rag: ramalama serve --rag doesn't take into account of RAG #1163
Description
When running ramalama serve with --rag option, I was expecting that RAG-enabled chatbot is available on the served port, but it actually exposes the yet to be augumented llama-server:
$ ramalama --image quay.io/ramalama/ramalama-rag rag README.md localhost/myrag:0.1
$ ramalama --debug --image quay.io/ramalama/ramalama-rag serve --rag localhost/myrag:0.1 deepseek
Working directory: None
Ignore stderr: False
Ignore all: False
Command finished with return code: 0
Checking if 8080 is available
serving on port 8080
exec_cmd: podman run --rm -i --label ai.ramalama --name ramalama_d6eTz3XYFa --env=HOME=/tmp --init --security-opt=label=disable --cap-drop=all --security-opt=no-new-privileges --label ai.ramalama.model=ollama://deepseek-r1 --label ai.ramalama.engine=podman --label ai.ramalama.runtime=llama.cpp --label ai.ramalama.port=8080 --label ai.ramalama.command=serve --pull=newer -t -p 8080:8080 --device /dev/dri --device /dev/kfd -e HIP_VISIBLE_DEVICES=0 --mount=type=image,source=localhost/myrag:0.1,destination=/rag,rw=true --mount=type=bind,src=/home/ueno/.local/share/ramalama/models/ollama/deepseek-r1:latest,destination=/mnt/models/model.file,ro quay.io/ramalama/ramalama-rag:0.7 bash -c nohup llama-server --port 8080 --model /mnt/models/model.file --alias deepseek-r1 --ctx-size 2048 --temp 0.8 --jinja -v -ngl 0 --threads 8 --host 0.0.0.0 &> /tmp/llama-server.log & rag_framework run /rag/vector.db
> What's the goal of RamaLama project?
The RamaLama project's goal is to create a robust, secure, and scalable platform for serving large AI models, [...]
$ ramalama client http://127.0.0.1:8080
> What's the goal of RamaLama project?
RamaLama is a cutting-edge research project focused on exploring consciousness, [...]It seems, even with ramalama serve, rag_framework is called with "run" subcommand and just exposing the port listened by the internal llama-server process.