Add llama2-uncensored LLM

[ErdemOzgen]

• Added ollama with docker container
• GPTVulnerabilityReportGenerator : Retrieves an OpenAI API key and sets the model to 'gpt-3.5-turbo'. If the API key is not available, it initializes an Ollama instance for local use.
• GPTAttackSuggestionGenerator: Similar to the GPTVulnerabilityReportGenerator, it retrieves an OpenAI API key and sets the model to 'gpt-3.5-turbo', or initializes an Ollama instance if the API key is not available.
• Changed makefile make up and make down sections.
• After docker container up it will run command to pull llama2-uncensored llm model and serve model.

This PR fixes #1035.

[yogeshojha]

Wow, excited to test this out!

[SubGlitch1]

instead of this it would be easier to integrate #1114 (respectfully) as that would give users the chance to host any model for themselves or to connect to a server that does.

[ErdemOzgen]

@SubGlitch1, the proposed PR updates the URLs for the LLM model. It's important to note that the responses from ChatGPT and local LLMs may differ slightly, potentially leading to parsing errors in the code. However, modifying the model URL in this PR is feasible.

While people might choose different machines for hosting and serving LLMs, it was a deliberate design choice for my site to require local LLM configurations from sources like environment variables or a config file. Although these aren't included in the initial proof of concept, we plan to integrate environment variables or config files once the first PoC is successfully implemented.

[SubGlitch1]

@ErdemOzgen

you are absolutely right saying that the responses from custom apis might differ. that is why i proposed to add support for text-generation-webui as that projects api has the same scheme as openai. either way the options dont have to rule each other out an integrated llm is also useful. good luck

[psyray]

I close this one in favor of #1116 because it's based on master branch of @ErdemOzgen fork
#1035 (comment)