Merge pull request #40 from debuggerone/examples

Add CoT example

Author: debuggerone

README.md

@@ -2,7 +2,7 @@
 
 AgentM is a library of "Micro Agents" that make it easy to add reliable intelligence to any application. The philosophy behind AgentM is that "Agents" should be mostly comprised of deterministic code with a sprinkle of LLM-powered intelligence mixed in. Many of the existing Agent frameworks place the LLM at the center of the application as an orchestrator that calls a collection of tools. In an AgentM application, your code is the orchestrator, and you only call a micro agent when you need to perform a task that requires intelligence.
 
-The initial draft of AgentM was written in JavaScript, but a Python version is available.
+Check out the [JavaScript version](https://github.com/Stevenic/agentm-js) to get a sense of AgentM's current feature set.
 
 ## Installation
 
@@ -70,3 +70,35 @@ Grape,Orange
 Original list: ['Apple', 'Orange', 'Banana', 'Grape', 'Pineapple']
 Sorted list: ['Apple', 'Banana', 'Orange', 'Grape', 'Pineapple']
 ```
+
+### Example 3: Chain of Thought
+
+Run the **chain of thought** example to see how AgentM solves problems using a step-by-step reasoning approach:
+
+```bash
+python examples/chain_of_thought_example.py
+```
+
+#### Sample Output:
+```bash
+Question: What is the square root of 144?
+Chain of Thought Reasoning: To find the square root of 144 step-by-step, follow these steps:
+
+1. **Understanding Square Roots**: The square root of a number is a value that, when multiplied by itself, gives that number. For example, if x is the square root of y, then x * x = y.
+
+2. **Identifying the Number**: In this case, we need to find the square root of 144.
+
+3. **Finding Factors**: We'll look for a number that, when multiplied by itself, equals 144. 
+
+4. **Testing Numbers**: 
+   - Let's start with smaller numbers:
+     - 12 * 12 = 144 (this is the answer)
+   
+5. **Conclusion**: The square root of 144 is 12.
+
+Thus, the final answer is:
+
+\[
+\sqrt{144} = 12
+\]
+```

examples/chain_of_thought_example.py

@@ -0,0 +1,13 @@
+import asyncio
+from core.chain_of_thought_agent import ChainOfThoughtAgent
+
+async def run_chain_of_thought_example():
+    question = 'What is the square root of 144?'
+    agent = ChainOfThoughtAgent(question=question)
+    result = await agent.chain_of_thought()
+
+    print("Question:", question)
+    print("Chain of Thought Reasoning:", result)
+
+if __name__ == "__main__":
+    asyncio.run(run_chain_of_thought_example())

src/core/chain_of_thought_agent.py

@@ -0,0 +1,21 @@
+import asyncio
+from typing import List
+from .openai_api import OpenAIClient
+
+class ChainOfThoughtAgent:
+    def __init__(self, question: str, max_tokens: int = 1000, temperature: float = 0.0):
+        self.question = question
+        self.max_tokens = max_tokens
+        self.temperature = temperature
+        self.openai_client = OpenAIClient()
+
+    async def chain_of_thought(self) -> str:
+        system_prompt = "You are an assistant tasked with solving problems using the 'chain of thought' reasoning process."
+        user_prompt = f"Solve the following problem step-by-step: {self.question}"
+
+        response = await self.openai_client.complete_chat([
+            {"role": "system", "content": system_prompt},
+            {"role": "user", "content": user_prompt}
+        ], max_tokens=self.max_tokens)
+
+        return response.strip()