🛠️Technical Overview

Oryza Agent is a custom Reasoning and Action (ReAct) agent built using LangGraph and LangChain.

The agent uses a cyclic graph architecture where:

1. Input is processed through a state graph

2. The model makes decisions about tool usage

3. Tools are executed as needed

4. Results are fed back to the model for further processing

graph TD
    A[Input State] --> B[StateGraph]
    B --> C[Model Node]
    C --> D{Route Output}
    D -->|Has Tool Calls| E[Tool Node]
    D -->|No Tool Calls| F[End]
    E --> C

Tech Stack

• Python

• LangGraph: Manages the agent's workflow and state

• LangChain: Provides the foundation for LLM interactions

• OpenAI GPT-4: Default language model for agent interactions

@dataclass(kw_only=True)
class Configuration:
    system_prompt: str
    model: str = "openai/gpt-4"
    max_search_results: int = 10
    max_tweets_to_process: int = 20
    min_engagement_rate: float = 0.01
    # ... additional configuration fields