When to import from "langchain" and when to import from "@langchain/core"

[nnduc1994]

Checked other resources

• I added a very descriptive title to this question.
• I searched the LangChain documentation with the integrated search.
• I used the GitHub search to find a similar question and didn't find it.

Commit to Help

• I commit to help with one of those options 👆

Example Code

import { PDFLoader } from "langchain/document_loaders/fs/pdf";
import { MemoryVectorStore } from "langchain/vectorstores/memory";
import { RecursiveCharacterTextSplitter } from "langchain/text_splitter";
import { OpenAIEmbeddings } from "langchain/embeddings/openai";
import { PromptTemplate } from "@langchain/core/prompts";
import { RunnableSequence } from "@langchain/core/runnables";
import { StringOutputParser } from "@langchain/core/output_parsers"

Description

I am trying to understand when to use import from langchain and when to import from @langchain/core. Examples on the internet and in the doc have a mix of import from both packages.

Also even though the documents mention

The @langchain/core package contains base abstractions that the rest of the LangChain ecosystem uses, along with the LangChain Expression Language. It is automatically installed along with langchain, but can also be used separately. Install with:

Looks like I still need to npm install @langchain/core separately.

System Info

[email protected] | MIT | deps: 18 | versions: 255
MAC
NODE 18

[dosubot[bot]]

The distinction between using imports from langchain versus @langchain/core and the necessity of installing @langchain/core separately, despite it being automatically installed with langchain, is rooted in the modular architecture of the LangChain ecosystem.

1. When to use imports from langchain versus @langchain/core:

   • You should use langchain for general functionalities provided by the LangChain library. This package is designed to offer a broad set of features and utilities for working with language models and integrations.
   • You should use @langchain/core when you need access to the base abstractions that the rest of the LangChain ecosystem uses. This includes the LangChain Expression Language and foundational components that are extended by other packages within the ecosystem.

2. Why it seems necessary to install @langchain/core separately:

   • The necessity to install @langchain/core separately arises from the need to ensure version consistency across the LangChain ecosystem, especially when using integration packages like @langchain/google-genai or @langchain/community. These integration packages depend on @langchain/core for their base functionalities.
   • To ensure that all integrations and their types interact properly and to avoid potential conflicts or issues arising from version mismatches, it is recommended to explicitly specify the version of @langchain/core in your project's package.json using the "resolutions", "overrides", or the equivalent field for your package manager. This guarantees that the same version of @langchain/core is used across all dependencies, providing a stable and consistent foundation for your project.

The documentation excerpt clearly outlines the installation process and the rationale behind the modular approach to managing dependencies within the LangChain ecosystem. This approach allows developers to tailor their installation to their specific needs, ensuring compatibility and functionality across different parts of the LangChain library and its integrations.

As for the specific functionalities provided by the langchain package that are not available in the @langchain/core package, I wasn't able to find an answer within the repository. However, the general rule of thumb is that langchain provides a broader set of features and utilities for working with language models and integrations, while @langchain/core provides the base abstractions used by the rest of the LangChain ecosystem.

I hope this clarifies your question. If you need further assistance, feel free to ask.

Sources

• docs/core_docs/docs/get_started/installation.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.

If you want to continue the conversation, start your reply with @dosu-bot.