexport class BabyAGI extends BaseChain implements BabyAGIInputs {\n  static lc_name() {\n    return \"BabyAGI\";\n  }\n\n  taskList: Task[];\n\n  creationChain: BaseChain;\n\n  prioritizationChain: BaseChain;\n\n  executionChain: BaseChain;\n\n  taskIDCounter: number;\n\n  vectorstore: VectorStoreInterface;\n\n  maxIterations: number;\n\n  constructor({\n    creationChain,\n    prioritizationChain,\n    executionChain,\n    vectorstore,\n    maxIterations = 100,\n    verbose,\n    callbacks,\n  }: BabyAGIInputs) {\n    super(undefined, verbose, callbacks);\n    this.taskList = [];\n    this.creationChain = creationChain;\n    this.prioritizationChain = prioritizationChain;\n    this.executionChain = executionChain;\n    this.taskIDCounter = 1;\n    this.vectorstore = vectorstore;\n    this.maxIterations = maxIterations;\n  }\n\n  _chainType() {\n    return \"BabyAGI\" as const;\n  }\n\n  get inputKeys() {\n    return [\"objective\", \"firstTask\"];\n  }\n\n  get outputKeys() {\n    return [];\n  }\n\n  /**\n   * Adds a task to the task list.\n   * @param task The task to be added.\n   * @returns Promise resolving to void.\n   */\n  async addTask(task: Task) {\n    this.taskList.push(task);\n  }\n\n  /**\n   * Prints the current task list to the console.\n   * @returns void\n   */\n  printTaskList() {\n    console.log(\"\\x1b[95m\\x1b[1m\\n*****TASK LIST*****\\n\\x1b[0m\\x1b[0m\");\n    for (const t of this.taskList) {\n      console.log(`${t.taskID}: ${t.taskName}`);\n    }\n  }\n\n  /**\n   * Prints the next task to the console.\n   * @param task The next task to be printed.\n   * @returns void\n   */\n  printNextTask(task: Task) {\n    console.log(\"\\x1b[92m\\x1b[1m\\n*****NEXT TASK*****\\n\\x1b[0m\\x1b[0m\");\n    console.log(`${task.taskID}: ${task.taskName}`);\n  }\n\n  /**\n   * Prints the result of a task to the console.\n   * @param result The result of the task.\n   * @returns void\n   */\n  printTaskResult(result: string) {\n    console.log(\"\\x1b[93m\\x1b[1m\\n*****TASK RESULT*****\\n\\x1b[0m\\x1b[0m\");\n    console.log(result.trim());\n  }\n\n  /**\n   * Generates the next tasks based on the result of the previous task, the\n   * task description, and the objective.\n   * @param result The result of the previous task.\n   * @param task_description The description of the task.\n   * @param objective The objective of the task.\n   * @param runManager Optional CallbackManagerForChainRun instance.\n   * @returns Promise resolving to an array of tasks without taskID.\n   */\n  async getNextTasks(\n    result: string,\n    task_description: string,\n    objective: string,\n    runManager?: CallbackManagerForChainRun\n  ): Promise<Optional<Task, \"taskID\">[]> {\n    const taskNames = this.taskList.map((t) => t.taskName);\n    const incomplete_tasks = taskNames.join(\", \");\n    const { [this.creationChain.outputKeys[0]]: text } =\n      await this.creationChain.call(\n        {\n          result,\n          task_description,\n          incomplete_tasks,\n          objective,\n        },\n        runManager?.getChild()\n      );\n    const newTasks = (text as string).split(\"\\n\");\n    return newTasks\n      .filter((taskName) => taskName.trim())\n      .map((taskName) => ({ taskName }));\n  }\n\n  /**\n   * Prioritizes the tasks based on the current task ID and the objective.\n   * @param thisTaskID The ID of the current task.\n   * @param objective The objective of the task.\n   * @param runManager Optional CallbackManagerForChainRun instance.\n   * @returns Promise resolving to an array of prioritized tasks.\n   */\n  async prioritizeTasks(\n    thisTaskID: number,\n    objective: string,\n    runManager?: CallbackManagerForChainRun\n  ) {\n    const taskNames = this.taskList.map((t) => t.taskName);\n    const nextTaskID = thisTaskID + 1;\n    const { [this.prioritizationChain.outputKeys[0]]: text } =\n      await this.prioritizationChain.call(\n        {\n          task_names: taskNames.join(\", \"),\n          next_task_id: String(nextTaskID),\n          objective,\n        },\n        runManager?.getChild()\n      );\n    const newTasks = (text as string).trim().split(\"\\n\");\n    const prioritizedTaskList = [];\n    for (const taskString of newTasks) {\n      const taskParts = taskString.trim().split(\".\", 2);\n      if (taskParts.length === 2) {\n        const taskID = taskParts[0].trim();\n        const taskName = taskParts[1].trim();\n        prioritizedTaskList.push({ taskID, taskName });\n      }\n    }\n    return prioritizedTaskList;\n  }\n\n  /**\n   * Retrieves the top tasks that are most similar to the given query.\n   * @param query The query to search for.\n   * @param k The number of top tasks to retrieve.\n   * @returns Promise resolving to an array of top tasks.\n   */\n  async getTopTasks(query: string, k = 5) {\n    const results = await this.vectorstore.similaritySearch(query, k);\n    if (!results) {\n      return [];\n    }\n    return results.map((item) => String(item.metadata.task));\n  }\n\n  /**\n   * Executes a task based on the objective and the task description.\n   * @param objective The objective of the task.\n   * @param task The task to be executed.\n   * @param runManager Optional CallbackManagerForChainRun instance.\n   * @returns Promise resolving to the result of the task execution as a string.\n   */\n  async executeTask(\n    objective: string,\n    task: string,\n    runManager?: CallbackManagerForChainRun\n  ) {\n    const context = await this.getTopTasks(objective);\n    const { [this.executionChain.outputKeys[0]]: text } =\n      await this.executionChain.call(\n        {\n          objective,\n          context: context.join(\"\\n\"),\n          task,\n        },\n        runManager?.getChild()\n      );\n    return text as string;\n  }\n\n  async _call(\n    { objective, firstTask = \"Make a todo list\" }: ChainValues,\n    runManager?: CallbackManagerForChainRun\n  ) {\n    this.taskList = [];\n    this.taskIDCounter = 1;\n    await this.addTask({ taskID: \"1\", taskName: firstTask });\n\n    let numIters = 0;\n    while (numIters < this.maxIterations && this.taskList.length > 0) {\n      this.printTaskList();\n\n      const task = this.taskList.shift()!;\n      this.printNextTask(task);\n\n      const result = await this.executeTask(\n        objective,\n        task.taskName,\n        runManager\n      );\n      const thisTaskID = parseInt(task.taskID, 10);\n      this.printTaskResult(result);\n\n      await this.vectorstore.addDocuments([\n        new Document({\n          pageContent: result,\n          metadata: { task: task.taskName },\n        }),\n      ]);\n\n      const newTasks = await this.getNextTasks(\n        result,\n        task.taskName,\n        objective,\n        runManager\n      );\n      for (const newTask of newTasks) {\n        this.taskIDCounter += 1;\n        newTask.taskID = this.taskIDCounter.toFixed();\n        await this.addTask(newTask as Task);\n      }\n      this.taskList = await this.prioritizeTasks(\n        thisTaskID,\n        objective,\n        runManager\n      );\n\n      numIters += 1;\n    }\n    return {};\n  }\n\n  serialize(): SerializedBaseChain {\n    throw new Error(\"Method not implemented.\");\n  }\n\n  /**\n   * Static method to create a new BabyAGI instance from a\n   * BaseLanguageModel.\n   * @param llm BaseLanguageModel instance used to generate a new BabyAGI instance.\n   * @param vectorstore VectorStore instance used to store and retrieve vectors.\n   * @param executionChain Optional BaseChain instance used to execute tasks.\n   * @param verbose Optional boolean indicating whether to log verbose output.\n   * @param callbacks Optional callbacks to be used during the execution of tasks.\n   * @param rest Optional additional parameters.\n   * @returns A new instance of BabyAGI.\n   */\n  static fromLLM({\n    llm,\n    vectorstore,\n    executionChain,\n    verbose,\n    callbacks,\n    ...rest\n  }: Optional<\n    BabyAGIInputs,\n    \"executionChain\" | \"creationChain\" | \"prioritizationChain\"\n  > & { llm: BaseLanguageModelInterface }) {\n    const creationChain = TaskCreationChain.fromLLM({\n      llm,\n      verbose,\n      callbacks,\n    });\n    const prioritizationChain = TaskPrioritizationChain.fromLLM({\n      llm,\n      verbose,\n      callbacks,\n    });\n    return new BabyAGI({\n      creationChain,\n      prioritizationChain,\n      executionChain:\n        executionChain ||\n        TaskExecutionChain.fromLLM({ llm, verbose, callbacks }),\n      vectorstore,\n      verbose,\n      callbacks,\n      ...rest,\n    });\n  }\n}\n"],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;;AAwDA,IAAa,UAAb,MAAa,gBAAgBA,uBAAmC;CAC9D,OAAO,UAAU;AACf,SAAO;CACR;CAED;CAEA;CAEA;CAEA;CAEA;CAEA;CAEA;CAEA,YAAY,EACV,eACA,qBACA,gBACA,aACA,gBAAgB,KAChB,SACA,WACc,EAAE;EAChB,MAAM,QAAW,SAAS,UAAU;EACpC,KAAK,WAAW,CAAE;EAClB,KAAK,gBAAgB;EACrB,KAAK,sBAAsB;EAC3B,KAAK,iBAAiB;EACtB,KAAK,gBAAgB;EACrB,KAAK,cAAc;EACnB,KAAK,gBAAgB;CACtB;CAED,aAAa;AACX,SAAO;CACR;CAED,IAAI,YAAY;AACd,SAAO,CAAC,aAAa,WAAY;CAClC;CAED,IAAI,aAAa;AACf,SAAO,CAAE;CACV;;;;;;CAOD,MAAM,QAAQC,MAAY;EACxB,KAAK,SAAS,KAAK,KAAK;CACzB;;;;;CAMD,gBAAgB;EACd,QAAQ,IAAI,uDAAuD;AACnE,OAAK,MAAM,KAAK,KAAK,UACnB,QAAQ,IAAI,GAAG,EAAE,OAAO,EAAE,EAAE,EAAE,UAAU,CAAC;CAE5C;;;;;;CAOD,cAAcA,MAAY;EACxB,QAAQ,IAAI,uDAAuD;EACnE,QAAQ,IAAI,GAAG,KAAK,OAAO,EAAE,EAAE,KAAK,UAAU,CAAC;CAChD;;;;;;CAOD,gBAAgBC,QAAgB;EAC9B,QAAQ,IAAI,yDAAyD;EACrE,QAAQ,IAAI,OAAO,MAAM,CAAC;CAC3B;;;;;;;;;;CAWD,MAAM,aACJA,QACAC,kBACAC,WACAC,YACqC;EACrC,MAAM,YAAY,KAAK,SAAS,IAAI,CAAC,MAAM,EAAE,SAAS;EACtD,MAAM,mBAAmB,UAAU,KAAK,KAAK;EAC7C,MAAM,EAAE,CAAC,KAAK,cAAc,WAAW,KAAK,MAAM,GAChD,MAAM,KAAK,cAAc,KACvB;GACE;GACA;GACA;GACA;EACD,GACD,YAAY,UAAU,CACvB;EACH,MAAM,WAAY,KAAgB,MAAM,KAAK;AAC7C,SAAO,SACJ,OAAO,CAAC,aAAa,SAAS,MAAM,CAAC,CACrC,IAAI,CAAC,cAAc,EAAE,SAAU,GAAE;CACrC;;;;;;;;CASD,MAAM,gBACJC,YACAF,WACAC,YACA;EACA,MAAM,YAAY,KAAK,SAAS,IAAI,CAAC,MAAM,EAAE,SAAS;EACtD,MAAM,aAAa,aAAa;EAChC,MAAM,EAAE,CAAC,KAAK,oBAAoB,WAAW,KAAK,MAAM,GACtD,MAAM,KAAK,oBAAoB,KAC7B;GACE,YAAY,UAAU,KAAK,KAAK;GAChC,cAAc,OAAO,WAAW;GAChC;EACD,GACD,YAAY,UAAU,CACvB;EACH,MAAM,WAAY,KAAgB,MAAM,CAAC,MAAM,KAAK;EACpD,MAAM,sBAAsB,CAAE;AAC9B,OAAK,MAAM,cAAc,UAAU;GACjC,MAAM,YAAY,WAAW,MAAM,CAAC,MAAM,KAAK,EAAE;AACjD,OAAI,UAAU,WAAW,GAAG;IAC1B,MAAM,SAAS,UAAU,GAAG,MAAM;IAClC,MAAM,WAAW,UAAU,GAAG,MAAM;IACpC,oBAAoB,KAAK;KAAE;KAAQ;IAAU,EAAC;GAC/C;EACF;AACD,SAAO;CACR;;;;;;;CAQD,MAAM,YAAYE,OAAe,IAAI,GAAG;EACtC,MAAM,UAAU,MAAM,KAAK,YAAY,iBAAiB,OAAO,EAAE;AACjE,MAAI,CAAC,QACH,QAAO,CAAE;AAEX,SAAO,QAAQ,IAAI,CAAC,SAAS,OAAO,KAAK,SAAS,KAAK,CAAC;CACzD;;;;;;;;CASD,MAAM,YACJH,WACAI,MACAH,YACA;EACA,MAAM,UAAU,MAAM,KAAK,YAAY,UAAU;EACjD,MAAM,EAAE,CAAC,KAAK,eAAe,WAAW,KAAK,MAAM,GACjD,MAAM,KAAK,eAAe,KACxB;GACE;GACA,SAAS,QAAQ,KAAK,KAAK;GAC3B;EACD,GACD,YAAY,UAAU,CACvB;AACH,SAAO;CACR;CAED,MAAM,MACJ,EAAE,WAAW,YAAY,oBAAiC,EAC1DA,YACA;EACA,KAAK,WAAW,CAAE;EAClB,KAAK,gBAAgB;EACrB,MAAM,KAAK,QAAQ;GAAE,QAAQ;GAAK,UAAU;EAAW,EAAC;EAExD,IAAI,WAAW;AACf,SAAO,WAAW,KAAK,iBAAiB,KAAK,SAAS,SAAS,GAAG;GAChE,KAAK,eAAe;GAEpB,MAAM,OAAO,KAAK,SAAS,OAAO;GAClC,KAAK,cAAc,KAAK;GAExB,MAAM,SAAS,MAAM,KAAK,YACxB,WACA,KAAK,UACL,WACD;GACD,MAAM,aAAa,SAAS,KAAK,QAAQ,GAAG;GAC5C,KAAK,gBAAgB,OAAO;GAE5B,MAAM,KAAK,YAAY,aAAa,CAClC,IAAII,oCAAS;IACX,aAAa;IACb,UAAU,EAAE,MAAM,KAAK,SAAU;GAClC,EACF,EAAC;GAEF,MAAM,WAAW,MAAM,KAAK,aAC1B,QACA,KAAK,UACL,WACA,WACD;AACD,QAAK,MAAM,WAAW,UAAU;IAC9B,KAAK,iBAAiB;IACtB,QAAQ,SAAS,KAAK,cAAc,SAAS;IAC7C,MAAM,KAAK,QAAQ,QAAgB;GACpC;GACD,KAAK,WAAW,MAAM,KAAK,gBACzB,YACA,WACA,WACD;GAED,YAAY;EACb;AACD,SAAO,CAAE;CACV;CAED,YAAiC;AAC/B,QAAM,IAAI,MAAM;CACjB;;;;;;;;;;;;CAaD,OAAO,QAAQ,EACb,KACA,aACA,gBACA,SACA,UACA,GAAG,MAIkC,EAAE;EACvC,MAAM,gBAAgBC,wCAAkB,QAAQ;GAC9C;GACA;GACA;EACD,EAAC;EACF,MAAM,sBAAsBC,oDAAwB,QAAQ;GAC1D;GACA;GACA;EACD,EAAC;AACF,SAAO,IAAI,QAAQ;GACjB;GACA;GACA,gBACE,kBACAC,0CAAmB,QAAQ;IAAE;IAAK;IAAS;GAAW,EAAC;GACzD;GACA;GACA;GACA,GAAG;EACJ;CACF;AACF"}
+{"version":3,"file":"agent.cjs","names":["BaseChain","Document","TaskCreationChain","TaskPrioritizationChain","TaskExecutionChain"],"sources":["../../../src/experimental/babyagi/agent.ts"],"sourcesContent":["import type { BaseLanguageModelInterface } from \"@langchain/core/language_models/base\";\nimport type { VectorStoreInterface } from \"@langchain/core/vectorstores\";\nimport { Document } from \"@langchain/core/documents\";\nimport { ChainValues } from \"@langchain/core/utils/types\";\nimport { CallbackManagerForChainRun } from \"@langchain/core/callbacks/manager\";\nimport { BaseChain, ChainInputs } from \"../../chains/base.js\";\nimport { SerializedBaseChain } from \"../../chains/serde.js\";\nimport { Optional } from \"../../types/type-utils.js\";\nimport { TaskCreationChain } from \"./task_creation.js\";\nimport { TaskExecutionChain } from \"./task_execution.js\";\nimport { TaskPrioritizationChain } from \"./task_prioritization.js\";\n\n/**\n * Interface defining the structure of a task. A task has a `taskID` and a\n * `taskName`.\n */\nexport interface Task {\n  taskID: string;\n  taskName: string;\n}\n\n/**\n * Interface defining the structure of the inputs for the `BabyAGI` class.\n * It extends the `ChainInputs` interface, omitting the 'memory' and\n * 'callbackManager' properties, and adds properties specific to\n * `BabyAGI`.\n */\nexport interface BabyAGIInputs extends Omit<\n  ChainInputs,\n  \"memory\" | \"callbackManager\"\n> {\n  creationChain: BaseChain;\n  prioritizationChain: BaseChain;\n  executionChain: BaseChain;\n  vectorstore: VectorStoreInterface;\n  maxIterations?: number;\n}\n\n/**\n * Class responsible for managing tasks, including their creation,\n * prioritization, and execution. It uses three chains for these\n * operations: `creationChain`, `prioritizationChain`, and\n * `executionChain`.\n * @example\n * ```typescript\n * const babyAGI = BabyAGI.fromLLM({\n *   llm: new OpenAI({ temperature: 0 }),\n *   vectorstore: new MemoryVectorStore(new OpenAIEmbeddings()),\n *   maxIterations: 3,\n * });\n *\n * const result = await babyAGI.call({\n *   objective: \"Write a weather report for SF today\",\n * });\n * ```\n */\nexport class BabyAGI extends BaseChain implements BabyAGIInputs {\n  static lc_name() {\n    return \"BabyAGI\";\n  }\n\n  taskList: Task[];\n\n  creationChain: BaseChain;\n\n  prioritizationChain: BaseChain;\n\n  executionChain: BaseChain;\n\n  taskIDCounter: number;\n\n  vectorstore: VectorStoreInterface;\n\n  maxIterations: number;\n\n  constructor({\n    creationChain,\n    prioritizationChain,\n    executionChain,\n    vectorstore,\n    maxIterations = 100,\n    verbose,\n    callbacks,\n  }: BabyAGIInputs) {\n    super(undefined, verbose, callbacks);\n    this.taskList = [];\n    this.creationChain = creationChain;\n    this.prioritizationChain = prioritizationChain;\n    this.executionChain = executionChain;\n    this.taskIDCounter = 1;\n    this.vectorstore = vectorstore;\n    this.maxIterations = maxIterations;\n  }\n\n  _chainType() {\n    return \"BabyAGI\" as const;\n  }\n\n  get inputKeys() {\n    return [\"objective\", \"firstTask\"];\n  }\n\n  get outputKeys() {\n    return [];\n  }\n\n  /**\n   * Adds a task to the task list.\n   * @param task The task to be added.\n   * @returns Promise resolving to void.\n   */\n  async addTask(task: Task) {\n    this.taskList.push(task);\n  }\n\n  /**\n   * Prints the current task list to the console.\n   * @returns void\n   */\n  printTaskList() {\n    console.log(\"\\x1b[95m\\x1b[1m\\n*****TASK LIST*****\\n\\x1b[0m\\x1b[0m\");\n    for (const t of this.taskList) {\n      console.log(`${t.taskID}: ${t.taskName}`);\n    }\n  }\n\n  /**\n   * Prints the next task to the console.\n   * @param task The next task to be printed.\n   * @returns void\n   */\n  printNextTask(task: Task) {\n    console.log(\"\\x1b[92m\\x1b[1m\\n*****NEXT TASK*****\\n\\x1b[0m\\x1b[0m\");\n    console.log(`${task.taskID}: ${task.taskName}`);\n  }\n\n  /**\n   * Prints the result of a task to the console.\n   * @param result The result of the task.\n   * @returns void\n   */\n  printTaskResult(result: string) {\n    console.log(\"\\x1b[93m\\x1b[1m\\n*****TASK RESULT*****\\n\\x1b[0m\\x1b[0m\");\n    console.log(result.trim());\n  }\n\n  /**\n   * Generates the next tasks based on the result of the previous task, the\n   * task description, and the objective.\n   * @param result The result of the previous task.\n   * @param task_description The description of the task.\n   * @param objective The objective of the task.\n   * @param runManager Optional CallbackManagerForChainRun instance.\n   * @returns Promise resolving to an array of tasks without taskID.\n   */\n  async getNextTasks(\n    result: string,\n    task_description: string,\n    objective: string,\n    runManager?: CallbackManagerForChainRun\n  ): Promise<Optional<Task, \"taskID\">[]> {\n    const taskNames = this.taskList.map((t) => t.taskName);\n    const incomplete_tasks = taskNames.join(\", \");\n    const { [this.creationChain.outputKeys[0]]: text } =\n      await this.creationChain.call(\n        {\n          result,\n          task_description,\n          incomplete_tasks,\n          objective,\n        },\n        runManager?.getChild()\n      );\n    const newTasks = (text as string).split(\"\\n\");\n    return newTasks\n      .filter((taskName) => taskName.trim())\n      .map((taskName) => ({ taskName }));\n  }\n\n  /**\n   * Prioritizes the tasks based on the current task ID and the objective.\n   * @param thisTaskID The ID of the current task.\n   * @param objective The objective of the task.\n   * @param runManager Optional CallbackManagerForChainRun instance.\n   * @returns Promise resolving to an array of prioritized tasks.\n   */\n  async prioritizeTasks(\n    thisTaskID: number,\n    objective: string,\n    runManager?: CallbackManagerForChainRun\n  ) {\n    const taskNames = this.taskList.map((t) => t.taskName);\n    const nextTaskID = thisTaskID + 1;\n    const { [this.prioritizationChain.outputKeys[0]]: text } =\n      await this.prioritizationChain.call(\n        {\n          task_names: taskNames.join(\", \"),\n          next_task_id: String(nextTaskID),\n          objective,\n        },\n        runManager?.getChild()\n      );\n    const newTasks = (text as string).trim().split(\"\\n\");\n    const prioritizedTaskList = [];\n    for (const taskString of newTasks) {\n      const taskParts = taskString.trim().split(\".\", 2);\n      if (taskParts.length === 2) {\n        const taskID = taskParts[0].trim();\n        const taskName = taskParts[1].trim();\n        prioritizedTaskList.push({ taskID, taskName });\n      }\n    }\n    return prioritizedTaskList;\n  }\n\n  /**\n   * Retrieves the top tasks that are most similar to the given query.\n   * @param query The query to search for.\n   * @param k The number of top tasks to retrieve.\n   * @returns Promise resolving to an array of top tasks.\n   */\n  async getTopTasks(query: string, k = 5) {\n    const results = await this.vectorstore.similaritySearch(query, k);\n    if (!results) {\n      return [];\n    }\n    return results.map((item) => String(item.metadata.task));\n  }\n\n  /**\n   * Executes a task based on the objective and the task description.\n   * @param objective The objective of the task.\n   * @param task The task to be executed.\n   * @param runManager Optional CallbackManagerForChainRun instance.\n   * @returns Promise resolving to the result of the task execution as a string.\n   */\n  async executeTask(\n    objective: string,\n    task: string,\n    runManager?: CallbackManagerForChainRun\n  ) {\n    const context = await this.getTopTasks(objective);\n    const { [this.executionChain.outputKeys[0]]: text } =\n      await this.executionChain.call(\n        {\n          objective,\n          context: context.join(\"\\n\"),\n          task,\n        },\n        runManager?.getChild()\n      );\n    return text as string;\n  }\n\n  async _call(\n    { objective, firstTask = \"Make a todo list\" }: ChainValues,\n    runManager?: CallbackManagerForChainRun\n  ) {\n    this.taskList = [];\n    this.taskIDCounter = 1;\n    await this.addTask({ taskID: \"1\", taskName: firstTask });\n\n    let numIters = 0;\n    while (numIters < this.maxIterations && this.taskList.length > 0) {\n      this.printTaskList();\n\n      const task = this.taskList.shift()!;\n      this.printNextTask(task);\n\n      const result = await this.executeTask(\n        objective,\n        task.taskName,\n        runManager\n      );\n      const thisTaskID = parseInt(task.taskID, 10);\n      this.printTaskResult(result);\n\n      await this.vectorstore.addDocuments([\n        new Document({\n          pageContent: result,\n          metadata: { task: task.taskName },\n        }),\n      ]);\n\n      const newTasks = await this.getNextTasks(\n        result,\n        task.taskName,\n        objective,\n        runManager\n      );\n      for (const newTask of newTasks) {\n        this.taskIDCounter += 1;\n        newTask.taskID = this.taskIDCounter.toFixed();\n        await this.addTask(newTask as Task);\n      }\n      this.taskList = await this.prioritizeTasks(\n        thisTaskID,\n        objective,\n        runManager\n      );\n\n      numIters += 1;\n    }\n    return {};\n  }\n\n  serialize(): SerializedBaseChain {\n    throw new Error(\"Method not implemented.\");\n  }\n\n  /**\n   * Static method to create a new BabyAGI instance from a\n   * BaseLanguageModel.\n   * @param llm BaseLanguageModel instance used to generate a new BabyAGI instance.\n   * @param vectorstore VectorStore instance used to store and retrieve vectors.\n   * @param executionChain Optional BaseChain instance used to execute tasks.\n   * @param verbose Optional boolean indicating whether to log verbose output.\n   * @param callbacks Optional callbacks to be used during the execution of tasks.\n   * @param rest Optional additional parameters.\n   * @returns A new instance of BabyAGI.\n   */\n  static fromLLM({\n    llm,\n    vectorstore,\n    executionChain,\n    verbose,\n    callbacks,\n    ...rest\n  }: Optional<\n    BabyAGIInputs,\n    \"executionChain\" | \"creationChain\" | \"prioritizationChain\"\n  > & { llm: BaseLanguageModelInterface }) {\n    const creationChain = TaskCreationChain.fromLLM({\n      llm,\n      verbose,\n      callbacks,\n    });\n    const prioritizationChain = TaskPrioritizationChain.fromLLM({\n      llm,\n      verbose,\n      callbacks,\n    });\n    return new BabyAGI({\n      creationChain,\n      prioritizationChain,\n      executionChain:\n        executionChain ||\n        TaskExecutionChain.fromLLM({ llm, verbose, callbacks }),\n      vectorstore,\n      verbose,\n      callbacks,\n      ...rest,\n    });\n  }\n}\n"],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;;AAwDA,IAAa,UAAb,MAAa,gBAAgBA,uBAAmC;CAC9D,OAAO,UAAU;AACf,SAAO;;CAGT;CAEA;CAEA;CAEA;CAEA;CAEA;CAEA;CAEA,YAAY,EACV,eACA,qBACA,gBACA,aACA,gBAAgB,KAChB,SACA,aACgB;AAChB,QAAM,QAAW,SAAS,UAAU;AACpC,OAAK,WAAW,EAAE;AAClB,OAAK,gBAAgB;AACrB,OAAK,sBAAsB;AAC3B,OAAK,iBAAiB;AACtB,OAAK,gBAAgB;AACrB,OAAK,cAAc;AACnB,OAAK,gBAAgB;;CAGvB,aAAa;AACX,SAAO;;CAGT,IAAI,YAAY;AACd,SAAO,CAAC,aAAa,YAAY;;CAGnC,IAAI,aAAa;AACf,SAAO,EAAE;;;;;;;CAQX,MAAM,QAAQ,MAAY;AACxB,OAAK,SAAS,KAAK,KAAK;;;;;;CAO1B,gBAAgB;AACd,UAAQ,IAAI,uDAAuD;AACnE,OAAK,MAAM,KAAK,KAAK,SACnB,SAAQ,IAAI,GAAG,EAAE,OAAO,IAAI,EAAE,WAAW;;;;;;;CAS7C,cAAc,MAAY;AACxB,UAAQ,IAAI,uDAAuD;AACnE,UAAQ,IAAI,GAAG,KAAK,OAAO,IAAI,KAAK,WAAW;;;;;;;CAQjD,gBAAgB,QAAgB;AAC9B,UAAQ,IAAI,yDAAyD;AACrE,UAAQ,IAAI,OAAO,MAAM,CAAC;;;;;;;;;;;CAY5B,MAAM,aACJ,QACA,kBACA,WACA,YACqC;EAErC,MAAM,mBADY,KAAK,SAAS,KAAK,MAAM,EAAE,SAAS,CACnB,KAAK,KAAK;EAC7C,MAAM,GAAG,KAAK,cAAc,WAAW,KAAK,SAC1C,MAAM,KAAK,cAAc,KACvB;GACE;GACA;GACA;GACA;GACD,EACD,YAAY,UAAU,CACvB;AAEH,SADkB,KAAgB,MAAM,KAAK,CAE1C,QAAQ,aAAa,SAAS,MAAM,CAAC,CACrC,KAAK,cAAc,EAAE,UAAU,EAAE;;;;;;;;;CAUtC,MAAM,gBACJ,YACA,WACA,YACA;EACA,MAAM,YAAY,KAAK,SAAS,KAAK,MAAM,EAAE,SAAS;EACtD,MAAM,aAAa,aAAa;EAChC,MAAM,GAAG,KAAK,oBAAoB,WAAW,KAAK,SAChD,MAAM,KAAK,oBAAoB,KAC7B;GACE,YAAY,UAAU,KAAK,KAAK;GAChC,cAAc,OAAO,WAAW;GAChC;GACD,EACD,YAAY,UAAU,CACvB;EACH,MAAM,WAAY,KAAgB,MAAM,CAAC,MAAM,KAAK;EACpD,MAAM,sBAAsB,EAAE;AAC9B,OAAK,MAAM,cAAc,UAAU;GACjC,MAAM,YAAY,WAAW,MAAM,CAAC,MAAM,KAAK,EAAE;AACjD,OAAI,UAAU,WAAW,GAAG;IAC1B,MAAM,SAAS,UAAU,GAAG,MAAM;IAClC,MAAM,WAAW,UAAU,GAAG,MAAM;AACpC,wBAAoB,KAAK;KAAE;KAAQ;KAAU,CAAC;;;AAGlD,SAAO;;;;;;;;CAST,MAAM,YAAY,OAAe,IAAI,GAAG;EACtC,MAAM,UAAU,MAAM,KAAK,YAAY,iBAAiB,OAAO,EAAE;AACjE,MAAI,CAAC,QACH,QAAO,EAAE;AAEX,SAAO,QAAQ,KAAK,SAAS,OAAO,KAAK,SAAS,KAAK,CAAC;;;;;;;;;CAU1D,MAAM,YACJ,WACA,MACA,YACA;EACA,MAAM,UAAU,MAAM,KAAK,YAAY,UAAU;EACjD,MAAM,GAAG,KAAK,eAAe,WAAW,KAAK,SAC3C,MAAM,KAAK,eAAe,KACxB;GACE;GACA,SAAS,QAAQ,KAAK,KAAK;GAC3B;GACD,EACD,YAAY,UAAU,CACvB;AACH,SAAO;;CAGT,MAAM,MACJ,EAAE,WAAW,YAAY,sBACzB,YACA;AACA,OAAK,WAAW,EAAE;AAClB,OAAK,gBAAgB;AACrB,QAAM,KAAK,QAAQ;GAAE,QAAQ;GAAK,UAAU;GAAW,CAAC;EAExD,IAAI,WAAW;AACf,SAAO,WAAW,KAAK,iBAAiB,KAAK,SAAS,SAAS,GAAG;AAChE,QAAK,eAAe;GAEpB,MAAM,OAAO,KAAK,SAAS,OAAO;AAClC,QAAK,cAAc,KAAK;GAExB,MAAM,SAAS,MAAM,KAAK,YACxB,WACA,KAAK,UACL,WACD;GACD,MAAM,aAAa,SAAS,KAAK,QAAQ,GAAG;AAC5C,QAAK,gBAAgB,OAAO;AAE5B,SAAM,KAAK,YAAY,aAAa,CAClC,IAAIC,mCAAS;IACX,aAAa;IACb,UAAU,EAAE,MAAM,KAAK,UAAU;IAClC,CAAC,CACH,CAAC;GAEF,MAAM,WAAW,MAAM,KAAK,aAC1B,QACA,KAAK,UACL,WACA,WACD;AACD,QAAK,MAAM,WAAW,UAAU;AAC9B,SAAK,iBAAiB;AACtB,YAAQ,SAAS,KAAK,cAAc,SAAS;AAC7C,UAAM,KAAK,QAAQ,QAAgB;;AAErC,QAAK,WAAW,MAAM,KAAK,gBACzB,YACA,WACA,WACD;AAED,eAAY;;AAEd,SAAO,EAAE;;CAGX,YAAiC;AAC/B,QAAM,IAAI,MAAM,0BAA0B;;;;;;;;;;;;;CAc5C,OAAO,QAAQ,EACb,KACA,aACA,gBACA,SACA,WACA,GAAG,QAIoC;AAWvC,SAAO,IAAI,QAAQ;GACjB,eAXoBC,wCAAkB,QAAQ;IAC9C;IACA;IACA;IACD,CAAC;GAQA,qBAP0BC,oDAAwB,QAAQ;IAC1D;IACA;IACA;IACD,CAAC;GAIA,gBACE,kBACAC,0CAAmB,QAAQ;IAAE;IAAK;IAAS;IAAW,CAAC;GACzD;GACA;GACA;GACA,GAAG;GACJ,CAAC"}

package/dist/experimental/babyagi/agent.d.cts

@@ -7,7 +7,6 @@ import { CallbackManagerForChainRun } from "@langchain/core/callbacks/manager";
 import { VectorStoreInterface } from "@langchain/core/vectorstores";
 
 //#region src/experimental/babyagi/agent.d.ts
-
 /**
  * Interface defining the structure of a task. A task has a `taskID` and a
  * `taskName`.

package/dist/experimental/babyagi/agent.d.cts.map

@@ -1 +1 @@
-{"version":3,"file":"agent.d.cts","names":["BaseLanguageModelInterface","VectorStoreInterface","ChainValues","CallbackManagerForChainRun","BaseChain","ChainInputs","SerializedBaseChain","Optional","Task","BabyAGIInputs","Omit","BabyAGI","creationChain","prioritizationChain","executionChain","vectorstore","maxIterations","verbose","callbacks","Promise","objective","firstTask","llm"],"sources":["../../../src/experimental/babyagi/agent.d.ts"],"sourcesContent":["import type { BaseLanguageModelInterface } from \"@langchain/core/language_models/base\";\nimport type { VectorStoreInterface } from \"@langchain/core/vectorstores\";\nimport { ChainValues } from \"@langchain/core/utils/types\";\nimport { CallbackManagerForChainRun } from \"@langchain/core/callbacks/manager\";\nimport { BaseChain, ChainInputs } from \"../../chains/base.js\";\nimport { SerializedBaseChain } from \"../../chains/serde.js\";\nimport { Optional } from \"../../types/type-utils.js\";\n/**\n * Interface defining the structure of a task. A task has a `taskID` and a\n * `taskName`.\n */\nexport interface Task {\n    taskID: string;\n    taskName: string;\n}\n/**\n * Interface defining the structure of the inputs for the `BabyAGI` class.\n * It extends the `ChainInputs` interface, omitting the 'memory' and\n * 'callbackManager' properties, and adds properties specific to\n * `BabyAGI`.\n */\nexport interface BabyAGIInputs extends Omit<ChainInputs, \"memory\" | \"callbackManager\"> {\n    creationChain: BaseChain;\n    prioritizationChain: BaseChain;\n    executionChain: BaseChain;\n    vectorstore: VectorStoreInterface;\n    maxIterations?: number;\n}\n/**\n * Class responsible for managing tasks, including their creation,\n * prioritization, and execution. It uses three chains for these\n * operations: `creationChain`, `prioritizationChain`, and\n * `executionChain`.\n * @example\n * ```typescript\n * const babyAGI = BabyAGI.fromLLM({\n *   llm: new OpenAI({ temperature: 0 }),\n *   vectorstore: new MemoryVectorStore(new OpenAIEmbeddings()),\n *   maxIterations: 3,\n * });\n *\n * const result = await babyAGI.call({\n *   objective: \"Write a weather report for SF today\",\n * });\n * ```\n */\nexport declare class BabyAGI extends BaseChain implements BabyAGIInputs {\n    static lc_name(): string;\n    taskList: Task[];\n    creationChain: BaseChain;\n    prioritizationChain: BaseChain;\n    executionChain: BaseChain;\n    taskIDCounter: number;\n    vectorstore: VectorStoreInterface;\n    maxIterations: number;\n    constructor({ creationChain, prioritizationChain, executionChain, vectorstore, maxIterations, verbose, callbacks }: BabyAGIInputs);\n    _chainType(): \"BabyAGI\";\n    get inputKeys(): string[];\n    get outputKeys(): never[];\n    /**\n     * Adds a task to the task list.\n     * @param task The task to be added.\n     * @returns Promise resolving to void.\n     */\n    addTask(task: Task): Promise<void>;\n    /**\n     * Prints the current task list to the console.\n     * @returns void\n     */\n    printTaskList(): void;\n    /**\n     * Prints the next task to the console.\n     * @param task The next task to be printed.\n     * @returns void\n     */\n    printNextTask(task: Task): void;\n    /**\n     * Prints the result of a task to the console.\n     * @param result The result of the task.\n     * @returns void\n     */\n    printTaskResult(result: string): void;\n    /**\n     * Generates the next tasks based on the result of the previous task, the\n     * task description, and the objective.\n     * @param result The result of the previous task.\n     * @param task_description The description of the task.\n     * @param objective The objective of the task.\n     * @param runManager Optional CallbackManagerForChainRun instance.\n     * @returns Promise resolving to an array of tasks without taskID.\n     */\n    getNextTasks(result: string, task_description: string, objective: string, runManager?: CallbackManagerForChainRun): Promise<Optional<Task, \"taskID\">[]>;\n    /**\n     * Prioritizes the tasks based on the current task ID and the objective.\n     * @param thisTaskID The ID of the current task.\n     * @param objective The objective of the task.\n     * @param runManager Optional CallbackManagerForChainRun instance.\n     * @returns Promise resolving to an array of prioritized tasks.\n     */\n    prioritizeTasks(thisTaskID: number, objective: string, runManager?: CallbackManagerForChainRun): Promise<{\n        taskID: string;\n        taskName: string;\n    }[]>;\n    /**\n     * Retrieves the top tasks that are most similar to the given query.\n     * @param query The query to search for.\n     * @param k The number of top tasks to retrieve.\n     * @returns Promise resolving to an array of top tasks.\n     */\n    getTopTasks(query: string, k?: number): Promise<string[]>;\n    /**\n     * Executes a task based on the objective and the task description.\n     * @param objective The objective of the task.\n     * @param task The task to be executed.\n     * @param runManager Optional CallbackManagerForChainRun instance.\n     * @returns Promise resolving to the result of the task execution as a string.\n     */\n    executeTask(objective: string, task: string, runManager?: CallbackManagerForChainRun): Promise<string>;\n    _call({ objective, firstTask }: ChainValues, runManager?: CallbackManagerForChainRun): Promise<{}>;\n    serialize(): SerializedBaseChain;\n    /**\n     * Static method to create a new BabyAGI instance from a\n     * BaseLanguageModel.\n     * @param llm BaseLanguageModel instance used to generate a new BabyAGI instance.\n     * @param vectorstore VectorStore instance used to store and retrieve vectors.\n     * @param executionChain Optional BaseChain instance used to execute tasks.\n     * @param verbose Optional boolean indicating whether to log verbose output.\n     * @param callbacks Optional callbacks to be used during the execution of tasks.\n     * @param rest Optional additional parameters.\n     * @returns A new instance of BabyAGI.\n     */\n    static fromLLM({ llm, vectorstore, executionChain, verbose, callbacks, ...rest }: Optional<BabyAGIInputs, \"executionChain\" | \"creationChain\" | \"prioritizationChain\"> & {\n        llm: BaseLanguageModelInterface;\n    }): BabyAGI;\n}\n//# sourceMappingURL=agent.d.ts.map"],"mappings":";;;;;;;;;;;;AAWA;AAUA;AAA4CK,UAV3BG,IAAAA,CAU2BH;EACzBD,MAAAA,EAAAA,MAAAA;EACMA,QAAAA,EAAAA,MAAAA;;;;AAFkB;AAyB3C;;;AAIyBA,UA7BRK,aAAAA,SAAsBC,IA6BdN,CA7BmBC,WA6BnBD,EAAAA,QAAAA,GAAAA,iBAAAA,CAAAA,CAAAA;EACLA,aAAAA,EA7BDA,SA6BCA;EAEHH,mBAAAA,EA9BQG,SA8BRH;EAECW,cAAAA,EA/BER,SA+BFQ;EAAeC,WAAAA,EA9BhBZ,oBA8BgBY;EAAqBC,aAAAA,CAAAA,EAAAA,MAAAA;;;;;;;;;;;;;;;;;;;;AA+DlBZ,cAxEfS,OAAAA,SAAgBP,SAAAA,YAAqBK,aAwEtBP,CAAAA;EAA0BC,OAAAA,OAAAA,CAAAA,CAAAA,EAAAA,MAAAA;EAA6BgB,QAAAA,EAtE7EX,IAsE6EW,EAAAA;EAC1Eb,aAAAA,EAtEEF,SAsEFE;EAYIgB,mBAAAA,EAjFIlB,SAiFJkB;EAAKP,cAAAA,EAhFNX,SAgFMW;EAAaD,aAAAA,EAAAA,MAAAA;EAAgBG,WAAAA,EA9EtChB,oBA8EsCgB;EAASC,aAAAA,EAAAA,MAAAA;EAA+BT,WAAAA,CAAAA;IAAAA,aAAAA;IAAAA,mBAAAA;IAAAA,cAAAA;IAAAA,WAAAA;IAAAA,aAAAA;IAAAA,OAAAA;IAAAA;EAAAA,CAAAA,EA5EyBA,aA4EzBA;EAATF,UAAAA,CAAAA,CAAAA,EAAAA,SAAAA;EACzEP,IAAAA,SAAAA,CAAAA,CAAAA,EAAAA,MAAAA,EAAAA;EACLW,IAAAA,UAAAA,CAAAA,CAAAA,EAAAA,KAAAA,EAAAA;EAvF6BP;;AAAkC;;;gBAkBrDI,OAAOW;;;;;;;;;;;sBAWDX;;;;;;;;;;;;;;;;yFAgBmEL,6BAA6BgB,QAAQZ,SAASC;;;;;;;;sEAQjEL,6BAA6BgB;;;;;;;;;;0CAUzDA;;;;;;;;4DAQkBhB,6BAA6BgB;;;;KACvDjB,0BAA0BC,6BAA6BgB;eAC1Eb;;;;;;;;;;;;;;;;;;;KAYqEC,SAASE;SAClFT;MACLW"}
+{"version":3,"file":"agent.d.cts","names":[],"sources":["../../../src/experimental/babyagi/agent.ts"],"mappings":";;;;;;;;;;;AAgBA;;UAAiB,IAAA;EACf,MAAA;EACA,QAAA;AAAA;;;;;;;UASe,aAAA,SAAsB,IAAA,CACrC,WAAA;EAGA,aAAA,EAAe,SAAA;EACf,mBAAA,EAAqB,SAAA;EACrB,cAAA,EAAgB,SAAA;EAChB,WAAA,EAAa,oBAAA;EACb,aAAA;AAAA;;;;;;;;;;;AAqBF;;;;;;;;cAAa,OAAA,SAAgB,SAAA,YAAqB,aAAA;EAAA,OACzC,OAAA,CAAA;EAIP,QAAA,EAAU,IAAA;EAEV,aAAA,EAAe,SAAA;EAEf,mBAAA,EAAqB,SAAA;EAErB,cAAA,EAAgB,SAAA;EAEhB,aAAA;EAEA,WAAA,EAAa,oBAAA;EAEb,aAAA;EAEA,WAAA,CAAA;IACE,aAAA;IACA,mBAAA;IACA,cAAA;IACA,WAAA;IACA,aAAA;IACA,OAAA;IACA;EAAA,GACC,aAAA;EAWH,UAAA,CAAA;EAAA,IAII,SAAA,CAAA;EAAA,IAIA,UAAA,CAAA;EA0DO;;;;;EAjDL,OAAA,CAAQ,IAAA,EAAM,IAAA,GAAI,OAAA;EAgIiB;;;;EAxHzC,aAAA,CAAA;EAyIyC;;;;;EA7HzC,aAAA,CAAc,IAAA,EAAM,IAAA;EAmMlB;;;;;EAzLF,eAAA,CAAgB,MAAA;EArFgC;;;;;;;;;EAmG1C,YAAA,CACJ,MAAA,UACA,gBAAA,UACA,SAAA,UACA,UAAA,GAAa,0BAAA,GACZ,OAAA,CAAQ,QAAA,CAAS,IAAA;EA/FpB;;;;;;;EAyHM,eAAA,CACJ,UAAA,UACA,SAAA,UACA,UAAA,GAAa,0BAAA,GAA0B,OAAA;;;;EAhHvC;;;;;;EAgJI,WAAA,CAAY,KAAA,UAAe,CAAA,YAAK,OAAA;EA7IpC;;;;;;;EA4JI,WAAA,CACJ,SAAA,UACA,IAAA,UACA,UAAA,GAAa,0BAAA,GAA0B,OAAA;EAenC,KAAA,CAAA;IACF,SAAA;IAAW;EAAA,GAAkC,WAAA,EAC/C,UAAA,GAAa,0BAAA,GAA0B,OAAA;EAkDzC,SAAA,CAAA,GAAa,mBAAA;EAnMO;;;;;;;;;;;EAAA,OAkNb,OAAA,CAAA;IACL,GAAA;IACA,WAAA;IACA,cAAA;IACA,OAAA;IACA,SAAA;IAAA,GACG;EAAA,GACF,QAAA,CACD,aAAA;IAEI,GAAA,EAAK,0BAAA;EAAA,IAA4B,OAAA;AAAA"}

package/dist/experimental/babyagi/agent.d.ts

@@ -7,7 +7,6 @@ import { ChainValues } from "@langchain/core/utils/types";
 import { VectorStoreInterface } from "@langchain/core/vectorstores";
 
 //#region src/experimental/babyagi/agent.d.ts
-
 /**
  * Interface defining the structure of a task. A task has a `taskID` and a
  * `taskName`.

package/dist/experimental/babyagi/agent.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"agent.d.ts","names":["BaseLanguageModelInterface","VectorStoreInterface","ChainValues","CallbackManagerForChainRun","BaseChain","ChainInputs","SerializedBaseChain","Optional","Task","BabyAGIInputs","Omit","BabyAGI","creationChain","prioritizationChain","executionChain","vectorstore","maxIterations","verbose","callbacks","Promise","objective","firstTask","llm"],"sources":["../../../src/experimental/babyagi/agent.d.ts"],"sourcesContent":["import type { BaseLanguageModelInterface } from \"@langchain/core/language_models/base\";\nimport type { VectorStoreInterface } from \"@langchain/core/vectorstores\";\nimport { ChainValues } from \"@langchain/core/utils/types\";\nimport { CallbackManagerForChainRun } from \"@langchain/core/callbacks/manager\";\nimport { BaseChain, ChainInputs } from \"../../chains/base.js\";\nimport { SerializedBaseChain } from \"../../chains/serde.js\";\nimport { Optional } from \"../../types/type-utils.js\";\n/**\n * Interface defining the structure of a task. A task has a `taskID` and a\n * `taskName`.\n */\nexport interface Task {\n    taskID: string;\n    taskName: string;\n}\n/**\n * Interface defining the structure of the inputs for the `BabyAGI` class.\n * It extends the `ChainInputs` interface, omitting the 'memory' and\n * 'callbackManager' properties, and adds properties specific to\n * `BabyAGI`.\n */\nexport interface BabyAGIInputs extends Omit<ChainInputs, \"memory\" | \"callbackManager\"> {\n    creationChain: BaseChain;\n    prioritizationChain: BaseChain;\n    executionChain: BaseChain;\n    vectorstore: VectorStoreInterface;\n    maxIterations?: number;\n}\n/**\n * Class responsible for managing tasks, including their creation,\n * prioritization, and execution. It uses three chains for these\n * operations: `creationChain`, `prioritizationChain`, and\n * `executionChain`.\n * @example\n * ```typescript\n * const babyAGI = BabyAGI.fromLLM({\n *   llm: new OpenAI({ temperature: 0 }),\n *   vectorstore: new MemoryVectorStore(new OpenAIEmbeddings()),\n *   maxIterations: 3,\n * });\n *\n * const result = await babyAGI.call({\n *   objective: \"Write a weather report for SF today\",\n * });\n * ```\n */\nexport declare class BabyAGI extends BaseChain implements BabyAGIInputs {\n    static lc_name(): string;\n    taskList: Task[];\n    creationChain: BaseChain;\n    prioritizationChain: BaseChain;\n    executionChain: BaseChain;\n    taskIDCounter: number;\n    vectorstore: VectorStoreInterface;\n    maxIterations: number;\n    constructor({ creationChain, prioritizationChain, executionChain, vectorstore, maxIterations, verbose, callbacks }: BabyAGIInputs);\n    _chainType(): \"BabyAGI\";\n    get inputKeys(): string[];\n    get outputKeys(): never[];\n    /**\n     * Adds a task to the task list.\n     * @param task The task to be added.\n     * @returns Promise resolving to void.\n     */\n    addTask(task: Task): Promise<void>;\n    /**\n     * Prints the current task list to the console.\n     * @returns void\n     */\n    printTaskList(): void;\n    /**\n     * Prints the next task to the console.\n     * @param task The next task to be printed.\n     * @returns void\n     */\n    printNextTask(task: Task): void;\n    /**\n     * Prints the result of a task to the console.\n     * @param result The result of the task.\n     * @returns void\n     */\n    printTaskResult(result: string): void;\n    /**\n     * Generates the next tasks based on the result of the previous task, the\n     * task description, and the objective.\n     * @param result The result of the previous task.\n     * @param task_description The description of the task.\n     * @param objective The objective of the task.\n     * @param runManager Optional CallbackManagerForChainRun instance.\n     * @returns Promise resolving to an array of tasks without taskID.\n     */\n    getNextTasks(result: string, task_description: string, objective: string, runManager?: CallbackManagerForChainRun): Promise<Optional<Task, \"taskID\">[]>;\n    /**\n     * Prioritizes the tasks based on the current task ID and the objective.\n     * @param thisTaskID The ID of the current task.\n     * @param objective The objective of the task.\n     * @param runManager Optional CallbackManagerForChainRun instance.\n     * @returns Promise resolving to an array of prioritized tasks.\n     */\n    prioritizeTasks(thisTaskID: number, objective: string, runManager?: CallbackManagerForChainRun): Promise<{\n        taskID: string;\n        taskName: string;\n    }[]>;\n    /**\n     * Retrieves the top tasks that are most similar to the given query.\n     * @param query The query to search for.\n     * @param k The number of top tasks to retrieve.\n     * @returns Promise resolving to an array of top tasks.\n     */\n    getTopTasks(query: string, k?: number): Promise<string[]>;\n    /**\n     * Executes a task based on the objective and the task description.\n     * @param objective The objective of the task.\n     * @param task The task to be executed.\n     * @param runManager Optional CallbackManagerForChainRun instance.\n     * @returns Promise resolving to the result of the task execution as a string.\n     */\n    executeTask(objective: string, task: string, runManager?: CallbackManagerForChainRun): Promise<string>;\n    _call({ objective, firstTask }: ChainValues, runManager?: CallbackManagerForChainRun): Promise<{}>;\n    serialize(): SerializedBaseChain;\n    /**\n     * Static method to create a new BabyAGI instance from a\n     * BaseLanguageModel.\n     * @param llm BaseLanguageModel instance used to generate a new BabyAGI instance.\n     * @param vectorstore VectorStore instance used to store and retrieve vectors.\n     * @param executionChain Optional BaseChain instance used to execute tasks.\n     * @param verbose Optional boolean indicating whether to log verbose output.\n     * @param callbacks Optional callbacks to be used during the execution of tasks.\n     * @param rest Optional additional parameters.\n     * @returns A new instance of BabyAGI.\n     */\n    static fromLLM({ llm, vectorstore, executionChain, verbose, callbacks, ...rest }: Optional<BabyAGIInputs, \"executionChain\" | \"creationChain\" | \"prioritizationChain\"> & {\n        llm: BaseLanguageModelInterface;\n    }): BabyAGI;\n}\n//# sourceMappingURL=agent.d.ts.map"],"mappings":";;;;;;;;;;;;AAWA;AAUA;AAA4CK,UAV3BG,IAAAA,CAU2BH;EACzBD,MAAAA,EAAAA,MAAAA;EACMA,QAAAA,EAAAA,MAAAA;;;;AAFkB;AAyB3C;;;AAIyBA,UA7BRK,aAAAA,SAAsBC,IA6BdN,CA7BmBC,WA6BnBD,EAAAA,QAAAA,GAAAA,iBAAAA,CAAAA,CAAAA;EACLA,aAAAA,EA7BDA,SA6BCA;EAEHH,mBAAAA,EA9BQG,SA8BRH;EAECW,cAAAA,EA/BER,SA+BFQ;EAAeC,WAAAA,EA9BhBZ,oBA8BgBY;EAAqBC,aAAAA,CAAAA,EAAAA,MAAAA;;;;;;;;;;;;;;;;;;;;AA+DlBZ,cAxEfS,OAAAA,SAAgBP,SAAAA,YAAqBK,aAwEtBP,CAAAA;EAA0BC,OAAAA,OAAAA,CAAAA,CAAAA,EAAAA,MAAAA;EAA6BgB,QAAAA,EAtE7EX,IAsE6EW,EAAAA;EAC1Eb,aAAAA,EAtEEF,SAsEFE;EAYIgB,mBAAAA,EAjFIlB,SAiFJkB;EAAKP,cAAAA,EAhFNX,SAgFMW;EAAaD,aAAAA,EAAAA,MAAAA;EAAgBG,WAAAA,EA9EtChB,oBA8EsCgB;EAASC,aAAAA,EAAAA,MAAAA;EAA+BT,WAAAA,CAAAA;IAAAA,aAAAA;IAAAA,mBAAAA;IAAAA,cAAAA;IAAAA,WAAAA;IAAAA,aAAAA;IAAAA,OAAAA;IAAAA;EAAAA,CAAAA,EA5EyBA,aA4EzBA;EAATF,UAAAA,CAAAA,CAAAA,EAAAA,SAAAA;EACzEP,IAAAA,SAAAA,CAAAA,CAAAA,EAAAA,MAAAA,EAAAA;EACLW,IAAAA,UAAAA,CAAAA,CAAAA,EAAAA,KAAAA,EAAAA;EAvF6BP;;AAAkC;;;gBAkBrDI,OAAOW;;;;;;;;;;;sBAWDX;;;;;;;;;;;;;;;;yFAgBmEL,6BAA6BgB,QAAQZ,SAASC;;;;;;;;sEAQjEL,6BAA6BgB;;;;;;;;;;0CAUzDA;;;;;;;;4DAQkBhB,6BAA6BgB;;;;KACvDjB,0BAA0BC,6BAA6BgB;eAC1Eb;;;;;;;;;;;;;;;;;;;KAYqEC,SAASE;SAClFT;MACLW"}
+{"version":3,"file":"agent.d.ts","names":[],"sources":["../../../src/experimental/babyagi/agent.ts"],"mappings":";;;;;;;;;;;AAgBA;;UAAiB,IAAA;EACf,MAAA;EACA,QAAA;AAAA;;;;;;;UASe,aAAA,SAAsB,IAAA,CACrC,WAAA;EAGA,aAAA,EAAe,SAAA;EACf,mBAAA,EAAqB,SAAA;EACrB,cAAA,EAAgB,SAAA;EAChB,WAAA,EAAa,oBAAA;EACb,aAAA;AAAA;;;;;;;;;;;AAqBF;;;;;;;;cAAa,OAAA,SAAgB,SAAA,YAAqB,aAAA;EAAA,OACzC,OAAA,CAAA;EAIP,QAAA,EAAU,IAAA;EAEV,aAAA,EAAe,SAAA;EAEf,mBAAA,EAAqB,SAAA;EAErB,cAAA,EAAgB,SAAA;EAEhB,aAAA;EAEA,WAAA,EAAa,oBAAA;EAEb,aAAA;EAEA,WAAA,CAAA;IACE,aAAA;IACA,mBAAA;IACA,cAAA;IACA,WAAA;IACA,aAAA;IACA,OAAA;IACA;EAAA,GACC,aAAA;EAWH,UAAA,CAAA;EAAA,IAII,SAAA,CAAA;EAAA,IAIA,UAAA,CAAA;EA0DO;;;;;EAjDL,OAAA,CAAQ,IAAA,EAAM,IAAA,GAAI,OAAA;EAgIiB;;;;EAxHzC,aAAA,CAAA;EAyIyC;;;;;EA7HzC,aAAA,CAAc,IAAA,EAAM,IAAA;EAmMlB;;;;;EAzLF,eAAA,CAAgB,MAAA;EArFgC;;;;;;;;;EAmG1C,YAAA,CACJ,MAAA,UACA,gBAAA,UACA,SAAA,UACA,UAAA,GAAa,0BAAA,GACZ,OAAA,CAAQ,QAAA,CAAS,IAAA;EA/FpB;;;;;;;EAyHM,eAAA,CACJ,UAAA,UACA,SAAA,UACA,UAAA,GAAa,0BAAA,GAA0B,OAAA;;;;EAhHvC;;;;;;EAgJI,WAAA,CAAY,KAAA,UAAe,CAAA,YAAK,OAAA;EA7IpC;;;;;;;EA4JI,WAAA,CACJ,SAAA,UACA,IAAA,UACA,UAAA,GAAa,0BAAA,GAA0B,OAAA;EAenC,KAAA,CAAA;IACF,SAAA;IAAW;EAAA,GAAkC,WAAA,EAC/C,UAAA,GAAa,0BAAA,GAA0B,OAAA;EAkDzC,SAAA,CAAA,GAAa,mBAAA;EAnMO;;;;;;;;;;;EAAA,OAkNb,OAAA,CAAA;IACL,GAAA;IACA,WAAA;IACA,cAAA;IACA,OAAA;IACA,SAAA;IAAA,GACG;EAAA,GACF,QAAA,CACD,aAAA;IAEI,GAAA,EAAK,0BAAA;EAAA,IAA4B,OAAA;AAAA"}