Abstracted LLM interaction and added unit tests

[branhoff]

• Introduced a new module llm_interaction.py to abstract the interaction with language models (LLMs) and embeddings
  • Added functions get_llm and get_embeddings to retrieve LLMs and embeddings based on environment variables
  • Supported OpenAI's GPT models and embeddings by default, with the ability to extend to other providers
• Updated util.py to use the abstracted LLM interaction module instead of directly calling OpenAI's APIs
• Modified form_routes.py to rename the response variable from chat_gpt_response to llm_response
• Added unit tests for the llm_interaction module in tests/test_llm_interaction.py
  • Tested the behavior of get_llm and get_embeddings functions under different scenarios
• Updated the README with information on the application architecture, LLM configuration, and testing instructions
• Improved code formatting and added docstrings to adhere to the Google style guide
• Updated .gitignore to exclude general temp files and Mac-specific files