e9e6fadb1d
- Fix ChatBubble to handle non-string content with String() wrapper - Fix API route to use generateText for non-streaming requests - Add @ai-sdk/openai-compatible for non-OpenAI providers (DeepSeek, etc.) - Use Chat Completions API instead of Responses API for compatible providers - Update ChatBubble tests and fix component exports to kebab-case - Remove stale PascalCase ChatBubble.tsx file
89 lines
5.3 KiB
Markdown
89 lines
5.3 KiB
Markdown
# Test Design: Epic 1 - Active Listening
|
|
|
|
**Epic:** 1 (Core Chat & Teacher Agent)
|
|
**Scope:** Epic-Level
|
|
**Date:** 2026-01-25
|
|
**Author:** QA Architect (AI)
|
|
|
|
## 1. Risk Assessment
|
|
|
|
### Identified Risks
|
|
|
|
| Risk ID | Category | Title | Description | Probability (1-3) | Impact (1-3) | Score | Action |
|
|
| :-------- | :------- | :------------------------------- | :-------------------------------------------------------------------------------------------------- | :---------------- | :----------- | :---- | :----------- |
|
|
| **R-1.1** | DATA | **Data Loss on Storage Failure** | User's chat session is lost due to IndexedDB quota exceeded, browser clearing, or schema mismatch. | 2 (Possible) | 3 (Critical) | **6** | **MITIGATE** |
|
|
| **R-1.2** | SEC | **API Key Leakage** | User's BYOD API Key is logged, exported in history, or sent to a non-provider endpoint. | 1 (Unlikely) | 3 (Critical) | 3 | MONITOR |
|
|
| **R-1.3** | PERF | **High Chat Latency** | "Teacher" agent response exceeds 3s, breaking the conversational "venting" flow. | 2 (Possible) | 2 (Degraded) | 4 | MONITOR |
|
|
| **R-1.4** | BUS | **Poor Intent Detection** | AI fails to distinguish "Venting" from "Insight", annoying the user with wrong mode. | 2 (Possible) | 2 (Degraded) | 4 | MONITOR |
|
|
| **R-1.5** | TECH | **Offline State Inconsistency** | App fails to load history or queue messages when device is offline (Service Worker/IndexedDB fail). | 2 (Possible) | 2 (Degraded) | 4 | MONITOR |
|
|
|
|
### Mitigation Strategies (High Risks)
|
|
|
|
**R-1.1: Data Loss on Storage Failure (Score 6)**
|
|
* **Mitigation:** Implement robust error handling around all Dexie operations. Add a "Quota Exceeded" UI warning. Ensure schema versioning is tested.
|
|
* **Owner:** Dev Team
|
|
* **Validation:** Unit tests for `ChatService` storage failures; E2E test for persistence across reloads.
|
|
|
|
---
|
|
|
|
## 2. Test Coverage Plan
|
|
|
|
### Acceptance Criteria Mapping
|
|
|
|
| Story | ID | Scenario | Level | Priority | Risk Link |
|
|
| :------ | :---- | :------------------------------------------------------ | :---------- | :------- | :-------- |
|
|
| **1.1** | 1.1.1 | New user sees initialized empty state (DB created) | Component | P1 | - |
|
|
| **1.1** | 1.1.2 | Sent message is saved to IndexedDB | Integration | **P0** | R-1.1 |
|
|
| **1.1** | 1.1.3 | Chat history persists after page reload | E2E | **P0** | R-1.1 |
|
|
| **1.1** | 1.1.4 | App loads history while offline | E2E | P1 | R-1.5 |
|
|
| **1.2** | 1.2.1 | UI renders "Morning Mist" theme bubbles | Component | P2 | - |
|
|
| **1.2** | 1.2.2 | Auto-scroll to bottom on new message | Component | P2 | - |
|
|
| **1.2** | 1.2.3 | "Teacher is typing..." indicator appears during wait | Component | P2 | R-1.3 |
|
|
| **1.3** | 1.3.1 | AI Classifies "Venting" vs "Insight" correctly (Mocked) | Unit | P1 | R-1.4 |
|
|
| **1.3** | 1.3.2 | Client sends request to custom Provider URL | Integration | **P0** | R-1.2 |
|
|
| **1.3** | 1.3.3 | API Key retrieved from secure storage (not hardcoded) | Unit | **P0** | R-1.2 |
|
|
| **1.3** | 1.3.4 | Response time < 3s (Performance Check) | E2E | P3 | R-1.3 |
|
|
| **1.4** | 1.4.1 | "Fast Track" button skips probing questions | E2E | P1 | - |
|
|
| **1.4** | 1.4.2 | Fast Track triggers immediate draft generation | Integration | P1 | - |
|
|
|
|
### Test Levels Strategy
|
|
|
|
* **Unit Tests (Vitest):**
|
|
* Focus on `ChatService` logic (DB interactions).
|
|
* Focus on `PromptEngine` (constructing prompts from templates).
|
|
* Focus on `SettingsService` (secure key storage/retrieval).
|
|
* **Component Tests (React Testing Library / Storybook):**
|
|
* `ChatBubble`: Verify styling (User vs AI).
|
|
* `ChatWindow`: Verify scroll behavior and typing indicators.
|
|
* `FastTrackToggle`: Verify state change.
|
|
* **Integration/E2E (Playwright):**
|
|
* **Critical Path (P0):** User configures Key -> Starts Chat -> Sends Message -> Verifies Persistence.
|
|
* **Offline Path (P1):** Load app offline -> verify history visible.
|
|
|
|
---
|
|
|
|
## 3. Execution Plan
|
|
|
|
### Smoke Tests (Pre-Merge)
|
|
1. **Unit:** All `ChatService` tests (Persistence logic).
|
|
2. **E2E:** "Happy Path" - User can send a message and see it appear.
|
|
|
|
### Regression Suite (Nightly)
|
|
1. **E2E:** Full persistence check (Reload page).
|
|
2. **E2E:** Offline mode loading.
|
|
3. **Performance:** Measure Time-to-First-Token (TTFT) simulation.
|
|
|
|
### Resource Estimates
|
|
* **P0 Scenarios:** 4 tests (approx. 4 hours implementation).
|
|
* **P1 Scenarios:** 4 tests (approx. 3 hours implementation).
|
|
* **P2/P3 Scenarios:** 3 tests (approx. 1 hour implementation).
|
|
* **Total Effort:** ~1 day.
|
|
|
|
---
|
|
|
|
## 4. Quality Gate Criteria
|
|
|
|
* **Pass Rate:** 100% on P0 tests.
|
|
* **Coverage:** 100% Unit coverage on `services/chat-service.ts`.
|
|
* **Mitigation:** `R-1.1` (Data Loss) must have verified error handling tests.
|