This Boilerplate is extended by the Next.js Enterprise Boilerplate ("https://github.com/Blazity/next-enterprise"). I've added langchain server into this template.

You can check detailed information in ("https://github.com/Blazity/next-enterprise")

🎯 Getting Started

To get started with this boilerplate, follow these steps:

0. Prerequisits:

• Setup homebrew for mac; https://brew.sh/
• Setup pyenv for multiple python environments; https://github.com/pyenv/pyenv

  brew install pyenv 

  • use miniconda3-latest from pyenv;
  • Afterwards all python packages are managed with conda
• Setup gh to interact with github; https://cli.github.com/manual/

  brew install gh 

1. Setup dedicated venv for langchain

conda create -n langchain
conda activate langchain

pip install -U poetry #langserv uses poetry to manage python dependencies https://python-poetry.org/

pip install -U langchain-core langchain-community "langserve[all]" langchain-cli langsmith

2. Use gh to fork; this will make your repo public;

gh repo fork uf-hobi-informatics-lab/next-enterprise-langserv --fork-name <your_app_name> --clone

If you want to have a private "fork", do this

git clone --bare https://github.com/uf-hobi-informatics-lab/next-enterprise-langserv.git <your_app_name>
cd <your_app_name>

gh repo create <your_app_name> --private ## create a empty repo under your own account; assuming you have already run gh auth, etc.
git push --mirror <your_github_link/your_app_name>

cd ..
rm -rf <your_app_name> #delete the old one

git clone <your_github_link/your_app_name>
cd <your_app_name>

## To pull updates from the public templates (in case this gets updated, and there are features you want to merge into your private repo
git remote add template https://github.com/uf-hobi-informatics-lab/next-enterprise-langserv
git pull template main

3. Run it

You need two terminal. In terminal 1, run frontend

  yarn dev

In terminal 2, run backend

conda activate langchain # make sure you use the right python venv

cd langserv 
uvicorn app.server:app --reload

4. Use it

• Open http://localhost:3000 to see the web frontend.

• Open http://localhost:8000 to see the langchain server.

5. MISC

• Please strict follow https://www.conventionalcommits.org/en/v1.0.0/#summary; and this project uses a git hook to enforce conventional commits. To install the git hook, run the following command in the root directory of the project:

brew install pre-commit
pre-commit install -t commit-msg

• It uses shadcn-ui from https://ui.shadcn.com/docs; by default it installed

npx shadcn-ui@latest add form button label dropdown-menu badge drawer input select textarea tooltip scroll-area

Frontend

• Take a look at "playground.tsx", which has set up dark model
• It aslo links to the backend with some example codes on both the frontend and backend side
  • You can access through http://localhost:3000/playground
• We use next.config.mjx rewrites to route calls to the backend

{
  source: "/backend/:path*",
  destination: process.env.NODE_ENV === "development" ? "http://127.0.0.1:8000/:path*" : "/"
}

• In your front, endpoint can be written as

 const endpoint = `backend/playground/get`
 const response = await fetch(endpoint, {
   method: "POST",
   headers: {
     "Content-Type": "application/json",
   },
   body: JSON.stringify({ url: data.message }),
 })

• It calls "http://127.0.0.1:8000/playground/get"

langserv

• Server side structure; read these https://fastapi.tiangolo.com/tutorial/bigger-applications/;

- app/
	- routers/                       -----API routes
	- schemas/                       ----- Pydantic models
	- services/                      ----- Business logic
	- utils/                         ----- Utils
	- playground/                    -----Playground
	- dependencies.py                ----- dependencies across modules when needed
	- server.py                      ----- main file
	- cli.py                         -----commandline
	- exc.py                         -----Exception handler
- tests/
- docs/
	- README.md
- pyproject.toml                   ----- peotry config; python dependency management
- README.md
- .env                             ----- dotenv (python-dotenv) for environment variables

• Operate under the langserv folder

cd ./langserv

• If you want to add a python package

poetry add langchain-openai

• [Deprecated] The packages are not really used; as it breaks some of the dependencies;
• If you want to add a package using existing templates; https://python.langchain.com/v0.2/docs/templates/;

  • e.g., the neo4j-advanced-rag; https://python.langchain.com/v0.2/docs/templates/neo4j-advanced-rag/

    langchain app add neo4j-advanced-rag

  • And add the following code to your server.py file:

    from neo4j_advanced_rag import chain as neo4j_advanced_chain
    
    add_routes(app, neo4j_advanced_chain, path="/neo4j-advanced-rag")

  • And set environment variables

    export OPENAI_API_KEY=<YOUR_OPENAI_API_KEY>
    export NEO4J_URI=<YOUR_NEO4J_URI>
    export NEO4J_USERNAME=<YOUR_NEO4J_USERNAME>
    export NEO4J_PASSWORD=<YOUR_NEO4J_PASSWORD>

  • We can see all templates at http://127.0.0.1:8000/docs

  • We can access the playground at http://127.0.0.1:8000/neo4j-advanced-rag/playground