Add support for optional LLM model parameter in content processing

[sammydeprez]

Instead of defining and being limited to the models supported in the library via

• llm_model_name: str
• api_key_label: str
  parameters.

I have added the option to add an llm_model (a langchain BaseLanguageModel object)
This gives the user the options to use any provider (supported by Langchain) without making changes in the podcastfy library

So instead of this

audio_file = generate_podcast(
    urls=["https://en.wikipedia.org/wiki/Artificial_intelligence"],
    llm_model_name="gpt-4-turbo",
    api_key_label="OPENAI_API_KEY"
)

The user can now do this:

# Define a llm
from langchain_openai.chat_models import AzureChatOpenAI
llm_model = AzureChatOpenAI(model="gpt-4.1")

# Use the model
audio_file = generate_podcast(
    urls=["https://en.wikipedia.org/wiki/Artificial_intelligence"],
    llm_model=llm_model
)