Skip to main content

Progress tracking

Tracking progress across all stages.

Timeline overview​

  • 2025‑09‑15 → 2025‑09‑18: Phase 1 Foundation built. ~24h over 4 days. See costs & time for ongoing tracking.

Stage 1​

Foundation for personal coach — 1/11 phases complete.

Phase 1 — Foundation (completed)​

Core Infrastructure:

  • Strava app + OAuth via ngrok
  • Local Postgres running
  • Prisma schema (activities, splits)
  • Fetch & store activities
  • CSV export
  • Thumbnails (first pass)
  • Details page (AU 24‑hour)
  • Prompt frame: Assumptions → Request → Safeguards
  • Cost/time logged

Real-time sync:

  • Webhook: Strava → <30s arrival — automatic push from Strava to database
  • Token refresh service (pre‑expiry window, friendly fail)
  • Dedupe on re‑sync (stravaId + time window)

Data model hardening:

  • Normalised split schema (no fabrication) — shows real distances like "0.08 km"
  • Edit/delete audit trails — complete token and webhook history
  • Activity edit/delete with audit flags (user, timestamp, edited flag)

Thumbnails & performance:

  • Auto thumbnail generation with map route and km markers
  • Background thumbnail worker with retry/cleanup
  • Exclude heavy fields from list endpoints
  • Sub-50ms API responses (p95 < 50ms achieved)
  • 95% reduction in client requests (lazy loading)

UI/UX core:

  • Mobile‑first feed optimised for Safari iPhone 14 Pro Max
  • Mobile-first dark theme
  • Activity filtering (title, type, distance, duration, source, date range)
  • Filter persistence (remembered when exiting/returning to app)
  • Menu system with activities, statistics, tools, partners sections
  • Strava athlete name display at bottom

Admin system:

  • 7 admin endpoints: webhooks, activity-jobs, thumbnails, best-efforts, sync-streams, resync
  • Admin monitoring APIs for ingest → worker → DB observability

Code quality:

  • Fix lint errors for production readiness
  • Add documentation sync validation script (npm run docs:check)
  • Implement comprehensive code quality tooling (Prettier, ESLint, Husky, lint-staged)
  • Set up automated Git hooks for pre-commit quality checks
  • Enforce conventional commit message format
  • Configure modern ESLint v9 with TypeScript and Next.js rules
  • Implement Jest testing framework with React Testing Library
  • Set up GitLab CI/CD pipeline with comprehensive quality checks
  • Add coverage reporting and quality validation scripts
  • Configure automated testing in pre-commit hooks
  • Resolve GitLab CI compatibility issues (Node.js 20, native binaries, TailwindCSS v3)
  • Implement 5-stage automated pipeline (.pre, validate, test, quality, security)
  • Add comprehensive debugging and error handling for CI environment
  • Configure Jest with coverage artifacts (JUnit, Cobertura) for GitLab integration
  • Set up pre-push hooks to prevent formatting issues reaching CI
  • Implement dynamic documentation sync checking with Node.js file operations

Phase 2 — Advanced analytics and best efforts​

  • GPS-based calculation infrastructure ready
  • Automatic best efforts calculation (400m, 800m, 1km, 1mi, 5km, 10km, 21.1km, 42.2km, 50km)
  • Activity ranking system for each distance
  • Time window filters (week, month, 3 months, year, 2 years, all-time)
  • Best efforts page in menu

Phase 3 — Statistics dashboard​

  • Statistics page (landscape orientation)
  • Bar graphs with top buttons: Distance, Duration, Calories, Max Distance, Avg Speed, Altitude
  • Time frame buttons: Year, Month, Week, Day
  • Dynamic scaling and scrolling for iPhone screen
  • Remember last selections (filter + timeframe persistence)
  • Filter by activity type (running, walking, cycling for future)

Phase 4 — Activity management​

  • Activity deletion from database with audit trail
  • API endpoint to list differences between database and Strava
  • Options to add back: one, several, or all missing activities
  • Handle two types: activities not in database, activities modified in Strava
  • Activity editing (title, distance, etc.) with audit flags
  • Prevent future syncs from overwriting edited records

Phase 5 — UI refinements and components​

  • Activity count display shows "X activities (Y filtered)" when filtered
  • Clear/refresh filter functionality
  • Export to CSV from filtered results
  • Enhance GUI using shadcn/ui Calendar and Popover components
  • Create consistent button-based triggers with icons
  • Provide visual feedback for selected states
  • Handle empty/invalid states explicitly

Phase 6 — Multi-user auth and permissions​

  • Unified User model with multi-tenancy
  • Create unified User model supporting multiple roles (ADMIN, ATHLETE, COACH)
  • Database model relationships: Single source of identity (one record per person)
  • UnifiedAuthAdapter for authentication
  • Roles and permissions system implemented
  • Session management
  • Complete user and session CRUD interfaces

Phase 7 — Data import and migration​

  • Add import tools for batch client/session data (template + validation) from Training Peaks

Phase 8 — Monitoring and alerting​

  • Background health monitoring and admin views
  • Extend health monitoring with alerting and notifications

Phase 9 — Testing and quality​

  • Write unit tests for utilities
  • Add integration tests for components
  • Include visual regression tests
  • Verify cross-browser compatibility

Phase 10 — Infra and GitOps foundation​

  • k8s manifests (dev/prod)
  • FluxCD wired to GitLab (GitHub mirror)
  • Secrets strategy (SOPS/Sealed)
  • Backups + tested restore

Phase 11 — Security and reliability​

  • Auth layer (admin vs user)
  • Error taxonomy + helpful messages
  • Structured logs + basic alerts
  • Data retention knobs

Stage 2: iOS App​

Apple HealthKit integration.

  • SwiftUI client consuming backend APIs
  • Apple HealthKit permissions and sync (read/write)
  • User auth flow on mobile
  • Native charts and best-efforts display

Stage 3​

Health metrics (replace HealthFit, Athlytic, WaterMinder, My Weight).

  • Sleep, HRV, RHR, oxygen saturation ingestion
  • Recovery, exertion, hydration logging
  • Weight tracking and graphing
  • Rolling sleep debt and consistency metrics

Stage 4: Nutrition basics​

Replace MyFitnessPal for simple use.

  • Manual food entry (calories, sugar)
  • Daily and weekly summaries
  • Export to CSV/JSON

Stage 5: Apple watch app​

Replace NRC + Strava.

  • Native run tracking (distance, pace, HR, GPS)
  • Music track logging during runs
  • Guided workouts
  • Offline sync to backend

Open decisions​

  • Whether to backfill full TP history now or later
  • Image storage path (NFS vs S3‑compatible)
  • Stats/graphing lib for web (keep it light)

Risks and mitigations​

  • AI edits can break code → lock Cursor to Claude 4 Sonnet; manual apply only; backups on.
  • Strava rate limits → throttle jobs, exponential backoff, diff‑based re‑ingest.
  • K8s complexity → start minimal, GitOps everything, test restores weekly.

Costs​

Recent improvements (stage 1 MVP)​

Advanced sync filters​

  • Smart filter interface - click "Sync Activities" to access comprehensive sync options.
  • Date range filtering - sync only activities within specific time periods using date pickers.
  • Quick action buttons - "Sync New" or "Force Update All" in one click
  • Configurable thresholds - adjust early termination from 5 to 999+ activities.
  • Activity limits - control processing volume from 50 to 1000 activities per sync.
  • Force update option - reprocess existing activities with latest Strava data.
  • Accessibility improvements - proper form labels, IDs, and screen reader compatibility.
  • Visual consistency - grouped controls, consistent spacing, and rounded button styling.

Automatic activity import​

  • Fully automated sync - activities import automatically every 4 hours.
  • Set it and forget it - no more manual "sync" button clicks needed.
  • Configurable thresholds - customise auto-sync behaviour via environment variables.
  • Intelligent early termination - stops processing when no new activities found.
  • Complete activity data - imports splits, thumbnails, and performance metrics automatically.

Database-first authentication architecture​

  • Pure PostgreSQL storage - 100% database-first token management with zero file dependencies.
  • Automatic token refresh - database-managed refresh with comprehensive audit trail.
  • Legacy code eliminated - removed 300+ lines of file-based fallback code for cleaner architecture.
  • Multi-user foundation - unique athlete IDs ready for future user expansion.
  • Cross-device reliability - database persistence works across all devices and browsers.
  • Performance optimised - direct database queries, no file I/O blocking.
  • Simplified codebase - single authentication code path, easier maintenance.

Mobile-first UI optimisation​

  • Mobile-optimised design - dark theme with professional card layouts and navigation.
  • Full-width activity cards - maximises screen real estate usage.
  • Reduced vertical spacing - fits more activities on screen (3x improvement in density).
  • One-tap navigation - entire activity row is clickable for instant access to details.
  • Streamlined content - removed average HR and running icons for cleaner appearance.
  • Compact thumbnails - optimised 56x56px size with proper aspect ratio.
  • Better spacing hierarchy - consistent padding and margins throughout.

Enterprise-grade audit trail (September 2025)​

  • Complete token history - maintains full audit trail of all Strava authentication events.
  • Security compliance - tracks token creation, refresh, and revocation timestamps.
  • Usage analytics - comprehensive connection patterns and authentication lifecycle data.
  • Soft deletion - preserves historical records instead of permanent deletion for compliance.
  • Multi-record per athlete - supports multiple token records per user with isActive flag.
  • Seamless migration - removed unique constraints to enable audit trail without breaking existing functionality.
  • Database optimisation - efficient queries with [athleteId, isActive] composite indexing.
  • Enhanced security - complete visibility into authentication events for security auditing.

Major performance optimisation (September 2025)​

  • 95% reduction in initial HTTP requests - eliminated 40+ unnecessary map requests on page load.
  • Lazy loading implementation - thumbnails load only when visible using Intersection Observer API.
  • Progressive image loading - 50px preload margin with smart viewport detection.
  • Instant page loads - 411ms DOM ready time, 998ms full load (down from several seconds).
  • Efficient bandwidth usage - 3MB total transfer for 50 activities (was 10MB+ previously).
  • Smart request prioritisation - core app loads first, images load as user scrolls.
  • Memory optimisation - automatic observer cleanup prevents memory leaks.
  • Mobile performance boost - particularly benefits slow mobile connections.

Enhanced splits system​

  • Database schema extension - normalised split storage with real distance/duration fields.
  • Accurate partial splits - displays true values like "0.08 km" and "0:38" for final laps.
  • Full-page splits view - dedicated /activities/[id]/splits page with professional table layout.
  • Improved column structure - Lap, Distance (km), Duration (min:sec), Pace (min/km), HR (bpm).
  • Centered headers with units - clean typography with units displayed in subheaders.
  • Larger data rows - 18px font size for better readability on mobile devices.
  • Database migration completed - 47 activities backfilled with accurate split data.
  • API optimisation - direct database queries instead of raw JSON parsing.
  • Target infra run‑rate: <$20/mo self‑hosted.