https://github.com/flowbrg/homer
Homer is a small academic project of a RAG application using Ollama and Langgraph as well as streamlit for the UI.
https://github.com/flowbrg/homer
langchain langgraph-python llm rag
Last synced: about 2 months ago
JSON representation
Homer is a small academic project of a RAG application using Ollama and Langgraph as well as streamlit for the UI.
- Host: GitHub
- URL: https://github.com/flowbrg/homer
- Owner: flowbrg
- Created: 2025-06-06T15:25:14.000Z (9 months ago)
- Default Branch: main
- Last Pushed: 2025-11-19T20:20:16.000Z (4 months ago)
- Last Synced: 2025-11-19T21:03:53.578Z (4 months ago)
- Topics: langchain, langgraph-python, llm, rag
- Language: Python
- Homepage:
- Size: 7.76 MB
- Stars: 1
- Watchers: 1
- Forks: 1
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Homer
Homer is a small academic project of a RAG application using Ollama and Langgraph as well as streamlit for the UI.
This project is designed to be executed on relatively low-performance laptops; therefore, computationally intensive tasks are delegated to a server (specifically, an Ollama instance operating on a non-default IP address).
## Environment
In PowerShell:
```
git pull https://github.com/flowbrg/homer.git
cd ./homer
python -m venv .venv
.\.venv\Scripts\Activate.ps1
pip install -r ./requirements.txt
```
Bash:
```
git pull https://github.com/flowbrg/homer.git
cd ./homer
python -m venv .venv
source .\.venv\bin\activate
pip install -r ./requirements.txt
```
You can use [uv](https://github.com/astral-sh/uv)
```
cd ./homer
uv venv
source .\.venv\bin\activate # or .\.venv\Scripts\Activate.ps1
uv sync
uv run .\src\main.py
```
## Dependencies
- [Ollama](https://ollama.com)
## TODO
- Refactor the configuration (persistent configuration using json);
- Improve the RAG-as-a-tool agent;
- Rework the report agent;
- Explore microsoft [markitdown](https://github.com/microsoft/markitdown) and google's [langextract](https://github.com/google/langextract) and replace the current vibe coded parser module;
- Migrate from streamlit probably;