TRON backend

An agentic RAG-powered coding assistant connected with your repository

• Read the thesis here

Tron is a fork of the elysia repository. Elysia did not support importing documents with the local weaviate cluster so I tried my best to add the chunk, embed and import documents on the local weaviate cluster. I basically combined Verba and Elysia.

• text pdfs supported
• unstructured pdfs supported (pdfs that have texts embedded into images and require OCR)
• txt files, md, mdx, json files supported

here you can find the original elysia readme

Check out also the TRON FRONTEND and Forked elysia frontend

Tron architecture

setup TRON

python -m venv env

./env/Scripts/activate

# install dependencies

python -m pip install -e .

# start TRON (elysia backend)

elysia start

Setup docprocessor

• Open new terminal

cd docprocessor

• create new environment

.env docprocessor

# get unstructured api key free online, you will have 7 days of OCR
UNSTRUCTURED_API_KEY=xxx
UNSTRUCTURED_API_URL=https://api.unstructuredapp.io/general/v0/general
#UNSTRUCTURED_API_URL=http://localhost:8070/general/v0/general

# if docker
#UNSTRUCTURED_API_URL=http://unstructured-api:8070/general/v0/general

python -m venv env
./env/Scripts/activate
python -m pip install -e .
docprocessor start

Setup local weaviate cluster

• go to weaviate cluster folder

docker compose up

now docker will create a container for weaviate cluster and a container for text2vec-transformers

using with ollama

simply go to config page inside TRON, click on the options and create these new options (click search bar and then create)

needs testing. Try gpt oss

• also checkout https://weaviate.github.io/elysia/setting_up/#local-model-integration-via-ollama