frag.jetzt Langchain (AI Provider)

This service serves all AI related functions for the webapp frag.jetzt. See git.thm.de/arsnova/frag.jetzt or github.com/thm-projects/frag.jetzt.

Installation

You need to use git.thm.de/arsnova/frag.jetzt-docker-orchestration to run this app either locally (./run-local.sh), with docker (./run-docker.sh) or manually (using langchain serve).

When you have Poetry installed, simply type poetry install.

Install libmagic (or libmagic-dev), if necessary.

To run all evaluations and tests, you need to install celadon inside the train directory. See here for more information.

cd train
git clone https://huggingface.co/PleIAs/celadon

If you need to install LangChain CLI:

pip install -U langchain-cli

How you build docker image for local run (./run-docker.sh):

./build-local.sh

Name		Name	Last commit message	Last commit date
Latest commit History 21 Commits
app		app
eval		eval
packages		packages
train		train
.gitignore		.gitignore
.gitlab-ci.yml		.gitlab-ci.yml
Dockerfile		Dockerfile
Dockerfile.kaniko		Dockerfile.kaniko
LICENSE		LICENSE
README.md		README.md
build-local.sh		build-local.sh
eval_rag_and_moderation.ipynb		eval_rag_and_moderation.ipynb
poetry.lock		poetry.lock
pyproject.toml		pyproject.toml
run-docker.sh		run-docker.sh
run-local.sh		run-local.sh