152 lines
8.6 KiB
Markdown
152 lines
8.6 KiB
Markdown
# System Diagram Alignment
|
|
|
|
## Purpose
|
|
|
|
This document maps the legacy Pegasus/Jibo cloud `system_diagram.png` architecture to the current OpenJibo `1.0.19` cloud.
|
|
|
|
Use it to keep release planning grounded in three views:
|
|
|
|
- where we were (legacy design intent)
|
|
- where we are (current hosted `.NET` implementation)
|
|
- where we are headed (next architecture slices)
|
|
|
|
As-of date: `2026-05-07`
|
|
|
|
## Diagram Inputs
|
|
|
|
- Legacy system architecture: `C:\Projects\jibo\pegasus\resources\system_diagram.png`
|
|
- Legacy generic skill scaffold: `C:\Projects\jibo\pegasus\packages\template-skill\docs\TemplateSkill.png`
|
|
- Legacy listener state machine: `C:\Projects\jibo\sdk\packages\skills-service-manager\resources\state-diagrams\glsm.png`
|
|
|
|
## Template Skill Verdict
|
|
|
|
The template-skill diagram is a generic scaffold, not a production behavior contract.
|
|
|
|
Evidence:
|
|
|
|
- `C:\Projects\jibo\pegasus\packages\template-skill\src\TemplateSkill.ts` is a starter graph (`Intent Split` -> `Do MIM` -> `Complete` -> `Done`).
|
|
- `C:\Projects\jibo\pegasus\packages\template-skill\src\nodes\MemoSplitNode.ts` uses placeholder memo validation (`SomeThing`).
|
|
|
|
Conclusion: do not treat template-skill flow as a port target. Treat it as a shape reference only.
|
|
|
|
## System Diagram Mapping
|
|
|
|
| Legacy block | OpenJibo `1.0.19` equivalent | Current gap / opportunity |
|
|
| --- | --- | --- |
|
|
| `Auth` | [JiboCloudProtocolService.cs](../src/Jibo.Cloud/dotnet/src/Jibo.Cloud.Application/Services/JiboCloudProtocolService.cs) (`CreateHubToken`, `CreateAccessToken`, account handlers) | move from in-memory/session stubs to durable tenant/account identity services |
|
|
| `Loop` | [JiboCloudProtocolService.cs](../src/Jibo.Cloud/dotnet/src/Jibo.Cloud.Application/Services/JiboCloudProtocolService.cs) (`HandleLoop`) + [InMemoryCloudStateStore.cs](../src/Jibo.Cloud/dotnet/src/Jibo.Cloud.Infrastructure/Persistence/InMemoryCloudStateStore.cs) | richer loop/member lifecycle and onboarding flows |
|
|
| `Hub` | [JiboWebSocketService.cs](../src/Jibo.Cloud/dotnet/src/Jibo.Cloud.Application/Services/JiboWebSocketService.cs) + [WebSocketTurnFinalizationService.cs](../src/Jibo.Cloud/dotnet/src/Jibo.Cloud.Application/Services/WebSocketTurnFinalizationService.cs) | split hub responsibilities into clearer protocol, routing, and orchestration boundaries |
|
|
| `ASR Handler` | STT strategy selection in [WebSocketTurnFinalizationService.cs](../src/Jibo.Cloud/dotnet/src/Jibo.Cloud.Application/Services/WebSocketTurnFinalizationService.cs) + DI in [ServiceCollectionExtensions.cs](../src/Jibo.Cloud/dotnet/src/Jibo.Cloud.Infrastructure/DependencyInjection/ServiceCollectionExtensions.cs) | short-turn reliability, managed STT comparison, and better low-signal/noise handling |
|
|
| `Parser / Robust Parser` | rule-based intent resolution in [JiboInteractionService.cs](../src/Jibo.Cloud/dotnet/src/Jibo.Cloud.Application/Services/JiboInteractionService.cs) + focused state machines (personal report/chitchat) | deeper phrase import from Pegasus intents/entities plus ambiguity guardrails |
|
|
| `Skill Router` | [JiboInteractionService.cs](../src/Jibo.Cloud/dotnet/src/Jibo.Cloud.Application/Services/JiboInteractionService.cs) decision switch and local skill payload shaping | external skill routing config and safer declarative intent mapping |
|
|
| `Proactivity Selector` | weighted candidate selection in [JiboInteractionService.cs](../src/Jibo.Cloud/dotnet/src/Jibo.Cloud.Application/Services/JiboInteractionService.cs) + pending-offer session state in [WebSocketTurnFinalizationService.cs](../src/Jibo.Cloud/dotnet/src/Jibo.Cloud.Application/Services/WebSocketTurnFinalizationService.cs) | externalized proactivity catalog, cooldown policy, and broader category coverage |
|
|
| `Presence / Identity Context` | runtime context passthrough in [ProtocolToTurnContextMapper.cs](../src/Jibo.Cloud/dotnet/src/Jibo.Cloud.Application/Services/ProtocolToTurnContextMapper.cs) and turn metadata handling in [WebSocketTurnFinalizationService.cs](../src/Jibo.Cloud/dotnet/src/Jibo.Cloud.Application/Services/WebSocketTurnFinalizationService.cs) | normalize `runtime.perception` fields (`speaker`, `peoplePresent`, focused person) for greeting/proactivity policy decisions |
|
|
| `Skill Registry` | implicit in current code/routing | formal registry abstraction for local/cloud capabilities and manifest metadata |
|
|
| `History` | tenant-scoped memory store in [InMemoryPersonalMemoryStore.cs](../src/Jibo.Cloud/dotnet/src/Jibo.Cloud.Infrastructure/Persistence/InMemoryPersonalMemoryStore.cs) | durable multi-tenant persistence and history timeline/query support |
|
|
| `Lasso` provider aggregation | partial provider integration via weather provider wiring in [ServiceCollectionExtensions.cs](../src/Jibo.Cloud/dotnet/src/Jibo.Cloud.Infrastructure/DependencyInjection/ServiceCollectionExtensions.cs) | full aggregation service for weather/news/calendar/knowledge inputs |
|
|
| `Proactivity Catalog` | in-code candidate lists/weights | explicit catalog service with tuned weights and operator controls |
|
|
| `Audio Logs` | file telemetry sinks in infrastructure telemetry | hosted indexed capture/retention for multi-operator analysis |
|
|
|
|
## GLSM Listener Flow Alignment (`2026-05-06`)
|
|
|
|
Captured source:
|
|
|
|
- `C:\Projects\jibo\sdk\packages\skills-service-manager\resources\state-diagrams\glsm.png`
|
|
|
|
First OpenJibo support slice (implemented):
|
|
|
|
- explicit derived listener phases are now emitted in cloud diagnostics:
|
|
- `HJ_LISTENING`
|
|
- `LISTENING`
|
|
- `WAIT_LISTEN_FINISHED`
|
|
- `DISPATCH_DIALOG`
|
|
- `PROCESS_LISTENER_QUEUE`
|
|
- turn telemetry now records `glsm_phase_transition` with previous/next state and trigger
|
|
- websocket telemetry now includes `glsmPhase` on binary, context, and turn-processed events
|
|
- stale pending-listen recovery is now implemented:
|
|
- when a pending `LISTEN` stays open long enough with no context/audio, a new hotphrase listen can recover the stuck state before continuing
|
|
|
|
Current parity boundary:
|
|
|
|
- this slice focuses on listener lifecycle observability plus stuck-listen recovery
|
|
- deeper explicit parity states from GLSM (`Interrupt Listeners`, `Handle Launch Parse`, `Handle Global Parse`, `Dispatch Dialog` sub-branches) are next candidates once this capture-driven slice is validated live
|
|
|
|
## Where We Were
|
|
|
|
Legacy cloud design was service-oriented around:
|
|
|
|
- hub orchestration
|
|
- parser robustness
|
|
- skill routing
|
|
- proactivity selection
|
|
- history/memory and provider aggregation
|
|
|
|
It emphasized a personality-rich surface while still being operationally observable.
|
|
|
|
## Where We Are
|
|
|
|
OpenJibo `1.0.19` is a functional hosted `.NET` modular monolith with:
|
|
|
|
- protocol compatibility paths for HTTP and websocket robot flows
|
|
- deterministic intent routing plus state-machine slices
|
|
- tenant-scoped memory foundation
|
|
- first proactivity baseline
|
|
- first external weather provider integration
|
|
|
|
This is the right shape for rapid parity plus safe incremental growth.
|
|
|
|
## Where We Are Headed
|
|
|
|
Near-term architecture evolution should preserve current shipping velocity:
|
|
|
|
1. Expand parser coverage and ambiguity guardrails from Pegasus phrase corpora.
|
|
2. Externalize proactivity policy and category catalogs.
|
|
3. Move memory from in-memory to durable multi-tenant backing stores.
|
|
4. Add stronger observability around STT, parser decisions, and follow-up turn state.
|
|
5. Build a focused aggregation layer (Lasso-like) for multi-provider content.
|
|
|
|
## Charm Preservation Rules
|
|
|
|
To keep Jibo's charm while modernizing the platform:
|
|
|
|
- keep MIM/ESML and expressive animation hooks as first-class outputs
|
|
- keep deterministic command-vs-question behavior for personality reliability
|
|
- layer richer provider data behind stable personality and gesture patterns
|
|
- prefer small source-backed slices over broad rewrites
|
|
|
|
## Queued Next `1.0.19` Task
|
|
|
|
The next queued implementation task is:
|
|
|
|
- `Dialog parsing expansion and ambiguity guardrails`
|
|
|
|
Tracking anchors:
|
|
|
|
- [release-1.0.19-plan.md](release-1.0.19-plan.md)
|
|
- [feature-backlog.md](feature-backlog.md)
|
|
|
|
Primary objective:
|
|
|
|
- import Pegasus parser intent phrases/entities to improve intent confidence while preserving command-vs-question personality behavior.
|
|
|
|
## Greetings And Presence Track (`2026-05-07`)
|
|
|
|
A dedicated presence-aware greetings plan is now captured for the next personality slice, grounded in Pegasus `@be/greetings` state, identity, and proactive policy behavior.
|
|
|
|
Reference:
|
|
|
|
- [greetings-presence-plan.md](greetings-presence-plan.md)
|
|
|
|
## Personal Report Parity Track (`2026-05-07`)
|
|
|
|
Personal report parity planning is now captured with a source-anchored implementation sequence for:
|
|
|
|
- weather visual/personality parity
|
|
- live news provider path
|
|
- commute provider path
|
|
- calendar/report coverage matrix
|
|
|
|
Reference:
|
|
|
|
- [personal-report-parity-plan.md](personal-report-parity-plan.md)
|