feat(lightspeed): show expandable card for deep thinking responses

[debsmita1]

Hey, I just made a Pull Request!

Resolves:
https://issues.redhat.com/browse/RHIDP-11420

Adds an expandable card when the model responds with reasoning

deppthinking.mov

Test setup:

pull a model that supports deep thinking

ollama pull deepseek-r1:7b

replace run.yaml in lightspeed-stack.yaml with this

version: 2
image_name: llama-stack-examples
apis:
  - agents
  - inference
  - safety
  - tool_runtime
  - vector_io

# external_providers_dir: /Users/kjeeyar/Documents/lightspeed-providers/providers.d

providers:
  tool_runtime:
    - provider_id: model-context-protocol
      provider_type: remote::model-context-protocol
      config: {}
  inference:
    - provider_id: ollama
      provider_type: remote::vllm
      config:
        url: http://localhost:11434/v1
        api_token: dummy
    - provider_id: sentence-transformers
      provider_type: inline::sentence-transformers
      config: {}

  safety:
    - provider_id: llama-guard
      provider_type: inline::llama-guard
      config:
        model: llama3.2:3b
    - provider_id: llama-guard
      provider_type: inline::llama-guard
      config:
        model: deepseek-r1:7b
  agents:
    - provider_id: meta-reference
      provider_type: inline::meta-reference
      config:
        persistence_store:
          type: sqlite
          db_path: ${env.SQLITE_STORE_DIR:=~/.llama/distributions/custom}/agents_store.db
        responses_store:
          type: sqlite
          db_path: ${env.SQLITE_STORE_DIR:=~/.llama/distributions/custom}/responses_store.db
  vector_io:
    - provider_id: faiss
      provider_type: inline::faiss
      config:
        kvstore:
          type: sqlite
          namespace: null
          db_path: ${env.SQLITE_STORE_DIR:=~/.llama/distributions/custom}/faiss_store.db

metadata_store:
  type: sqlite
  db_path: ${env.SQLITE_STORE_DIR:=~/.llama/distributions/custom}/registry.db
  namespace: null

inference_store:
  type: sqlite
  db_path: ${env.SQLITE_STORE_DIR:=~/.llama/distributions/custom}/inference_store.db

models:
  - metadata: {}
    model_id: llama3.2:3b
    provider_id: ollama
    provider_model_id: llama3.2:3b
  - metadata: {}
    model_id: deepseek-r1:7b
    provider_id: ollama
    provider_model_id: deepseek-r1:7b
  - model_id: sentence-transformers/all-mpnet-base-v2
    metadata:
      embedding_dimension: 768
    model_type: embedding
    provider_id: sentence-transformers
    provider_model_id: sentence-transformers/all-mpnet-base-v2

shields: []
  # - shield_id: lightspeed_question_validity-shield
  #   provider_id: lightspeed_question_validity

server:
  port: 8321

and llama_stack url in lightspeed-stack.yaml

url: http://localhost:8321

Note: not all AI models support tool calling , like this deepseek model. So, ensure that you do not have mcp server configured in lightspeed-stack.yaml

✔️ Checklist

• A changeset describing the change and affected packages. (more info)
• Added or Updated documentation
• Tests for new functionality and regression tests for bug fixes
• Screenshots attached (for UI changes)

[rhdh-gh-app]

Changed Packages

Package Name	Package Path	Changeset Bump	Current Version
@red-hat-developer-hub/backstage-plugin-lightspeed	workspaces/lightspeed/plugins/lightspeed	patch	v1.1.2
app	workspaces/quickstart/packages/app	none	v0.0.0

[its-mitesh-kumar]

@debsmita1 In the recording I can see the Thinking block is below the main content, but I think it should be above the main content ?

[debsmita1]

@debsmita1 In the recording I can see the Thinking block is below the main content, but I think it should be above the main content ?

This is a Patternfly component that I am using for rendering the deepthinking content

[rohitkrai03]

Overall looks good. Just one comment - shouldn't the thinking component be rendered at the top even if the main content is available?

[debsmita1]

Fixed it!

cc @rohitkrai03 @its-mitesh-kumar

[sonarqubecloud]

Quality Gate passed

Issues
2 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[its-mitesh-kumar]

Code looks good to me . Though not able to verify it locally. Adding lgtm based on screen recording.
/lgtm
/approve

[rohitkrai03]

/approve