The planner is the BRouter client in this architecture — route
compute, rate limiting, and the eventual move of BRouter off-box all
live there. The journal had two places that called BRouter directly
(demo-bot and /api/v1/routes/compute) and both were broken on prod
because the journal service has no BROUTER_URL.
Fix:
- Extend the planner's /api/route with an optional `format: "gpx"`
that returns BRouter's raw GPX (for server-to-server callers that
don't need the way-tag enriched GeoJSON).
- Point the journal demo-bot at `PLANNER_URL/api/route` with
`format: "gpx"` instead of calling BRouter directly.
- Point /api/v1/routes/compute at `PLANNER_URL/api/route` instead of
BRouter directly.
Journal no longer needs BROUTER_URL at all. When BRouter moves to a
dedicated host later, only the planner changes.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The same-origin check compared the browser's Origin header against
new URL(request.url), but inside the Planner container request.url
is http://planner:3001/... while the browser sends Origin
https://planner.trails.cool — so production always 403'd.
Trust Caddy's X-Forwarded-Host and X-Forwarded-Proto headers when
present (both are set by Caddy's reverse_proxy directive by default)
to reconstruct the external origin the browser actually connected to.
Falls back to request.url for dev and any direct (non-proxied) access.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Near-identical viewports from different collaborators (slightly
different pan/zoom because Yjs awareness doesn't enforce pixel-exact
alignment) previously produced byte-different Overpass queries and
therefore different server-side cache keys. Each client missed cache
and triggered its own upstream fetch.
buildQuery now quantizes the bbox to a 0.01° grid (~1 km), expanding
outward (south/west floor, north/east ceil) so the viewport is always
covered, and formats coords with 3 decimals for a stable string.
Trade-offs:
- Slightly larger query bbox: ≤ one grid cell (≈ 1 km) of padding on
each side. For a zoom-12 viewport (~11 km wide) that's <10% extra
data; bounded by the existing `[maxsize:1048576]` + `out 100` limits.
- Coordinates in the query are now 3-decimal strings regardless of the
caller's precision — existing buildQuery test updated to match.
Two new tests cover the cache-alignment property and the outward-
expansion invariant. Hit ratio impact will be visible on the Grafana
"Overpass Cache Hit Ratio" stat added in the prior PR.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds four metrics surfaced via the existing /metrics endpoint:
- overpass_cache_events_total{result=hit|miss|coalesced}
- overpass_cache_size (gauge)
- overpass_upstream_duration_seconds (histogram)
- overpass_upstream_requests_total{status} — covers 2xx/4xx/5xx and
an explicit "error" label for network-level failures
Grafana dashboards/planner.json gets a new row:
- Upstream health (5m 2xx success ratio, red <90%, green >99%)
- Cache hit ratio
- Cache size
- Upstream p95 latency
Plus timeseries panels for cache event breakdown, upstream status
over time, and p50/p95/p99 upstream latency.
The success-rate formula uses clamp_min(..., 1) on the denominator so
it doesn't NaN when traffic is zero.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds apps/planner/app/routes/api.overpass.ts — a same-origin, rate-limited
server-side proxy that forwards Overpass QL to the upstream endpoint
configured via OVERPASS_URL (default: overpass.private.coffee).
Motivation:
- private.coffee's best-practices require a meaningful User-Agent
identifying the project. Browsers cannot set User-Agent on fetch()
(forbidden header), so the request has to originate server-side.
- Collaborative sessions commonly have N clients pan/zoom the same map,
so the same bbox query arrives multiple times within seconds.
- Setting up the proxy now lets us swap OVERPASS_URL to a self-hosted
Overpass later without client changes.
What the proxy does:
- Sets User-Agent "trails.cool Planner (https://trails.cool; legal@trails.cool)"
- Enforces same-origin via the Origin header
- Rate-limits per client IP (120 req/min)
- In-memory LRU cache of upstream responses keyed on the form-encoded body
(TTL 10 min, max 200 entries)
- Coalesces concurrent misses for the same key so N simultaneous clients
in one session incur exactly one upstream call
Client change: apps/planner/app/lib/overpass.ts POSTs to /api/overpass
instead of iterating over public Overpass endpoints. Fallback list
removed; resilience is now the proxy's responsibility.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Switch the Planner's primary Overpass endpoint from
overpass.kumi.systems to overpass.private.coffee. private.coffee is a
privacy-focused public Overpass instance that doesn't log queries,
which fits the Planner's privacy-first principle better than the
previous default. Keep overpass-api.de as the silent fallback for
resilience while the self-host-overpass change is in flight.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- sentry.client.ts: add stopSentryClient() that awaits Sentry.close() so
the SDK tears down when a user logs out. Hub methods become no-ops
until initSentryClient() is called again on next login.
- root.tsx: call stopSentryClient() from the user-effect when user === null.
- Add sentry-config to journal + planner Dockerfiles (was missing after
#237 introduced the new workspace package, breaking the Dockerfile
package check).
- Footer: replace inner <nav> with <div> — nested nav landmarks broke
the journal "nav bar shows on all pages" e2e test (two navigation
roles on the page).
- e2e auth: tick the required ToS checkbox in the shared registerUser
helper so passkey registration tests pass.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Introduces @trails-cool/sentry-config with three helpers that encode
the common Sentry.init options:
- nodeSentryConfig(appContext) for the two HTTP servers
- browserSentryConfig(appContext, env) for the Journal client
- mobileSentryConfig(__DEV__) for the React Native app
- drop404s beforeSend for servers
Also fixes three inconsistencies found in the audit:
- Planner server was missing sendDefaultPii: false (IPs could leak)
- Planner entry.server.tsx didn't call Sentry.captureException on SSR
render errors; Journal's SSR did
- Sentry.init location now matches across apps (both in server.ts;
Journal previously had it in app/entry.server.tsx, which meant init
ran lazily on first request rather than at server startup)
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
When Sentry init runs but is disabled (local dev / CI), log a debug
message so developers can see Sentry *would* send in production.
Also explicitly set replaysSessionSampleRate/replaysOnErrorSampleRate
to 0 in the client and mobile configs. Replay integration isn't
installed, so this is documentation — it prevents future accidental
enablement if an integration is ever added.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- Journal: Impressum (§5 TMG), Terms of Service, GDPR privacy page
- Registration flow: required ToS checkbox, termsAcceptedAt persisted
- Alpha banner on Journal (always visible); alpha badge on Planner hero
- Footer with legal links + GitHub on both apps
- Reduce PII: sendDefaultPii=false, Sentry init only after login (Journal)
- Drop Sentry from Planner (no login, no consent possible)
- Drop localStorage caching from i18n client detection
- Sync SSR i18n resources each request so locale edits HMR without restart
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds a script that verifies every packages/*/package.json is listed in
all app Dockerfiles. Catches the recurring issue where adding a new
workspace package breaks production because the Dockerfile deps stage
doesn't know about it.
Also fixes two pre-existing missing packages in the planner Dockerfile
(api, map-core) caught by the new check.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Both Dockerfiles were missing packages/jobs/package.json in the deps stage,
causing pnpm install to skip pg-boss. Also copy app/jobs directories into
the runtime stage so job handlers are available at runtime.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add @trails-cool/jobs package wrapping pg-boss for durable background job
execution using the existing PostgreSQL database. Wire up planner session
expiry as the first scheduled job (hourly, 7-day TTL) — this was previously
dead code that never ran. Add placeholder worker to journal for future
Komoot import and federation jobs.
Infrastructure: grant grafana_reader access to pgboss schema, add job queue
health panels to Service Health dashboard, add failed job alert.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Fix blind spot where Caddy 502 errors were invisible: the dashboards and
alerts queried caddy_http_response_duration_seconds (upstream responses only),
missing 502s that Caddy generates itself. Switch to
caddy_http_request_duration_seconds (server-level, all responses).
Add Journal Grafana dashboard with: 502 rate, response codes, request rate
by route, latency percentiles, container restarts/memory/CPU, Node.js event
loop lag and heap, and Loki log panels for errors and Caddy 5xx entries.
Add color coding to Caddy status code panel (green=2xx, blue=3xx, yellow=4xx,
red=5xx). Add log-based error rate panel to the overview dashboard.
New alerts: container restart loop, PostgreSQL connections > 80, application
crash log detection (Loki), and Caddy 502 rate.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Previously only 2-3 workspaces participated in each turbo task. This
expands test, lint, and typecheck to cover all 11 workspaces with
parallel execution and caching.
Test pipeline:
- Move from single root vitest to per-workspace test scripts via turbo
- Shared vitest config (vitest.shared.ts) with passWithNoTests
- Per-workspace configs re-export the shared base
- Mobile uses Jest (jest-expo + React Native Testing Library)
- Add node-environment GPX tests verifying linkedom fallback
- Add i18n mobile init tests
- Exclude apps/mobile from root vitest (uses Jest separately)
Lint pipeline:
- Add eslint lint script to all 9 previously unlinted workspaces
- Standardize all scripts to "eslint ." with shared root config
- Add .expo/ to global ESLint ignores
- Fix lint errors: unused imports in api/types, export parseGpx
Typecheck pipeline:
- Add "typecheck": "tsc" to all 8 packages
- Add @types/node (catalog) to gpx and db packages
- Fix mobile app.config.ts: remove deprecated experiments.monorepo
and newArchEnabled (both default in Expo SDK 55)
- Add allowImportingTsExtensions to mobile tsconfig
Shared package compatibility (mobile-app Phase 1.3):
- Add linkedom as explicit dependency to @trails-cool/gpx
- Add initI18nMobile() export to @trails-cool/i18n
- Confirm @trails-cool/types is pure interfaces (no DOM deps)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
New renderer-agnostic package with zero dependencies:
- Tile configs (base layers, overlay layers)
- Color palettes (surface, highway, smoothness, tracktype, cycleway,
bikeroute, elevation, maxspeed) — 8 color maps + 3 color functions
- POI category definitions (9 categories with Overpass queries, icons,
colors, profile mappings)
- Z-index layering constants
- Snap distance constant
Updated 15 consuming files to import from @trails-cool/map-core.
Deleted poi-categories.ts and z-index.ts from the Planner (fully moved).
packages/map re-exports tile configs from map-core for backwards compat.
All 117 tests pass, no user-facing changes.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Participant list: hidden on mobile (sm:block)
- Undo/redo buttons: hidden on mobile (sm:flex)
- Connection status: hidden on mobile (sm:inline)
- Header: no-wrap, tighter padding on mobile
- Username input: text-base on mobile to prevent iOS auto-zoom
Keeps profile selector and export button visible on all sizes.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
maximum-scale=1 + user-scalable=no prevents the browser UI from
zooming when pinching. Map zoom (Leaflet) still works since it
handles touch events independently.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- viewport-fit=cover enables safe area insets in Safari
- Header: pt-[max(0.5rem,env(safe-area-inset-top))] clears the notch
- Elevation chart: pb-[max(0.5rem,env(safe-area-inset-bottom))] clears
the home indicator bar
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Single touch + drag: scrub the highlight point along the chart,
map dot follows in real-time
- Tap: pan map to that point (10px threshold for movement)
- Two-finger touch: range select — area between fingers highlighted,
on release zooms map to fit that route section
- touch-none CSS + preventDefault prevents page scrolling on chart
Spec updated with mobile touch interaction scenarios.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Three new interactions:
1. Route hover → chart: Invisible interactive polyline (weight 20,
opacity 0) detects mouse hover on the route, computes cumulative
distance, and highlights the corresponding position on the chart.
2. Chart click → map pan: Clicking the elevation chart pans the map
to center on that point along the route. Uses 5px threshold to
distinguish from drag.
3. Chart drag-select → map zoom: Dragging a range on the chart shows
a blue selection overlay, then zooms the map to fit the route
coordinates in that range. "Reset zoom" button appears to restore
the full route view.
State flow avoids feedback loops: external highlights (from map hover)
set the chart crosshair without re-emitting onHover back to the map.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Clicking the chart title (e.g. "Road Type Profile") opens the
corresponding OSM wiki page. Plain, elevation, and grade modes
remain as plain text since they're not OSM tag-based.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The dummyUsage trick needs a separate assign per tag:
assign dummyUsage2 = maxspeed=
(not appended to the existing smoothness= line)
- Dockerfile: Patch all BRouter profiles to include maxspeed in
WayTags output via separate assign dummyUsage2
- brouter.ts: Extract maxspeed with direction handling
(maxspeed:forward/backward + reversedirection awareness)
- ColoredRoute: maxspeed color mode (green ≤30, yellow ≤50,
orange ≤70, red ≤100, dark red 100+)
- ElevationChart: maxspeed chart rendering, legend, hover "X km/h"
- PlannerMap: maxspeeds state + Yjs read + prop passing
- i18n: EN "Speed Limit" / DE "Tempolimit"
- Mock fixtures: maxspeed data for E2E tests
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
E2E fixes:
- Color mode select locator uses option[value='highway'] filter
instead of fragile page.locator("select").last()
Removed maxspeed color mode:
- BRouter does not expose maxspeed in WayTags output despite the tag
existing in its lookup table. The dummyUsage profile trick only
affects cost calculation, not tiledesc output.
- Can revisit if BRouter adds maxspeed to WayTags in a future version
or if we fork/patch the BRouter profiles more deeply.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
E2E fixes:
- Color mode select locator now uses option[value='highway'] filter
instead of fragile page.locator("select").last()
Speed limit color mode:
- Extract maxspeed tags from BRouter tiledesc (with forward/backward
direction handling following brouter-web's pattern)
- Color route and chart by speed limit: green ≤30, yellow ≤50,
orange ≤70, red ≤100, dark red 100+ km/h
- Legend with speed thresholds, hover shows "X km/h"
- i18n: EN "Speed Limit" / DE "Tempolimit"
- Mock fixtures include maxspeed data
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Extract highway=* tags from BRouter tiledesc messages alongside surface
data and add a new "Road Type" color mode that colors the route polyline
and elevation chart by OSM highway classification (cycleway, residential,
path, etc.). Includes color palette, legend, hover labels, i18n (EN+DE),
unit tests for tag extraction, and E2E tests.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Full notes lifecycle:
- GPX: description field in GpxData, <metadata><desc> in generate/parse
- Export: Plan GPX and Save to Journal include notes as description
- Journal: updateRoute extracts description from GPX, stores on route
- Reimport: Edit in Planner passes notes via URL params → Yjs Y.Text
- Drop import: GPX with <desc> restores notes in session
Spec updated: session-notes gains GPX export, Journal sync, and
reimport requirements.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Legend shows actual surface types present in the route (up to 6)
with colored swatches instead of generic text
- Hover label shows surface type name (e.g. "34m · 12.3km · asphalt")
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The chart was falling through to plain blue when surface mode was
selected. Now reads surface data from Yjs routeData and draws
segments colored by road surface type, matching the route colors
on the map.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Legends sit between the title and the mode selector dropdown:
- Grade: colored swatches with percentage thresholds
- Elevation: gradient bar from low (green) to high (red)
- Surface: text hint about surface coloring
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Shows grade thresholds with color swatches: <3% green, <6% yellow,
<10% orange, <15% red, 15%+ dark red. Only visible in grade mode.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Moved color mode dropdown from header bar to inline with the
elevation chart title — closer to the visualization it controls
- Removed ColorModeToggle component from SessionView header
- Added "Grade" color mode: colors route and elevation chart by
steepness (green <3%, yellow <6%, orange <10%, red <15%, dark red 15%+)
- Grade hover label shows percentage (e.g. "+4.2%")
- Fixed useTranslation namespace in ElevationChart ("planner")
- i18n: Added grade keys for en + de
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Initialize CodeMirror with existing Y.Text content (doc: yjs.notes.toString())
so text is visible immediately, not just after the next remote change
- Use dedicated Y.UndoManager for notes instead of sharing with waypoints
undo manager (different tracked origins caused conflicts)
- Use yUndoManagerKeymap for proper Yjs-aware undo/redo
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
y-codemirror.next uses the awareness 'cursor' field for text editor
cursor positions. Our map cursor tracker was using the same field name,
causing Invalid LatLng errors when the receiving participant tried to
render a CodeMirror cursor object as map coordinates.
Renamed to 'mapCursor' to avoid the collision.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
When a route has overnight waypoints, the export dropdown shows a
third option: "Export Day Segments" which downloads one GPX file per
day. Each file is named day-N-start-name.gpx with the day's track
segment. Only visible when the route has multiple days.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Right-clicking a waypoint now shows a context menu with:
- "Mark as overnight stop" / "Remove overnight stop" (middle waypoints only)
- "Delete" (all waypoints)
Previously first/last waypoints were deleted on right-click and
middle waypoints toggled overnight, with no visual indication of
which action would happen.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
When clicking near the existing route to add a waypoint (including POI
snap), find the closest route segment and insert after it. Falls back
to appending at the end if the click is far from the route (~200m
threshold in degrees).
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- leaflet.markercluster: Dynamic import with fallback to plain layer
group. Clusters POI markers at low zoom, ungroups at zoom 15+.
- Base layer sync: baselayerchange event writes to Yjs, new
participants load the selected base layer on connect.
- Both tile overlays and base layer now persist across participants
and crash recovery.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
OverlaySync component listens to Leaflet overlayadd/overlayremove
events and writes enabled overlay IDs to Yjs routeData. On connect,
loads initial overlay state from Yjs. LayersControl.Overlay uses
checked prop to reflect the synced state.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
useYjsPoiSync: Bidirectional sync between local POI state and Yjs
routeData. When a participant enables/disables POI categories, the
change propagates to all connected clients. Initial state loaded from
Yjs on connect. Crash recovery included via existing Yjs localStorage
snapshot.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
POI markers should render below waypoint markers so waypoints are
always easy to click and drag.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
useProfileDefaults hook observes profile changes in Yjs and merges
relevant POI categories into the enabled set. Skips initial load to
respect existing state. Cycling profiles → bike infra, hiking →
shelter + viewpoints.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
