langchain openai tools agent chain using the new langserve playground

Author: Paul

.env.example

@@ -0,0 +1,4 @@
+OPENAI_API_KEY=sk-
+LANGCHAIN_API_KEY=
+LANGCHAIN_TRACING_V2=Tue
+LANGCHAIN_PROJECT=helloagent

.gitignore

@@ -0,0 +1 @@
+__pycache__

Dockerfile

@@ -0,0 +1,21 @@
+FROM python:3.11-slim
+
+RUN pip install poetry==1.6.1
+
+RUN poetry config virtualenvs.create false
+
+WORKDIR /code
+
+COPY ./pyproject.toml ./README.md ./poetry.lock* ./
+
+COPY ./package[s] ./packages
+
+RUN poetry install  --no-interaction --no-ansi --no-root
+
+COPY ./app ./app
+
+RUN poetry install --no-interaction --no-ansi
+
+EXPOSE 8080
+
+CMD exec uvicorn app.server:app --host 0.0.0.0 --port 8080

README.md

@@ -0,0 +1,79 @@
+# helloagent
+
+## Installation
+
+Install the LangChain CLI if you haven't yet
+
+```bash
+pip install -U langchain-cli
+```
+
+## Adding packages
+
+```bash
+# adding packages from 
+# https://github.com/langchain-ai/langchain/tree/master/templates
+langchain app add $PROJECT_NAME
+
+# adding custom GitHub repo packages
+langchain app add --repo $OWNER/$REPO
+# or with whole git string (supports other git providers):
+# langchain app add git+https://github.com/hwchase17/chain-of-verification
+
+# with a custom api mount point (defaults to `/{package_name}`)
+langchain app add $PROJECT_NAME --api_path=/my/custom/path/rag
+```
+
+Note: you remove packages by their api path
+
+```bash
+langchain app remove my/custom/path/rag
+```
+
+## Setup LangSmith (Optional)
+LangSmith will help us trace, monitor and debug LangChain applications. 
+LangSmith is currently in private beta, you can sign up [here](https://smith.langchain.com/). 
+If you don't have access, you can skip this section
+
+
+```shell
+export LANGCHAIN_TRACING_V2=true
+export LANGCHAIN_API_KEY=<your-api-key>
+export LANGCHAIN_PROJECT=<your-project>  # if not specified, defaults to "default"
+```
+
+## Launch LangServe
+
+```bash
+langchain serve
+```
+
+## Running in Docker
+
+This project folder includes a Dockerfile that allows you to easily build and host your LangServe app.
+
+### Building the Image
+
+To build the image, you simply:
+
+```shell
+docker build . -t my-langserve-app
+```
+
+If you tag your image with something other than `my-langserve-app`,
+note it for use in the next step.
+
+### Running the Image Locally
+
+To run the image, you'll need to include any environment variables
+necessary for your application.
+
+In the below example, we inject the `OPENAI_API_KEY` environment
+variable with the value set in my local environment
+(`$OPENAI_API_KEY`)
+
+We also expose port 8080 with the `-p 8080:8080` option.
+
+```shell
+docker run -e OPENAI_API_KEY=$OPENAI_API_KEY -p 8080:8080 my-langserve-app
+```

app/__init__.py

@@ -0,0 +1,34 @@
+from langchain_openai import ChatOpenAI
+from dotenv import load_dotenv, find_dotenv
+from typing import List, Union
+from langchain_core.messages import HumanMessage, AIMessage, SystemMessage
+from langchain.pydantic_v1 import BaseModel, Field
+from langchain.prompts import ChatPromptTemplate, MessagesPlaceholder
+from langchain_community.tools import DuckDuckGoSearchRun
+from langchain.prompts import ChatPromptTemplate, MessagesPlaceholder, SystemMessagePromptTemplate, HumanMessagePromptTemplate
+from langchain.agents import create_openai_tools_agent, AgentExecutor
+from langchain_core.runnables import RunnableLambda
+
+load_dotenv(find_dotenv())
+
+prompt = ChatPromptTemplate.from_messages([
+    SystemMessagePromptTemplate.from_template("You are a helpful assistant who makes fun of Swift developers."),
+    MessagesPlaceholder(variable_name="chat_history", optional=True),
+    HumanMessagePromptTemplate.from_template("{input}"),
+    MessagesPlaceholder(variable_name="agent_scratchpad")
+])
+
+gpt4 = ChatOpenAI(model="gpt-4-0125-preview", temperature=0)
+
+agent_tools = [DuckDuckGoSearchRun()]
+langchain_agent = create_openai_tools_agent(gpt4, agent_tools, prompt)
+agent_executor = AgentExecutor(agent=langchain_agent, tools=agent_tools, return_intermediate_steps=True)
+
+class Input(BaseModel):
+    chat_history: List[Union[HumanMessage, AIMessage, SystemMessage]] = Field(default_factory=list)
+    input: str
+
+def parse_agent_output(agent_output):
+    return agent_output["output"]
+
+chain = (agent_executor | RunnableLambda(parse_agent_output)).with_types(input_type=Input, output_type=str)

app/helloagent.py

@@ -0,0 +1,19 @@
+from fastapi import FastAPI
+from fastapi.responses import RedirectResponse
+from langserve import add_routes
+from app.helloagent import chain
+
+app = FastAPI()
+
+
+@app.get("/")
+async def redirect_root_to_docs():
+    return RedirectResponse("/docs")
+
+
+add_routes(app, chain, path="/helloagent", playground_type="chat")
+
+if __name__ == "__main__":
+    import uvicorn
+
+    uvicorn.run(app, host="0.0.0.0", port=8000)