Add a built-in tool + example tutorial for using LiteLLMModel with DeepSeek provider

examples/llm/README.md

@@ -0,0 +1,59 @@
+# 📈 Company Info Agent (Powered by LiteLLMModel and DeepSeek)
+
+This example demonstrates how to build a company information agent using the [ISEK](https://github.com/isekOS/ISEK) framework. The agent integrates the `deepseek` large language model via  [`LiteLLMModel`](** https://github.com/isekOS/ISEK/tree/main/isek/models/litellm**)and can fetch stock codes and company information for companies.
+
+---
+
+## 🚀 Features
+
+- Query company info by **company name**
+- Retrieve **stock code** (e.g. `TSLA`)
+- Fetch **basic stock info** and **company details** using tools
+- Uses `DeepSeek-chat` LLM via [`LiteLLMModel`](** https://github.com/isekOS/ISEK/tree/main/isek/models/litellm**)
+- Fully extensible with other models and tools via ISEK
+
+---
+
+## 🧠 How It Works
+
+The agent uses the following components:
+
+-  [`LiteLLMModel`](** https://github.com/isekOS/ISEK/tree/main/isek/models/litellm**): Wrapper for any OpenAI-compatible LLM (e.g., DeepSeek, GPT-4, Claude, etc.)
+- [`base_info_tools`]: A collection of tools to retrieve stock codes, stock information, and company details.
+- `IsekAgent`: Manages tool usage and reasoning based on instructions.
+
+---
+
+## 🧩 Set Up Environment
+
+Before you try this example, don't forget to modify your .env file:
+
+```bash
+DEEPSEEK_API_KEY=your_deepseek_apikay
+DEEPSEEK_BASE_URL=https://api.deepseek.com/v1
+```
+
+## 🔄 Other Example Configs
+
+You can easily switch to other models and providers using [`LiteLLMModel`](** https://github.com/isekOS/ISEK/tree/main/isek/models/litellm**). Here are some common configurations:
+
+#### ✅ DeepSeek via Ollama (local deployment)
+
+```python
+LiteLLMModel(
+    provider="ollama",
+    model_id="deepseek-chat",
+    base_url="http://localhost:11434",
+    api_env_key=None  # No API key needed for local use
+)
+```
+
+#### ✅ Claude (Anthropic)
+
+```python
+LiteLLMModel(
+    provider="anthropic",
+    model_id="claude-3-opus-20240229",
+    api_env_key="ANTHROPIC_API_KEY"
+)
+```

examples/llm/deepseek_tool_agent.py

@@ -0,0 +1,27 @@
+from isek.agent.isek_agent import IsekAgent
+from isek.models.litellm import LiteLLMModel
+from isek.models.base import SimpleMessage
+from isek.tools.finance_toolkit.get_company_base_info import company_base_info_tools
+
+
+import dotenv
+dotenv.load_dotenv()
+
+agent = IsekAgent(
+    name="A-Share Company Info Agent",
+    model=LiteLLMModel(provider = "deepseek", model_id="deepseek/deepseek-chat"),
+    tools=[company_base_info_tools],
+    description="An assistant that finds stock info and company info given a company name",
+    instructions=["Be polite", 
+                  "Always first retrieve a numeric stock code (e.g. '600519') and a company 'About' or official website URL base on the company name.",
+                  "Then use the appropriate tool with the stock code to retrieve basic stock information.",
+                  "If a valid 'About' or official website URL is available, use the appropriate tool to fetch additional company details and return it.",
+                  "Always return concise and structured information.",
+                  "Only make tool calls when needed.",
+                  "If a valid stock code or official company URL cannot be found, inform the user explicitly and avoid making unnecessary tool calls."],
+    success_criteria="User receives the correct stock info and basic company information for the given company name.",
+    debug_mode=True
+)
+
+agent.print_response("hello")
+agent.print_response("Give me the base info of Apple company, including its stock info and company info.")

isek/tools/finance_toolkit/get_company_base_info.py

@@ -0,0 +1,48 @@
+from isek.tools.toolkit import Toolkit
+import efinance as ef
+
+import requests
+from bs4 import BeautifulSoup
+
+
+def get_stock_info(stock_code: str):
+    """get base info of stock base on stock code"""
+
+    return ef.stock.get_base_info(stock_code)
+
+
+def get_company_info(url: str):
+    """fetch additional company details base this url"""
+    headers = {
+        "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/91.0.4472.124 Safari/537.36"
+    }
+
+    try:
+        response = requests.get(url, headers=headers)
+        response.encoding = "utf-8"
+        soup = BeautifulSoup(response.text, "html.parser")
+        return soup.get_text(strip=True)
+    except Exception:
+        return ""
+
+
+# Create toolkit with debug enabled
+company_base_info_tools = Toolkit(
+    name="stock_info_tool",
+    tools=[get_stock_info, get_company_info],
+    instructions="Fetch company information based on its stock code and the company's 'About' page or official website URL.",
+    debug=True,
+)
+
+
+# Optionally, for demonstration, call list_functions and execute_function in debug mode
+if __name__ == "__main__":
+    company_base_info_tools.list_functions()
+    stock_info = company_base_info_tools.execute_function(
+        "get_stock_info", stock_code="00020"
+    )
+    company_info = company_base_info_tools.execute_function(
+        "get_company_info", url="https://www.sensetime.com/cn/about-index#1"
+    )
+    print("Stock Info:", stock_info)
+    print("Company Info:", company_info)