Files
pedwfrontend/lib/domain/case-lifecycle/README.md
T
Robert Bond 9595ad86df Merged PR 2374: addeding domain layer extrraction
# Summary

This PR introduces a **Case Lifecycle Domain Boundary** to centralize lifecycle decision logic and reduce coupling within the appeals application.

The work is **behaviour-preserving** and introduces no intentional changes to business rules, CRM integrations, translations, dashboards, API routes, or user-facing functionality.

## What was added

New lifecycle boundary:

```text
lib/domain/case-lifecycle/
```

Key responsibilities extracted:

- Specialist process normalization
- Appeal type mapping
- Specialist process stage override mapping
- Stage case-type resolution
- Stage catalogue lookup
- Closed-case status recognition
- Lifecycle stage index resolution
- Lifecycle stage status assignment

## Behaviour preserved

Characterization tests were added before each extraction to preserve:

- Appeal type mapping and aliases
- Specialist process handling
- Lifecycle stage progression
- Closed-case handling
- Status assignment (`complete`, `in-progress`, `not-started`)
- Existing ROW behaviour
- Existing `statuscode` lifecycle semantics

Closed-case recognition remains unchanged for:

```text
1000
5
6
846040013
846040059
846040060
```

## Documentation

Added:

```text
lib/domain/case-lifecycle/README.md
```

Documenting:

- Boundary ownership
- Non-goals
- Lifecycle invariants
- Known architectural constraints
- Future extraction roadmap

## Testing

Added lifecycle characterization coverage for:

- Stage wrapper behaviour
- Specialist process normalization
- Appeal type mapping
- Specialist process stage mapping
- Stage case-type resolution
- Stage catalogue lookup
- Progress behaviour
- Closed-case status handling
- Stage index resolution
- Stage status assignment

## Validation

- Lifecycle characterization tests passed
- `npm run lint` passed with no errors

## Out of Scope

No changes to:

- Stage catalogue ownership
- Representation eligibility
- Dashboard calculations
- CRM/OData queries
- API routes
- Redux state
- EN/CY translations
- Event visibility logic

## Risk

**Low risk**

The refactor was delivered through small, characterization-first slices with no functional changes intended.

Related work items: #23527
2026-06-08 09:34:30 +00:00

151 lines
4.8 KiB
Markdown

# Case Lifecycle Domain Boundary
## Purpose
This directory contains a narrow **case lifecycle decision boundary**, not a full domain model.
Its current purpose is to centralise **behaviour-preserving interpretation logic** that was previously embedded in UI-adjacent helpers, while keeping live behaviour unchanged during refactor.
This boundary should be treated as an incremental extraction seam for read-only lifecycle interpretation, not as a place to redesign lifecycle behaviour.
## Current ownership
### `getLifecycleStagesForCase(...)`
- thin wrapper around the existing stage/progress helper
- preserves existing output shape for current consumers
- does not define new lifecycle semantics
### `normalizeSpecialistProcess(...)`
- raw CRM field fallback only
- preserves current canonical and misspelled field fallback behaviour
- does not validate or reinterpret CRM values
### `mapAppealType(...)`
- current appeal-type-to-case-type mapping
- current alias handling
- current unknown input behaviour
### `mapSpecialistProcessStageType(...)`
- specialist-process stage override mapping
- preserves current placeholder-like CRM specialist process values
- preserves current override target keys and fallback behaviour
### `getStageCaseTypeKey(...)`
- current stage case-type resolution decision
- combines:
- appeal type mapping
- specialist-process stage override mapping
- alias normalization
- current fallback behaviour
### `getStagesForCaseTypeKey(...)`
- current stage catalogue lookup only
- preserves:
- direct key lookup
- string indirection resolution
- fallback to `[]`
- does not own stage catalogue contents
## Stage catalogue ownership and lookup
Stage catalogue arrays still live in:
- `components/case/summary/utils/caseStagesByAppealType.js`
The domain boundary currently owns **lookup behaviour only**:
- direct key lookup
- string indirection resolution
- fallback to `[]`
`getStagesForCaseTypeKey(...)` must not be treated as stage catalogue ownership yet.
Lazy access in `getStagesForCaseTypeKey(...)` is intentional:
- it avoids circular dependency issues while stage arrays remain in the legacy helper
- the domain helper must not eagerly import catalogue contents while ownership remains outside the domain boundary
Any future catalogue migration must be a separate, explicitly scoped slice.
Until catalogue ownership moves:
- stage array contents remain the source of truth in the legacy helper
- the domain lookup helper must preserve reference identity
- the domain lookup helper must preserve current fallback behaviour
## Explicit non-goals
This boundary does **not** currently own:
- stage catalogue arrays
- progress/status calculation
- closed-case handling
- representation eligibility
- dashboard layout or calculations
- translations / EN-CY labels
- CRM/OData query construction
- API routes
- Redux state shape
- upload, submit, finalisation, or notification logic
## Behaviour-preservation invariants
Future slices must preserve:
- current stage IDs
- translation keys
- alias behaviour
- unknown input behaviour
- current fallback semantics
- current specialist-process override semantics
- status-vs-stage ambiguity
- current closed-case behaviour
## Known oddities to preserve
1. `statuscode` is currently passed into stage/progress logic.
- Do not rename or reinterpret this without a separate behaviour-change decision.
2. Specialist-process overrides are intentionally narrow.
- They currently apply only through the existing appeal-type-based resolution path.
- Do not broaden this behaviour during refactor.
3. Some specialist process values appear placeholder-like.
- Do not “correct” them without CRM/product confirmation.
4. Some ROW inputs currently produce non-intuitive stage output.
- Preserve characterized behaviour unless explicitly changing business rules.
5. The test harnesses are VM-based and import-sensitive.
- Future extractions must update harness injections carefully and avoid broad rewrites.
## Testing expectations
Future slices should run:
```bash
node tests/phase22/case-lifecycle-stage-case-type-resolution.test.cjs
node tests/phase22/case-lifecycle-specialist-process-stage-mapping.test.cjs
node tests/phase22/case-lifecycle-appeal-type-mapping.test.cjs
node tests/phase22/case-lifecycle-specialist-process-normalization.test.cjs
node tests/phase22/case-lifecycle-stage-wrapper.test.cjs
node tests/phase22/representation-build-reps-arr-rules.test.cjs
npm run lint
```
## Future extraction order
Recommended future order:
1. stage catalogue lookup helper
2. stage catalogue ownership only after enough characterization coverage
3. progress/status calculation extraction
4. additional read-only consumers
5. representation eligibility only as a separate policy slice after characterization