docs: expand README with features, deploy URLs, and FMCSA test matrix

Author: FelixBC

README.md

@@ -3,40 +3,57 @@
 Trip planner for property-carrying truck drivers that auto-generates
 FMCSA-compliant ELD daily logs and route maps from a single trip input.
 
-**Live Demo:** https://spotter-eld-gamma.vercel.app
+**Live Demo:** https://spotter-eld-gamma.vercel.app  
+**Backend:** https://spotter-eld-production.up.railway.app
 
 ## What It Does
 
 Enter four fields — current location, pickup, dropoff, and cycle hours
 used — and get back a fully planned, legally-compliant trip schedule with:
 
-- Interactive route map with mandatory stop markers
-- Day-by-day timeline view with color-coded duty status
+- Interactive route map with click-to-set location pins
+- Day-by-day timeline with color-coded duty status bars
+- DOT 11h check per day showing driving since last 10h reset
+- Cross-day continuous rest pills (10h reset ✓ and 34h restart ✓)
 - FMCSA Driver's Daily Log sheets (faithful paper form replica)
-- Automatic HOS rule enforcement (11h limit, 14h window, 30-min break,
-  70-hour cycle, 34-hour restart)
+- Hours displayed in hours and minutes (not decimal) for DOT readability
+- Automatic HOS rule enforcement:
+  - 11-hour driving limit per duty window
+  - 14-hour on-duty window
+  - 30-minute mandatory break after 8 cumulative driving hours
+  - 10-hour consecutive rest requirement
+  - 70-hour/8-day cycle limit
+  - 34-hour cycle restart
 
 ## Stack
 
 - **Backend:** Django 5 + Django REST Framework → Railway
 - **Frontend:** React 18 + Vite + TypeScript + Tailwind CSS → Vercel
 - **Maps:** Leaflet + OpenRouteService
-- **Testing:** pytest (26 tests, all passing)
+- **Testing:** pytest (37 tests, all passing)
 
 ## Architecture
 
 The HOS simulator (`backend/simulator/`) is pure Python with zero Django
 dependency. It takes a `TripInput` and returns a `TripPlanResult` —
 Django is a thin REST wrapper around it.
 
+The frontend compliance display (`TimelineView`) walks all events as one
+continuous timestamp stream using real ISO timestamps, never resetting
+the driving accumulator at calendar midnight. Rest runs are detected by
+exact timestamp contiguity so cross-midnight rest periods are correctly
+merged before the 10h threshold check.
+
 See `docs/` for full specifications:
+
 - `BRIEF.md` — product vision
 - `HOS_RULES.md` — FMCSA rules enforced
 - `ALGORITHM.md` — simulation decisions and edge cases
 
 ## Running Locally
 
 **Backend:**
+
 ```bash
 cd backend
 pip install -r requirements.txt
@@ -46,6 +63,7 @@ python manage.py runserver
 ```
 
 **Frontend:**
+
 ```bash
 cd frontend
 npm install
@@ -54,19 +72,35 @@ npm run dev
 ```
 
 **Tests:**
+
 ```bash
 cd backend
 pytest simulator/tests/ -v
 ```
 
+## Test Scenarios
+
+The test suite covers eight explicit FMCSA compliance scenarios:
+
+| # | Route | Cycle h | Rule exercised |
+|---|-------|---------|----------------|
+| 1 | Chicago → St. Louis → Memphis | 0 | 11h driving limit, single window |
+| 2 | Chicago → Indianapolis → Nashville | 0 | 30-min break after 8h driving |
+| 3 | Dallas → Oklahoma City → Kansas City | 60 | Cross-midnight 10h rest detection |
+| 4 | Los Angeles → Las Vegas → Salt Lake City | 0 | 14h window before 11h driving |
+| 5 | Chicago → Denver → Los Angeles | 65 | 70h cycle exhaustion masks 11h limit |
+| 6 | Chicago → Dallas → Los Angeles | 68 | 34h restart scheduled and credited |
+| 7 | Atlanta → Birmingham → Nashville | 0 | Two legal windows on same calendar day |
+| 8 | New York → Chicago → Los Angeles | 0 | Full stress test, all rules fire |
+
 ## Known Limitations
 
 - Sleeper berth 7+3 split not implemented (full 10h reset only)
 - Cycle hours use flat balance (no rolling 8-day window falloff)
 - Map stop positions use linear interpolation (not real highway routing)
 - Home terminal timezone hardcoded to America/Chicago
 
-## Assessment Notes
+## About
 
-Built for Spotter AI Full Stack Developer assessment.
+Built for Spotter AI Full Stack Developer assessment.  
 Time spent: ~16 hours over 2 days.