Memory module is export as /memory/index

chunf commented 5 months ago

Checked other resources

[X] I added a very descriptive title to this issue.
[X] I searched the LangChain.js documentation with the integrated search.
[X] I used the GitHub search to find a similar question and didn't find it.
[X] I am sure that this is a bug in LangChain.js rather than my code.
[X] The bug is not resolved by updating to the latest stable version of LangChain (or the specific integration package).

Example Code

import { ConversationSummaryBufferMemory } from 'langchain/memory'

Error Message and Stack Trace (if applicable)

Error: Package subpath './memory' is not defined by "exports" in node_modules\langchain\package.json

Description

I am trying to use ConversationSummaryBufferMemory by importing it from 'langchain/memory'. The issue is fix if I modify line 1233 of package.json of langchain to "./memory": { "types": { "import": "./memory/index.d.ts", "require": "./memory/index.d.cts", "default": "./memory/index.d.ts" }, "import": "./memory/index.js", "require": "./memory/index.cjs" },

System Info

Node v20.9.0 "langchain": "^0.2.1",

dosubot[bot] commented 5 months ago

To resolve the issue with importing ConversationSummaryBufferMemory from 'langchain/memory', you should use the following import statement:

import { ConversationSummaryBufferMemory } from 'langchain/memory/index';

Here is an example usage:

import { OpenAI, ChatOpenAI } from "@langchain/openai";
import { ConversationSummaryBufferMemory } from "langchain/memory/index";
import { ConversationChain } from "langchain/chains";
import {
  ChatPromptTemplate,
  HumanMessagePromptTemplate,
  MessagesPlaceholder,
  SystemMessagePromptTemplate,
} from "@langchain/core/prompts";

// summary buffer memory
const memory = new ConversationSummaryBufferMemory({
  llm: new OpenAI({ model: "gpt-3.5-turbo-instruct", temperature: 0 }),
  maxTokenLimit: 10,
});

await memory.saveContext({ input: "hi" }, { output: "whats up" });
await memory.saveContext({ input: "not much you" }, { output: "not much" });
const history = await memory.loadMemoryVariables({});
console.log({ history });

// We can also get the history as a list of messages (this is useful if you are using this with a chat prompt).
const chatPromptMemory = new ConversationSummaryBufferMemory({
  llm: new ChatOpenAI({ model: "gpt-3.5-turbo", temperature: 0 }),
  maxTokenLimit: 10,
  returnMessages: true,
});
await chatPromptMemory.saveContext({ input: "hi" }, { output: "whats up" });
await chatPromptMemory.saveContext(
  { input: "not much you" },
  { output: "not much" }
);

// We can also utilize the predict_new_summary method directly.
const messages = await chatPromptMemory.chatHistory.getMessages();
const previous_summary = "";
const predictSummary = await chatPromptMemory.predictNewSummary(
  messages,
  previous_summary
);
console.log(JSON.stringify(predictSummary));

// Using in a chain
const chatPrompt = ChatPromptTemplate.fromMessages([
  SystemMessagePromptTemplate.fromTemplate(
    "The following is a friendly conversation between a human and an AI. The AI is talkative and provides lots of specific details from its context. If the AI does not know the answer to a question, it truthfully says it does not know."
  ),
  new MessagesPlaceholder("history"),
  HumanMessagePromptTemplate.fromTemplate("{input}"),
]);

const model = new ChatOpenAI({ temperature: 0.9, verbose: true });
const chain = new ConversationChain({
  llm: model,
  memory: chatPromptMemory,
  prompt: chatPrompt,
});

const res1 = await chain.invoke({ input: "Hi, what's up?" });
console.log({ res1 });

const res2 = await chain.invoke({
  input: "Just working on writing some documentation!",
});
console.log({ res2 });

const res3 = await chain.invoke({
  input: "For LangChain! Have you heard of it?",
});
console.log({ res3 });

const res4 = await chain.invoke({
  input:
    "That's not the right one, although a lot of people confuse it for that!",
});
console.log({ res4 });

This should resolve the issue without needing to modify the package.json file.

#### Details

**Sources** - [I'm getting this Error [ERR_PACKAGE_PATH_NOT_EXPORTED]: Package subpath './dist/tools/json' is not defined by "exports" in node_modules\langchain\package.json (working with typescript code)](https://github.com/langchain-ai/langchainjs/discussions/4858) - [Package.json import issues](https://github.com/langchain-ai/langchainjs/issues/3041) - [Cannot find module 'langchain/stores/message/mongodb'](https://github.com/langchain-ai/langchainjs/issues/2667) - [Package subpath './document_loaders/web/youtube' is not defined by "exports"](https://github.com/langchain-ai/langchainjs/issues/2416) - [Issue importing './llms' submodule](https://github.com/langchain-ai/langchainjs/issues/253) - [langchain/src/memory/index.ts](https://github.com/langchain-ai/langchainjs/blob/main/langchain/src/memory/index.ts) - [examples/src/memory/summary_buffer.ts](https://github.com/langchain-ai/langchainjs/blob/main/examples/src/memory/summary_buffer.ts) - [docs/core_docs/docs/versions/v0_2.mdx](https://github.com/langchain-ai/langchainjs/blob/main/docs/core_docs/docs/versions/v0_2.mdx) - [langchain/package.json](https://github.com/langchain-ai/langchainjs/blob/main/langchain/package.json) - [docs/core_docs/docs/versions/release_policy.mdx](https://github.com/langchain-ai/langchainjs/blob/main/docs/core_docs/docs/versions/release_policy.mdx)
**About Dosu** This response is meant to be useful and save you time. It is **not meant to be a precise solution**, but rather a starting point for your own research. Help me be more useful! Please leave a 👍 if this is helpful and 👎 if it is irrelevant.

To continue the conversation, mention @dosu.

jacoblee93 commented 5 months ago

Very sorry about that, just shipped a fix to readd the old entrypoint. It is live in langchain@0.2.2

langchain-ai / langchainjs