Skip to content

share-flow-sequence.md

Latest commit

 

History

History
99 lines (80 loc) · 3.88 KB

share-flow-sequence.md

File metadata and controls

99 lines (80 loc) · 3.88 KB

Share to Expert / Moderator Mode Library — Sequence Diagrams

1. Expert Share (Share to Expert Library)

sequenceDiagram
    autonumber
    participant User
    participant ExpertMgmt as ExpertManagement
    participant API as topicExpertsApi
    participant Backend as topic_experts API
    participant Workspace as workspace/topics/{id}
    participant Libs as libs/experts/topiclab_shared
    participant Reload as reload_expert_specs

    User->>ExpertMgmt: Click "Share to Expert Library"
    ExpertMgmt->>User: confirm dialog
    User->>ExpertMgmt: Confirm
    ExpertMgmt->>API: share(topicId, expert.name)
    API->>Backend: POST /topics/{id}/experts/{name}/share

    Backend->>Backend: Validate topic exists
    Backend->>Backend: Validate not built-in (source≠default)
    Backend->>Workspace: Read agents/{name}/role.md
    Backend->>Workspace: Read config/experts_metadata.json
    Workspace-->>Backend: role content + metadata

    Backend->>Libs: Write {name}.md
    Backend->>Libs: Update meta.json (experts entry)
    Backend->>Reload: reload_expert_specs()
    Reload->>Libs: load_aggregated_experts_meta
    Reload->>Reload: EXPERT_SPECS.clear() + update()
    Backend->>Backend: invalidate_libs_cache()

    Backend-->>API: 200 { message, expert_name }
    API-->>ExpertMgmt: Success
    ExpertMgmt->>ExpertMgmt: loadExperts() refresh topic experts
    ExpertMgmt->>User: handleApiSuccess "Shared to platform"
Loading

2. Moderator Mode Share (Share to Moderator Mode Library)

sequenceDiagram
    autonumber
    participant User
    participant TopicConfig as TopicConfigTabs
    participant ShareDialog
    participant API as moderatorModesApi
    participant Backend as moderator_modes API
    participant Workspace as workspace/topics/{id}/config
    participant Libs as libs/moderator_modes/topiclab_shared
    participant Reload as reload_moderator_modes

    User->>TopicConfig: Click "Share to Moderator Mode Library" in custom mode dialog
    TopicConfig->>ShareDialog: Open share dialog
    User->>ShareDialog: Enter mode_id, name, description
    User->>ShareDialog: Click "Confirm Share"
    ShareDialog->>API: share(topicId, { mode_id, name?, description? })
    API->>Backend: POST /topics/{id}/moderator-mode/share

    Backend->>Workspace: Read config/moderator_mode.json
    Backend->>Backend: Validate mode_id=custom and custom_prompt exists
    Backend->>Backend: Validate mode_id does not conflict with built-in

    Backend->>Libs: Write {mode_id}.md (custom_prompt content)
    Backend->>Libs: Update meta.json (modes entry)
    Backend->>Reload: reload_moderator_modes()
    Reload->>Libs: get_modes_and_common
    Reload->>Reload: PRESET_MODES.clear() + update()
    Backend->>Backend: invalidate_libs_cache()

    Backend-->>API: 200 { message, mode_id }
    API-->>ShareDialog: Success
    ShareDialog->>ShareDialog: Close dialog
    ShareDialog->>API: listAssignable() refresh moderator modes
    ShareDialog->>User: handleApiSuccess "Shared to Moderator Mode Library"
Loading

3. Key Steps Summary

Step Expert Share Moderator Mode Share
Data source workspace/topics/{id}/agents/{name}/role.md custom_prompt from workspace/topics/{id}/config/moderator_mode.json
Write target libs/experts/topiclab_shared/{name}.md libs/moderator_modes/topiclab_shared/{mode_id}.md
Metadata topiclab_shared/meta.json → experts topiclab_shared/meta.json → modes
Category category: topiclab category: topiclab
In-memory reload reload_expert_specs() in-place updates EXPERT_SPECS reload_moderator_modes() in-place updates PRESET_MODES
Cache invalidate_libs_cache() Same

4. Persistence

With Docker, libs is mounted as ./backend/libs. User-shared content written to topiclab_shared/ persists on the host; it remains available after container restarts.