Add support for Ollama as a local inference backend

[Rage997]

It would be great to add Ollama as an alternative backend to OpenAI. This would let users run GitNexus with local LLMs, improving privacy, avoiding API costs, and enabling offline use.

[abhigyanpatwari]

Great idea. Will do

[Phum7895]

hi

[abhigyanpatwari]

Hi. Sorry this is taking long since I am working an embedings pipeline too. Combining RAG + Graph RAG for better retireval while working in browser constraints

[Rage997]

@abhigyanpatwari no need to apologise mate. It's great that you are doing this for free and open source keep the good work up!

[abhigyanpatwari]

Ollama provider support is added. But the agent through ollama behaves weird, might need custom implementation instead of langchain agent for this.

[abhigyanpatwari]

Ollama provider support is added. But the agent through ollama behaves weird, might need custom implementation instead of langchain agent for this.