satellite-provider/_docs/04_refactoring/03-code-quality-refactoring/baseline_metrics.md

Phase 0 — Baseline Metrics (03-code-quality-refactoring)

Date: 2026-05-10 Mode: Automatic (user-discovered scope) Goal (user request): Assess current code, identify bad practices, produce steps to eliminate them.

Scope

• Full codebase, all 8 production/test projects.
• No specific feature change: this is a quality-focused refactor cycle, after the recent coupling refactor (02-coupling-refactoring) closed all five Architecture-compliance findings.
• Behavior preservation required: smoke + unit suite must remain green.

Refactoring Goals (Phase 0a)

These are the exit criteria that Phase 6 (Verification) will check against. They are deliberately framed in terms of categories of bad practice rather than fixed file-level changes — the precise list is produced by Phase 1.

1. Code-smell inventory: every C# source file in the production projects (Api, Common, DataAccess, Services.*) is reviewed for: error handling gaps, scope discipline (long methods / SRP violations), magic numbers, hidden side effects, async-correctness (async void, missing ConfigureAwait / CancellationToken), thread-safety in hosted services, and duplication across the three Services.* siblings.
2. Inventory delivered as list-of-changes.md with severity (Low / Medium / High), risk, and dependencies — same template used for 01-testability-refactoring and 02-coupling-refactoring, so the change list is directly consumable by the implement skill.
3. Roadmap (Phase 2) ranks the entries by ROI (impact ÷ risk×effort) and groups them into batches that fit the 2–5-point complexity policy from the user's tracker rules.
4. No regression in unit (37) or integration smoke (5 scenarios) suites at Phase 6 verification.
5. Open question for Phase 0 BLOCKING gate: does the user want the run to execute (Phases 3–7) or stop after roadmap (Quick Assessment, Phases 0–2 only)?

Constraints

• No public HTTP route changes. /api/satellite/tiles/latlon, /api/satellite/tiles/{z}/{x}/{y}, /api/satellite/request, /api/satellite/route shapes are stable.
• No database schema changes (would require migration sequencing and is out of scope).
• No DI-graph re-ordering that would change observable startup behavior of the hosted services (RegionProcessingService, RouteProcessingService).
• Architecture Vision (_docs/02_document/architecture.md H2) is authoritative. Any proposed change that conflicts with it must be flagged, not silently accepted.

Current Code Topology

Project	LOC (.cs, excl. bin/obj)	Files	Notes
SatelliteProvider.Api	380	1 (Program.cs)	Minimal-API endpoints + DI wiring
SatelliteProvider.Common	474	25	DTOs, interfaces, configs, GeoUtils
SatelliteProvider.DataAccess	614	11	Repositories, Entities, Migrations
SatelliteProvider.Services.TileDownloader	620	3	TileService, GoogleMapsDownloaderV2, DI ext.
SatelliteProvider.Services.RegionProcessing	564	4	RegionService, hosted service, queue, DI ext.
SatelliteProvider.Services.RouteManagement	1055	3	RouteService, RouteProcessingService, DI ext. ← largest unit
SatelliteProvider.Tests	1221	6	xUnit unit tests
SatelliteProvider.IntegrationTests	1434	10	Console runner, integration scenarios

Total production C# LOC: 3 707 across 47 files (Api + Common + DataAccess + 3 × Services). Total test C# LOC: 2 655 across 16 files.

Test Coverage Baseline (carried forward + verified)

• Unit suite (SatelliteProvider.Tests): 37 [Fact]/[Theory] annotations (FINAL_report of 02-coupling claimed 40 — small discrepancy; will reconcile in Phase 6 against dotnet test --list-tests).
• Integration smoke (scripts/run-tests.sh --smoke): 5 scenarios, last green run 2026-05-10, wall-clock 111.86 s (_docs/03_implementation/test_run_step7.md).
• Coverage % not measured (no Coverlet wired up; would add as a Phase 0 sub-task only if the user wants it gating verification).

Architecture Compliance Baseline (carried forward)

_docs/02_document/architecture_compliance_baseline.md — all 5 findings Resolved by epic AZ-309. F1, F2 (High) and F3, F4 (Medium) and F5 (Low). No new architecture findings have been opened since.

Complexity / Static-Analysis Tools

Not present in this repository:

• No Coverlet / dotCover / fine-grained coverage runner.
• No dotnet format --verify-no-changes gate.
• No SonarLint / Roslyn-analyzer ruleset beyond defaults.
• No cyclomatic-complexity tool wired up.

These gaps are themselves candidate findings for Phase 1 (tooling category).

Functionality Inventory

Endpoint / Hosted Service	Owner Project	Status	Smoke Coverage
GET /api/satellite/tiles/{z}/{x}/{y}	Api → Services.TileDownloader	live	✓ (TileTests.RunGetTileByLatLonTest)
GET /api/satellite/tiles/latlon	Api → Services.TileDownloader	live	✓
GET /api/satellite/tiles/mgrs	Api	stub (returns 501)	not exercised
POST /api/satellite/request	Api → Services.RegionProcessing	live	✓ (RegionTests.RunRegionProcessingTest_200m_Zoom18)
GET /api/satellite/region/{id}	Api → Services.RegionProcessing	live	✓
POST /api/satellite/route	Api → Services.RouteManagement	live	✓ (BasicRouteTests.RunSimpleRouteTest, ExtendedRouteTests.RunRouteWithTilesZipTest)
GET /api/satellite/route/{id}	Api → Services.RouteManagement	live	✓
POST /api/satellite/upload	Api	stub (returns 501)	not exercised
RegionProcessingService (BackgroundService)	Services.RegionProcessing	live	exercised via region scenarios
RouteProcessingService (BackgroundService)	Services.RouteManagement	live	exercised via route scenarios

Self-Verification (Phase 0)

• RUN_DIR created with auto-incremented prefix 03-.
• Goals captured (5 numbered goals above).
• Constraints captured (no HTTP-shape change, no schema change, no DI reorder, Architecture Vision honored).
• Current topology measured by LOC and file count.
• Test coverage baseline noted; coverage % flagged as not measured.
• Architecture compliance baseline noted (carried from 02-coupling FINAL_report).
• Functionality inventory complete (10 endpoints + 2 hosted services).
• Tool gaps explicitly listed so Phase 1 can decide whether to fold them in.

Open Question for User (BLOCKING gate)

See the Phase 0 summary presented in chat — two decisions:

1. Confirm goals and constraints above.
2. Choose: full refactor run (Phases 0–7) vs. Quick Assessment (Phases 0–2 only — produces list-of-changes.md + refactoring_roadmap.md and stops).