82c88379d4d4e2b121ad31bfcad3e01d23093c4d
13
Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
  Logan CusanoandClaude Opus 5
|
82c88379d4 |
Stop an incident lying about what it is and where it is
An incident header had two independently last-write-wins halves, and in the
2026-08-20 dump both were wrong at once. `b9b4f392` opened on a suspect search
at 80 Grasslands Road; it was labelled "100 South Mosher" (its third call),
pinned at `Westmed` (its second), and titled after the label. Five of six
incidents were pinned somewhere other than the place they claimed to be.
Location and pin are now one value
----------------------------------
`_resolve_location_pair()` computes `location`, `location_coords` and the new
`location_coords_source` together, and `_update_incident`/`_create_incident`/
`_create_master_incident` always write all three. There is no longer a code
path that can move one and leave another behind — including the cross-system
master, which used to take its label from the parent and its pin from the call.
The pin now carries the label it was geocoded from. `_verified_pin()` returns
it only when that source still matches the incident's current label; anything
else is dropped. That includes every pre-existing incident, whose pin has no
recorded source and therefore cannot be reconciled — which is the right
outcome, since the dump says 5 in 6 of those are wrong. A missing pin reads as
missing data; a wrong pin reads as fact, and this is a map people may act on.
An incident also keeps the first place it was given rather than the latest.
Later mentions still accumulate in `location_mentions` (what the map path is
drawn from); they just don't rename the incident's own location. The one
permitted change is filling in a pin the incident never had, from a later call
naming the exact same label — geocoding needs the node position, a quota and a
response, so the same address genuinely does fail once and resolve later.
"49" is not a place
-------------------
`clean_location()` rejects any string with no two-letter word in it, applied at
extraction (intelligence.py, before the geocoder and before the call document)
and again at the correlator's context boundary. `9d376ffe` carried
`location: "49"` from "Fire received. Flames from 49." — a box number — and its
summary asserted "A fire incident was reported at location 49". Nothing
validated that field at all, so it would have recurred.
Title: the founding event, escalation only
------------------------------------------
The title was re-derived from the newest classified call, so `f5190670` was
named after the thirteenth of its thirteen events. It now names the call that
opened the incident, recorded in `title_tag`/`title_severity`, and can only be
replaced by a call of strictly higher severity.
Three candidates were considered:
* Newest call (status quo) — rejected. The same incident has a different name
at different times, so a user who saw it in the rail cannot find it again,
and the name is decided by radio timing rather than by the event.
* Highest severity alone — rejected as the sole rule. Severity has four
levels and most traffic sits on one of them, so ties are the common case
and the tiebreak degrades to "newest" — the defect it was meant to fix.
* Founding event, escalated by strictly-greater severity — chosen. An
incident's identity is the event that opened it, so that is its default
name and it is stable for the incident's whole life. The single case where
the header MUST change is the one where the situation got worse: a check
condition that becomes a structure fire is a structure fire, and the
worst-first rail, the "Major only" filter and the map colour all exist so
that is never missed. Requiring strictly-greater makes it monotonic, the
same contract `_max_severity` already gives the severity field: routine
chatter can never take the name back.
A summary-level title regenerated as a whole was rejected outright: it needs an
LLM call per incident, AI flags are off in production, and every incident today
would have no title at all.
Two renames survive, because neither replaces an event name: filling in the
placeholder title of an incident that opened on a call with no content tags
("Police — Ch 1"), and re-rendering the same event once the incident learns its
address. Incidents created before this change have no `title_tag`, so their
existing title is treated as the founding one rather than handed to whichever
call links next.
Interaction with the caps from
|
||
|
Logan Cusano
|
c7be6416f2 |
Surface LLM correlation fields in debug view; fix unit-continuity path
/admin/debug/correlation stripped corr_consensus and the corr_llm_* fields
that upload.py's consensus correlator writes onto the call doc, making it
the one tool built to answer "is the LLM correlation tier alive" unable to
answer it (2026-08-19 dump had to infer LLM state from commit dates instead
of reading it off the data). admin.py's _call_summary() now includes
corr_consensus, corr_llm_reasoning, corr_llm_action, corr_rules_action.
The unit-continuity correlation path never wrote corr_matched_units, unlike
fast/single and fast/disambig, so the debug view showed null for a match
that was in fact unit-driven by construction. Now populated unconditionally
on that path (server-26#16).
Also traced the negative corr_incident_idle_min (-4.1 observed) to its root
cause: the re-correlation sweep anchors `now` to the linking call's own
started_at, and that back-dated value was being written straight into the
incident's updated_at, letting it land before the incident's own
started_at. Added _floor_at_started_at() so updated_at can never precede
started_at. (commit
|
||
|
Logan Cusano
|
8fbfe7d6de |
Make a failed deploy impossible to miss, and a wildcard CORS harmless
Two unrelated-looking problems with the same shape: a dangerous state that
looked fine from the outside.
DEPLOY (server-26#21). The Deploy job failed on fifteen consecutive pushes
between 2026-08-18 and 08-20 and nobody noticed for two days, because the
build job was green and a red run is only visible to someone who opens Gitea.
Production served 08-18 code the whole time -- including the entire frontend
redesign, chunks 2 through 8. Three changes:
* The health check now asserts WHICH build answered, not just that something
did. CI bakes the commit into the image (Dockerfile ARG/ENV GIT_SHA) and
/health reports it, so a deploy that "succeeds" while the previous
container keeps running now fails. Liveness alone could never have caught
this.
* The image pull retries once after a prune. The actual failure was
containerd unable to extract a layer -- "failed to Lchown ... no such file
or directory" -- a corrupted entry in the snapshot store, which a prune
clears. A second failure after pruning is a real problem (check the VM's
disk) and still stops the deploy.
* A notify-failure job POSTs to DEPLOY_ALERT_WEBHOOK when anything in the
workflow fails. Unset means skip quietly, not fail.
CORS (server-26#20). allow_origins=["*"] with allow_credentials=True is not
the permissive-but-harmless setting it reads as. Starlette does not reject the
pair -- it reflects the caller's Origin back and still sends
Access-Control-Allow-Credentials: true, so the effective policy is "any
origin, WITH credentials", the opposite of what a wildcard normally means.
Rather than trust every deployment to remember CORS_ORIGINS, the pair is now
unrepresentable: a wildcard forces allow_credentials off and logs an ERROR
naming the variable to set. Correctly configured deployments that name their
origins are unaffected and keep credentialed requests.
Severity honestly: low today. c2-core is bearer-auth, and browsers do not
attach bearer tokens cross-origin the way they attach cookies. This is a
misconfiguration waiting for the day something starts trusting a cookie.
Also adds firebase_admin.auth.UserRecord and the list/update/create/delete_user
names to the conftest stub. routers/users.py annotates with UserRecord at
import time, so without it importing app.main failed at collection -- which is
why nothing had ever tested anything wired at app level, CORS included.
Tests: 5 new in test_cors_policy.py, covering the pure policy function, the
middleware actually mounted on the app (so re-hardcoding allow_credentials=True
fails here), and the presence of the build stamp.
Closes logan/server-26#20
Closes logan/server-26#21
|
||
|
Logan CusanoandClaude Opus 5
|
33a247d306 |
Stop thin calls fusing a work shift into one incident (server-26#22)
The 2026-08-20 production dump had 4 of 6 sampled incidents as junk chains,
the worst being f5190670: 68 calls over 4h09m, 44 units, 12 tags, at least
13 genuinely distinct events. 58 of 133 linked calls took the fast/thin
path, which is the one path that attaches a call with no fit test at all.
Three defects combined to produce that, and all three are fixed here.
1. What counted as thin was wrong.
is_thin_call was "not units and not vehicles and not coords". A real
dispatch qualified as thin whenever no unit ID parsed and the geocode
failed - six of them did in that dump, including "All units head over to
the powerhouse, 55 Hyman Hills Road ... she's 87 years old", a brand new
job that attached to the four-hour chain and then overwrote its location
and its title. A call is now substantive if it carries tags, a location
string, a severity above routine, or is a reassignment; only genuinely
content-free housekeeping ("10-4", "Copy") stays thin. Those calls now go
through _call_fits_incident like everything else, which on a dispatch
backbone with no positive signal means they open their own incident or
orphan rather than merging.
The reassignment clause closes a self-defeating guard: upload.py blanks
units when dispatch pulls a unit onto a NEW job, specifically to stop
unit-overlap chaining - and blanking units made the call thin, routing it
to the only path with no fit check. The guard produced the merge it
existed to prevent.
2. The thin path was bounded on dispatch channels only.
Every other talkgroup fell through to "thin_pool = tg_recent": any
incident idle up to tg_fast_path_idle_minutes (90), no single-candidate
requirement, no fit test. The 30-second tier-1 / single-candidate tier-2
structure now applies to all channels. Non-dispatch gets its own window,
TG_THIN_IDLE_MINUTES=15, rather than sharing the dispatch value: a
tactical channel really is dedicated to one scene so it earns longer, but
15 sits inside the 20-minute tactical-default window already used in
_call_fits_incident, so the no-evidence path is never more permissive than
the fit-tested path on the same channel.
Recency gates now compare the magnitude of the idle, not the signed value.
The re-correlation sweep anchors "now" to the call's own started_at, so
idle goes negative routinely - incident 9d376ffe recorded
corr_incident_idle_min: -4.1 - and every "idle <= window" test in this
module reads True for a negative number. Those gates had silently stopped
bounding anything for exactly the calls the sweep re-examines.
3. Nothing capped an incident's total size.
Every fit test in the correlator is pairwise: does this call belong with
that incident. Each of f5190670's 68 links was individually arguable; the
mistake was the accumulated shape, which no pairwise rule can see. Two
hard caps now remove an incident from the candidate pool entirely, before
any path can choose it - including the LLM tier, which reads the same
ctx lists.
INCIDENT_MAX_DURATION_MINUTES=120. The one incident in that dump that was
genuinely a single event ran 63 minutes (06:15 wrong-way driver to 07:18
closeout), so the cap has to clear an hour with real headroom. The four
junk chains ran 3h41m, 3h43m, 4h05m and 4h09m, so it has to sit well under
three hours. 120 also equals correlation_window_hours: the location and
slow paths already refuse a candidate older than that, and the fast path
was the only one exempt, so this removes an inconsistency rather than
inventing a number.
INCIDENT_MAX_CALLS=40. A backstop for a burst that fills up inside the
duration cap, not the primary bound. The worst chain averaged ~16
calls/hour while absorbing an entire dispatch backbone, so 40 calls in
under two hours means one incident is eating most of the channel. Set
deliberately above any plausible single-incident call volume (a
multi-alarm fire on its own tactical channel) so this cap errs toward
keeping real incidents whole and lets the duration cap do the cutting.
Capping is not truncation: the incident keeps every call it has and still
auto-resolves on the normal idle sweep. It just stops being a candidate.
Every ambiguous call here was resolved toward a separate incident rather
than a merge. A wrongly-separate incident is visibly wrong and can be
merged later; a wrongly-merged one silently corrupts every unit, tag,
severity and map pin on the incident it joined, and poisons the AI
summary written from them. The cost is some acknowledgements orphaning
instead of riding along on an incident, which is a small, visible loss.
Deliberately NOT changed, since both push toward more merging while the
current failure mode is entirely over-merging (every incident in the dump
has exactly one "new" call; there is no over-splitting left to trade
against):
- unit-overlap positive feedback on shared dispatch channels, which is
now bounded by the caps rather than fixed at its root
- the sweep retry budget expiring before the target incident exists
Tests: 31 new cases in tests/test_correlator_merge_caps.py, including a
replay of the f5190670 night - 13 unrelated jobs at their real offsets,
plus roster unit traffic and acknowledgements every two minutes. Without
the caps that traffic still builds a 125-call incident spanning 244
minutes; with the old thinness test on top, 153 calls over 247 minutes in
3 incidents. With this commit it is 13 incidents, largest 40 calls over 80
minutes. Each new case was checked to fail when the behaviour it covers is
reverted. Suite: 138 passed.
No AI feature flag was touched; correlation stays off in production.
Closes logan/server-26#22
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
|
||
|
Logan Cusano
|
a250c29e3c |
Add AI provider degradation registry and alerting (server-26#14)
Three AI dependency failures in one night (retired Gemini model IDs, depleted Gemini balance, unpayable OpenAI account) each surfaced only as a single ERROR log line that nobody was watching. Add app/internal/ai_health.py, a shared in-memory registry that transcription.py and llm_correlator.py report into on every call (success and failure), distinguishing permanent conditions (dead model, dead billing) which alert immediately from transient ones (rate limits, network blips) which only alert after they persist. Alerts POST once per degradation episode and once on recovery to an optional Discord webhook (AI_ALERT_WEBHOOK_URL), reusing alerter.py's httpx pattern. State is exposed unauthenticated at GET /health/ai alongside the existing /health. Closes logan/server-26#14 |
||
|
Logan CusanoandClaude Opus 5
|
6dfa5bc66d |
fix: repair 10 stale tests in test_mqtt_handler.py and test_node_sweeper.py
All 10 failures were tests that had drifted behind the product code, not
regressions in it. Diagnosed each individually:
test_mqtt_handler.py:
- test_checkin_creates_new_node, test_checkin_new_node_defaults_lat_lon:
unpacked 4 positional args from doc_set.call_args[0], but
fstore.doc_set(collection, doc_id, data, merge=False) always passes
merge as a kwarg, so only 3 positional args are ever recorded. Fixed
the unpack to 3.
- test_call_start_creates_call_doc, test_call_start_uses_now_when_started_at_missing:
mocked fstore.doc_get, but _on_call_start looks the node up via the
cached fstore.doc_get_cached (added when Firestore reads were cut to
stay in the free tier). The unmocked doc_get_cached returned a bare
MagicMock, which isn't awaitable. Mocked doc_get_cached instead; also
fixed the same 4-vs-3 positional-arg unpack on doc_set's merge=False call.
- test_call_end_updates_status_and_times, test_call_end_sets_audio_url_when_present:
mocked fstore.doc_update, but _on_call_end now writes via
fstore.doc_set(merge=True) (see the "Fix Upload 404 warning" commit —
doc_update raised "No document to update" when call_end arrived before
call_start). Also calls doc_get_cached to stamp org_id. Mocked
doc_get_cached and asserted against doc_set instead of doc_update.
test_node_sweeper.py:
- test_stale_online_node_marked_offline, test_stale_recording_node_marked_offline,
test_tz_naive_last_seen_is_handled, test_only_stale_nodes_updated_in_batch:
_sweep() now calls app.routers.tokens.release_token(node_id) for every
node it marks offline (added in
|
||
|
Logan Cusano
|
70d63abeaa |
Re-evaluate incident severity on link, stamp resolved_at at every resolution site
#17: severity was written once at _create_incident and never touched again, so an incident that opened routine and escalated to a working fire stayed routine forever. _update_incident now merges call_severity into the incident via _max_severity() on every link. Severity is monotonic: it only ever rises, never falls. An incident briefly assessed "major" genuinely was major at that moment; a later, calmer-sounding call is evidence the situation is winding down, not that the earlier read was wrong. status/resolved_at exist to retire an incident — severity should stay as the high-water mark so the worst-first rail, "Major only" filter, and map colouring never bury a call that was genuinely major. See _max_severity's docstring in incident_correlator.py for the full argument. #18: none of the resolution sites wrote resolved_at, so an incident's lifespan couldn't be reconstructed for the history-scrub feature. Added resolved_at alongside status="resolved" at all six sites that flip it: - incident_correlator.py _update_incident (signal-based: units all cleared) - incident_correlator.py maybe_resolve_parent (master auto-resolve) - summarizer.py _stale_sweep (90-minute auto-resolve) - upload.py, both scene-resolution loops (single- and multi-scene) - calls.py reprocess/correction path (_update_incident's signal-resolve and maybe_resolve_parent's master-resolve weren't named in the issue's four call sites, but they set status the same way and were missing resolved_at too.) No backfill: existing resolved incidents keep resolved_at = null, which means "resolved before this field existed," not "never resolved." Backfilling from updated_at would be a guess dressed up as data. Tests: added to tests/test_correlator_gate.py, which needs no Firestore for the pure _max_severity cases and patches fstore for the _update_incident/ maybe_resolve_parent writes. Covers the escalation case (routine -> major), the no-downgrade case, and resolved_at on both the signal-resolve and master-resolve paths. 52/52 passing in that file; 83 passed / 10 pre-existing failures for drb-c2-core overall (baseline was 69/10 — the +14 is exactly the new tests, no regressions). Fixes #17, #18. |
||
|
Logan CusanoandClaude Opus 5
|
c09cb72f66 |
Compare unit IDs by normalised key, not exact string
Dispatch audio names the same unit several ways within one conversation, and
every comparison in the correlator used exact string equality, so a follow-up
transmission from a unit already on an incident simply failed to find it. With
the creation gate no longer letting routine traffic open its own incident,
these stopped becoming junk incidents and started becoming orphans instead --
which is how they became visible. In the 01:05Z dump, five of eighteen orphans
were calls belonging to an incident that was open at that moment:
"K-9A2" vs "K-9-A-2" punctuation
"5-1-6" vs "516" digits read out individually
"37" vs "37th Post" ordinal plus role word
"11-Victor" vs "11 Victor" hyphen vs space
_normalize_unit lowercases, drops punctuation and role words (post/unit/car),
strips ordinal suffixes, and joins the remaining tokens, so each pair above
collapses to one key. All six comparison sites now go through it: the two
fast-path debug reporters, unit-continuity candidate selection and its
reassignment check, the cross-talkgroup 2+ shared-unit test, and the
disambiguation scorer.
What it deliberately does NOT do is match a bare district letter -- "Adam" is
not treated as "6-Adam". Every district has an Adam, and collapsing them would
merge unrelated incidents across districts. That leaves a couple of the
observed orphans unlinked, which is the right trade: a missed link leaves an
orphan the re-correlation sweep retries three times, while a false link
corrupts an incident permanently and nothing walks it back.
Two smaller things fall out of the shared helper. Matches are reported as the
original spoken strings rather than the normalised keys, so corr_matched_units
stays readable in the debug view. And a unit made only of role words ("Post")
would normalise to the empty string and then compare equal to every other such
unit, so it falls back to the raw text -- tested, because that failure would be
silent and would merge aggressively.
Adds 13 cases: each observed pair, five pairs that must stay distinct, the
empty-key guard, match reporting, and an end-to-end check that the K-9A2 call
now links where it previously orphaned. 38 pass.
No new environment variables, so CI deploys this without an ansible run.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
|
||
|
Logan CusanoandClaude Opus 5
|
7b5258cfdf |
Halve the tier-2 thin-call window, from 10 minutes to 5
With over-creation fixed, the incidents that remain are readable enough to judge, and the ones that still do not make sense all fail the same way. A content-free call attaches to the single active incident on its talkgroup if that incident has been idle under tg_dispatch_thin_idle_minutes, and at 10 minutes that is long enough for the channel to have moved on to something else. In the 00:30Z dump a "72 at Holland Station" incident absorbed a Grand Central train-crew meet 9.6 minutes later, and a status check absorbed a records lookup at 9.7. Being the only candidate is not evidence. It means the channel was quiet, which is exactly when guessing is weakest -- the single-candidate rule was meant to avoid picking wrongly among several, not to license a match no other signal supports. Every correct thin attach in that dump was <= 3.4 minutes idle and every wrong one was >= 8.2, so 5 separates them with room on both sides. Real back-and-forth is unaffected: it runs through the 30-second tier-1 path, and the observed conversational replies sit near zero. Tests pin both sides of the new boundary at 4.9 and 5.1 minutes so a later change to this number has to be deliberate. 23 pass. Also corrects a DEFERRED.md entry written earlier today. It claimed nothing ever closes an incident that goes quiet; summarizer.py has run a stale sweep at incident_auto_resolve_minutes (90) the whole time. The 37 open incidents were caused by over-creation, not by a missing sweeper, and 90 minutes may be fine now -- worth rechecking on a fully post-fix dump before changing it. No new environment variables: tg_dispatch_thin_idle_minutes is a config.py default and is not templated into any .env, so CI deploys this without an ansible run. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> |
||
|
Logan CusanoandClaude Opus 5
|
96625fabd0 |
Stop ambient radio chatter from opening incidents, and refill the map
The 23:46Z correlation dump confirmed the severity gate fixed the problem it
was written for -- orphans fell from 69 to 16, and only three of those are
after the deploy boundary, two of them deliberate skips. Nothing on TG 9048
absorbs the channel any more; the largest post-deploy incident is four calls
over nine minutes and is genuinely one event.
It overcorrected. 37 of 50 incidents were open, most a single routine call.
The cause was the gate's own substance test, which counted `units` and
`location`. Radio protocol puts a unit ID in essentially every transmission
and a place name in most of them, so has_substance was true almost always and
the severity check never actually ran -- "11-Victor, 72 at Holland Station"
became its own permanent incident. Substance is now a vehicle, a geocode or a
tag: things the extractor found beyond who was speaking and where they stood.
Severity still opens an incident on its own, so nothing real is lost.
incident_type is now validated against the enum the prompt offers rather than
trusted. It is written straight through to incident.type and rendered as the
title, so a model that answered the severity question in the type field
produced an incident titled "Routine -- TGID 9563". Unrecognised values become
None and fall to the tag/severity path, which is what "unknown" already did.
The map was empty for a separate reason: geocoding accepted only ROOFTOP and
RANGE_INTERPOLATED. Dispatch names places the way people speak, and Google
returns GEOMETRIC_CENTER for exactly those forms -- intersections ("Lake
Street and Veterans Memorial Drive") and named POIs ("Brewster Station").
Requiring a street address discarded nearly every real dispatch location and
left only numbered addresses plotted, which is why the July incidents have
coordinates and none since do. GEOMETRIC_CENTER is now accepted; APPROXIMATE
is still rejected, since a region centroid is what an ungeocodable string
degrades to. Note this is necessary but may not be sufficient -- if
GOOGLE_MAPS_API_KEY is unset on the host the map stays empty regardless, and
that has not been checked from here.
Two things found and deliberately not fixed, both in DEFERRED.md. One call can
still land in two incidents, because upload.py correlates each extracted scene
independently and the model over-split one conversation; multi-scene is
intentional, so that is prompt tuning rather than a code change. And nothing
closes an incident that merely goes quiet -- signal-resolution and master
auto-resolve both exist, but a one-call incident nobody clears stays active
forever. That wanted the over-creation fixed first so a time-based sweeper
would not just paper over it.
Gate tests updated: units and location alone must now orphan, and the case
that matters most is kept explicit -- units with a real severity still open an
incident. 17 pass. No new environment variables, so CI deploys this without an
ansible run.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
|
||
|
Logan Cusano
|
6d5eb4c5f2 |
Let severity, not incident_type, decide what becomes an incident
The 2026-08-16 correlation dump showed two failures that looked unrelated and were the same bug. TG 9048 held one incident of 28 calls spanning 49 minutes -- a prisoner transport, a drone retrieval, a records lookup and a canvass, glued together -- while 32 other calls on that same channel stayed permanently orphaned. Creating an incident required a concrete incident_type. Nothing on a transit police channel produced one: the extraction prompt said to prefer "other" when uncertain, extraction then collapsed "other" to None, and the tag-based fallback had no tags to work with because administrative traffic carries none. So the channel could never open a SECOND incident. Every later call funnelled into whichever incident happened to exist first, and every call too substantial for the thin path had nowhere to go at all. The two symptoms were the same missing value seen from opposite ends. Severity now decides incident-worthiness. It is a better fit for the question being asked -- "is this a real event?" -- than a service label ever was, and unlike incident_type it is always present. The prompt defines four levels with no escape hatch (routine/minor/moderate/major, "unknown" is gone) and calls skipped for a too-short transcript are still recorded as routine, because downstream code reads a missing severity as "not processed yet" rather than "nothing happened". Anything above routine, or carrying any extracted content, opens an incident under the neutral "other" type. "other" is also kept as a real classification now -- rail operations and public works genuinely are not police, fire or EMS. Separately, thin calls no longer refresh updated_at; they write last_thin_at. updated_at drives every recency gate in the fast path, so each "10-4" was resetting the idle clock on whatever it attached to, keeping that incident inside the gate for as long as anyone kept acknowledging. An incident now ages from its last substantive call. This is what made the 49-minute incident possible even once buckets existed, so it is fixed independently rather than being left to the gate change. The re-correlation sweep also now honours skip_reason. /upload has always refused to correlate garbage and too-short transcripts, but the sweep did not apply the same filter, so those fragments came back minutes later through the thin path and attached to whatever was most recent -- a second, quieter route into the same over-merge. Adds tests/test_correlator_gate.py (15 cases), the first tests against incident_correlator.py in its 1,517-line history. tests/conftest.py stubs firebase-admin only when it is genuinely absent, so the container's real SDK is never shadowed; this is what makes the correlator importable in the dev venv. That stub also made test_mqtt_handler and test_node_sweeper collectable for the first time, revealing 10 pre-existing failures in them -- test-vs-code drift, untouched here and catalogued in DEFERRED.md. No new environment variables, so CI deploys this without an ansible run. |
||
|
Logan Cusano
|
97013e1505 |
Stop Whisper hallucinations and dedupe recordings across nodes
Two independent sources of garbage in the AI pipeline, both visible in the 2026-08-16 correlation dump. 1. Hallucinated transcripts. The Whisper prompt opened with an enumerated run of ten-codes: 10-4, 10-23, 10-20, 10-97 and so on. Whisper treats prompt text as preceding transcript, so on noisy or silent audio it continued the series, emitting transcripts that count upward from 10-4 to 10-99. The existing no_speech_prob filter could not catch these: the model is highly confident in text it invented by continuing a pattern. The prompt no longer contains a series to extend, and _is_degenerate() rejects the three shapes this failure takes: ascending ten-code runs, one phrase looping, and near-identical segments across a whole recording. Verified against 13 transcripts from production: all four known hallucinations rejected, all nine real ones kept, including terse traffic containing legitimate codes. 2. Duplicate recordings. node-002 and node-PI-2 both cover TG 9048 and both uploaded the same transmissions, ~1.1s apart. Nine pairs appeared in one dump. Each was transcribed, billed and correlated twice, and the resulting incident listed two units where there was one. Canonical selection is by earliest started_at, tie-broken on call_id, NOT by upload order: upload order varies with encode time and network latency, so it would make the authoritative recording non-deterministic. Call documents are created from MQTT call_start before uploads arrive, so both nodes independently reach the same verdict. The loser keeps its audio (it may be the cleaner capture) but is excluded from STT, correlation, the re-correlation sweep and the orphan debug view. Also fixes _sync_transcribe returning a bare None when OPENAI_API_KEY is missing, where the caller unpacks two values. A missing key surfaced as a misleading "Transcription failed" instead of the real warning. Adds tests/test_dedup.py (15 cases). dedup.py reaches Firestore through an injected callable so it stays importable without firebase-admin present. |
||
|
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. |