feat: simple react agent structure

[DagaBhai]

What does this PR do?

contributing to #99

adds a simple autoconfigure agent module

files added:

• agent.py - Agent class which takes the yaml file and sets the shared dict with the llm
• llm.py - Client lets the user add any llm they want with a customized function for the llm to call
• tool.py - Scaffold of some of the basic tools (to be worked on)
• node.py - based on the Node class which extends ThinkNode, ActNode, ObserveNode all 3 consists of 3 stages - [pre, exec, post]
• ThinkNode - uses the LLM to decide what to do next
• ActNode - executes tool calls
• ObserveNode - records the result and feeds it back into the loop

Test

from fenn.agents.agent import Agent
from fenn.agents.llm import Client
from google import genai
import yaml

# load config to get model name
with open("agent.yaml") as f:
    config = yaml.safe_load(f)

gemini_client = genai.Client(api_key="")
model = config["llm"]["model"]  # ← from yaml

def my_fn(messages, **kwargs):
    system = next((m["content"] for m in messages if m["role"] == "system"), "")

    history = []
    for msg in messages:
        if msg["role"] == "system":
            continue
        history.append({
            "role": "user" if msg["role"] == "user" else "model",
            "parts": [{"text": msg["content"]}]
        })

    if system and history:
        history[0]["parts"][0]["text"] = f"{system}\n\n{history[0]['parts'][0]['text']}"

    response = gemini_client.models.generate_content(
        model=model,
        contents=history
    )
    return response.text

gpt = Client()
gpt.register(my_fn)

agent = Agent(config="agent.yaml", llm=gpt)
print(agent.run("What is 123 * 456?"))

Note

• some llms need the model in the client init and some while calling so it may not read the model from the yaml file for example
• openai

from openai import OpenAI

client = OpenAI(
    api_key="your-key"
)

response = client.chat.completions.create(
    model="gpt-4.1",
    messages=[
        {"role": "user", "content": "Hello"}
    ]
)

• langchain_google_genai

from langchain_google_genai import ChatGoogleGenerativeAI

llm = ChatGoogleGenerativeAI(
    model="gemini-2.5-flash",
    google_api_key="YOUR_API_KEY"
)

response = llm.invoke("Hello")
print(response.content)

[blkdmr]

Great work! Thank you very much for your support!

[DagaBhai]

Great work! Thank you very much for your support!

thank you