search

James4Ever0 / prometheous

AI generated documentation and RAG
https://james4ever0.github.io/prometheous_doc
6 stars 0 forks source link

How to select local folder dataset to integrate the search paths locally #1

Open Tortoise17 opened 7 months ago

Tortoise17 commented 7 months ago

I have seen this repo as a nice excellent work. I want to ask that after the installation, how could this be linked to the local offline disc search so that it could be later optimized by this workflow. If you could guide would be a great help in start.

James4Ever0 commented 7 months ago

If I interpreted it correctly, you mean how to generate document for a new repository, right?

Before you run the below script, you may want to setup LLM service via this script. You need to install LiteLLM and Ollama.

You can check this documentation generation script. There are a few environment variables that you might want to change, in order to specify other LLM endpoints, the project path to document and the location of document webpage storage.

Tortoise17 commented 7 months ago

Repository means like i have database folder with some AI based vectors, or database folders with metadata. If you also understood it correctly like this, YES. I will start working on it and will ask for help if needed more. Thank you again. This is really a nice concept.

James4Ever0 commented 7 months ago

After running documentation generation steps above, you may execute text embedding script and perform RAG. Notice this will create the vector database folder under the documentation folder.

Tortoise17 commented 7 months ago

Dear @James4Ever0 conda unfortunately cannot install how it is mentioned in documentation. it throws the error EnvironmentLocationNotFound error. If you can have a look what is the error and how to make it install.

James4Ever0 commented 7 months ago

This is not a conda project. No need to install. I believe I have told you some key steps necessary to run this project. If you need further assistance, I offer paid support.

Tortoise17 commented 7 months ago

@James4Ever0 Thank you. I have one more question. As I understood, it is comitting into the git repo and trying to document the project. Is it also possible to use just big database and not store on git and just store locally in a folder to optimize search in a big pool. Like searching millions of files and millions of sentences and interlinking that both things with each time search in making search optimization of the local database? If you can guide me some more about it would be a great help.

Tortoise17 commented 7 months ago

@James4Ever0 Also second question, where is the executing shellscript which will open a search bar like web browser or so which will take intou account our text entered for search.

James4Ever0 commented 7 months ago

The questions you asked have their answers. You can ask ChatGPT instead of me if you want to ask for free.

This repository is provided as is, and it might or might not be updated, depending on my research focus.

Tortoise17 commented 6 months ago

@James4Ever0 Your repo is really great. I would like to apply it to the metadata of some projects to make or generate some documentation. I hope this should work that way .. Excellent work.