@langchain/classic 1.0.17 → 1.0.18

package/dist/agents/chat_convo/index.js

@@ -1,9 +1,9 @@
+import { DEFAULT_PREFIX, DEFAULT_SUFFIX, PREFIX_END, TEMPLATE_TOOL_RESPONSE } from "./prompt.js";
 import { Agent } from "../agent.js";
 import { LLMChain } from "../../chains/llm_chain.js";
-import { DEFAULT_PREFIX, DEFAULT_SUFFIX, PREFIX_END, TEMPLATE_TOOL_RESPONSE } from "./prompt.js";
 import { ChatConversationalAgentOutputParserWithRetries } from "./outputParser.js";
-import { ChatPromptTemplate, HumanMessagePromptTemplate, MessagesPlaceholder, SystemMessagePromptTemplate, renderTemplate } from "@langchain/core/prompts";
 import { AIMessage, HumanMessage } from "@langchain/core/messages";
+import { ChatPromptTemplate, HumanMessagePromptTemplate, MessagesPlaceholder, SystemMessagePromptTemplate, renderTemplate } from "@langchain/core/prompts";
 
 //#region src/agents/chat_convo/index.ts
 /**
@@ -84,10 +84,8 @@ var ChatConversationalAgent = class ChatConversationalAgent extends Agent {
 		const humanMessage = args?.humanMessage ?? DEFAULT_SUFFIX;
 		const toolStrings = tools.map((tool) => `${tool.name}: ${tool.description}`).join("\n");
 		const toolNames = tools.map((tool) => tool.name);
-		const outputParser = args?.outputParser ?? ChatConversationalAgent.getDefaultOutputParser({ toolNames });
-		const formatInstructions = outputParser.getFormatInstructions({ toolNames });
 		const renderedHumanMessage = renderTemplate(humanMessage, "f-string", {
-			format_instructions: formatInstructions,
+			format_instructions: (args?.outputParser ?? ChatConversationalAgent.getDefaultOutputParser({ toolNames })).getFormatInstructions({ toolNames }),
 			tools: toolStrings
 		});
 		const messages = [
@@ -113,17 +111,15 @@ var ChatConversationalAgent = class ChatConversationalAgent extends Agent {
 			llm,
 			toolNames: tools.map((tool) => tool.name)
 		});
-		const prompt = ChatConversationalAgent.createPrompt(tools, {
-			...args,
-			outputParser
-		});
-		const chain = new LLMChain({
-			prompt,
-			llm,
-			callbacks: args?.callbacks ?? args?.callbackManager
-		});
 		return new ChatConversationalAgent({
-			llmChain: chain,
+			llmChain: new LLMChain({
+				prompt: ChatConversationalAgent.createPrompt(tools, {
+					...args,
+					outputParser
+				}),
+				llm,
+				callbacks: args?.callbacks ?? args?.callbackManager
+			}),
 			outputParser,
 			allowedTools: tools.map((t) => t.name)
 		});

package/dist/agents/chat_convo/index.js.map

@@ -1 +1 @@
-{"version":3,"file":"index.js","names":["input: ChatConversationalAgentInput","tools: ToolInterface[]","steps: AgentStep[]","thoughts: BaseMessage[]","fields?: OutputParserArgs & {\n      toolNames: string[];\n    }","args?: ChatConversationalCreatePromptArgs","llm: BaseLanguageModelInterface","args?: ChatConversationalCreatePromptArgs & AgentArgs"],"sources":["../../../src/agents/chat_convo/index.ts"],"sourcesContent":["import type { BaseLanguageModelInterface } from \"@langchain/core/language_models/base\";\nimport type { ToolInterface } from \"@langchain/core/tools\";\nimport {\n  ChatPromptTemplate,\n  HumanMessagePromptTemplate,\n  MessagesPlaceholder,\n  SystemMessagePromptTemplate,\n  renderTemplate,\n} from \"@langchain/core/prompts\";\nimport type { AgentStep } from \"@langchain/core/agents\";\nimport {\n  type BaseMessage,\n  HumanMessage,\n  AIMessage,\n} from \"@langchain/core/messages\";\nimport { LLMChain } from \"../../chains/llm_chain.js\";\nimport { Optional } from \"../../types/type-utils.js\";\nimport { Agent, AgentArgs, OutputParserArgs } from \"../agent.js\";\nimport { AgentActionOutputParser, AgentInput } from \"../types.js\";\nimport { ChatConversationalAgentOutputParserWithRetries } from \"./outputParser.js\";\nimport {\n  PREFIX_END,\n  DEFAULT_PREFIX,\n  DEFAULT_SUFFIX,\n  TEMPLATE_TOOL_RESPONSE,\n} from \"./prompt.js\";\n\n/**\n * Interface defining the structure of arguments used to create a prompt\n * for the ChatConversationalAgent class.\n */\nexport interface ChatConversationalCreatePromptArgs {\n  /** String to put after the list of tools. */\n  systemMessage?: string;\n  /** String to put before the list of tools. */\n  humanMessage?: string;\n  /** List of input variables the final prompt will expect. */\n  inputVariables?: string[];\n  /** Output parser to use for formatting. */\n  outputParser?: AgentActionOutputParser;\n}\n\n/**\n * Type that extends the AgentInput interface for the\n * ChatConversationalAgent class, making the outputParser property\n * optional.\n */\nexport type ChatConversationalAgentInput = Optional<AgentInput, \"outputParser\">;\n\n/**\n * Agent for the MRKL chain.\n * @augments Agent\n */\nexport class ChatConversationalAgent extends Agent {\n  static lc_name() {\n    return \"ChatConversationalAgent\";\n  }\n\n  lc_namespace = [\"langchain\", \"agents\", \"chat_convo\"];\n\n  declare ToolType: ToolInterface;\n\n  constructor(input: ChatConversationalAgentInput) {\n    const outputParser =\n      input.outputParser ?? ChatConversationalAgent.getDefaultOutputParser();\n    super({ ...input, outputParser });\n  }\n\n  _agentType() {\n    return \"chat-conversational-react-description\" as const;\n  }\n\n  observationPrefix() {\n    return \"Observation: \";\n  }\n\n  llmPrefix() {\n    return \"Thought:\";\n  }\n\n  _stop(): string[] {\n    return [\"Observation:\"];\n  }\n\n  static validateTools(tools: ToolInterface[]) {\n    const descriptionlessTool = tools.find((tool) => !tool.description);\n    if (descriptionlessTool) {\n      const msg =\n        `Got a tool ${descriptionlessTool.name} without a description.` +\n        ` This agent requires descriptions for all tools.`;\n      throw new Error(msg);\n    }\n  }\n\n  /**\n   * Constructs the agent scratchpad based on the agent steps. It returns an\n   * array of base messages representing the thoughts of the agent.\n   * @param steps The agent steps to construct the scratchpad from.\n   * @returns An array of base messages representing the thoughts of the agent.\n   */\n  async constructScratchPad(steps: AgentStep[]): Promise<BaseMessage[]> {\n    const thoughts: BaseMessage[] = [];\n    for (const step of steps) {\n      thoughts.push(new AIMessage(step.action.log));\n      thoughts.push(\n        new HumanMessage(\n          renderTemplate(TEMPLATE_TOOL_RESPONSE, \"f-string\", {\n            observation: step.observation,\n          })\n        )\n      );\n    }\n    return thoughts;\n  }\n\n  /**\n   * Returns the default output parser for the ChatConversationalAgent\n   * class. It takes optional fields as arguments to customize the output\n   * parser.\n   * @param fields Optional fields to customize the output parser.\n   * @returns The default output parser for the ChatConversationalAgent class.\n   */\n  static getDefaultOutputParser(\n    fields?: OutputParserArgs & {\n      toolNames: string[];\n    }\n  ): AgentActionOutputParser {\n    if (fields?.llm) {\n      return ChatConversationalAgentOutputParserWithRetries.fromLLM(\n        fields.llm,\n        {\n          toolNames: fields.toolNames,\n        }\n      );\n    }\n    return new ChatConversationalAgentOutputParserWithRetries({\n      toolNames: fields?.toolNames,\n    });\n  }\n\n  /**\n   * Create prompt in the style of the ChatConversationAgent.\n   *\n   * @param tools - List of tools the agent will have access to, used to format the prompt.\n   * @param args - Arguments to create the prompt with.\n   * @param args.systemMessage - String to put before the list of tools.\n   * @param args.humanMessage - String to put after the list of tools.\n   * @param args.outputParser - Output parser to use for formatting.\n   */\n  static createPrompt(\n    tools: ToolInterface[],\n    args?: ChatConversationalCreatePromptArgs\n  ) {\n    const systemMessage = (args?.systemMessage ?? DEFAULT_PREFIX) + PREFIX_END;\n    const humanMessage = args?.humanMessage ?? DEFAULT_SUFFIX;\n    const toolStrings = tools\n      .map((tool) => `${tool.name}: ${tool.description}`)\n      .join(\"\\n\");\n    const toolNames = tools.map((tool) => tool.name);\n    const outputParser =\n      args?.outputParser ??\n      ChatConversationalAgent.getDefaultOutputParser({ toolNames });\n    const formatInstructions = outputParser.getFormatInstructions({\n      toolNames,\n    });\n    const renderedHumanMessage = renderTemplate(humanMessage, \"f-string\", {\n      format_instructions: formatInstructions,\n      tools: toolStrings,\n    });\n    const messages = [\n      SystemMessagePromptTemplate.fromTemplate(systemMessage),\n      new MessagesPlaceholder(\"chat_history\"),\n      HumanMessagePromptTemplate.fromTemplate(renderedHumanMessage),\n      new MessagesPlaceholder(\"agent_scratchpad\"),\n    ];\n    return ChatPromptTemplate.fromMessages(messages);\n  }\n\n  /**\n   * Creates an instance of the ChatConversationalAgent class from a\n   * BaseLanguageModel and a set of tools. It takes optional arguments to\n   * customize the agent.\n   * @param llm The BaseLanguageModel to create the agent from.\n   * @param tools The set of tools to create the agent from.\n   * @param args Optional arguments to customize the agent.\n   * @returns An instance of the ChatConversationalAgent class.\n   */\n  static fromLLMAndTools(\n    llm: BaseLanguageModelInterface,\n    tools: ToolInterface[],\n    args?: ChatConversationalCreatePromptArgs & AgentArgs\n  ) {\n    ChatConversationalAgent.validateTools(tools);\n    const outputParser =\n      args?.outputParser ??\n      ChatConversationalAgent.getDefaultOutputParser({\n        llm,\n        toolNames: tools.map((tool) => tool.name),\n      });\n    const prompt = ChatConversationalAgent.createPrompt(tools, {\n      ...args,\n      outputParser,\n    });\n    const chain = new LLMChain({\n      prompt,\n      llm,\n      callbacks: args?.callbacks ?? args?.callbackManager,\n    });\n    return new ChatConversationalAgent({\n      llmChain: chain,\n      outputParser,\n      allowedTools: tools.map((t) => t.name),\n    });\n  }\n}\n"],"mappings":";;;;;;;;;;;;AAqDA,IAAa,0BAAb,MAAa,gCAAgC,MAAM;CACjD,OAAO,UAAU;AACf,SAAO;CACR;CAED,eAAe;EAAC;EAAa;EAAU;CAAa;CAIpD,YAAYA,OAAqC;EAC/C,MAAM,eACJ,MAAM,gBAAgB,wBAAwB,wBAAwB;EACxE,MAAM;GAAE,GAAG;GAAO;EAAc,EAAC;CAClC;CAED,aAAa;AACX,SAAO;CACR;CAED,oBAAoB;AAClB,SAAO;CACR;CAED,YAAY;AACV,SAAO;CACR;CAED,QAAkB;AAChB,SAAO,CAAC,cAAe;CACxB;CAED,OAAO,cAAcC,OAAwB;EAC3C,MAAM,sBAAsB,MAAM,KAAK,CAAC,SAAS,CAAC,KAAK,YAAY;AACnE,MAAI,qBAAqB;GACvB,MAAM,MACJ,CAAC,WAAW,EAAE,oBAAoB,KAAK,uEAAuB,CACZ;AACpD,SAAM,IAAI,MAAM;EACjB;CACF;;;;;;;CAQD,MAAM,oBAAoBC,OAA4C;EACpE,MAAMC,WAA0B,CAAE;AAClC,OAAK,MAAM,QAAQ,OAAO;GACxB,SAAS,KAAK,IAAI,UAAU,KAAK,OAAO,KAAK;GAC7C,SAAS,KACP,IAAI,aACF,eAAe,wBAAwB,YAAY,EACjD,aAAa,KAAK,YACnB,EAAC,EAEL;EACF;AACD,SAAO;CACR;;;;;;;;CASD,OAAO,uBACLC,QAGyB;AACzB,MAAI,QAAQ,IACV,QAAO,+CAA+C,QACpD,OAAO,KACP,EACE,WAAW,OAAO,UACnB,EACF;AAEH,SAAO,IAAI,+CAA+C,EACxD,WAAW,QAAQ,UACpB;CACF;;;;;;;;;;CAWD,OAAO,aACLH,OACAI,MACA;EACA,MAAM,iBAAiB,MAAM,iBAAiB,kBAAkB;EAChE,MAAM,eAAe,MAAM,gBAAgB;EAC3C,MAAM,cAAc,MACjB,IAAI,CAAC,SAAS,GAAG,KAAK,KAAK,EAAE,EAAE,KAAK,aAAa,CAAC,CAClD,KAAK,KAAK;EACb,MAAM,YAAY,MAAM,IAAI,CAAC,SAAS,KAAK,KAAK;EAChD,MAAM,eACJ,MAAM,gBACN,wBAAwB,uBAAuB,EAAE,UAAW,EAAC;EAC/D,MAAM,qBAAqB,aAAa,sBAAsB,EAC5D,UACD,EAAC;EACF,MAAM,uBAAuB,eAAe,cAAc,YAAY;GACpE,qBAAqB;GACrB,OAAO;EACR,EAAC;EACF,MAAM,WAAW;GACf,4BAA4B,aAAa,cAAc;GACvD,IAAI,oBAAoB;GACxB,2BAA2B,aAAa,qBAAqB;GAC7D,IAAI,oBAAoB;EACzB;AACD,SAAO,mBAAmB,aAAa,SAAS;CACjD;;;;;;;;;;CAWD,OAAO,gBACLC,KACAL,OACAM,MACA;EACA,wBAAwB,cAAc,MAAM;EAC5C,MAAM,eACJ,MAAM,gBACN,wBAAwB,uBAAuB;GAC7C;GACA,WAAW,MAAM,IAAI,CAAC,SAAS,KAAK,KAAK;EAC1C,EAAC;EACJ,MAAM,SAAS,wBAAwB,aAAa,OAAO;GACzD,GAAG;GACH;EACD,EAAC;EACF,MAAM,QAAQ,IAAI,SAAS;GACzB;GACA;GACA,WAAW,MAAM,aAAa,MAAM;EACrC;AACD,SAAO,IAAI,wBAAwB;GACjC,UAAU;GACV;GACA,cAAc,MAAM,IAAI,CAAC,MAAM,EAAE,KAAK;EACvC;CACF;AACF"}
+{"version":3,"file":"index.js","names":[],"sources":["../../../src/agents/chat_convo/index.ts"],"sourcesContent":["import type { BaseLanguageModelInterface } from \"@langchain/core/language_models/base\";\nimport type { ToolInterface } from \"@langchain/core/tools\";\nimport {\n  ChatPromptTemplate,\n  HumanMessagePromptTemplate,\n  MessagesPlaceholder,\n  SystemMessagePromptTemplate,\n  renderTemplate,\n} from \"@langchain/core/prompts\";\nimport type { AgentStep } from \"@langchain/core/agents\";\nimport {\n  type BaseMessage,\n  HumanMessage,\n  AIMessage,\n} from \"@langchain/core/messages\";\nimport { LLMChain } from \"../../chains/llm_chain.js\";\nimport { Optional } from \"../../types/type-utils.js\";\nimport { Agent, AgentArgs, OutputParserArgs } from \"../agent.js\";\nimport { AgentActionOutputParser, AgentInput } from \"../types.js\";\nimport { ChatConversationalAgentOutputParserWithRetries } from \"./outputParser.js\";\nimport {\n  PREFIX_END,\n  DEFAULT_PREFIX,\n  DEFAULT_SUFFIX,\n  TEMPLATE_TOOL_RESPONSE,\n} from \"./prompt.js\";\n\n/**\n * Interface defining the structure of arguments used to create a prompt\n * for the ChatConversationalAgent class.\n */\nexport interface ChatConversationalCreatePromptArgs {\n  /** String to put after the list of tools. */\n  systemMessage?: string;\n  /** String to put before the list of tools. */\n  humanMessage?: string;\n  /** List of input variables the final prompt will expect. */\n  inputVariables?: string[];\n  /** Output parser to use for formatting. */\n  outputParser?: AgentActionOutputParser;\n}\n\n/**\n * Type that extends the AgentInput interface for the\n * ChatConversationalAgent class, making the outputParser property\n * optional.\n */\nexport type ChatConversationalAgentInput = Optional<AgentInput, \"outputParser\">;\n\n/**\n * Agent for the MRKL chain.\n * @augments Agent\n */\nexport class ChatConversationalAgent extends Agent {\n  static lc_name() {\n    return \"ChatConversationalAgent\";\n  }\n\n  lc_namespace = [\"langchain\", \"agents\", \"chat_convo\"];\n\n  declare ToolType: ToolInterface;\n\n  constructor(input: ChatConversationalAgentInput) {\n    const outputParser =\n      input.outputParser ?? ChatConversationalAgent.getDefaultOutputParser();\n    super({ ...input, outputParser });\n  }\n\n  _agentType() {\n    return \"chat-conversational-react-description\" as const;\n  }\n\n  observationPrefix() {\n    return \"Observation: \";\n  }\n\n  llmPrefix() {\n    return \"Thought:\";\n  }\n\n  _stop(): string[] {\n    return [\"Observation:\"];\n  }\n\n  static validateTools(tools: ToolInterface[]) {\n    const descriptionlessTool = tools.find((tool) => !tool.description);\n    if (descriptionlessTool) {\n      const msg =\n        `Got a tool ${descriptionlessTool.name} without a description.` +\n        ` This agent requires descriptions for all tools.`;\n      throw new Error(msg);\n    }\n  }\n\n  /**\n   * Constructs the agent scratchpad based on the agent steps. It returns an\n   * array of base messages representing the thoughts of the agent.\n   * @param steps The agent steps to construct the scratchpad from.\n   * @returns An array of base messages representing the thoughts of the agent.\n   */\n  async constructScratchPad(steps: AgentStep[]): Promise<BaseMessage[]> {\n    const thoughts: BaseMessage[] = [];\n    for (const step of steps) {\n      thoughts.push(new AIMessage(step.action.log));\n      thoughts.push(\n        new HumanMessage(\n          renderTemplate(TEMPLATE_TOOL_RESPONSE, \"f-string\", {\n            observation: step.observation,\n          })\n        )\n      );\n    }\n    return thoughts;\n  }\n\n  /**\n   * Returns the default output parser for the ChatConversationalAgent\n   * class. It takes optional fields as arguments to customize the output\n   * parser.\n   * @param fields Optional fields to customize the output parser.\n   * @returns The default output parser for the ChatConversationalAgent class.\n   */\n  static getDefaultOutputParser(\n    fields?: OutputParserArgs & {\n      toolNames: string[];\n    }\n  ): AgentActionOutputParser {\n    if (fields?.llm) {\n      return ChatConversationalAgentOutputParserWithRetries.fromLLM(\n        fields.llm,\n        {\n          toolNames: fields.toolNames,\n        }\n      );\n    }\n    return new ChatConversationalAgentOutputParserWithRetries({\n      toolNames: fields?.toolNames,\n    });\n  }\n\n  /**\n   * Create prompt in the style of the ChatConversationAgent.\n   *\n   * @param tools - List of tools the agent will have access to, used to format the prompt.\n   * @param args - Arguments to create the prompt with.\n   * @param args.systemMessage - String to put before the list of tools.\n   * @param args.humanMessage - String to put after the list of tools.\n   * @param args.outputParser - Output parser to use for formatting.\n   */\n  static createPrompt(\n    tools: ToolInterface[],\n    args?: ChatConversationalCreatePromptArgs\n  ) {\n    const systemMessage = (args?.systemMessage ?? DEFAULT_PREFIX) + PREFIX_END;\n    const humanMessage = args?.humanMessage ?? DEFAULT_SUFFIX;\n    const toolStrings = tools\n      .map((tool) => `${tool.name}: ${tool.description}`)\n      .join(\"\\n\");\n    const toolNames = tools.map((tool) => tool.name);\n    const outputParser =\n      args?.outputParser ??\n      ChatConversationalAgent.getDefaultOutputParser({ toolNames });\n    const formatInstructions = outputParser.getFormatInstructions({\n      toolNames,\n    });\n    const renderedHumanMessage = renderTemplate(humanMessage, \"f-string\", {\n      format_instructions: formatInstructions,\n      tools: toolStrings,\n    });\n    const messages = [\n      SystemMessagePromptTemplate.fromTemplate(systemMessage),\n      new MessagesPlaceholder(\"chat_history\"),\n      HumanMessagePromptTemplate.fromTemplate(renderedHumanMessage),\n      new MessagesPlaceholder(\"agent_scratchpad\"),\n    ];\n    return ChatPromptTemplate.fromMessages(messages);\n  }\n\n  /**\n   * Creates an instance of the ChatConversationalAgent class from a\n   * BaseLanguageModel and a set of tools. It takes optional arguments to\n   * customize the agent.\n   * @param llm The BaseLanguageModel to create the agent from.\n   * @param tools The set of tools to create the agent from.\n   * @param args Optional arguments to customize the agent.\n   * @returns An instance of the ChatConversationalAgent class.\n   */\n  static fromLLMAndTools(\n    llm: BaseLanguageModelInterface,\n    tools: ToolInterface[],\n    args?: ChatConversationalCreatePromptArgs & AgentArgs\n  ) {\n    ChatConversationalAgent.validateTools(tools);\n    const outputParser =\n      args?.outputParser ??\n      ChatConversationalAgent.getDefaultOutputParser({\n        llm,\n        toolNames: tools.map((tool) => tool.name),\n      });\n    const prompt = ChatConversationalAgent.createPrompt(tools, {\n      ...args,\n      outputParser,\n    });\n    const chain = new LLMChain({\n      prompt,\n      llm,\n      callbacks: args?.callbacks ?? args?.callbackManager,\n    });\n    return new ChatConversationalAgent({\n      llmChain: chain,\n      outputParser,\n      allowedTools: tools.map((t) => t.name),\n    });\n  }\n}\n"],"mappings":";;;;;;;;;;;;AAqDA,IAAa,0BAAb,MAAa,gCAAgC,MAAM;CACjD,OAAO,UAAU;AACf,SAAO;;CAGT,eAAe;EAAC;EAAa;EAAU;EAAa;CAIpD,YAAY,OAAqC;EAC/C,MAAM,eACJ,MAAM,gBAAgB,wBAAwB,wBAAwB;AACxE,QAAM;GAAE,GAAG;GAAO;GAAc,CAAC;;CAGnC,aAAa;AACX,SAAO;;CAGT,oBAAoB;AAClB,SAAO;;CAGT,YAAY;AACV,SAAO;;CAGT,QAAkB;AAChB,SAAO,CAAC,eAAe;;CAGzB,OAAO,cAAc,OAAwB;EAC3C,MAAM,sBAAsB,MAAM,MAAM,SAAS,CAAC,KAAK,YAAY;AACnE,MAAI,qBAAqB;GACvB,MAAM,MACJ,cAAc,oBAAoB,KAAK;AAEzC,SAAM,IAAI,MAAM,IAAI;;;;;;;;;CAUxB,MAAM,oBAAoB,OAA4C;EACpE,MAAM,WAA0B,EAAE;AAClC,OAAK,MAAM,QAAQ,OAAO;AACxB,YAAS,KAAK,IAAI,UAAU,KAAK,OAAO,IAAI,CAAC;AAC7C,YAAS,KACP,IAAI,aACF,eAAe,wBAAwB,YAAY,EACjD,aAAa,KAAK,aACnB,CAAC,CACH,CACF;;AAEH,SAAO;;;;;;;;;CAUT,OAAO,uBACL,QAGyB;AACzB,MAAI,QAAQ,IACV,QAAO,+CAA+C,QACpD,OAAO,KACP,EACE,WAAW,OAAO,WACnB,CACF;AAEH,SAAO,IAAI,+CAA+C,EACxD,WAAW,QAAQ,WACpB,CAAC;;;;;;;;;;;CAYJ,OAAO,aACL,OACA,MACA;EACA,MAAM,iBAAiB,MAAM,iBAAiB,kBAAkB;EAChE,MAAM,eAAe,MAAM,gBAAgB;EAC3C,MAAM,cAAc,MACjB,KAAK,SAAS,GAAG,KAAK,KAAK,IAAI,KAAK,cAAc,CAClD,KAAK,KAAK;EACb,MAAM,YAAY,MAAM,KAAK,SAAS,KAAK,KAAK;EAOhD,MAAM,uBAAuB,eAAe,cAAc,YAAY;GACpE,sBANA,MAAM,gBACN,wBAAwB,uBAAuB,EAAE,WAAW,CAAC,EACvB,sBAAsB,EAC5D,WACD,CAAC;GAGA,OAAO;GACR,CAAC;EACF,MAAM,WAAW;GACf,4BAA4B,aAAa,cAAc;GACvD,IAAI,oBAAoB,eAAe;GACvC,2BAA2B,aAAa,qBAAqB;GAC7D,IAAI,oBAAoB,mBAAmB;GAC5C;AACD,SAAO,mBAAmB,aAAa,SAAS;;;;;;;;;;;CAYlD,OAAO,gBACL,KACA,OACA,MACA;AACA,0BAAwB,cAAc,MAAM;EAC5C,MAAM,eACJ,MAAM,gBACN,wBAAwB,uBAAuB;GAC7C;GACA,WAAW,MAAM,KAAK,SAAS,KAAK,KAAK;GAC1C,CAAC;AAUJ,SAAO,IAAI,wBAAwB;GACjC,UANY,IAAI,SAAS;IACzB,QALa,wBAAwB,aAAa,OAAO;KACzD,GAAG;KACH;KACD,CAAC;IAGA;IACA,WAAW,MAAM,aAAa,MAAM;IACrC,CAAC;GAGA;GACA,cAAc,MAAM,KAAK,MAAM,EAAE,KAAK;GACvC,CAAC"}

package/dist/agents/chat_convo/outputParser.cjs

@@ -1,9 +1,9 @@
-const require_rolldown_runtime = require('../../_virtual/rolldown_runtime.cjs');
-const require_types = require('../types.cjs');
+const require_runtime = require('../../_virtual/_rolldown/runtime.cjs');
 const require_prompt = require('./prompt.cjs');
+const require_types = require('../types.cjs');
 const require_fix = require('../../output_parsers/fix.cjs');
-const __langchain_core_prompts = require_rolldown_runtime.__toESM(require("@langchain/core/prompts"));
-const __langchain_core_output_parsers = require_rolldown_runtime.__toESM(require("@langchain/core/output_parsers"));
+let _langchain_core_prompts = require("@langchain/core/prompts");
+let _langchain_core_output_parsers = require("@langchain/core/output_parsers");
 
 //#region src/agents/chat_convo/outputParser.ts
 /**
@@ -33,16 +33,14 @@ var ChatConversationalAgentOutputParser = class extends require_types.AgentActio
 		if (jsonOutput.includes("```json") || jsonOutput.includes("```")) {
 			const testString = jsonOutput.includes("```json") ? "```json" : "```";
 			const firstIndex = jsonOutput.indexOf(testString);
-			const actionInputIndex = jsonOutput.indexOf("action_input");
-			if (actionInputIndex > firstIndex) {
+			if (jsonOutput.indexOf("action_input") > firstIndex) {
 				jsonOutput = jsonOutput.slice(firstIndex + testString.length).trimStart();
 				const lastIndex = jsonOutput.lastIndexOf("```");
 				if (lastIndex !== -1) jsonOutput = jsonOutput.slice(0, lastIndex).trimEnd();
 			}
 		}
 		try {
-			const response = JSON.parse(jsonOutput);
-			const { action, action_input } = response;
+			const { action, action_input } = JSON.parse(jsonOutput);
 			if (action === "Final Answer") return {
 				returnValues: { output: action_input },
 				log: text
@@ -53,7 +51,7 @@ var ChatConversationalAgentOutputParser = class extends require_types.AgentActio
 				log: text
 			};
 		} catch (e) {
-			throw new __langchain_core_output_parsers.OutputParserException(`Failed to parse. Text: "${text}". Error: ${e}`);
+			throw new _langchain_core_output_parsers.OutputParserException(`Failed to parse. Text: "${text}". Error: ${e}`);
 		}
 	}
 	/**
@@ -63,7 +61,7 @@ var ChatConversationalAgentOutputParser = class extends require_types.AgentActio
 	* @returns Format instructions as a string.
 	*/
 	getFormatInstructions() {
-		return (0, __langchain_core_prompts.renderTemplate)(require_prompt.FORMAT_INSTRUCTIONS, "f-string", { tool_names: this.toolNames.join(", ") });
+		return (0, _langchain_core_prompts.renderTemplate)(require_prompt.FORMAT_INSTRUCTIONS, "f-string", { tool_names: this.toolNames.join(", ") });
 	}
 };
 /**
@@ -93,7 +91,7 @@ var ChatConversationalAgentOutputParserWithRetries = class ChatConversationalAge
 	*/
 	getFormatInstructions(options) {
 		if (options.raw) return require_prompt.FORMAT_INSTRUCTIONS;
-		return (0, __langchain_core_prompts.renderTemplate)(require_prompt.FORMAT_INSTRUCTIONS, "f-string", { tool_names: options.toolNames.join(", ") });
+		return (0, _langchain_core_prompts.renderTemplate)(require_prompt.FORMAT_INSTRUCTIONS, "f-string", { tool_names: options.toolNames.join(", ") });
 	}
 	/**
 	* Parses the given text into an AgentAction or AgentFinish object.
@@ -116,10 +114,9 @@ var ChatConversationalAgentOutputParserWithRetries = class ChatConversationalAge
 	*/
 	static fromLLM(llm, options) {
 		const baseParser = options.baseParser ?? new ChatConversationalAgentOutputParser({ toolNames: options.toolNames ?? [] });
-		const outputFixingParser = require_fix.OutputFixingParser.fromLLM(llm, baseParser);
 		return new ChatConversationalAgentOutputParserWithRetries({
 			baseParser,
-			outputFixingParser,
+			outputFixingParser: require_fix.OutputFixingParser.fromLLM(llm, baseParser),
 			toolNames: options.toolNames
 		});
 	}

package/dist/agents/chat_convo/outputParser.cjs.map

@@ -1 +1 @@
-{"version":3,"file":"outputParser.cjs","names":["AgentActionOutputParser","fields: { toolNames: string[] }","text: string","OutputParserException","FORMAT_INSTRUCTIONS","fields: ChatConversationalAgentOutputParserArgs","options: ChatConversationalAgentOutputParserFormatInstructionsOptions","llm: BaseLanguageModelInterface","options: Omit<ChatConversationalAgentOutputParserArgs, \"outputFixingParser\">","OutputFixingParser"],"sources":["../../../src/agents/chat_convo/outputParser.ts"],"sourcesContent":["import type { BaseLanguageModelInterface } from \"@langchain/core/language_models/base\";\nimport type { AgentAction, AgentFinish } from \"@langchain/core/agents\";\nimport {\n  FormatInstructionsOptions,\n  OutputParserException,\n} from \"@langchain/core/output_parsers\";\nimport { renderTemplate } from \"@langchain/core/prompts\";\nimport { AgentActionOutputParser } from \"../types.js\";\nimport { FORMAT_INSTRUCTIONS } from \"./prompt.js\";\nimport { OutputFixingParser } from \"../../output_parsers/fix.js\";\n\nexport type ChatConversationalAgentOutputParserFormatInstructionsOptions =\n  FormatInstructionsOptions & {\n    toolNames: string[];\n    raw?: boolean;\n  };\n\n/**\n * Class that represents an output parser for the ChatConversationalAgent\n * class. It extends the AgentActionOutputParser class and provides\n * methods for parsing the output of the MRKL chain into agent actions.\n */\nexport class ChatConversationalAgentOutputParser extends AgentActionOutputParser {\n  lc_namespace = [\"langchain\", \"agents\", \"chat_convo\"];\n\n  private toolNames: string[];\n\n  constructor(fields: { toolNames: string[] }) {\n    super(...arguments);\n    this.toolNames = fields.toolNames;\n  }\n\n  /**\n   * Parses the given text into an AgentAction or AgentFinish object. If an\n   * output fixing parser is defined, uses it to parse the text.\n   * @param text Text to parse.\n   * @returns Promise that resolves to an AgentAction or AgentFinish object.\n   */\n  async parse(text: string): Promise<AgentAction | AgentFinish> {\n    let jsonOutput = text.trim();\n    if (jsonOutput.includes(\"```json\") || jsonOutput.includes(\"```\")) {\n      const testString = jsonOutput.includes(\"```json\") ? \"```json\" : \"```\";\n      const firstIndex = jsonOutput.indexOf(testString);\n      const actionInputIndex = jsonOutput.indexOf(\"action_input\");\n      if (actionInputIndex > firstIndex) {\n        jsonOutput = jsonOutput\n          .slice(firstIndex + testString.length)\n          .trimStart();\n        const lastIndex = jsonOutput.lastIndexOf(\"```\");\n        if (lastIndex !== -1) {\n          jsonOutput = jsonOutput.slice(0, lastIndex).trimEnd();\n        }\n      }\n    }\n\n    try {\n      const response = JSON.parse(jsonOutput);\n\n      const { action, action_input } = response;\n\n      if (action === \"Final Answer\") {\n        return { returnValues: { output: action_input }, log: text };\n      }\n      return { tool: action, toolInput: action_input, log: text };\n    } catch (e) {\n      throw new OutputParserException(\n        `Failed to parse. Text: \"${text}\". Error: ${e}`\n      );\n    }\n  }\n\n  /**\n   * Returns the format instructions as a string. If the 'raw' option is\n   * true, returns the raw FORMAT_INSTRUCTIONS.\n   * @param options Options for getting the format instructions.\n   * @returns Format instructions as a string.\n   */\n  getFormatInstructions(): string {\n    return renderTemplate(FORMAT_INSTRUCTIONS, \"f-string\", {\n      tool_names: this.toolNames.join(\", \"),\n    });\n  }\n}\n\nexport type ChatConversationalAgentOutputParserArgs = {\n  baseParser?: ChatConversationalAgentOutputParser;\n  outputFixingParser?: OutputFixingParser<AgentAction | AgentFinish>;\n  toolNames?: string[];\n};\n\n/**\n * Class that represents an output parser with retries for the\n * ChatConversationalAgent class. It extends the AgentActionOutputParser\n * class and provides methods for parsing the output of the MRKL chain\n * into agent actions with retry functionality.\n */\nexport class ChatConversationalAgentOutputParserWithRetries extends AgentActionOutputParser {\n  lc_namespace = [\"langchain\", \"agents\", \"chat_convo\"];\n\n  private baseParser: ChatConversationalAgentOutputParser;\n\n  private outputFixingParser?: OutputFixingParser<AgentAction | AgentFinish>;\n\n  private toolNames: string[] = [];\n\n  constructor(fields: ChatConversationalAgentOutputParserArgs) {\n    super(fields);\n    this.toolNames = fields.toolNames ?? this.toolNames;\n    this.baseParser =\n      fields?.baseParser ??\n      new ChatConversationalAgentOutputParser({ toolNames: this.toolNames });\n    this.outputFixingParser = fields?.outputFixingParser;\n  }\n\n  /**\n   * Returns the format instructions as a string.\n   * @returns Format instructions as a string.\n   */\n  getFormatInstructions(\n    options: ChatConversationalAgentOutputParserFormatInstructionsOptions\n  ): string {\n    if (options.raw) {\n      return FORMAT_INSTRUCTIONS;\n    }\n    return renderTemplate(FORMAT_INSTRUCTIONS, \"f-string\", {\n      tool_names: options.toolNames.join(\", \"),\n    });\n  }\n\n  /**\n   * Parses the given text into an AgentAction or AgentFinish object.\n   * @param text Text to parse.\n   * @returns Promise that resolves to an AgentAction or AgentFinish object.\n   */\n  async parse(text: string): Promise<AgentAction | AgentFinish> {\n    if (this.outputFixingParser !== undefined) {\n      return this.outputFixingParser.parse(text);\n    }\n    return this.baseParser.parse(text);\n  }\n\n  /**\n   * Static method to create a new\n   * ChatConversationalAgentOutputParserWithRetries from a BaseLanguageModelInterface\n   * and options. If no base parser is provided in the options, a new\n   * ChatConversationalAgentOutputParser is created. An OutputFixingParser\n   * is also created from the BaseLanguageModelInterface and the base parser.\n   * @param llm BaseLanguageModelInterface instance used to create the OutputFixingParser.\n   * @param options Options for creating the ChatConversationalAgentOutputParserWithRetries instance.\n   * @returns A new instance of ChatConversationalAgentOutputParserWithRetries.\n   */\n  static fromLLM(\n    llm: BaseLanguageModelInterface,\n    options: Omit<ChatConversationalAgentOutputParserArgs, \"outputFixingParser\">\n  ): ChatConversationalAgentOutputParserWithRetries {\n    const baseParser =\n      options.baseParser ??\n      new ChatConversationalAgentOutputParser({\n        toolNames: options.toolNames ?? [],\n      });\n    const outputFixingParser = OutputFixingParser.fromLLM(llm, baseParser);\n    return new ChatConversationalAgentOutputParserWithRetries({\n      baseParser,\n      outputFixingParser,\n      toolNames: options.toolNames,\n    });\n  }\n}\n"],"mappings":";;;;;;;;;;;;;AAsBA,IAAa,sCAAb,cAAyDA,sCAAwB;CAC/E,eAAe;EAAC;EAAa;EAAU;CAAa;CAEpD,AAAQ;CAER,YAAYC,QAAiC;EAC3C,MAAM,GAAG,UAAU;EACnB,KAAK,YAAY,OAAO;CACzB;;;;;;;CAQD,MAAM,MAAMC,MAAkD;EAC5D,IAAI,aAAa,KAAK,MAAM;AAC5B,MAAI,WAAW,SAAS,UAAU,IAAI,WAAW,SAAS,MAAM,EAAE;GAChE,MAAM,aAAa,WAAW,SAAS,UAAU,GAAG,YAAY;GAChE,MAAM,aAAa,WAAW,QAAQ,WAAW;GACjD,MAAM,mBAAmB,WAAW,QAAQ,eAAe;AAC3D,OAAI,mBAAmB,YAAY;IACjC,aAAa,WACV,MAAM,aAAa,WAAW,OAAO,CACrC,WAAW;IACd,MAAM,YAAY,WAAW,YAAY,MAAM;AAC/C,QAAI,cAAc,IAChB,aAAa,WAAW,MAAM,GAAG,UAAU,CAAC,SAAS;GAExD;EACF;AAED,MAAI;GACF,MAAM,WAAW,KAAK,MAAM,WAAW;GAEvC,MAAM,EAAE,QAAQ,cAAc,GAAG;AAEjC,OAAI,WAAW,eACb,QAAO;IAAE,cAAc,EAAE,QAAQ,aAAc;IAAE,KAAK;GAAM;AAE9D,UAAO;IAAE,MAAM;IAAQ,WAAW;IAAc,KAAK;GAAM;EAC5D,SAAQ,GAAG;AACV,SAAM,IAAIC,sDACR,CAAC,wBAAwB,EAAE,KAAK,UAAU,EAAE,GAAG;EAElD;CACF;;;;;;;CAQD,wBAAgC;AAC9B,sDAAsBC,oCAAqB,YAAY,EACrD,YAAY,KAAK,UAAU,KAAK,KAAK,CACtC,EAAC;CACH;AACF;;;;;;;AAcD,IAAa,iDAAb,MAAa,uDAAuDJ,sCAAwB;CAC1F,eAAe;EAAC;EAAa;EAAU;CAAa;CAEpD,AAAQ;CAER,AAAQ;CAER,AAAQ,YAAsB,CAAE;CAEhC,YAAYK,QAAiD;EAC3D,MAAM,OAAO;EACb,KAAK,YAAY,OAAO,aAAa,KAAK;EAC1C,KAAK,aACH,QAAQ,cACR,IAAI,oCAAoC,EAAE,WAAW,KAAK,UAAW;EACvE,KAAK,qBAAqB,QAAQ;CACnC;;;;;CAMD,sBACEC,SACQ;AACR,MAAI,QAAQ,IACV,QAAOF;AAET,sDAAsBA,oCAAqB,YAAY,EACrD,YAAY,QAAQ,UAAU,KAAK,KAAK,CACzC,EAAC;CACH;;;;;;CAOD,MAAM,MAAMF,MAAkD;AAC5D,MAAI,KAAK,uBAAuB,OAC9B,QAAO,KAAK,mBAAmB,MAAM,KAAK;AAE5C,SAAO,KAAK,WAAW,MAAM,KAAK;CACnC;;;;;;;;;;;CAYD,OAAO,QACLK,KACAC,SACgD;EAChD,MAAM,aACJ,QAAQ,cACR,IAAI,oCAAoC,EACtC,WAAW,QAAQ,aAAa,CAAE,EACnC;EACH,MAAM,qBAAqBC,+BAAmB,QAAQ,KAAK,WAAW;AACtE,SAAO,IAAI,+CAA+C;GACxD;GACA;GACA,WAAW,QAAQ;EACpB;CACF;AACF"}
+{"version":3,"file":"outputParser.cjs","names":["AgentActionOutputParser","OutputParserException","FORMAT_INSTRUCTIONS","OutputFixingParser"],"sources":["../../../src/agents/chat_convo/outputParser.ts"],"sourcesContent":["import type { BaseLanguageModelInterface } from \"@langchain/core/language_models/base\";\nimport type { AgentAction, AgentFinish } from \"@langchain/core/agents\";\nimport {\n  FormatInstructionsOptions,\n  OutputParserException,\n} from \"@langchain/core/output_parsers\";\nimport { renderTemplate } from \"@langchain/core/prompts\";\nimport { AgentActionOutputParser } from \"../types.js\";\nimport { FORMAT_INSTRUCTIONS } from \"./prompt.js\";\nimport { OutputFixingParser } from \"../../output_parsers/fix.js\";\n\nexport type ChatConversationalAgentOutputParserFormatInstructionsOptions =\n  FormatInstructionsOptions & {\n    toolNames: string[];\n    raw?: boolean;\n  };\n\n/**\n * Class that represents an output parser for the ChatConversationalAgent\n * class. It extends the AgentActionOutputParser class and provides\n * methods for parsing the output of the MRKL chain into agent actions.\n */\nexport class ChatConversationalAgentOutputParser extends AgentActionOutputParser {\n  lc_namespace = [\"langchain\", \"agents\", \"chat_convo\"];\n\n  private toolNames: string[];\n\n  constructor(fields: { toolNames: string[] }) {\n    super(...arguments);\n    this.toolNames = fields.toolNames;\n  }\n\n  /**\n   * Parses the given text into an AgentAction or AgentFinish object. If an\n   * output fixing parser is defined, uses it to parse the text.\n   * @param text Text to parse.\n   * @returns Promise that resolves to an AgentAction or AgentFinish object.\n   */\n  async parse(text: string): Promise<AgentAction | AgentFinish> {\n    let jsonOutput = text.trim();\n    if (jsonOutput.includes(\"```json\") || jsonOutput.includes(\"```\")) {\n      const testString = jsonOutput.includes(\"```json\") ? \"```json\" : \"```\";\n      const firstIndex = jsonOutput.indexOf(testString);\n      const actionInputIndex = jsonOutput.indexOf(\"action_input\");\n      if (actionInputIndex > firstIndex) {\n        jsonOutput = jsonOutput\n          .slice(firstIndex + testString.length)\n          .trimStart();\n        const lastIndex = jsonOutput.lastIndexOf(\"```\");\n        if (lastIndex !== -1) {\n          jsonOutput = jsonOutput.slice(0, lastIndex).trimEnd();\n        }\n      }\n    }\n\n    try {\n      const response = JSON.parse(jsonOutput);\n\n      const { action, action_input } = response;\n\n      if (action === \"Final Answer\") {\n        return { returnValues: { output: action_input }, log: text };\n      }\n      return { tool: action, toolInput: action_input, log: text };\n    } catch (e) {\n      throw new OutputParserException(\n        `Failed to parse. Text: \"${text}\". Error: ${e}`\n      );\n    }\n  }\n\n  /**\n   * Returns the format instructions as a string. If the 'raw' option is\n   * true, returns the raw FORMAT_INSTRUCTIONS.\n   * @param options Options for getting the format instructions.\n   * @returns Format instructions as a string.\n   */\n  getFormatInstructions(): string {\n    return renderTemplate(FORMAT_INSTRUCTIONS, \"f-string\", {\n      tool_names: this.toolNames.join(\", \"),\n    });\n  }\n}\n\nexport type ChatConversationalAgentOutputParserArgs = {\n  baseParser?: ChatConversationalAgentOutputParser;\n  outputFixingParser?: OutputFixingParser<AgentAction | AgentFinish>;\n  toolNames?: string[];\n};\n\n/**\n * Class that represents an output parser with retries for the\n * ChatConversationalAgent class. It extends the AgentActionOutputParser\n * class and provides methods for parsing the output of the MRKL chain\n * into agent actions with retry functionality.\n */\nexport class ChatConversationalAgentOutputParserWithRetries extends AgentActionOutputParser {\n  lc_namespace = [\"langchain\", \"agents\", \"chat_convo\"];\n\n  private baseParser: ChatConversationalAgentOutputParser;\n\n  private outputFixingParser?: OutputFixingParser<AgentAction | AgentFinish>;\n\n  private toolNames: string[] = [];\n\n  constructor(fields: ChatConversationalAgentOutputParserArgs) {\n    super(fields);\n    this.toolNames = fields.toolNames ?? this.toolNames;\n    this.baseParser =\n      fields?.baseParser ??\n      new ChatConversationalAgentOutputParser({ toolNames: this.toolNames });\n    this.outputFixingParser = fields?.outputFixingParser;\n  }\n\n  /**\n   * Returns the format instructions as a string.\n   * @returns Format instructions as a string.\n   */\n  getFormatInstructions(\n    options: ChatConversationalAgentOutputParserFormatInstructionsOptions\n  ): string {\n    if (options.raw) {\n      return FORMAT_INSTRUCTIONS;\n    }\n    return renderTemplate(FORMAT_INSTRUCTIONS, \"f-string\", {\n      tool_names: options.toolNames.join(\", \"),\n    });\n  }\n\n  /**\n   * Parses the given text into an AgentAction or AgentFinish object.\n   * @param text Text to parse.\n   * @returns Promise that resolves to an AgentAction or AgentFinish object.\n   */\n  async parse(text: string): Promise<AgentAction | AgentFinish> {\n    if (this.outputFixingParser !== undefined) {\n      return this.outputFixingParser.parse(text);\n    }\n    return this.baseParser.parse(text);\n  }\n\n  /**\n   * Static method to create a new\n   * ChatConversationalAgentOutputParserWithRetries from a BaseLanguageModelInterface\n   * and options. If no base parser is provided in the options, a new\n   * ChatConversationalAgentOutputParser is created. An OutputFixingParser\n   * is also created from the BaseLanguageModelInterface and the base parser.\n   * @param llm BaseLanguageModelInterface instance used to create the OutputFixingParser.\n   * @param options Options for creating the ChatConversationalAgentOutputParserWithRetries instance.\n   * @returns A new instance of ChatConversationalAgentOutputParserWithRetries.\n   */\n  static fromLLM(\n    llm: BaseLanguageModelInterface,\n    options: Omit<ChatConversationalAgentOutputParserArgs, \"outputFixingParser\">\n  ): ChatConversationalAgentOutputParserWithRetries {\n    const baseParser =\n      options.baseParser ??\n      new ChatConversationalAgentOutputParser({\n        toolNames: options.toolNames ?? [],\n      });\n    const outputFixingParser = OutputFixingParser.fromLLM(llm, baseParser);\n    return new ChatConversationalAgentOutputParserWithRetries({\n      baseParser,\n      outputFixingParser,\n      toolNames: options.toolNames,\n    });\n  }\n}\n"],"mappings":";;;;;;;;;;;;;AAsBA,IAAa,sCAAb,cAAyDA,sCAAwB;CAC/E,eAAe;EAAC;EAAa;EAAU;EAAa;CAEpD,AAAQ;CAER,YAAY,QAAiC;AAC3C,QAAM,GAAG,UAAU;AACnB,OAAK,YAAY,OAAO;;;;;;;;CAS1B,MAAM,MAAM,MAAkD;EAC5D,IAAI,aAAa,KAAK,MAAM;AAC5B,MAAI,WAAW,SAAS,UAAU,IAAI,WAAW,SAAS,MAAM,EAAE;GAChE,MAAM,aAAa,WAAW,SAAS,UAAU,GAAG,YAAY;GAChE,MAAM,aAAa,WAAW,QAAQ,WAAW;AAEjD,OADyB,WAAW,QAAQ,eAAe,GACpC,YAAY;AACjC,iBAAa,WACV,MAAM,aAAa,WAAW,OAAO,CACrC,WAAW;IACd,MAAM,YAAY,WAAW,YAAY,MAAM;AAC/C,QAAI,cAAc,GAChB,cAAa,WAAW,MAAM,GAAG,UAAU,CAAC,SAAS;;;AAK3D,MAAI;GAGF,MAAM,EAAE,QAAQ,iBAFC,KAAK,MAAM,WAAW;AAIvC,OAAI,WAAW,eACb,QAAO;IAAE,cAAc,EAAE,QAAQ,cAAc;IAAE,KAAK;IAAM;AAE9D,UAAO;IAAE,MAAM;IAAQ,WAAW;IAAc,KAAK;IAAM;WACpD,GAAG;AACV,SAAM,IAAIC,qDACR,2BAA2B,KAAK,YAAY,IAC7C;;;;;;;;;CAUL,wBAAgC;AAC9B,qDAAsBC,oCAAqB,YAAY,EACrD,YAAY,KAAK,UAAU,KAAK,KAAK,EACtC,CAAC;;;;;;;;;AAgBN,IAAa,iDAAb,MAAa,uDAAuDF,sCAAwB;CAC1F,eAAe;EAAC;EAAa;EAAU;EAAa;CAEpD,AAAQ;CAER,AAAQ;CAER,AAAQ,YAAsB,EAAE;CAEhC,YAAY,QAAiD;AAC3D,QAAM,OAAO;AACb,OAAK,YAAY,OAAO,aAAa,KAAK;AAC1C,OAAK,aACH,QAAQ,cACR,IAAI,oCAAoC,EAAE,WAAW,KAAK,WAAW,CAAC;AACxE,OAAK,qBAAqB,QAAQ;;;;;;CAOpC,sBACE,SACQ;AACR,MAAI,QAAQ,IACV,QAAOE;AAET,qDAAsBA,oCAAqB,YAAY,EACrD,YAAY,QAAQ,UAAU,KAAK,KAAK,EACzC,CAAC;;;;;;;CAQJ,MAAM,MAAM,MAAkD;AAC5D,MAAI,KAAK,uBAAuB,OAC9B,QAAO,KAAK,mBAAmB,MAAM,KAAK;AAE5C,SAAO,KAAK,WAAW,MAAM,KAAK;;;;;;;;;;;;CAapC,OAAO,QACL,KACA,SACgD;EAChD,MAAM,aACJ,QAAQ,cACR,IAAI,oCAAoC,EACtC,WAAW,QAAQ,aAAa,EAAE,EACnC,CAAC;AAEJ,SAAO,IAAI,+CAA+C;GACxD;GACA,oBAHyBC,+BAAmB,QAAQ,KAAK,WAAW;GAIpE,WAAW,QAAQ;GACpB,CAAC"}

package/dist/agents/chat_convo/outputParser.d.cts.map

@@ -1 +1 @@
-{"version":3,"file":"outputParser.d.cts","names":["BaseLanguageModelInterface","AgentAction","AgentFinish","FormatInstructionsOptions","AgentActionOutputParser","OutputFixingParser","ChatConversationalAgentOutputParserFormatInstructionsOptions","ChatConversationalAgentOutputParser","Promise","ChatConversationalAgentOutputParserArgs","ChatConversationalAgentOutputParserWithRetries","Omit"],"sources":["../../../src/agents/chat_convo/outputParser.d.ts"],"sourcesContent":["import type { BaseLanguageModelInterface } from \"@langchain/core/language_models/base\";\nimport type { AgentAction, AgentFinish } from \"@langchain/core/agents\";\nimport { FormatInstructionsOptions } from \"@langchain/core/output_parsers\";\nimport { AgentActionOutputParser } from \"../types.js\";\nimport { OutputFixingParser } from \"../../output_parsers/fix.js\";\nexport type ChatConversationalAgentOutputParserFormatInstructionsOptions = FormatInstructionsOptions & {\n    toolNames: string[];\n    raw?: boolean;\n};\n/**\n * Class that represents an output parser for the ChatConversationalAgent\n * class. It extends the AgentActionOutputParser class and provides\n * methods for parsing the output of the MRKL chain into agent actions.\n */\nexport declare class ChatConversationalAgentOutputParser extends AgentActionOutputParser {\n    lc_namespace: string[];\n    private toolNames;\n    constructor(fields: {\n        toolNames: string[];\n    });\n    /**\n     * Parses the given text into an AgentAction or AgentFinish object. If an\n     * output fixing parser is defined, uses it to parse the text.\n     * @param text Text to parse.\n     * @returns Promise that resolves to an AgentAction or AgentFinish object.\n     */\n    parse(text: string): Promise<AgentAction | AgentFinish>;\n    /**\n     * Returns the format instructions as a string. If the 'raw' option is\n     * true, returns the raw FORMAT_INSTRUCTIONS.\n     * @param options Options for getting the format instructions.\n     * @returns Format instructions as a string.\n     */\n    getFormatInstructions(): string;\n}\nexport type ChatConversationalAgentOutputParserArgs = {\n    baseParser?: ChatConversationalAgentOutputParser;\n    outputFixingParser?: OutputFixingParser<AgentAction | AgentFinish>;\n    toolNames?: string[];\n};\n/**\n * Class that represents an output parser with retries for the\n * ChatConversationalAgent class. It extends the AgentActionOutputParser\n * class and provides methods for parsing the output of the MRKL chain\n * into agent actions with retry functionality.\n */\nexport declare class ChatConversationalAgentOutputParserWithRetries extends AgentActionOutputParser {\n    lc_namespace: string[];\n    private baseParser;\n    private outputFixingParser?;\n    private toolNames;\n    constructor(fields: ChatConversationalAgentOutputParserArgs);\n    /**\n     * Returns the format instructions as a string.\n     * @returns Format instructions as a string.\n     */\n    getFormatInstructions(options: ChatConversationalAgentOutputParserFormatInstructionsOptions): string;\n    /**\n     * Parses the given text into an AgentAction or AgentFinish object.\n     * @param text Text to parse.\n     * @returns Promise that resolves to an AgentAction or AgentFinish object.\n     */\n    parse(text: string): Promise<AgentAction | AgentFinish>;\n    /**\n     * Static method to create a new\n     * ChatConversationalAgentOutputParserWithRetries from a BaseLanguageModelInterface\n     * and options. If no base parser is provided in the options, a new\n     * ChatConversationalAgentOutputParser is created. An OutputFixingParser\n     * is also created from the BaseLanguageModelInterface and the base parser.\n     * @param llm BaseLanguageModelInterface instance used to create the OutputFixingParser.\n     * @param options Options for creating the ChatConversationalAgentOutputParserWithRetries instance.\n     * @returns A new instance of ChatConversationalAgentOutputParserWithRetries.\n     */\n    static fromLLM(llm: BaseLanguageModelInterface, options: Omit<ChatConversationalAgentOutputParserArgs, \"outputFixingParser\">): ChatConversationalAgentOutputParserWithRetries;\n}\n//# sourceMappingURL=outputParser.d.ts.map"],"mappings":";;;;;;;KAKYM,4DAAAA,GAA+DH;;EAA/DG,GAAAA,CAAAA,EAAAA,OAAAA;AASZ,CAAA;;;;;AAAwF;AAqB5EG,cArBSF,mCAAAA,SAA4CH,uBAAAA,CAqBd;EAClCG,YAAAA,EAAAA,MAAAA,EAAAA;EAC2BN,QAAAA,SAAAA;EAAcC,WAAAA,CAAAA,MAAAA,EAAAA;IAAjCG,SAAAA,EAAAA,MAAAA,EAAAA;EAAkB,CAAA;EAStBK;;;;;;EA2BGV,KAAAA,CAAAA,IAAAA,EAAAA,MAAAA,CAAAA,EA/CCQ,OA+CDR,CA/CSC,WA+CTD,GA/CuBE,WA+CvBF,CAAAA;EAA0CS;;;;AA3BiC;;;;KAXvFA,uCAAAA;eACKF;uBACQF,mBAAmBJ,cAAcC;;;;;;;;;cASrCQ,8CAAAA,SAAuDN,uBAAAA;;;;;sBAKpDK;;;;;iCAKWH;;;;;;uBAMVE,QAAQP,cAAcC;;;;;;;;;;;sBAWvBF,qCAAqCW,KAAKF,iEAAiEC"}
+{"version":3,"file":"outputParser.d.cts","names":[],"sources":["../../../src/agents/chat_convo/outputParser.ts"],"mappings":";;;;;;;KAWY,4DAAA,GACV,yBAAA;EACE,SAAA;EACA,GAAA;AAAA;;;;;;cAQS,mCAAA,SAA4C,uBAAA;EACvD,YAAA;EAAA,QAEQ,SAAA;EAER,WAAA,CAAY,MAAA;IAAU,SAAA;EAAA;EAW2B;;;;;;EAA3C,KAAA,CAAM,IAAA,WAAe,OAAA,CAAQ,WAAA,GAAc,WAAA;EAbzC;;;;;;EAoDR,qBAAA,CAAA;AAAA;AAAA,KAOU,uCAAA;EACV,UAAA,GAAa,mCAAA;EACb,kBAAA,GAAqB,kBAAA,CAAmB,WAAA,GAAc,WAAA;EACtD,SAAA;AAAA;;;;;;;cASW,8CAAA,SAAuD,uBAAA;EAClE,YAAA;EAAA,QAEQ,UAAA;EAAA,QAEA,kBAAA;EAAA,QAEA,SAAA;EAER,WAAA,CAAY,MAAA,EAAQ,uCAAA;EAnBoB;;;;EAgCxC,qBAAA,CACE,OAAA,EAAS,4DAAA;EAvBA;;;;;EAsCL,KAAA,CAAM,IAAA,WAAe,OAAA,CAAQ,WAAA,GAAc,WAAA;EAAA;;;;;;;;;;EAAA,OAiB1C,OAAA,CACL,GAAA,EAAK,0BAAA,EACL,OAAA,EAAS,IAAA,CAAK,uCAAA,0BACb,8CAAA;AAAA"}

package/dist/agents/chat_convo/outputParser.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"outputParser.d.ts","names":["BaseLanguageModelInterface","AgentAction","AgentFinish","FormatInstructionsOptions","AgentActionOutputParser","OutputFixingParser","ChatConversationalAgentOutputParserFormatInstructionsOptions","ChatConversationalAgentOutputParser","Promise","ChatConversationalAgentOutputParserArgs","ChatConversationalAgentOutputParserWithRetries","Omit"],"sources":["../../../src/agents/chat_convo/outputParser.d.ts"],"sourcesContent":["import type { BaseLanguageModelInterface } from \"@langchain/core/language_models/base\";\nimport type { AgentAction, AgentFinish } from \"@langchain/core/agents\";\nimport { FormatInstructionsOptions } from \"@langchain/core/output_parsers\";\nimport { AgentActionOutputParser } from \"../types.js\";\nimport { OutputFixingParser } from \"../../output_parsers/fix.js\";\nexport type ChatConversationalAgentOutputParserFormatInstructionsOptions = FormatInstructionsOptions & {\n    toolNames: string[];\n    raw?: boolean;\n};\n/**\n * Class that represents an output parser for the ChatConversationalAgent\n * class. It extends the AgentActionOutputParser class and provides\n * methods for parsing the output of the MRKL chain into agent actions.\n */\nexport declare class ChatConversationalAgentOutputParser extends AgentActionOutputParser {\n    lc_namespace: string[];\n    private toolNames;\n    constructor(fields: {\n        toolNames: string[];\n    });\n    /**\n     * Parses the given text into an AgentAction or AgentFinish object. If an\n     * output fixing parser is defined, uses it to parse the text.\n     * @param text Text to parse.\n     * @returns Promise that resolves to an AgentAction or AgentFinish object.\n     */\n    parse(text: string): Promise<AgentAction | AgentFinish>;\n    /**\n     * Returns the format instructions as a string. If the 'raw' option is\n     * true, returns the raw FORMAT_INSTRUCTIONS.\n     * @param options Options for getting the format instructions.\n     * @returns Format instructions as a string.\n     */\n    getFormatInstructions(): string;\n}\nexport type ChatConversationalAgentOutputParserArgs = {\n    baseParser?: ChatConversationalAgentOutputParser;\n    outputFixingParser?: OutputFixingParser<AgentAction | AgentFinish>;\n    toolNames?: string[];\n};\n/**\n * Class that represents an output parser with retries for the\n * ChatConversationalAgent class. It extends the AgentActionOutputParser\n * class and provides methods for parsing the output of the MRKL chain\n * into agent actions with retry functionality.\n */\nexport declare class ChatConversationalAgentOutputParserWithRetries extends AgentActionOutputParser {\n    lc_namespace: string[];\n    private baseParser;\n    private outputFixingParser?;\n    private toolNames;\n    constructor(fields: ChatConversationalAgentOutputParserArgs);\n    /**\n     * Returns the format instructions as a string.\n     * @returns Format instructions as a string.\n     */\n    getFormatInstructions(options: ChatConversationalAgentOutputParserFormatInstructionsOptions): string;\n    /**\n     * Parses the given text into an AgentAction or AgentFinish object.\n     * @param text Text to parse.\n     * @returns Promise that resolves to an AgentAction or AgentFinish object.\n     */\n    parse(text: string): Promise<AgentAction | AgentFinish>;\n    /**\n     * Static method to create a new\n     * ChatConversationalAgentOutputParserWithRetries from a BaseLanguageModelInterface\n     * and options. If no base parser is provided in the options, a new\n     * ChatConversationalAgentOutputParser is created. An OutputFixingParser\n     * is also created from the BaseLanguageModelInterface and the base parser.\n     * @param llm BaseLanguageModelInterface instance used to create the OutputFixingParser.\n     * @param options Options for creating the ChatConversationalAgentOutputParserWithRetries instance.\n     * @returns A new instance of ChatConversationalAgentOutputParserWithRetries.\n     */\n    static fromLLM(llm: BaseLanguageModelInterface, options: Omit<ChatConversationalAgentOutputParserArgs, \"outputFixingParser\">): ChatConversationalAgentOutputParserWithRetries;\n}\n//# sourceMappingURL=outputParser.d.ts.map"],"mappings":";;;;;;;KAKYM,4DAAAA,GAA+DH;;EAA/DG,GAAAA,CAAAA,EAAAA,OAAAA;AASZ,CAAA;;;;;AAAwF;AAqB5EG,cArBSF,mCAAAA,SAA4CH,uBAAAA,CAqBd;EAClCG,YAAAA,EAAAA,MAAAA,EAAAA;EAC2BN,QAAAA,SAAAA;EAAcC,WAAAA,CAAAA,MAAAA,EAAAA;IAAjCG,SAAAA,EAAAA,MAAAA,EAAAA;EAAkB,CAAA;EAStBK;;;;;;EA2BGV,KAAAA,CAAAA,IAAAA,EAAAA,MAAAA,CAAAA,EA/CCQ,OA+CDR,CA/CSC,WA+CTD,GA/CuBE,WA+CvBF,CAAAA;EAA0CS;;;;AA3BiC;;;;KAXvFA,uCAAAA;eACKF;uBACQF,mBAAmBJ,cAAcC;;;;;;;;;cASrCQ,8CAAAA,SAAuDN,uBAAAA;;;;;sBAKpDK;;;;;iCAKWH;;;;;;uBAMVE,QAAQP,cAAcC;;;;;;;;;;;sBAWvBF,qCAAqCW,KAAKF,iEAAiEC"}
+{"version":3,"file":"outputParser.d.ts","names":[],"sources":["../../../src/agents/chat_convo/outputParser.ts"],"mappings":";;;;;;;KAWY,4DAAA,GACV,yBAAA;EACE,SAAA;EACA,GAAA;AAAA;;;;;;cAQS,mCAAA,SAA4C,uBAAA;EACvD,YAAA;EAAA,QAEQ,SAAA;EAER,WAAA,CAAY,MAAA;IAAU,SAAA;EAAA;EAW2B;;;;;;EAA3C,KAAA,CAAM,IAAA,WAAe,OAAA,CAAQ,WAAA,GAAc,WAAA;EAbzC;;;;;;EAoDR,qBAAA,CAAA;AAAA;AAAA,KAOU,uCAAA;EACV,UAAA,GAAa,mCAAA;EACb,kBAAA,GAAqB,kBAAA,CAAmB,WAAA,GAAc,WAAA;EACtD,SAAA;AAAA;;;;;;;cASW,8CAAA,SAAuD,uBAAA;EAClE,YAAA;EAAA,QAEQ,UAAA;EAAA,QAEA,kBAAA;EAAA,QAEA,SAAA;EAER,WAAA,CAAY,MAAA,EAAQ,uCAAA;EAnBoB;;;;EAgCxC,qBAAA,CACE,OAAA,EAAS,4DAAA;EAvBA;;;;;EAsCL,KAAA,CAAM,IAAA,WAAe,OAAA,CAAQ,WAAA,GAAc,WAAA;EAAA;;;;;;;;;;EAAA,OAiB1C,OAAA,CACL,GAAA,EAAK,0BAAA,EACL,OAAA,EAAS,IAAA,CAAK,uCAAA,0BACb,8CAAA;AAAA"}

package/dist/agents/chat_convo/outputParser.js

@@ -1,5 +1,5 @@
-import { AgentActionOutputParser } from "../types.js";
 import { FORMAT_INSTRUCTIONS } from "./prompt.js";
+import { AgentActionOutputParser } from "../types.js";
 import { OutputFixingParser } from "../../output_parsers/fix.js";
 import { renderTemplate } from "@langchain/core/prompts";
 import { OutputParserException } from "@langchain/core/output_parsers";
@@ -32,16 +32,14 @@ var ChatConversationalAgentOutputParser = class extends AgentActionOutputParser
 		if (jsonOutput.includes("```json") || jsonOutput.includes("```")) {
 			const testString = jsonOutput.includes("```json") ? "```json" : "```";
 			const firstIndex = jsonOutput.indexOf(testString);
-			const actionInputIndex = jsonOutput.indexOf("action_input");
-			if (actionInputIndex > firstIndex) {
+			if (jsonOutput.indexOf("action_input") > firstIndex) {
 				jsonOutput = jsonOutput.slice(firstIndex + testString.length).trimStart();
 				const lastIndex = jsonOutput.lastIndexOf("```");
 				if (lastIndex !== -1) jsonOutput = jsonOutput.slice(0, lastIndex).trimEnd();
 			}
 		}
 		try {
-			const response = JSON.parse(jsonOutput);
-			const { action, action_input } = response;
+			const { action, action_input } = JSON.parse(jsonOutput);
 			if (action === "Final Answer") return {
 				returnValues: { output: action_input },
 				log: text
@@ -115,10 +113,9 @@ var ChatConversationalAgentOutputParserWithRetries = class ChatConversationalAge
 	*/
 	static fromLLM(llm, options) {
 		const baseParser = options.baseParser ?? new ChatConversationalAgentOutputParser({ toolNames: options.toolNames ?? [] });
-		const outputFixingParser = OutputFixingParser.fromLLM(llm, baseParser);
 		return new ChatConversationalAgentOutputParserWithRetries({
 			baseParser,
-			outputFixingParser,
+			outputFixingParser: OutputFixingParser.fromLLM(llm, baseParser),
 			toolNames: options.toolNames
 		});
 	}

package/dist/agents/chat_convo/outputParser.js.map

@@ -1 +1 @@
-{"version":3,"file":"outputParser.js","names":["fields: { toolNames: string[] }","text: string","fields: ChatConversationalAgentOutputParserArgs","options: ChatConversationalAgentOutputParserFormatInstructionsOptions","llm: BaseLanguageModelInterface","options: Omit<ChatConversationalAgentOutputParserArgs, \"outputFixingParser\">"],"sources":["../../../src/agents/chat_convo/outputParser.ts"],"sourcesContent":["import type { BaseLanguageModelInterface } from \"@langchain/core/language_models/base\";\nimport type { AgentAction, AgentFinish } from \"@langchain/core/agents\";\nimport {\n  FormatInstructionsOptions,\n  OutputParserException,\n} from \"@langchain/core/output_parsers\";\nimport { renderTemplate } from \"@langchain/core/prompts\";\nimport { AgentActionOutputParser } from \"../types.js\";\nimport { FORMAT_INSTRUCTIONS } from \"./prompt.js\";\nimport { OutputFixingParser } from \"../../output_parsers/fix.js\";\n\nexport type ChatConversationalAgentOutputParserFormatInstructionsOptions =\n  FormatInstructionsOptions & {\n    toolNames: string[];\n    raw?: boolean;\n  };\n\n/**\n * Class that represents an output parser for the ChatConversationalAgent\n * class. It extends the AgentActionOutputParser class and provides\n * methods for parsing the output of the MRKL chain into agent actions.\n */\nexport class ChatConversationalAgentOutputParser extends AgentActionOutputParser {\n  lc_namespace = [\"langchain\", \"agents\", \"chat_convo\"];\n\n  private toolNames: string[];\n\n  constructor(fields: { toolNames: string[] }) {\n    super(...arguments);\n    this.toolNames = fields.toolNames;\n  }\n\n  /**\n   * Parses the given text into an AgentAction or AgentFinish object. If an\n   * output fixing parser is defined, uses it to parse the text.\n   * @param text Text to parse.\n   * @returns Promise that resolves to an AgentAction or AgentFinish object.\n   */\n  async parse(text: string): Promise<AgentAction | AgentFinish> {\n    let jsonOutput = text.trim();\n    if (jsonOutput.includes(\"```json\") || jsonOutput.includes(\"```\")) {\n      const testString = jsonOutput.includes(\"```json\") ? \"```json\" : \"```\";\n      const firstIndex = jsonOutput.indexOf(testString);\n      const actionInputIndex = jsonOutput.indexOf(\"action_input\");\n      if (actionInputIndex > firstIndex) {\n        jsonOutput = jsonOutput\n          .slice(firstIndex + testString.length)\n          .trimStart();\n        const lastIndex = jsonOutput.lastIndexOf(\"```\");\n        if (lastIndex !== -1) {\n          jsonOutput = jsonOutput.slice(0, lastIndex).trimEnd();\n        }\n      }\n    }\n\n    try {\n      const response = JSON.parse(jsonOutput);\n\n      const { action, action_input } = response;\n\n      if (action === \"Final Answer\") {\n        return { returnValues: { output: action_input }, log: text };\n      }\n      return { tool: action, toolInput: action_input, log: text };\n    } catch (e) {\n      throw new OutputParserException(\n        `Failed to parse. Text: \"${text}\". Error: ${e}`\n      );\n    }\n  }\n\n  /**\n   * Returns the format instructions as a string. If the 'raw' option is\n   * true, returns the raw FORMAT_INSTRUCTIONS.\n   * @param options Options for getting the format instructions.\n   * @returns Format instructions as a string.\n   */\n  getFormatInstructions(): string {\n    return renderTemplate(FORMAT_INSTRUCTIONS, \"f-string\", {\n      tool_names: this.toolNames.join(\", \"),\n    });\n  }\n}\n\nexport type ChatConversationalAgentOutputParserArgs = {\n  baseParser?: ChatConversationalAgentOutputParser;\n  outputFixingParser?: OutputFixingParser<AgentAction | AgentFinish>;\n  toolNames?: string[];\n};\n\n/**\n * Class that represents an output parser with retries for the\n * ChatConversationalAgent class. It extends the AgentActionOutputParser\n * class and provides methods for parsing the output of the MRKL chain\n * into agent actions with retry functionality.\n */\nexport class ChatConversationalAgentOutputParserWithRetries extends AgentActionOutputParser {\n  lc_namespace = [\"langchain\", \"agents\", \"chat_convo\"];\n\n  private baseParser: ChatConversationalAgentOutputParser;\n\n  private outputFixingParser?: OutputFixingParser<AgentAction | AgentFinish>;\n\n  private toolNames: string[] = [];\n\n  constructor(fields: ChatConversationalAgentOutputParserArgs) {\n    super(fields);\n    this.toolNames = fields.toolNames ?? this.toolNames;\n    this.baseParser =\n      fields?.baseParser ??\n      new ChatConversationalAgentOutputParser({ toolNames: this.toolNames });\n    this.outputFixingParser = fields?.outputFixingParser;\n  }\n\n  /**\n   * Returns the format instructions as a string.\n   * @returns Format instructions as a string.\n   */\n  getFormatInstructions(\n    options: ChatConversationalAgentOutputParserFormatInstructionsOptions\n  ): string {\n    if (options.raw) {\n      return FORMAT_INSTRUCTIONS;\n    }\n    return renderTemplate(FORMAT_INSTRUCTIONS, \"f-string\", {\n      tool_names: options.toolNames.join(\", \"),\n    });\n  }\n\n  /**\n   * Parses the given text into an AgentAction or AgentFinish object.\n   * @param text Text to parse.\n   * @returns Promise that resolves to an AgentAction or AgentFinish object.\n   */\n  async parse(text: string): Promise<AgentAction | AgentFinish> {\n    if (this.outputFixingParser !== undefined) {\n      return this.outputFixingParser.parse(text);\n    }\n    return this.baseParser.parse(text);\n  }\n\n  /**\n   * Static method to create a new\n   * ChatConversationalAgentOutputParserWithRetries from a BaseLanguageModelInterface\n   * and options. If no base parser is provided in the options, a new\n   * ChatConversationalAgentOutputParser is created. An OutputFixingParser\n   * is also created from the BaseLanguageModelInterface and the base parser.\n   * @param llm BaseLanguageModelInterface instance used to create the OutputFixingParser.\n   * @param options Options for creating the ChatConversationalAgentOutputParserWithRetries instance.\n   * @returns A new instance of ChatConversationalAgentOutputParserWithRetries.\n   */\n  static fromLLM(\n    llm: BaseLanguageModelInterface,\n    options: Omit<ChatConversationalAgentOutputParserArgs, \"outputFixingParser\">\n  ): ChatConversationalAgentOutputParserWithRetries {\n    const baseParser =\n      options.baseParser ??\n      new ChatConversationalAgentOutputParser({\n        toolNames: options.toolNames ?? [],\n      });\n    const outputFixingParser = OutputFixingParser.fromLLM(llm, baseParser);\n    return new ChatConversationalAgentOutputParserWithRetries({\n      baseParser,\n      outputFixingParser,\n      toolNames: options.toolNames,\n    });\n  }\n}\n"],"mappings":";;;;;;;;;;;;AAsBA,IAAa,sCAAb,cAAyD,wBAAwB;CAC/E,eAAe;EAAC;EAAa;EAAU;CAAa;CAEpD,AAAQ;CAER,YAAYA,QAAiC;EAC3C,MAAM,GAAG,UAAU;EACnB,KAAK,YAAY,OAAO;CACzB;;;;;;;CAQD,MAAM,MAAMC,MAAkD;EAC5D,IAAI,aAAa,KAAK,MAAM;AAC5B,MAAI,WAAW,SAAS,UAAU,IAAI,WAAW,SAAS,MAAM,EAAE;GAChE,MAAM,aAAa,WAAW,SAAS,UAAU,GAAG,YAAY;GAChE,MAAM,aAAa,WAAW,QAAQ,WAAW;GACjD,MAAM,mBAAmB,WAAW,QAAQ,eAAe;AAC3D,OAAI,mBAAmB,YAAY;IACjC,aAAa,WACV,MAAM,aAAa,WAAW,OAAO,CACrC,WAAW;IACd,MAAM,YAAY,WAAW,YAAY,MAAM;AAC/C,QAAI,cAAc,IAChB,aAAa,WAAW,MAAM,GAAG,UAAU,CAAC,SAAS;GAExD;EACF;AAED,MAAI;GACF,MAAM,WAAW,KAAK,MAAM,WAAW;GAEvC,MAAM,EAAE,QAAQ,cAAc,GAAG;AAEjC,OAAI,WAAW,eACb,QAAO;IAAE,cAAc,EAAE,QAAQ,aAAc;IAAE,KAAK;GAAM;AAE9D,UAAO;IAAE,MAAM;IAAQ,WAAW;IAAc,KAAK;GAAM;EAC5D,SAAQ,GAAG;AACV,SAAM,IAAI,sBACR,CAAC,wBAAwB,EAAE,KAAK,UAAU,EAAE,GAAG;EAElD;CACF;;;;;;;CAQD,wBAAgC;AAC9B,SAAO,eAAe,qBAAqB,YAAY,EACrD,YAAY,KAAK,UAAU,KAAK,KAAK,CACtC,EAAC;CACH;AACF;;;;;;;AAcD,IAAa,iDAAb,MAAa,uDAAuD,wBAAwB;CAC1F,eAAe;EAAC;EAAa;EAAU;CAAa;CAEpD,AAAQ;CAER,AAAQ;CAER,AAAQ,YAAsB,CAAE;CAEhC,YAAYC,QAAiD;EAC3D,MAAM,OAAO;EACb,KAAK,YAAY,OAAO,aAAa,KAAK;EAC1C,KAAK,aACH,QAAQ,cACR,IAAI,oCAAoC,EAAE,WAAW,KAAK,UAAW;EACvE,KAAK,qBAAqB,QAAQ;CACnC;;;;;CAMD,sBACEC,SACQ;AACR,MAAI,QAAQ,IACV,QAAO;AAET,SAAO,eAAe,qBAAqB,YAAY,EACrD,YAAY,QAAQ,UAAU,KAAK,KAAK,CACzC,EAAC;CACH;;;;;;CAOD,MAAM,MAAMF,MAAkD;AAC5D,MAAI,KAAK,uBAAuB,OAC9B,QAAO,KAAK,mBAAmB,MAAM,KAAK;AAE5C,SAAO,KAAK,WAAW,MAAM,KAAK;CACnC;;;;;;;;;;;CAYD,OAAO,QACLG,KACAC,SACgD;EAChD,MAAM,aACJ,QAAQ,cACR,IAAI,oCAAoC,EACtC,WAAW,QAAQ,aAAa,CAAE,EACnC;EACH,MAAM,qBAAqB,mBAAmB,QAAQ,KAAK,WAAW;AACtE,SAAO,IAAI,+CAA+C;GACxD;GACA;GACA,WAAW,QAAQ;EACpB;CACF;AACF"}
+{"version":3,"file":"outputParser.js","names":[],"sources":["../../../src/agents/chat_convo/outputParser.ts"],"sourcesContent":["import type { BaseLanguageModelInterface } from \"@langchain/core/language_models/base\";\nimport type { AgentAction, AgentFinish } from \"@langchain/core/agents\";\nimport {\n  FormatInstructionsOptions,\n  OutputParserException,\n} from \"@langchain/core/output_parsers\";\nimport { renderTemplate } from \"@langchain/core/prompts\";\nimport { AgentActionOutputParser } from \"../types.js\";\nimport { FORMAT_INSTRUCTIONS } from \"./prompt.js\";\nimport { OutputFixingParser } from \"../../output_parsers/fix.js\";\n\nexport type ChatConversationalAgentOutputParserFormatInstructionsOptions =\n  FormatInstructionsOptions & {\n    toolNames: string[];\n    raw?: boolean;\n  };\n\n/**\n * Class that represents an output parser for the ChatConversationalAgent\n * class. It extends the AgentActionOutputParser class and provides\n * methods for parsing the output of the MRKL chain into agent actions.\n */\nexport class ChatConversationalAgentOutputParser extends AgentActionOutputParser {\n  lc_namespace = [\"langchain\", \"agents\", \"chat_convo\"];\n\n  private toolNames: string[];\n\n  constructor(fields: { toolNames: string[] }) {\n    super(...arguments);\n    this.toolNames = fields.toolNames;\n  }\n\n  /**\n   * Parses the given text into an AgentAction or AgentFinish object. If an\n   * output fixing parser is defined, uses it to parse the text.\n   * @param text Text to parse.\n   * @returns Promise that resolves to an AgentAction or AgentFinish object.\n   */\n  async parse(text: string): Promise<AgentAction | AgentFinish> {\n    let jsonOutput = text.trim();\n    if (jsonOutput.includes(\"```json\") || jsonOutput.includes(\"```\")) {\n      const testString = jsonOutput.includes(\"```json\") ? \"```json\" : \"```\";\n      const firstIndex = jsonOutput.indexOf(testString);\n      const actionInputIndex = jsonOutput.indexOf(\"action_input\");\n      if (actionInputIndex > firstIndex) {\n        jsonOutput = jsonOutput\n          .slice(firstIndex + testString.length)\n          .trimStart();\n        const lastIndex = jsonOutput.lastIndexOf(\"```\");\n        if (lastIndex !== -1) {\n          jsonOutput = jsonOutput.slice(0, lastIndex).trimEnd();\n        }\n      }\n    }\n\n    try {\n      const response = JSON.parse(jsonOutput);\n\n      const { action, action_input } = response;\n\n      if (action === \"Final Answer\") {\n        return { returnValues: { output: action_input }, log: text };\n      }\n      return { tool: action, toolInput: action_input, log: text };\n    } catch (e) {\n      throw new OutputParserException(\n        `Failed to parse. Text: \"${text}\". Error: ${e}`\n      );\n    }\n  }\n\n  /**\n   * Returns the format instructions as a string. If the 'raw' option is\n   * true, returns the raw FORMAT_INSTRUCTIONS.\n   * @param options Options for getting the format instructions.\n   * @returns Format instructions as a string.\n   */\n  getFormatInstructions(): string {\n    return renderTemplate(FORMAT_INSTRUCTIONS, \"f-string\", {\n      tool_names: this.toolNames.join(\", \"),\n    });\n  }\n}\n\nexport type ChatConversationalAgentOutputParserArgs = {\n  baseParser?: ChatConversationalAgentOutputParser;\n  outputFixingParser?: OutputFixingParser<AgentAction | AgentFinish>;\n  toolNames?: string[];\n};\n\n/**\n * Class that represents an output parser with retries for the\n * ChatConversationalAgent class. It extends the AgentActionOutputParser\n * class and provides methods for parsing the output of the MRKL chain\n * into agent actions with retry functionality.\n */\nexport class ChatConversationalAgentOutputParserWithRetries extends AgentActionOutputParser {\n  lc_namespace = [\"langchain\", \"agents\", \"chat_convo\"];\n\n  private baseParser: ChatConversationalAgentOutputParser;\n\n  private outputFixingParser?: OutputFixingParser<AgentAction | AgentFinish>;\n\n  private toolNames: string[] = [];\n\n  constructor(fields: ChatConversationalAgentOutputParserArgs) {\n    super(fields);\n    this.toolNames = fields.toolNames ?? this.toolNames;\n    this.baseParser =\n      fields?.baseParser ??\n      new ChatConversationalAgentOutputParser({ toolNames: this.toolNames });\n    this.outputFixingParser = fields?.outputFixingParser;\n  }\n\n  /**\n   * Returns the format instructions as a string.\n   * @returns Format instructions as a string.\n   */\n  getFormatInstructions(\n    options: ChatConversationalAgentOutputParserFormatInstructionsOptions\n  ): string {\n    if (options.raw) {\n      return FORMAT_INSTRUCTIONS;\n    }\n    return renderTemplate(FORMAT_INSTRUCTIONS, \"f-string\", {\n      tool_names: options.toolNames.join(\", \"),\n    });\n  }\n\n  /**\n   * Parses the given text into an AgentAction or AgentFinish object.\n   * @param text Text to parse.\n   * @returns Promise that resolves to an AgentAction or AgentFinish object.\n   */\n  async parse(text: string): Promise<AgentAction | AgentFinish> {\n    if (this.outputFixingParser !== undefined) {\n      return this.outputFixingParser.parse(text);\n    }\n    return this.baseParser.parse(text);\n  }\n\n  /**\n   * Static method to create a new\n   * ChatConversationalAgentOutputParserWithRetries from a BaseLanguageModelInterface\n   * and options. If no base parser is provided in the options, a new\n   * ChatConversationalAgentOutputParser is created. An OutputFixingParser\n   * is also created from the BaseLanguageModelInterface and the base parser.\n   * @param llm BaseLanguageModelInterface instance used to create the OutputFixingParser.\n   * @param options Options for creating the ChatConversationalAgentOutputParserWithRetries instance.\n   * @returns A new instance of ChatConversationalAgentOutputParserWithRetries.\n   */\n  static fromLLM(\n    llm: BaseLanguageModelInterface,\n    options: Omit<ChatConversationalAgentOutputParserArgs, \"outputFixingParser\">\n  ): ChatConversationalAgentOutputParserWithRetries {\n    const baseParser =\n      options.baseParser ??\n      new ChatConversationalAgentOutputParser({\n        toolNames: options.toolNames ?? [],\n      });\n    const outputFixingParser = OutputFixingParser.fromLLM(llm, baseParser);\n    return new ChatConversationalAgentOutputParserWithRetries({\n      baseParser,\n      outputFixingParser,\n      toolNames: options.toolNames,\n    });\n  }\n}\n"],"mappings":";;;;;;;;;;;;AAsBA,IAAa,sCAAb,cAAyD,wBAAwB;CAC/E,eAAe;EAAC;EAAa;EAAU;EAAa;CAEpD,AAAQ;CAER,YAAY,QAAiC;AAC3C,QAAM,GAAG,UAAU;AACnB,OAAK,YAAY,OAAO;;;;;;;;CAS1B,MAAM,MAAM,MAAkD;EAC5D,IAAI,aAAa,KAAK,MAAM;AAC5B,MAAI,WAAW,SAAS,UAAU,IAAI,WAAW,SAAS,MAAM,EAAE;GAChE,MAAM,aAAa,WAAW,SAAS,UAAU,GAAG,YAAY;GAChE,MAAM,aAAa,WAAW,QAAQ,WAAW;AAEjD,OADyB,WAAW,QAAQ,eAAe,GACpC,YAAY;AACjC,iBAAa,WACV,MAAM,aAAa,WAAW,OAAO,CACrC,WAAW;IACd,MAAM,YAAY,WAAW,YAAY,MAAM;AAC/C,QAAI,cAAc,GAChB,cAAa,WAAW,MAAM,GAAG,UAAU,CAAC,SAAS;;;AAK3D,MAAI;GAGF,MAAM,EAAE,QAAQ,iBAFC,KAAK,MAAM,WAAW;AAIvC,OAAI,WAAW,eACb,QAAO;IAAE,cAAc,EAAE,QAAQ,cAAc;IAAE,KAAK;IAAM;AAE9D,UAAO;IAAE,MAAM;IAAQ,WAAW;IAAc,KAAK;IAAM;WACpD,GAAG;AACV,SAAM,IAAI,sBACR,2BAA2B,KAAK,YAAY,IAC7C;;;;;;;;;CAUL,wBAAgC;AAC9B,SAAO,eAAe,qBAAqB,YAAY,EACrD,YAAY,KAAK,UAAU,KAAK,KAAK,EACtC,CAAC;;;;;;;;;AAgBN,IAAa,iDAAb,MAAa,uDAAuD,wBAAwB;CAC1F,eAAe;EAAC;EAAa;EAAU;EAAa;CAEpD,AAAQ;CAER,AAAQ;CAER,AAAQ,YAAsB,EAAE;CAEhC,YAAY,QAAiD;AAC3D,QAAM,OAAO;AACb,OAAK,YAAY,OAAO,aAAa,KAAK;AAC1C,OAAK,aACH,QAAQ,cACR,IAAI,oCAAoC,EAAE,WAAW,KAAK,WAAW,CAAC;AACxE,OAAK,qBAAqB,QAAQ;;;;;;CAOpC,sBACE,SACQ;AACR,MAAI,QAAQ,IACV,QAAO;AAET,SAAO,eAAe,qBAAqB,YAAY,EACrD,YAAY,QAAQ,UAAU,KAAK,KAAK,EACzC,CAAC;;;;;;;CAQJ,MAAM,MAAM,MAAkD;AAC5D,MAAI,KAAK,uBAAuB,OAC9B,QAAO,KAAK,mBAAmB,MAAM,KAAK;AAE5C,SAAO,KAAK,WAAW,MAAM,KAAK;;;;;;;;;;;;CAapC,OAAO,QACL,KACA,SACgD;EAChD,MAAM,aACJ,QAAQ,cACR,IAAI,oCAAoC,EACtC,WAAW,QAAQ,aAAa,EAAE,EACnC,CAAC;AAEJ,SAAO,IAAI,+CAA+C;GACxD;GACA,oBAHyB,mBAAmB,QAAQ,KAAK,WAAW;GAIpE,WAAW,QAAQ;GACpB,CAAC"}

package/dist/agents/chat_convo/prompt.cjs.map

@@ -1 +1 @@
-{"version":3,"file":"prompt.cjs","names":[],"sources":["../../../src/agents/chat_convo/prompt.ts"],"sourcesContent":["export const DEFAULT_PREFIX = `Assistant is a large language model trained by OpenAI.\n\nAssistant is designed to be able to assist with a wide range of tasks, from answering simple questions to providing in-depth explanations and discussions on a wide range of topics. As a language model, Assistant is able to generate human-like text based on the input it receives, allowing it to engage in natural-sounding conversations and provide responses that are coherent and relevant to the topic at hand.\n\nAssistant is constantly learning and improving, and its capabilities are constantly evolving. It is able to process and understand large amounts of text, and can use this knowledge to provide accurate and informative responses to a wide range of questions. Additionally, Assistant is able to generate its own text based on the input it receives, allowing it to engage in discussions and provide explanations and descriptions on a wide range of topics.\n\nOverall, Assistant is a powerful system that can help with a wide range of tasks and provide valuable insights and information on a wide range of topics. Whether you need help with a specific question or just want to have a conversation about a particular topic, Assistant is here to assist.`;\n\nexport const PREFIX_END = ` However, above all else, all responses must adhere to the format of RESPONSE FORMAT INSTRUCTIONS.`;\n\nexport const FORMAT_INSTRUCTIONS = `RESPONSE FORMAT INSTRUCTIONS\n----------------------------\n\nOutput a JSON markdown code snippet containing a valid JSON object in one of two formats:\n\n**Option 1:**\nUse this if you want the human to use a tool.\nMarkdown code snippet formatted in the following schema:\n\n\\`\\`\\`json\n{{{{\n    \"action\": string, // The action to take. Must be one of [{tool_names}]\n    \"action_input\": string // The input to the action. May be a stringified object.\n}}}}\n\\`\\`\\`\n\n**Option #2:**\nUse this if you want to respond directly and conversationally to the human. Markdown code snippet formatted in the following schema:\n\n\\`\\`\\`json\n{{{{\n    \"action\": \"Final Answer\",\n    \"action_input\": string // You should put what you want to return to user here and make sure to use valid json newline characters.\n}}}}\n\\`\\`\\`\n\nFor both options, remember to always include the surrounding markdown code snippet delimiters (begin with \"\\`\\`\\`json\" and end with \"\\`\\`\\`\")!\n`;\n\nexport const DEFAULT_SUFFIX = `TOOLS\n------\nAssistant can ask the user to use tools to look up information that may be helpful in answering the users original question. The tools the human can use are:\n\n{tools}\n\n{format_instructions}\n\nUSER'S INPUT\n--------------------\nHere is the user's input (remember to respond with a markdown code snippet of a json blob with a single action, and NOTHING else):\n\n{{input}}`;\n\nexport const TEMPLATE_TOOL_RESPONSE = `TOOL RESPONSE:\n---------------------\n{observation}\n\nUSER'S INPUT\n--------------------\n\nOkay, so what is the response to my last comment? If using information obtained from the tools you must mention it explicitly without mentioning the tool names - I have forgotten all TOOL RESPONSES! Remember to respond with a markdown code snippet of a json blob with a single action, and NOTHING else.`;\n"],"mappings":";;AAAA,MAAa,iBAAiB,CAAC;;;;;;mSAMoQ,CAAC;AAEpS,MAAa,aAAa,CAAC,kGAAkG,CAAC;AAE9H,MAAa,sBAAsB,CAAC;;;;;;;;;;;;;;;;;;;;;;;;;;;AA2BpC,CAAC;AAED,MAAa,iBAAiB,CAAC;;;;;;;;;;;;SAYtB,CAAC;AAEV,MAAa,yBAAyB,CAAC;;;;;;;8SAOuQ,CAAC"}
+{"version":3,"file":"prompt.cjs","names":[],"sources":["../../../src/agents/chat_convo/prompt.ts"],"sourcesContent":["export const DEFAULT_PREFIX = `Assistant is a large language model trained by OpenAI.\n\nAssistant is designed to be able to assist with a wide range of tasks, from answering simple questions to providing in-depth explanations and discussions on a wide range of topics. As a language model, Assistant is able to generate human-like text based on the input it receives, allowing it to engage in natural-sounding conversations and provide responses that are coherent and relevant to the topic at hand.\n\nAssistant is constantly learning and improving, and its capabilities are constantly evolving. It is able to process and understand large amounts of text, and can use this knowledge to provide accurate and informative responses to a wide range of questions. Additionally, Assistant is able to generate its own text based on the input it receives, allowing it to engage in discussions and provide explanations and descriptions on a wide range of topics.\n\nOverall, Assistant is a powerful system that can help with a wide range of tasks and provide valuable insights and information on a wide range of topics. Whether you need help with a specific question or just want to have a conversation about a particular topic, Assistant is here to assist.`;\n\nexport const PREFIX_END = ` However, above all else, all responses must adhere to the format of RESPONSE FORMAT INSTRUCTIONS.`;\n\nexport const FORMAT_INSTRUCTIONS = `RESPONSE FORMAT INSTRUCTIONS\n----------------------------\n\nOutput a JSON markdown code snippet containing a valid JSON object in one of two formats:\n\n**Option 1:**\nUse this if you want the human to use a tool.\nMarkdown code snippet formatted in the following schema:\n\n\\`\\`\\`json\n{{{{\n    \"action\": string, // The action to take. Must be one of [{tool_names}]\n    \"action_input\": string // The input to the action. May be a stringified object.\n}}}}\n\\`\\`\\`\n\n**Option #2:**\nUse this if you want to respond directly and conversationally to the human. Markdown code snippet formatted in the following schema:\n\n\\`\\`\\`json\n{{{{\n    \"action\": \"Final Answer\",\n    \"action_input\": string // You should put what you want to return to user here and make sure to use valid json newline characters.\n}}}}\n\\`\\`\\`\n\nFor both options, remember to always include the surrounding markdown code snippet delimiters (begin with \"\\`\\`\\`json\" and end with \"\\`\\`\\`\")!\n`;\n\nexport const DEFAULT_SUFFIX = `TOOLS\n------\nAssistant can ask the user to use tools to look up information that may be helpful in answering the users original question. The tools the human can use are:\n\n{tools}\n\n{format_instructions}\n\nUSER'S INPUT\n--------------------\nHere is the user's input (remember to respond with a markdown code snippet of a json blob with a single action, and NOTHING else):\n\n{{input}}`;\n\nexport const TEMPLATE_TOOL_RESPONSE = `TOOL RESPONSE:\n---------------------\n{observation}\n\nUSER'S INPUT\n--------------------\n\nOkay, so what is the response to my last comment? If using information obtained from the tools you must mention it explicitly without mentioning the tool names - I have forgotten all TOOL RESPONSES! Remember to respond with a markdown code snippet of a json blob with a single action, and NOTHING else.`;\n"],"mappings":";;AAAA,MAAa,iBAAiB;;;;;;;AAQ9B,MAAa,aAAa;AAE1B,MAAa,sBAAsB;;;;;;;;;;;;;;;;;;;;;;;;;;;;AA6BnC,MAAa,iBAAiB;;;;;;;;;;;;;AAc9B,MAAa,yBAAyB"}

package/dist/agents/chat_convo/prompt.js.map

@@ -1 +1 @@
-{"version":3,"file":"prompt.js","names":[],"sources":["../../../src/agents/chat_convo/prompt.ts"],"sourcesContent":["export const DEFAULT_PREFIX = `Assistant is a large language model trained by OpenAI.\n\nAssistant is designed to be able to assist with a wide range of tasks, from answering simple questions to providing in-depth explanations and discussions on a wide range of topics. As a language model, Assistant is able to generate human-like text based on the input it receives, allowing it to engage in natural-sounding conversations and provide responses that are coherent and relevant to the topic at hand.\n\nAssistant is constantly learning and improving, and its capabilities are constantly evolving. It is able to process and understand large amounts of text, and can use this knowledge to provide accurate and informative responses to a wide range of questions. Additionally, Assistant is able to generate its own text based on the input it receives, allowing it to engage in discussions and provide explanations and descriptions on a wide range of topics.\n\nOverall, Assistant is a powerful system that can help with a wide range of tasks and provide valuable insights and information on a wide range of topics. Whether you need help with a specific question or just want to have a conversation about a particular topic, Assistant is here to assist.`;\n\nexport const PREFIX_END = ` However, above all else, all responses must adhere to the format of RESPONSE FORMAT INSTRUCTIONS.`;\n\nexport const FORMAT_INSTRUCTIONS = `RESPONSE FORMAT INSTRUCTIONS\n----------------------------\n\nOutput a JSON markdown code snippet containing a valid JSON object in one of two formats:\n\n**Option 1:**\nUse this if you want the human to use a tool.\nMarkdown code snippet formatted in the following schema:\n\n\\`\\`\\`json\n{{{{\n    \"action\": string, // The action to take. Must be one of [{tool_names}]\n    \"action_input\": string // The input to the action. May be a stringified object.\n}}}}\n\\`\\`\\`\n\n**Option #2:**\nUse this if you want to respond directly and conversationally to the human. Markdown code snippet formatted in the following schema:\n\n\\`\\`\\`json\n{{{{\n    \"action\": \"Final Answer\",\n    \"action_input\": string // You should put what you want to return to user here and make sure to use valid json newline characters.\n}}}}\n\\`\\`\\`\n\nFor both options, remember to always include the surrounding markdown code snippet delimiters (begin with \"\\`\\`\\`json\" and end with \"\\`\\`\\`\")!\n`;\n\nexport const DEFAULT_SUFFIX = `TOOLS\n------\nAssistant can ask the user to use tools to look up information that may be helpful in answering the users original question. The tools the human can use are:\n\n{tools}\n\n{format_instructions}\n\nUSER'S INPUT\n--------------------\nHere is the user's input (remember to respond with a markdown code snippet of a json blob with a single action, and NOTHING else):\n\n{{input}}`;\n\nexport const TEMPLATE_TOOL_RESPONSE = `TOOL RESPONSE:\n---------------------\n{observation}\n\nUSER'S INPUT\n--------------------\n\nOkay, so what is the response to my last comment? If using information obtained from the tools you must mention it explicitly without mentioning the tool names - I have forgotten all TOOL RESPONSES! Remember to respond with a markdown code snippet of a json blob with a single action, and NOTHING else.`;\n"],"mappings":";AAAA,MAAa,iBAAiB,CAAC;;;;;;mSAMoQ,CAAC;AAEpS,MAAa,aAAa,CAAC,kGAAkG,CAAC;AAE9H,MAAa,sBAAsB,CAAC;;;;;;;;;;;;;;;;;;;;;;;;;;;AA2BpC,CAAC;AAED,MAAa,iBAAiB,CAAC;;;;;;;;;;;;SAYtB,CAAC;AAEV,MAAa,yBAAyB,CAAC;;;;;;;8SAOuQ,CAAC"}
+{"version":3,"file":"prompt.js","names":[],"sources":["../../../src/agents/chat_convo/prompt.ts"],"sourcesContent":["export const DEFAULT_PREFIX = `Assistant is a large language model trained by OpenAI.\n\nAssistant is designed to be able to assist with a wide range of tasks, from answering simple questions to providing in-depth explanations and discussions on a wide range of topics. As a language model, Assistant is able to generate human-like text based on the input it receives, allowing it to engage in natural-sounding conversations and provide responses that are coherent and relevant to the topic at hand.\n\nAssistant is constantly learning and improving, and its capabilities are constantly evolving. It is able to process and understand large amounts of text, and can use this knowledge to provide accurate and informative responses to a wide range of questions. Additionally, Assistant is able to generate its own text based on the input it receives, allowing it to engage in discussions and provide explanations and descriptions on a wide range of topics.\n\nOverall, Assistant is a powerful system that can help with a wide range of tasks and provide valuable insights and information on a wide range of topics. Whether you need help with a specific question or just want to have a conversation about a particular topic, Assistant is here to assist.`;\n\nexport const PREFIX_END = ` However, above all else, all responses must adhere to the format of RESPONSE FORMAT INSTRUCTIONS.`;\n\nexport const FORMAT_INSTRUCTIONS = `RESPONSE FORMAT INSTRUCTIONS\n----------------------------\n\nOutput a JSON markdown code snippet containing a valid JSON object in one of two formats:\n\n**Option 1:**\nUse this if you want the human to use a tool.\nMarkdown code snippet formatted in the following schema:\n\n\\`\\`\\`json\n{{{{\n    \"action\": string, // The action to take. Must be one of [{tool_names}]\n    \"action_input\": string // The input to the action. May be a stringified object.\n}}}}\n\\`\\`\\`\n\n**Option #2:**\nUse this if you want to respond directly and conversationally to the human. Markdown code snippet formatted in the following schema:\n\n\\`\\`\\`json\n{{{{\n    \"action\": \"Final Answer\",\n    \"action_input\": string // You should put what you want to return to user here and make sure to use valid json newline characters.\n}}}}\n\\`\\`\\`\n\nFor both options, remember to always include the surrounding markdown code snippet delimiters (begin with \"\\`\\`\\`json\" and end with \"\\`\\`\\`\")!\n`;\n\nexport const DEFAULT_SUFFIX = `TOOLS\n------\nAssistant can ask the user to use tools to look up information that may be helpful in answering the users original question. The tools the human can use are:\n\n{tools}\n\n{format_instructions}\n\nUSER'S INPUT\n--------------------\nHere is the user's input (remember to respond with a markdown code snippet of a json blob with a single action, and NOTHING else):\n\n{{input}}`;\n\nexport const TEMPLATE_TOOL_RESPONSE = `TOOL RESPONSE:\n---------------------\n{observation}\n\nUSER'S INPUT\n--------------------\n\nOkay, so what is the response to my last comment? If using information obtained from the tools you must mention it explicitly without mentioning the tool names - I have forgotten all TOOL RESPONSES! Remember to respond with a markdown code snippet of a json blob with a single action, and NOTHING else.`;\n"],"mappings":";AAAA,MAAa,iBAAiB;;;;;;;AAQ9B,MAAa,aAAa;AAE1B,MAAa,sBAAsB;;;;;;;;;;;;;;;;;;;;;;;;;;;;AA6BnC,MAAa,iBAAiB;;;;;;;;;;;;;AAc9B,MAAa,yBAAyB"}

package/dist/agents/executor.cjs

@@ -1,14 +1,14 @@
-const require_rolldown_runtime = require('../_virtual/rolldown_runtime.cjs');
+const require_runtime = require('../_virtual/_rolldown/runtime.cjs');
 const require_agent = require('./agent.cjs');
 const require_base = require('../chains/base.cjs');
-const __langchain_core_load_serializable = require_rolldown_runtime.__toESM(require("@langchain/core/load/serializable"));
-const __langchain_core_runnables = require_rolldown_runtime.__toESM(require("@langchain/core/runnables"));
-const __langchain_core_tools = require_rolldown_runtime.__toESM(require("@langchain/core/tools"));
-const __langchain_core_callbacks_manager = require_rolldown_runtime.__toESM(require("@langchain/core/callbacks/manager"));
-const __langchain_core_output_parsers = require_rolldown_runtime.__toESM(require("@langchain/core/output_parsers"));
+let _langchain_core_load_serializable = require("@langchain/core/load/serializable");
+let _langchain_core_runnables = require("@langchain/core/runnables");
+let _langchain_core_tools = require("@langchain/core/tools");
+let _langchain_core_callbacks_manager = require("@langchain/core/callbacks/manager");
+let _langchain_core_output_parsers = require("@langchain/core/output_parsers");
 
 //#region src/agents/executor.ts
-var AgentExecutorIterator = class extends __langchain_core_load_serializable.Serializable {
+var AgentExecutorIterator = class extends _langchain_core_load_serializable.Serializable {
 	lc_namespace = [
 		"langchain",
 		"agents",
@@ -32,10 +32,7 @@ var AgentExecutorIterator = class extends __langchain_core_load_serializable.Ser
 	/** Intended to be used as a setter method, needs to be async. */
 	async setFinalOutputs(value) {
 		this._finalOutputs = void 0;
-		if (value) {
-			const preparedOutputs = await this.agentExecutor.prepOutputs(this.inputs, value, true);
-			this._finalOutputs = preparedOutputs;
-		}
+		if (value) this._finalOutputs = await this.agentExecutor.prepOutputs(this.inputs, value, true);
 	}
 	runManager;
 	intermediateSteps = [];
@@ -71,8 +68,7 @@ var AgentExecutorIterator = class extends __langchain_core_load_serializable.Ser
 		this.reset();
 		while (true) try {
 			if (this.iterations === 0) await this.onFirstStep();
-			const result = await this._callNext();
-			yield result;
+			yield await this._callNext();
 		} catch (e) {
 			if ("message" in e && e.message.startsWith("Final outputs already reached: ")) {
 				if (!this.finalOutputs) throw e;
@@ -88,8 +84,7 @@ var AgentExecutorIterator = class extends __langchain_core_load_serializable.Ser
 	*/
 	async onFirstStep() {
 		if (this.iterations === 0) {
-			const callbackManager = await __langchain_core_callbacks_manager.CallbackManager.configure(this.callbacks ?? this.config?.callbacks, this.agentExecutor.callbacks, this.tags ?? this.config?.tags, this.agentExecutor.tags, this.metadata ?? this.config?.metadata, this.agentExecutor.metadata, { verbose: this.agentExecutor.verbose });
-			this.runManager = await callbackManager?.handleChainStart(this.agentExecutor.toJSON(), this.inputs, this.config?.runId, void 0, this.tags ?? this.config?.tags, this.metadata ?? this.config?.metadata, this.runName ?? this.config?.runName);
+			this.runManager = await (await _langchain_core_callbacks_manager.CallbackManager.configure(this.callbacks ?? this.config?.callbacks, this.agentExecutor.callbacks, this.tags ?? this.config?.tags, this.agentExecutor.tags, this.metadata ?? this.config?.metadata, this.agentExecutor.metadata, { verbose: this.agentExecutor.verbose }))?.handleChainStart(this.agentExecutor.toJSON(), this.inputs, this.config?.runId, void 0, this.tags ?? this.config?.tags, this.metadata ?? this.config?.metadata, this.runName ?? this.config?.runName);
 			if (this.config !== void 0) delete this.config.runId;
 		}
 	}
@@ -106,10 +101,10 @@ var AgentExecutorIterator = class extends __langchain_core_load_serializable.Ser
 	*/
 	async _processNextStepOutput(nextStepOutput, runManager) {
 		if ("returnValues" in nextStepOutput) {
-			const output$1 = await this.agentExecutor._return(nextStepOutput, this.intermediateSteps, runManager);
-			if (this.runManager) await this.runManager.handleChainEnd(output$1);
-			await this.setFinalOutputs(output$1);
-			return output$1;
+			const output = await this.agentExecutor._return(nextStepOutput, this.intermediateSteps, runManager);
+			if (this.runManager) await this.runManager.handleChainEnd(output);
+			await this.setFinalOutputs(output);
+			return output;
 		}
 		this.intermediateSteps = this.intermediateSteps.concat(nextStepOutput);
 		let output = {};
@@ -145,7 +140,7 @@ var AgentExecutorIterator = class extends __langchain_core_load_serializable.Ser
 * Tool that just returns the query.
 * Used for exception tracking.
 */
-var ExceptionTool = class extends __langchain_core_tools.Tool {
+var ExceptionTool = class extends _langchain_core_tools.Tool {
 	name = "_Exception";
 	description = "Exception tool";
 	async _call(query) {
@@ -208,7 +203,7 @@ var AgentExecutor = class AgentExecutor extends require_base.BaseChain {
 	constructor(input) {
 		let agent;
 		let returnOnlyOutputs = true;
-		if (__langchain_core_runnables.Runnable.isRunnable(input.agent)) {
+		if (_langchain_core_runnables.Runnable.isRunnable(input.agent)) {
 			if (require_agent.AgentRunnableSequence.isAgentRunnableSequence(input.agent)) if (input.agent.singleAction) agent = new require_agent.RunnableSingleActionAgent({
 				runnable: input.agent,
 				streamRunnable: input.agent.streamRunnable
@@ -282,7 +277,7 @@ var AgentExecutor = class AgentExecutor extends require_base.BaseChain {
 			try {
 				output = await this.agent.plan(steps, inputs, runManager?.getChild(), config);
 			} catch (e) {
-				if (e instanceof __langchain_core_output_parsers.OutputParserException) {
+				if (e instanceof _langchain_core_output_parsers.OutputParserException) {
 					let observation;
 					let text = e.message;
 					if (this.handleParsingErrors === true) if (e.sendToLLM) {
@@ -308,10 +303,10 @@ var AgentExecutor = class AgentExecutor extends require_base.BaseChain {
 				const tool = action.tool === "_Exception" ? new ExceptionTool() : toolsByName[action.tool?.toLowerCase()];
 				let observation;
 				try {
-					observation = tool ? await tool.invoke(action.toolInput, (0, __langchain_core_runnables.patchConfig)(config, { callbacks: runManager?.getChild() })) : `${action.tool} is not a valid tool, try another one.`;
+					observation = tool ? await tool.invoke(action.toolInput, (0, _langchain_core_runnables.patchConfig)(config, { callbacks: runManager?.getChild() })) : `${action.tool} is not a valid tool, try another one.`;
 					if (typeof observation !== "string") throw new Error("Received unsupported non-string response from tool call.");
 				} catch (e) {
-					if (e instanceof __langchain_core_tools.ToolInputParsingException) {
+					if (e instanceof _langchain_core_tools.ToolInputParsingException) {
 						if (this.handleParsingErrors === true) observation = "Invalid or incomplete tool input. Please try again.";
 						else if (typeof this.handleParsingErrors === "string") observation = this.handleParsingErrors;
 						else if (typeof this.handleParsingErrors === "function") observation = this.handleParsingErrors(e);
@@ -330,22 +325,20 @@ var AgentExecutor = class AgentExecutor extends require_base.BaseChain {
 			}));
 			steps.push(...newSteps);
 			const lastStep = steps[steps.length - 1];
-			const lastTool = toolsByName[lastStep.action.tool?.toLowerCase()];
-			if (lastTool?.returnDirect) return getOutput({
+			if (toolsByName[lastStep.action.tool?.toLowerCase()]?.returnDirect) return getOutput({
 				returnValues: { [this.agent.returnValues[0]]: lastStep.observation },
 				log: ""
 			});
 			iterations += 1;
 		}
-		const finish = await this.agent.returnStoppedResponse(this.earlyStoppingMethod, steps, inputs);
-		return getOutput(finish);
+		return getOutput(await this.agent.returnStoppedResponse(this.earlyStoppingMethod, steps, inputs));
 	}
 	async _takeNextStep(nameToolMap, inputs, intermediateSteps, runManager, config) {
 		let output;
 		try {
 			output = await this.agent.plan(intermediateSteps, inputs, runManager?.getChild(), config);
 		} catch (e) {
-			if (e instanceof __langchain_core_output_parsers.OutputParserException) {
+			if (e instanceof _langchain_core_output_parsers.OutputParserException) {
 				let observation;
 				let text = e.message;
 				if (this.handleParsingErrors === true) if (e.sendToLLM) {
@@ -376,7 +369,7 @@ var AgentExecutor = class AgentExecutor extends require_base.BaseChain {
 					observation = await tool.call(agentAction.toolInput, runManager?.getChild());
 					if (typeof observation !== "string") throw new Error("Received unsupported non-string response from tool call.");
 				} catch (e) {
-					if (e instanceof __langchain_core_tools.ToolInputParsingException) {
+					if (e instanceof _langchain_core_tools.ToolInputParsingException) {
 						if (this.handleParsingErrors === true) observation = "Invalid or incomplete tool input. Please try again.";
 						else if (typeof this.handleParsingErrors === "string") observation = this.handleParsingErrors;
 						else if (typeof this.handleParsingErrors === "function") observation = this.handleParsingErrors(e);
@@ -418,15 +411,14 @@ var AgentExecutor = class AgentExecutor extends require_base.BaseChain {
 		throw new Error(`Got unsupported early_stopping_method: ${earlyStoppingMethod}`);
 	}
 	async *_streamIterator(inputs, options) {
-		const agentExecutorIterator = new AgentExecutorIterator({
+		const iterator = new AgentExecutorIterator({
 			inputs,
 			agentExecutor: this,
 			config: options,
 			metadata: options?.metadata,
 			tags: options?.tags,
 			callbacks: options?.callbacks
-		});
-		const iterator = agentExecutorIterator.streamIterator();
+		}).streamIterator();
 		for await (const step of iterator) {
 			if (!step) continue;
 			yield step;