GenAI changes to allow it to run on Debian GNU/Linux unstable

[HaakonME]

I needed a few changes to make GenAI work on my Debian GNU/Linux unstable. If this is interesting to you, perhaps you can turn it into --profile debian-unstable option?

Cheers,
Haakon

[matthieuml]

Isn't docker-compose deprecated for docker compose ?

[HaakonME]

"docker compose" is the updated version from Docker, yes. However, in the latest version of Debian GNU/Linux and the Unstable branch, it is still a separate command for "docker" and "docker-compose". That is why I suggested, if possible, to create an extra profile, e.g. --profile debianunstable, rather than assume every Linux can use --profile linux. I do not know if this is possible, but it would be nice. As a general rule, "docker compose", with the space is the updated way. I hope this helps.

[matthieuml]

What I meant is we should encourage the use of docker compose. But if it is available on Debian Unstable, why not use it ?

Anyway, the changes you made should work even when using docker compose (the context: . is implicit in docker compose, don't know for docker-compose). The only thing I would do differently is creating another start script for the old docker-compose instead of changing the actual one.

I would still be interested in what error you had before, maybe there is another workaround!

[oskarhane]

Thanks!
What's the issue you're experiencing?

Also, the start.sh file should be removed so need to edit that (it's a leftover from early days).

[HaakonME]

@oskarhane , the issue is that there is a difference between official Docker and official Debian unstable packages for Docker, the latter contains docker without compose and a separate command called docker-compose. I am not sure if this means means Docker v1 vs Docker v2, but the PR and the .env below resolved the issue for me on Debian unstable, so maybe it can help someone else.

Note to someone else: using the LLM below will take quite some time to fetch in the background, probably several minutes - be patient.

#OPENAI_API_KEY=sk-...
#OLLAMA_BASE_URL=http://host.docker.internal:11434
OLLAMA_BASE_URL=http://llm:11434
NEO4J_URI=neo4j://database:7687
NEO4J_USERNAME=neo4j
NEO4J_PASSWORD=password
#LLM=llama2 #or any Ollama model tag, or gpt-4 or gpt-3.5
LLM=llama2-uncensored:7b
EMBEDDING_MODEL=sentence_transformer #or openai or ollama

LANGCHAIN_ENDPOINT="https://api.smith.langchain.com"
LANGCHAIN_TRACING_V2=true # false
LANGCHAIN_PROJECT=#your-project-name
LANGCHAIN_API_KEY=#your-api-key ls_...