Commit Graph
13 Commits
Author SHA1 Message Date
Logan CusanoandClaude Opus 5 58efdbd6eb Correct the transcript before anything reads it
Build & Deploy / Build & push images (push) Successful in 4m0s
Build & Deploy / Deploy to VM (push) Successful in 2m28s
Build & Deploy / Report a failed deploy (push) Skipped
Correction existed, but as a line in intelligence.py's EXTRACTION_PROMPT --
which put it in the wrong place twice over. The same model call that extracted
units, location and severity emitted the correction afterwards, so extraction
reasoned over text already known to be wrong; and it sat behind
correlation_enabled, so during a cost-controlled STT-only window nothing was
ever corrected at all. That is the normal state during development.

internal/transcript_correction.py is now its own pass, between the degenerate
filter and the Firestore write. It receives an already-produced transcript plus
a reference list, so unlike a Whisper prompt it has no series to extend -- the
distinction that keeps vocabulary out of the recogniser's prompt, where an
enumerated ten-code list once made it hallucinate ten-code runs.

Reference data is merged from the talkgroup and the system, TALKGROUP FIRST. A
system spanning several counties can have a talkgroup covering one
municipality, and that municipality's streets must not be buried under a
county-wide list. A single-municipality system is the degenerate case: populate
the system level and every talkgroup inherits it. Area context is now SET --
municipality, county, roads, landmarks, on both scopes -- rather than guessed
from talkgroup names, which is what vocabulary_learner did and which is close
to useless across multiple counties.

Segments are corrected too, not just the joined text. extract_scenes builds its
prompt from numbered segments whenever there is more than one, so a correction
that only fixed the transcript would have been discarded on exactly the
multi-transmission calls carrying the most content. Alignment is enforced: an
array of the wrong length or type is dropped whole, because scenes map back to
transmissions by index and a shifted array would misattribute audio silently.

Whisper is also retried once on degenerate output. Call e49ea32c produced a
56-word ten-code counting run on one attempt and ordinary speech on the next --
same clip, same temperature=0 -- so a hallucination is a coin-flip, and
discarding on the first bad roll threw away a recoverable transcript.

Two things found on the way:

PUT /systems/{id} wiped ten_codes on every save. The systems form sends only
{name, type, config}, and model_dump() wrote every omitted field as its default
over the top. Now exclude_unset. area_context would have been the next victim,
which is why it gets its own route alongside ten-codes rather than a field on
that payload.

Closes server-26#36.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-23 14:23:41 -04:00
Logan CusanoandClaude Opus 5 be79499635 Give the nav's dead links somewhere to land
Build & Deploy / Build & push images (push) Successful in 4m3s
Build & Deploy / Deploy to VM (push) Successful in 1m54s
Build & Deploy / Report a failed deploy (push) Skipped
Three of the app's routes were referenced but never existed, so the redesign's
navigation pointed at 404s from several directions.

/dashboard was the post-login and fallback redirect target in nine places --
login, onboarding, middleware, the admin/nodes/systems/tokens/settings guards,
and the marketing header -- but app/dashboard/ was never created. Signing in
normally dropped the user on a 404. The real signed-in home is "/", which
app/page.tsx already renders as LiveView for an authed user with an org, and
which the nav labels "Live"; all nine now point there.

Nav also linked /watch and /network, neither of which existed. /watch is the
alerts screen under its redesign name, so it re-exports app/alerts/page.tsx
and /alerts stays reachable for old links. /network is new: the "my equipment"
hub the redesign moved /nodes, /systems and /tokens behind and then never
built, which had left /systems and /tokens with no entry point in the UI at
all. Its hooks all run before the admin/operator guard, per d041c86.

Separately, the admin page's guard read isAdmin without authLoading, so every
cold load of /admin -- typed URL, hard refresh, bookmark -- redirected away
while the Firebase claims were still resolving. Admin was only reachable by
clicking through from an already-mounted page. Now it waits, like every other
guarded route does.

And /incidents no longer lies about an empty list: a failed Firestore query
leaves `incidents` empty just as a quiet night does, and the page was printing
"No incidents recorded yet" over the top of a missing-composite-index error.
useIncidents already returned `error`; the page just ignored it. It now renders
an ErrorBanner instead, so the undeployed indexes in server-26#13 read as a
failure rather than as silence on the radio.

Closes server-26#30, server-26#31. server-26#13 stays open -- the rules and
indexes still have to be pushed to the live project by hand.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-23 02:28:02 -04:00
Logan CusanoandClaude Opus 5 d041c8648d Run every hook before the admin guard on /nodes and /systems
Build & Deploy / Build & push images (push) Successful in 4m6s
Build & Deploy / Deploy to VM (push) Successful in 2m25s
Both pages crashed to a blank "client-side exception" screen in production.
React error #310: the useState calls sat *below* `if (authLoading || (!isAdmin
&& !isOperator)) return null`, so the first render returned before reaching
them and the next render, once auth resolved, ran more hooks than the previous
one. React tracks hooks by call order and refuses.

The guard itself is fine and stays where it is -- only the hook declarations
move above it. Behaviour is unchanged for a user who passes the guard, and a
user who fails it still renders nothing before the effect redirects them.

Found by walking the deployed site: /nodes and /systems were the only two
routes that failed outright rather than merely showing empty data. The empty
data everywhere else is the org_id backfill, which is a separate problem.

npx tsc --noEmit clean.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-18 23:25:24 -04:00
Logan 1f17b6c0d2 feat: add role-based user management, audit log, and session tracking
Introduces a full user management system with three roles (admin, operator,
viewer), an audit log, and per-session login history.

Backend:
- app/internal/audit.py: write_audit() helper → audit_log Firestore collection
- app/internal/auth.py: get_role() helper; require_admin_token accepts both
  legacy admin:true claim and new role:"admin" claim for backward compat
- app/routers/users.py: CRUD under /admin/users — list, create (returns
  one-time invite link), get (with sessions), patch role/nodes/name,
  disable, enable, delete; operator role requires ≥1 owned node
- app/routers/links.py: POST /auth/session records sign-in events to
  user_sessions Firestore collection
- app/routers/admin.py: GET /admin/audit paginated endpoint
- app/main.py: register users router

Frontend:
- AuthProvider: exposes role, isAdmin, isOperator, ownedNodeIds from claims
- Nav: role-gated links — viewers get dashboard/calls/incidents/map/alerts/
  trips; operators add nodes/systems/tokens; admins add admin
- admin/page.tsx: new Users tab (list table, create modal, inline edit panel
  with role/nodes editor, disable/enable/delete, login history) and Audit
  Log tab (paginated, color-coded actions)
- login/page.tsx: calls recordSession() on email and Google sign-in
- nodes, systems, tokens pages: role guards redirect viewers to dashboard
- profile/page.tsx: shows accurate role badge and label
- lib/types.ts: UserRole, UserRecord, UserSession, AuditEntry types
- lib/c2api.ts: user management methods + recordSession

Firestore collections added: user_profiles, audit_log, user_sessions
Firebase custom claims schema: { role, owned_node_ids, admin (legacy) }
2026-06-22 00:02:09 -04:00
Logan e55412d8c7 UI Updates
app/map/page.tsx

Removed IncidentCard component and the incidents grid below the map — the on-map sidebar inside MapView is the single display
Moved kiosk exit button from top-3 left-3 (overlapping zoom controls) to bottom-[5.5rem] left-3
components/MapView.tsx

Fixed popup "View incident →" link — adds stopPropagation() + window.location.href to prevent Leaflet intercepting the click
Added "View details →" link on each sidebar incident card so you can navigate from the map panel without opening a popup
Added "News Alerts" overlay layer (placeholder, ready for RSS/feed integration)
lib/types.ts

Added preferred_token_id?: string | null to SystemRecord
lib/c2api.ts

Added setPreferredToken(tokenId, systemId) calling PUT /tokens/{tokenId}/prefer/{systemId} (backend already existed)
app/systems/page.tsx

Added PreferredTokenPanel component — loads the token pool lazily on expand, shows radio buttons to set/clear the preferred token, displayed on each system card above the AI flags panel
2026-06-03 01:08:21 -04:00
Logan f65873d690 Fix TypeScript key prop error on SourceCallPlayer map
Wrap SourceCallPlayer in Fragment to avoid the broken JSX env treating
key as a component prop on the custom component.
2026-06-01 01:56:51 -04:00
Logan 913fe0cbee Add source call audio playback to vocabulary suggestions
When the induction loop proposes a new vocabulary term, it now records
which sampled call(s) most likely produced the suggestion. Admins see
a collapsible "▶ source" player under each pending term showing the
audio clip and transcript, so they can hear what was actually said
before approving or dismissing.

- vocabulary_learner: track sampled call docs, attach source_call_ids
  to each pending term via word-overlap search with fallback
- types: VocabularyPendingTerm.source_call_ids?: string[]
- c2api: add getCall(id) using existing GET /calls/{call_id} endpoint
- VocabularyPanel: SourceCallPlayer component — lazy-loads call on
  first expand, shows audio controls + transcript snippet
2026-06-01 01:45:03 -04:00
Logan fa5c53891c Add PD/Town name for TG import 2026-05-25 16:42:09 -04:00
Logan 4fc44dcc86 feat: map overhaul, kiosk mode, RR importer, duplicate system
Map (MapView.tsx):
- Fan/hand-of-cards marker clustering: groups nearby markers by pixel
  proximity (union-find), renders as rotated color cards showing all types
- Pulsing ring CSS animation on recording nodes (pulse-ring keyframe)
- Live incident overlay panel — right sidebar (desktop) / bottom drawer (mobile),
  clickable to flyTo incident location
- Auto-fit button (⤢) fits all markers in view with fitBounds
- "Live · Xs ago" timestamp badge (refreshes every 10s)
- Weather Radar layer (NEXRAD via Iowa Env Mesonet, no API key)
- ADS-B + Meshtastic placeholder layers (off by default)

Map page (map/page.tsx):
- Fullscreen / kiosk toggle: fixed z-50 overlay covers nav, map fills viewport
- lastUpdated tracking passed to MapView for Live timestamp

Systems page (systems/page.tsx):
- Duplicate System button: opens form pre-filled with Copy of <name>
- RadioReference HTML import: file upload → DOMParser validates .rrlblue
  structure, parses talkgroup categories, modal lets user select which
  categories to import, auto-maps RR tags to local tags (law→police, etc.)
2026-05-23 23:52:49 -04:00
Logan 4c3b1fcc84 UI Updates 2026-05-10 21:47:34 -04:00
Logan 640667c9f9 Implement per-system AI flags 2026-04-27 00:50:01 -04:00
Logan 338b946ba3 Start to learn vocab from talkgroups to improve accuracy of STT 2026-04-21 22:17:30 -04:00
Logan 2f0597c81b Initial commit — DRB server stack
Includes c2-core (FastAPI/MQTT/Firestore), discord-bot (slash commands),
frontend (Next.js admin UI), and mosquitto config.
2026-04-05 19:01:39 -04:00