Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/vinothpandian/ai-research-assistant
AI research assistant
https://github.com/vinothpandian/ai-research-assistant
Last synced: 13 days ago
JSON representation
AI research assistant
- Host: GitHub
- URL: https://github.com/vinothpandian/ai-research-assistant
- Owner: vinothpandian
- Created: 2023-12-28T06:34:52.000Z (about 1 year ago)
- Default Branch: main
- Last Pushed: 2024-04-03T15:59:53.000Z (11 months ago)
- Last Synced: 2024-12-15T12:19:10.372Z (2 months ago)
- Language: Python
- Size: 787 KB
- Stars: 1
- Watchers: 1
- Forks: 0
- Open Issues: 12
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# AI Research Assistant
---
# Setup
## Requirements
* Install `poetry` version 1.4.2
* Install `python` version 3.11.4
> Tip: You can use `asdf` ([link](https://asdf-vm.com/)) to manage your python and poetry versions.
## Install dependencies
Run `poetry install` to install the dependencies. This will create a poetry virtual environment and
install the dependencies in it. You can activate the virtual environment by running `poetry shell`.
# Development
## Configuration
You can configure the application by editing the copying the `config/dev.config.yaml` to `config.yaml` in
the root directory and filling in the values.
### Huggingface
* Change the type field of summarizer, embedding, qa in `config.yaml` to `huggingface`.
### Ollama
* Install `ollama` from [here](https://ollama.ai/). Only available for Linux and Mac.
* Change the type field of summarizer, embedding, qa in `config.yaml` to `ollama`.
* Set the correct
### OpenAI
* Set the key field of openai in `config.yaml` with
an [OpenAI API key](https://help.openai.com/en/articles/4936850-where-do-i-find-my-api-key).
* Change the type field of summarizer, embedding, qa in `config.yaml` to `openai`.
### Mixed
* You can mix and match the different types of summarizer, embedding, qa by setting the type field
can be one of `huggingface`, `ollama`, `openai`. Make sure the embedding dimension and the distance are
set correctly in `config.yaml`.
# Deployment
* Install `docker` version 20.10+ and run the docker daemon.
* Copy the `config/prod.config.yaml` to `config.yaml` and update the values based on the
[configuration](#configuration) section.
### Running the AI services
* Run AI server if you're using huggingface with the following command. It'll take a while to download and run the models.
```shell
make run_ray
```
* You can check the status of the ray cluster at `http://localhost:8265`.
* Run Ollama API service if you're using ollama with the following command:
```shell
make run_ollama_service
```
* If you're using openai, you don't need to do anything other than setting the key in `config.yaml` and
updating the type field to `openai`.
### Running the application
Run the application with the following command:
```shell
make run_all
```
It'll run the application at `http://localhost:8501`.