Improve support for LLM usage

[gunthercox]

This pull request removes the experimental model parameter and introduces a more consistent implementation that uses the LogicAdapter pattern.

Additionally, this introduces the concept that each logic adapter can identify itself as a "tool" that can be used by a defined LLM.

Examples

Basic LLM with tools:

bot = ChatBot('Bot',
    logic_adapters=[
        {
            'import_path': 'chatterbot.logic.OllamaLogicAdapter',
            'model': 'llama3.1',
            'logic_adapters_as_tools': [
                'chatterbot.logic.MathematicalEvaluation',
                'chatterbot.logic.TimeLogicAdapter',
                'chatterbot.logic.UnitConversion'
            ]
        }
    ]
)

Hybrid configuration (LLM + standalone adapters for consensus):

bot = ChatBot('Bot',
    logic_adapters=[
        {
            'import_path': 'chatterbot.logic.OllamaLogicAdapter',
            'model': 'llama3.1',
            'logic_adapters_as_tools': [
                'chatterbot.logic.MathematicalEvaluation',
                'chatterbot.logic.TimeLogicAdapter'
            ]
        },
        'chatterbot.logic.BestMatch',  # Participates in consensus voting
        'chatterbot.logic.SpecificResponseAdapter'
    ]
)

Tool-capable adapter with parameters:

bot = ChatBot('Bot',
    logic_adapters=[
        {
            'import_path': 'chatterbot.logic.OllamaLogicAdapter',
            'model': 'mistral',
            'host': 'http://localhost:11434',
            'force_native_tools': True,
            'logic_adapters_as_tools': [
                {
                    'import_path': 'chatterbot.logic.MathematicalEvaluation',
                    'language': 'ENG'  # Adapter-specific parameters
                },
                'chatterbot.logic.TimeLogicAdapter'
            ]
        }
    ]
)

[gunthercox]

This can also take a few other parameters depending on features the model supports. For example, thinking models:

response = chat(model="qwen3", messages=messages, tools=[...], think=True)

https://docs.ollama.com/capabilities/tool-calling#python