langchain 0.0.131 → 0.0.132

package/dist/experimental/babyagi/agent.js

@@ -3,6 +3,12 @@ import { Document } from "../../document.js";
 import { TaskCreationChain } from "./task_creation.js";
 import { TaskExecutionChain } from "./task_execution.js";
 import { TaskPrioritizationChain } from "./task_prioritization.js";
+/**
+ * Class responsible for managing tasks, including their creation,
+ * prioritization, and execution. It uses three chains for these
+ * operations: `creationChain`, `prioritizationChain`, and
+ * `executionChain`.
+ */
 export class BabyAGI extends BaseChain {
     static lc_name() {
         return "BabyAGI";
@@ -68,23 +74,51 @@ export class BabyAGI extends BaseChain {
     get outputKeys() {
         return [];
     }
+    /**
+     * Adds a task to the task list.
+     * @param task The task to be added.
+     * @returns Promise resolving to void.
+     */
     async addTask(task) {
         this.taskList.push(task);
     }
+    /**
+     * Prints the current task list to the console.
+     * @returns void
+     */
     printTaskList() {
         console.log("\x1b[95m\x1b[1m\n*****TASK LIST*****\n\x1b[0m\x1b[0m");
         for (const t of this.taskList) {
             console.log(`${t.taskID}: ${t.taskName}`);
         }
     }
+    /**
+     * Prints the next task to the console.
+     * @param task The next task to be printed.
+     * @returns void
+     */
     printNextTask(task) {
         console.log("\x1b[92m\x1b[1m\n*****NEXT TASK*****\n\x1b[0m\x1b[0m");
         console.log(`${task.taskID}: ${task.taskName}`);
     }
+    /**
+     * Prints the result of a task to the console.
+     * @param result The result of the task.
+     * @returns void
+     */
     printTaskResult(result) {
         console.log("\x1b[93m\x1b[1m\n*****TASK RESULT*****\n\x1b[0m\x1b[0m");
         console.log(result.trim());
     }
+    /**
+     * Generates the next tasks based on the result of the previous task, the
+     * task description, and the objective.
+     * @param result The result of the previous task.
+     * @param task_description The description of the task.
+     * @param objective The objective of the task.
+     * @param runManager Optional CallbackManagerForChainRun instance.
+     * @returns Promise resolving to an array of tasks without taskID.
+     */
     async getNextTasks(result, task_description, objective, runManager) {
         const taskNames = this.taskList.map((t) => t.taskName);
         const incomplete_tasks = taskNames.join(", ");
@@ -99,6 +133,13 @@ export class BabyAGI extends BaseChain {
             .filter((taskName) => taskName.trim())
             .map((taskName) => ({ taskName }));
     }
+    /**
+     * Prioritizes the tasks based on the current task ID and the objective.
+     * @param thisTaskID The ID of the current task.
+     * @param objective The objective of the task.
+     * @param runManager Optional CallbackManagerForChainRun instance.
+     * @returns Promise resolving to an array of prioritized tasks.
+     */
     async prioritizeTasks(thisTaskID, objective, runManager) {
         const taskNames = this.taskList.map((t) => t.taskName);
         const nextTaskID = thisTaskID + 1;
@@ -119,6 +160,12 @@ export class BabyAGI extends BaseChain {
         }
         return prioritizedTaskList;
     }
+    /**
+     * Retrieves the top tasks that are most similar to the given query.
+     * @param query The query to search for.
+     * @param k The number of top tasks to retrieve.
+     * @returns Promise resolving to an array of top tasks.
+     */
     async getTopTasks(query, k = 5) {
         const results = await this.vectorstore.similaritySearch(query, k);
         if (!results) {
@@ -126,6 +173,13 @@ export class BabyAGI extends BaseChain {
         }
         return results.map((item) => String(item.metadata.task));
     }
+    /**
+     * Executes a task based on the objective and the task description.
+     * @param objective The objective of the task.
+     * @param task The task to be executed.
+     * @param runManager Optional CallbackManagerForChainRun instance.
+     * @returns Promise resolving to the result of the task execution as a string.
+     */
     async executeTask(objective, task, runManager) {
         const context = await this.getTopTasks(objective);
         const { [this.executionChain.outputKeys[0]]: text } = await this.executionChain.call({
@@ -168,6 +222,17 @@ export class BabyAGI extends BaseChain {
     serialize() {
         throw new Error("Method not implemented.");
     }
+    /**
+     * Static method to create a new BabyAGI instance from a
+     * BaseLanguageModel.
+     * @param llm BaseLanguageModel instance used to generate a new BabyAGI instance.
+     * @param vectorstore VectorStore instance used to store and retrieve vectors.
+     * @param executionChain Optional BaseChain instance used to execute tasks.
+     * @param verbose Optional boolean indicating whether to log verbose output.
+     * @param callbacks Optional callbacks to be used during the execution of tasks.
+     * @param rest Optional additional parameters.
+     * @returns A new instance of BabyAGI.
+     */
     static fromLLM({ llm, vectorstore, executionChain, verbose, callbacks, ...rest }) {
         const creationChain = TaskCreationChain.fromLLM({
             llm,

package/dist/experimental/babyagi/task_creation.cjs

@@ -8,6 +8,16 @@ class TaskCreationChain extends llm_chain_js_1.LLMChain {
     static lc_name() {
         return "TaskCreationChain";
     }
+    /**
+     * Creates a new TaskCreationChain instance. It takes an object of type
+     * LLMChainInput as input, omitting the 'prompt' field. It uses the
+     * PromptTemplate class to create a new prompt based on the task creation
+     * template and the input variables. The new TaskCreationChain instance is
+     * then created with this prompt and the remaining fields from the input
+     * object.
+     * @param fields An object of type LLMChainInput, omitting the 'prompt' field.
+     * @returns A new instance of TaskCreationChain.
+     */
     static fromLLM(fields) {
         const taskCreationTemplate = `You are an task creation AI that uses the result of an execution agent` +
             ` to create new tasks with the following objective: {objective},` +

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

@@ -2,5 +2,15 @@ import { LLMChain, LLMChainInput } from "../../chains/llm_chain.js";
 /** Chain to generate tasks. */
 export declare class TaskCreationChain extends LLMChain {
     static lc_name(): string;
+    /**
+     * Creates a new TaskCreationChain instance. It takes an object of type
+     * LLMChainInput as input, omitting the 'prompt' field. It uses the
+     * PromptTemplate class to create a new prompt based on the task creation
+     * template and the input variables. The new TaskCreationChain instance is
+     * then created with this prompt and the remaining fields from the input
+     * object.
+     * @param fields An object of type LLMChainInput, omitting the 'prompt' field.
+     * @returns A new instance of TaskCreationChain.
+     */
     static fromLLM(fields: Omit<LLMChainInput, "prompt">): LLMChain;
 }

package/dist/experimental/babyagi/task_creation.js

@@ -5,6 +5,16 @@ export class TaskCreationChain extends LLMChain {
     static lc_name() {
         return "TaskCreationChain";
     }
+    /**
+     * Creates a new TaskCreationChain instance. It takes an object of type
+     * LLMChainInput as input, omitting the 'prompt' field. It uses the
+     * PromptTemplate class to create a new prompt based on the task creation
+     * template and the input variables. The new TaskCreationChain instance is
+     * then created with this prompt and the remaining fields from the input
+     * object.
+     * @param fields An object of type LLMChainInput, omitting the 'prompt' field.
+     * @returns A new instance of TaskCreationChain.
+     */
     static fromLLM(fields) {
         const taskCreationTemplate = `You are an task creation AI that uses the result of an execution agent` +
             ` to create new tasks with the following objective: {objective},` +

package/dist/experimental/babyagi/task_execution.cjs

@@ -8,6 +8,15 @@ class TaskExecutionChain extends llm_chain_js_1.LLMChain {
     static lc_name() {
         return "TaskExecutionChain";
     }
+    /**
+     * A static factory method that creates an instance of TaskExecutionChain.
+     * It constructs a prompt template for task execution, which is then used
+     * to create a new instance of TaskExecutionChain. The prompt template
+     * instructs an AI to perform a task based on a given objective, taking
+     * into account previously completed tasks.
+     * @param fields An object of type LLMChainInput, excluding the "prompt" field.
+     * @returns An instance of LLMChain.
+     */
     static fromLLM(fields) {
         const executionTemplate = `You are an AI who performs one task based on the following objective: ` +
             `{objective}.` +

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

@@ -2,5 +2,14 @@ import { LLMChain, LLMChainInput } from "../../chains/llm_chain.js";
 /** Chain to execute tasks. */
 export declare class TaskExecutionChain extends LLMChain {
     static lc_name(): string;
+    /**
+     * A static factory method that creates an instance of TaskExecutionChain.
+     * It constructs a prompt template for task execution, which is then used
+     * to create a new instance of TaskExecutionChain. The prompt template
+     * instructs an AI to perform a task based on a given objective, taking
+     * into account previously completed tasks.
+     * @param fields An object of type LLMChainInput, excluding the "prompt" field.
+     * @returns An instance of LLMChain.
+     */
     static fromLLM(fields: Omit<LLMChainInput, "prompt">): LLMChain;
 }

package/dist/experimental/babyagi/task_execution.js

@@ -5,6 +5,15 @@ export class TaskExecutionChain extends LLMChain {
     static lc_name() {
         return "TaskExecutionChain";
     }
+    /**
+     * A static factory method that creates an instance of TaskExecutionChain.
+     * It constructs a prompt template for task execution, which is then used
+     * to create a new instance of TaskExecutionChain. The prompt template
+     * instructs an AI to perform a task based on a given objective, taking
+     * into account previously completed tasks.
+     * @param fields An object of type LLMChainInput, excluding the "prompt" field.
+     * @returns An instance of LLMChain.
+     */
     static fromLLM(fields) {
         const executionTemplate = `You are an AI who performs one task based on the following objective: ` +
             `{objective}.` +

package/dist/experimental/babyagi/task_prioritization.cjs

@@ -8,6 +8,14 @@ class TaskPrioritizationChain extends llm_chain_js_1.LLMChain {
     static lc_name() {
         return "TaskPrioritizationChain";
     }
+    /**
+     * Static method to create a new TaskPrioritizationChain from a
+     * BaseLanguageModel. It generates a prompt using the PromptTemplate class
+     * and the task prioritization template, and returns a new instance of
+     * TaskPrioritizationChain.
+     * @param fields Object with fields used to initialize the chain, excluding the prompt.
+     * @returns A new instance of TaskPrioritizationChain.
+     */
     static fromLLM(fields) {
         const taskPrioritizationTemplate = `You are a task prioritization AI tasked with cleaning the formatting of ` +
             `and reprioritizing the following tasks: {task_names}.` +

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

@@ -2,5 +2,13 @@ import { LLMChain, LLMChainInput } from "../../chains/llm_chain.js";
 /** Chain to prioritize tasks. */
 export declare class TaskPrioritizationChain extends LLMChain {
     static lc_name(): string;
+    /**
+     * Static method to create a new TaskPrioritizationChain from a
+     * BaseLanguageModel. It generates a prompt using the PromptTemplate class
+     * and the task prioritization template, and returns a new instance of
+     * TaskPrioritizationChain.
+     * @param fields Object with fields used to initialize the chain, excluding the prompt.
+     * @returns A new instance of TaskPrioritizationChain.
+     */
     static fromLLM(fields: Omit<LLMChainInput, "prompt">): LLMChain;
 }

package/dist/experimental/babyagi/task_prioritization.js

@@ -5,6 +5,14 @@ export class TaskPrioritizationChain extends LLMChain {
     static lc_name() {
         return "TaskPrioritizationChain";
     }
+    /**
+     * Static method to create a new TaskPrioritizationChain from a
+     * BaseLanguageModel. It generates a prompt using the PromptTemplate class
+     * and the task prioritization template, and returns a new instance of
+     * TaskPrioritizationChain.
+     * @param fields Object with fields used to initialize the chain, excluding the prompt.
+     * @returns A new instance of TaskPrioritizationChain.
+     */
     static fromLLM(fields) {
         const taskPrioritizationTemplate = `You are a task prioritization AI tasked with cleaning the formatting of ` +
             `and reprioritizing the following tasks: {task_names}.` +

package/dist/experimental/generative_agents/generative_agent.cjs

@@ -4,6 +4,11 @@ exports.GenerativeAgent = void 0;
 const llm_chain_js_1 = require("../../chains/llm_chain.cjs");
 const index_js_1 = require("../../prompts/index.cjs");
 const base_js_1 = require("../../chains/base.cjs");
+/**
+ * Implementation of a generative agent that can learn and form new memories over
+ * time. It extends the BaseChain class, which is a generic
+ * sequence of calls to components, including other chains.
+ */
 class GenerativeAgent extends base_js_1.BaseChain {
     static lc_name() {
         return "GenerativeAgent";
@@ -96,12 +101,23 @@ class GenerativeAgent extends base_js_1.BaseChain {
         // this.dailySummaries = config.dailySummaries ?? this.dailySummaries;
     }
     // LLM methods
+    /**
+     * Parses a newline-separated string into a list of strings.
+     * @param text The string to parse.
+     * @returns An array of strings parsed from the input text.
+     */
     parseList(text) {
         // parse a newline-seperated string into a list of strings
         const lines = text.trim().split("\n");
         const result = lines.map((line) => line.replace(/^\s*\d+\.\s*/, "").trim());
         return result;
     }
+    /**
+     * Creates a new LLMChain with the given prompt and the agent's language
+     * model, verbosity, output key, and memory.
+     * @param prompt The prompt to use for the LLMChain.
+     * @returns A new LLMChain instance.
+     */
     chain(prompt) {
         const chain = new llm_chain_js_1.LLMChain({
             llm: this.llm,
@@ -112,6 +128,12 @@ class GenerativeAgent extends base_js_1.BaseChain {
         });
         return chain;
     }
+    /**
+     * Extracts the observed entity from the given observation.
+     * @param observation The observation to extract the entity from.
+     * @param runManager Optional CallbackManagerForChainRun instance.
+     * @returns The extracted entity as a string.
+     */
     async getEntityFromObservations(observation, runManager) {
         const prompt = index_js_1.PromptTemplate.fromTemplate("What is the observed entity in the following observation? {observation}" +
             "\nEntity=");
@@ -120,6 +142,13 @@ class GenerativeAgent extends base_js_1.BaseChain {
         }, runManager?.getChild("entity_extractor"));
         return result.output;
     }
+    /**
+     * Extracts the action of the given entity from the given observation.
+     * @param observation The observation to extract the action from.
+     * @param entityName The name of the entity to extract the action for.
+     * @param runManager Optional CallbackManagerForChainRun instance.
+     * @returns The extracted action as a string.
+     */
     async getEntityAction(observation, entityName, runManager) {
         const prompt = index_js_1.PromptTemplate.fromTemplate("What is the {entity} doing in the following observation? {observation}" +
             "\nThe {entity} is");
@@ -130,6 +159,12 @@ class GenerativeAgent extends base_js_1.BaseChain {
         const trimmedResult = result.output.trim();
         return trimmedResult;
     }
+    /**
+     * Summarizes memories that are most relevant to an observation.
+     * @param observation The observation to summarize related memories for.
+     * @param runManager Optional CallbackManagerForChainRun instance.
+     * @returns The summarized memories as a string.
+     */
     async summarizeRelatedMemories(observation, runManager) {
         // summarize memories that are most relevant to an observation
         const prompt = index_js_1.PromptTemplate.fromTemplate(`
@@ -212,6 +247,12 @@ Relevant context:`);
         const regex = new RegExp(`^${this.name} `);
         return text.replace(regex, "").trim();
     }
+    /**
+     * Generates a reaction to the given observation.
+     * @param observation The observation to generate a reaction for.
+     * @param now Optional current date.
+     * @returns A boolean indicating whether to continue the dialogue and the output string.
+     */
     async generateReaction(observation, now) {
         const callToActionTemplate = `Should {agent_name} react to the observation, and if so,` +
             ` what would be an appropriate reaction? Respond in one line.` +
@@ -225,6 +266,12 @@ Relevant context:`);
         });
         return [continue_dialogue, output];
     }
+    /**
+     * Generates a dialogue response to the given observation.
+     * @param observation The observation to generate a dialogue response for.
+     * @param now Optional current date.
+     * @returns A boolean indicating whether to continue the dialogue and the output string.
+     */
     async generateDialogueResponse(observation, now) {
         const callToActionTemplate = `What would ${this.name} say? To end the conversation, write: GOODBYE: "what to say". Otherwise to continue the conversation, write: SAY: "what to say next"\n\n`;
         const { output, continue_dialogue } = await this.call({
@@ -238,6 +285,14 @@ Relevant context:`);
     // Each dialog or response prompt includes a header
     // summarizing the agent's self-description. This is
     // updated periodically through probing it's memories
+    /**
+     * Gets the agent's summary, which includes the agent's name, age, traits,
+     * and a summary of the agent's core characteristics. The summary is
+     * updated periodically through probing the agent's memories.
+     * @param config Optional configuration object with current date and a boolean to force refresh.
+     * @param runManager Optional CallbackManagerForChainRun instance.
+     * @returns The agent's summary as a string.
+     */
     async getSummary(config, runManager) {
         const { now = new Date(), forceRefresh = false } = config ?? {};
         const sinceRefresh = Math.floor((now.getTime() - this.lastRefreshed.getTime()) / 1000);
@@ -258,6 +313,12 @@ Relevant context:`);
 Innate traits: ${this.traits}
 ${this.summary}`;
     }
+    /**
+     * Computes the agent's summary by summarizing the agent's core
+     * characteristics given the agent's relevant memories.
+     * @param runManager Optional CallbackManagerForChainRun instance.
+     * @returns The computed summary as a string.
+     */
     async computeAgentSummary(runManager) {
         const prompt = index_js_1.PromptTemplate.fromTemplate("How would you summarize {name}'s core characteristics given the following statements:\n" +
             "----------" +
@@ -272,6 +333,11 @@ ${this.summary}`;
         }, runManager?.getChild("compute_agent_summary"));
         return result.output.trim();
     }
+    /**
+     * Returns a full header of the agent's status, summary, and current time.
+     * @param config Optional configuration object with current date and a boolean to force refresh.
+     * @returns The full header as a string.
+     */
     getFullHeader(config = {}) {
         const { now = new Date(), forceRefresh = false } = config;
         // return a full header of the agent's status, summary, and current time.
@@ -286,6 +352,14 @@ ${this.summary}`;
         });
         return `${summary}\nIt is ${currentTimeString}.\n${this.name}'s status: ${this.status}`;
     }
+    /**
+     * Adds a memory to the agent's long-term memory.
+     * @param memoryContent The content of the memory to add.
+     * @param now Optional current date.
+     * @param metadata Optional metadata for the memory.
+     * @param callbacks Optional Callbacks instance.
+     * @returns The result of adding the memory to the agent's long-term memory.
+     */
     async addMemory(memoryContent, now, metadata, callbacks) {
         return this.longTermMemory.addMemory(memoryContent, now, metadata, callbacks);
     }

package/dist/experimental/generative_agents/generative_agent.d.ts

@@ -5,6 +5,11 @@ import { GenerativeAgentMemory } from "./generative_agent_memory.js";
 import { ChainValues } from "../../schema/index.js";
 import { BaseChain } from "../../chains/base.js";
 import { CallbackManagerForChainRun, Callbacks } from "../../callbacks/manager.js";
+/**
+ * Configuration for the GenerativeAgent class. Defines the character's
+ * name, optional age, permanent traits, status, verbosity, and summary
+ * refresh seconds.
+ */
 export type GenerativeAgentConfig = {
     name: string;
     age?: number;
@@ -13,6 +18,11 @@ export type GenerativeAgentConfig = {
     verbose?: boolean;
     summaryRefreshSeconds?: number;
 };
+/**
+ * Implementation of a generative agent that can learn and form new memories over
+ * time. It extends the BaseChain class, which is a generic
+ * sequence of calls to components, including other chains.
+ */
 export declare class GenerativeAgent extends BaseChain {
     static lc_name(): string;
     name: string;
@@ -29,23 +39,92 @@ export declare class GenerativeAgent extends BaseChain {
     get inputKeys(): string[];
     get outputKeys(): string[];
     constructor(llm: BaseLanguageModel, longTermMemory: GenerativeAgentMemory, config: GenerativeAgentConfig);
+    /**
+     * Parses a newline-separated string into a list of strings.
+     * @param text The string to parse.
+     * @returns An array of strings parsed from the input text.
+     */
     parseList(text: string): string[];
+    /**
+     * Creates a new LLMChain with the given prompt and the agent's language
+     * model, verbosity, output key, and memory.
+     * @param prompt The prompt to use for the LLMChain.
+     * @returns A new LLMChain instance.
+     */
     chain(prompt: PromptTemplate): LLMChain;
+    /**
+     * Extracts the observed entity from the given observation.
+     * @param observation The observation to extract the entity from.
+     * @param runManager Optional CallbackManagerForChainRun instance.
+     * @returns The extracted entity as a string.
+     */
     getEntityFromObservations(observation: string, runManager?: CallbackManagerForChainRun): Promise<string>;
+    /**
+     * Extracts the action of the given entity from the given observation.
+     * @param observation The observation to extract the action from.
+     * @param entityName The name of the entity to extract the action for.
+     * @param runManager Optional CallbackManagerForChainRun instance.
+     * @returns The extracted action as a string.
+     */
     getEntityAction(observation: string, entityName: string, runManager?: CallbackManagerForChainRun): Promise<string>;
+    /**
+     * Summarizes memories that are most relevant to an observation.
+     * @param observation The observation to summarize related memories for.
+     * @param runManager Optional CallbackManagerForChainRun instance.
+     * @returns The summarized memories as a string.
+     */
     summarizeRelatedMemories(observation: string, runManager?: CallbackManagerForChainRun): Promise<string>;
     _call(values: ChainValues, runManager?: CallbackManagerForChainRun): Promise<ChainValues>;
     private _cleanResponse;
+    /**
+     * Generates a reaction to the given observation.
+     * @param observation The observation to generate a reaction for.
+     * @param now Optional current date.
+     * @returns A boolean indicating whether to continue the dialogue and the output string.
+     */
     generateReaction(observation: string, now?: Date): Promise<[boolean, string]>;
+    /**
+     * Generates a dialogue response to the given observation.
+     * @param observation The observation to generate a dialogue response for.
+     * @param now Optional current date.
+     * @returns A boolean indicating whether to continue the dialogue and the output string.
+     */
     generateDialogueResponse(observation: string, now?: Date): Promise<[boolean, string]>;
+    /**
+     * Gets the agent's summary, which includes the agent's name, age, traits,
+     * and a summary of the agent's core characteristics. The summary is
+     * updated periodically through probing the agent's memories.
+     * @param config Optional configuration object with current date and a boolean to force refresh.
+     * @param runManager Optional CallbackManagerForChainRun instance.
+     * @returns The agent's summary as a string.
+     */
     getSummary(config?: {
         now?: Date;
         forceRefresh?: boolean;
     }, runManager?: CallbackManagerForChainRun): Promise<string>;
+    /**
+     * Computes the agent's summary by summarizing the agent's core
+     * characteristics given the agent's relevant memories.
+     * @param runManager Optional CallbackManagerForChainRun instance.
+     * @returns The computed summary as a string.
+     */
     computeAgentSummary(runManager?: CallbackManagerForChainRun): Promise<string>;
+    /**
+     * Returns a full header of the agent's status, summary, and current time.
+     * @param config Optional configuration object with current date and a boolean to force refresh.
+     * @returns The full header as a string.
+     */
     getFullHeader(config?: {
         now?: Date;
         forceRefresh?: boolean;
     }): string;
+    /**
+     * Adds a memory to the agent's long-term memory.
+     * @param memoryContent The content of the memory to add.
+     * @param now Optional current date.
+     * @param metadata Optional metadata for the memory.
+     * @param callbacks Optional Callbacks instance.
+     * @returns The result of adding the memory to the agent's long-term memory.
+     */
     addMemory(memoryContent: string, now?: Date, metadata?: Record<string, unknown>, callbacks?: Callbacks): Promise<ChainValues>;
 }