Commit bd015bdb authored by Miguel Matos's avatar Miguel Matos
Browse files

feat: configure the LLM provider in Helm, compose and Dockerfile 

Add aiAgent.llm.* and secrets.openaiApiKey to the chart, make the Groq key
optional, and pass the OpenAI-compatible variables through docker-compose.
parent fc5fdd39

ai_agent/Dockerfile

+7 −0
Changes for ai_agent/Dockerfile: 7 added lines, 0 removed lines.
Original line number Diff line number Diff line
@@ -4,9 +4,16 @@ ENV APP_HOST=0.0.0.0 
ENV APP_PORT=9013

ENV MCP_SERVER_URL=http://127.0.0.1:8004/mcp



ENV LLM_PROVIDER=groq



ENV GROQ_API_KEY=""

ENV GROQ_MODEL_NAME=openai/gpt-oss-20b



ENV OPENAI_BASE_URL=""

ENV OPENAI_MODEL_NAME=""

ENV OPENAI_API_KEY=""

ENV OPENAI_TIMEOUT=120



WORKDIR /app



COPY requirements.txt .

deployment/docker/README.md

+4 −0
Changes for deployment/docker/README.md: 4 added lines, 0 removed lines.
Original line number Diff line number Diff line
@@ -13,6 +13,10 @@ OEG_SERVICE_URL=http://your-oeg-api-url 
GROQ_MODEL_NAME=qwen/qwen3.8-27b

```



For an OpenAI-compatible endpoint such as Open WebUI, set `LLM_PROVIDER=openai`,

`OPENAI_BASE_URL` (`https://<host>/api`), `OPENAI_MODEL_NAME` and `OPENAI_API_KEY`; optional

`OPENAI_TIMEOUT` (`120`).



Optional, with the defaults the compose file uses: `REDIS_URL` (`redis://redis:6379/0`),

`MEMORY_WINDOW_TURNS` (`10`) and `MEMORY_TTL_SECONDS` (`3600`) for the agent's session memory.

deployment/docker/docker-compose.yaml

+6 −1
Changes for deployment/docker/docker-compose.yaml: 6 added lines, 1 removed line.
Original line number Diff line number Diff line
@@ -35,8 +35,13 @@ services: 
        condition: service_healthy

    environment:

      MCP_SERVER_URL: http://mcp:8004/mcp

      GROQ_API_KEY: ${GROQ_API_KEY}

      LLM_PROVIDER: ${LLM_PROVIDER:-groq}

      GROQ_API_KEY: ${GROQ_API_KEY:-}

      GROQ_MODEL_NAME: ${GROQ_MODEL_NAME:-qwen/qwen3.8-27b}

      OPENAI_BASE_URL: ${OPENAI_BASE_URL:-}

      OPENAI_MODEL_NAME: ${OPENAI_MODEL_NAME:-}

      OPENAI_API_KEY: ${OPENAI_API_KEY:-}

      OPENAI_TIMEOUT: ${OPENAI_TIMEOUT:-120}

      REDIS_URL: ${REDIS_URL:-redis://redis:6379/0}

      MEMORY_WINDOW_TURNS: ${MEMORY_WINDOW_TURNS:-10}

      MEMORY_TTL_SECONDS: ${MEMORY_TTL_SECONDS:-3600}

deployment/helm/README.md

+5 −1
Changes for deployment/helm/README.md: 5 added lines, 1 removed line.
Original line number Diff line number Diff line
@@ -10,7 +10,7 @@ waits for the MCP server to be healthy before starting (via initContainer). 


- Kubernetes cluster

- Helm 3+

- Groq API key

- Groq API key, or an OpenAI-compatible endpoint such as Open WebUI (see below)

- OEG service reachable from within the cluster



## Install
@@ -23,6 +23,10 @@ helm install ai2test ./deployment/helm/ai2 \ 
  --set mcp.oegUrl=http://<oeg-host>

```



For an OpenAI-compatible endpoint such as Open WebUI, use `--set aiAgent.llm.provider=openai`,

`--set aiAgent.llm.openai.baseUrl=https://<host>/api`, `--set aiAgent.llm.openai.model=<model>` and

`--set secrets.openaiApiKey=<key>`.



Without the umbrella chart there is no Keycloak; add `--set global.auth.enabled=false` to call

AI2 without a token.

deployment/helm/ai2/README.md

+8 −3
Changes for deployment/helm/ai2/README.md: 8 added lines, 3 removed lines.
Original line number Diff line number Diff line
@@ -2,7 +2,7 @@ 


Deploys AI2, the platform's natural-language assistant:



- **ai-agent**: takes a chat message, asks a Groq model what to do and calls the MCP server's

- **ai-agent**: takes a chat message, asks an LLM (Groq or an OpenAI-compatible endpoint) what to do and calls the MCP server's

  tools. The portal forwards `/api/ai2/chat` to it.

- **mcp**: MCP server whose tools call OEG's CAMARA APIs.

- **redis**: in-memory store for the agent's session memory. Nothing is persisted.
@@ -22,12 +22,17 @@ both components get `AUTH_ENABLED=false`. 
| `aiAgent.image.repository` / `tag` | Agent image | `labs.etsi.org:5050/oop/code/ai2/ai_agent`, `2.0.0` |

| `aiAgent.waitForMcpImage` | Init container image that waits for the MCP server | `curlimages/curl:8.11.1` |

| `aiAgent.groqModelName` | Groq model | `qwen/qwen3.8-27b` |

| `aiAgent.llm.provider` | `groq` or `openai` | `groq` |

| `aiAgent.memory.windowTurns` | Turns of a chat kept as context | `3` |

| `aiAgent.memory.ttlSeconds` | How long an idle chat is remembered | `3600` |

| `aiAgent.service.name` / `port` | Agent Service, the address the portal calls | `global.ai2.host` / `port` (`ai2-ai-agent`, `9013`) |

| `aiAgent.service.type` / `nodePort` | Agent Service type | `ClusterIP` |

| `secrets.groqApiKey` | Groq API key; required for the agent to answer | `""` |

| `secrets.existingSecret` | Existing Secret with the key under `GROQ_API_KEY`; `groqApiKey` is then ignored | `""` |

| `aiAgent.llm.openai.baseUrl` | OpenAI-compatible endpoint (Open WebUI: `https://<host>/api`); required when the provider is `openai` | `""` |

| `aiAgent.llm.openai.model` | Model name; required when the provider is `openai` | `""` |

| `aiAgent.llm.openai.timeoutSeconds` | Request timeout | `120` |

| `secrets.openaiApiKey` | API key, sent as a Bearer token; required when the provider is `openai` | `""` |

| `secrets.groqApiKey` | Groq API key; required when the provider is `groq` | `""` |

| `secrets.existingSecret` | Existing Secret with `GROQ_API_KEY` and/or `OPENAI_API_KEY`; the keys above are then ignored | `""` |

| `redis.enabled` | Deploy Redis; `false` turns session memory off | `true` |

| `redis.image` | Redis image | `redis:7.4-alpine` |