Add frames agent

[jbouder]

Reference Issues or PRs

What does this implement/fix?

Put a x in the boxes that apply

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds a feature)
• Breaking change (fix or feature that would cause existing features not to work as expected)
• Documentation Update
• Code style update (formatting, renaming)
• Refactoring (no functional changes, no API changes)
• Build related changes
• Other (please describe):

Testing

• Did you test the pull request locally?
• Did you add new tests?

Documentation

Access-centered content checklist

Text styling

• The content is written with plain language (where relevant).
• If there are headers, they use the proper header tags (with only one level-one header: H1 or # in markdown).
• All links describe where they link to (for example, check the Nebari website).
• This content adheres to the Nebari style guides.

Non-text content

• All content is represented as text (for example, images need alt text, and videos need captions or descriptive transcripts).
• If there are emojis, there are not more than three in a row.
• Don't use flashing GIFs or videos.
• If the content were to be read as plain text, it still makes sense, and no information is missing.

Any other comments?

[pmeier]

With #164 merged, all the MCP handling in this PR is superseded. The only thing left is setting the agent name and system prompt. Both can be set through the agent constructor and thus directly through the ravnar config. Meaning, unless we need anything beyond the MCP connection and the system prompt, this PR does not add anything.

resolved after merge

[pmeier]

@jbouder I cleaned up the branch after #164 was merged. As of now a valid config for this branch could look like

agents:
  static:
    frames:
      cls_or_fn: ravnar_nebari_chat.make_frames_agent
      params:
        agent:
          cls_or_fn: pydantic_ai.Agent
          params:
            model:
              cls_or_fn: pydantic_ai.models.google.GoogleModel
              params:
                model_name: gemini-flash-latest
                provider:
                  cls_or_fn: pydantic_ai.providers.google.GoogleProvider
                  params:
                    api_key: '{{ GOOGLE_API_KEY }}'
        mcp_url: "http://my-mcp-server"
        oidc_issuer: "http://keycloak-keycloakx-http.keycloak.svc.cluster.local:8080/realms/nebari"
        client_id: '{{ CLIENT_SECRET }}'
        client_secret: '{{ CLIENT_SECRET }}'

However, as explained here likely this branch is not needed. Here is how to get the exact same result on top of 0.0.14

agents:
  static:
    frames:
      cls_or_fn: ravnar.agents.PydanticAiAgentWrapper
      params:
        agent:
          cls_or_fn: pydantic_ai.Agent
          params:
            model:
              cls_or_fn: pydantic_ai.models.google.GoogleModel
              params:
                model_name: gemini-flash-latest
                provider:
                  cls_or_fn: pydantic_ai.providers.google.GoogleProvider
                  params:
                    api_key: '{{ GOOGLE_API_KEY }}'
            name: Nebari Frames Agent
            system_prompt: |
              You are an agent that answers the user's questions using the
              nebari-frames tools. Those tools are provided over MCP; inspect the
              tools available to you and call them to fetch data and perform
              actions rather than guessing.
  
              Prefer calling a tool over answering from memory whenever the answer
              depends on live data. When a tool fails or returns nothing useful,
              explain what you tried instead of inventing a result.
            toolsets:
              - cls_or_fn: ravnar_nebari_mcp.ImpersonatingMCPToolset
                params:
                  mcp_toolset_factory:
                    cls_or_fn: ravnar_nebari_mcp.bearer_token_mcp_toolset_factory
                    params:
                      url: http://my-mcp-server
                  impersonator:
                    cls_or_fn: ravnar_nebari_mcp.OIDCImpersonator
                    params:
                      issuer: http://keycloak-keycloakx-http.keycloak.svc.cluster.local:8080/realms/nebari
                      client_id: '{{ CLIENT_ID }}'
                      client_secret: '{{ IMPERSONATOR_CLIENT_SECRET }}'

[jbouder]

It works! Closing this PR.