queryGrouped was missing route_type, snr, rssi fields. The packet filter
language couldn't filter by route/snr/rssi since grouped packets didn't
have those fields.
Migration runs automatically on next startup — drops paths first (FK to
packets), then packets. Removes insertPacket(), insertPath(), all
prepared statements and references to both tables. Server-side type/
observer filtering also removed (client does it in-memory).
Saves ~2M rows (paths) + full packets table worth of disk.
During _loadNormalized(), observations load in DESC order so the first
observation processed is the LATEST. tx.observer_id was set from this
latest observation. Added post-load pass that finds the earliest
observation by timestamp and sets tx.observer_id/path_json to match.
When a later observation has an earlier timestamp, the transmission's
first_seen was updated but observer_id and path_json were not. This
caused the header row to show the wrong observer and path — whichever
MQTT message arrived first, not whichever observation was actually
earliest.
The insert() method had a second code path (for building rows from
packetData) that pushed observations without checking for duplicates.
Added the same dedup check (observer_id + path_json) that exists in
the other insert path and in the load-from-DB path.
The UNIQUE index on (hash, observer_id, path_json) didn't prevent
duplicates when path_json was NULL because SQLite treats NULL != NULL
for uniqueness. Fixed by:
1. Using COALESCE(path_json, '') in the UNIQUE index expression
2. Adding migration to clean up existing duplicate rows
3. Adding NULL-safe dedup checks in PacketStore load and insert paths
packets_v view uses observation IDs, not transmission IDs, so
getPacket(transmissionId) returned null. Skip the DB lookup entirely
and construct the row directly from the incoming packetData object
which already has all needed fields.
packets_v view uses transmission IDs, not packets table IDs.
insertPacket returns a packets table ID which doesn't exist in
packets_v, so getPacket returned null and new packets never got
indexed in memory. Use transmissionId from insertTransmission instead.
- Create packets_v SQL view joining transmissions+observations to match old packets schema
- Replace all SELECT FROM packets with packets_v in db.js, packet-store.js, server.js
- Update countPackets/countRecentPackets to query observations directly
- Update seed() to use insertTransmission instead of insertPacket
- Remove insertPacket from exports (no longer called)
- Keep packets table schema intact (not dropped yet, pending testing)
The previous approach matched nodes via data packet hashes seen by
regional observers — but mesh packets propagate everywhere, so nearly
every node matched every region (550/558).
New approach: _advertByObserver index tracks which observers saw each
node's ADVERT packets. ADVERTs are local broadcasts that indicate
physical presence, so they're the correct signal for geographic filtering.
Also fixes role counts to reflect filtered results, not global totals.
~26% of observations were duplicates from multi-broker MQTT ingestion.
Added UNIQUE index to prevent future dupes, INSERT OR IGNORE to skip
silently, and in-memory dedup check in packet-store.
After dedup migration, packet IDs from the legacy 'packets' table differ
from transmission IDs in the 'transmissions' table. URLs using numeric IDs
became invalid after restart when _loadNormalized() assigned different IDs.
Changes:
- All packet URLs now use 16-char hex hashes instead of numeric IDs
(#/packets/HASH instead of #/packet/ID)
- selectPacket() accepts hash parameter, uses hash-based URLs
- Copy Link generates hash-based URLs
- Search results link to hash-based URLs
- /api/packets/:id endpoint accepts both numeric IDs and 16-char hashes
- insert() now calls insertTransmission() to get stable transmission IDs
- Added db.getTransmission() for direct transmission table lookup
- Removed redundant byTransmission map (identical to byHash)
- All byTransmission references replaced with byHash
After dedup, table rows have transmission IDs but getById() maps
observation IDs. Added byTxId index so /api/packets/:id resolves
both observation and transmission IDs correctly.
- Add transmissions and observations schema to db.js init
- Add insertTransmission() function: upsert transmission by hash,
always insert observation row
- All 6 pktStore.insert() call sites in server.js now also call
db.insertTransmission() with try/catch (non-fatal on error)
- packet-store.js: add byTransmission Map index (hash → transmission
with observations array) for future M3 query migration
- Existing insertPacket() and all read paths unchanged
One method resolves name→pubkey, combines byNode index + text search.
Used in: query fast-path, combined-filter path, health endpoint,
analytics endpoint. Bulk-health still uses index-only (perf).
Zero SQLite reads from packets table. Every endpoint that previously
scanned packets now reads from the in-memory PacketStore.
Expected: subpaths from 1.6s to <100ms, topology from 700ms to <50ms,
RF from 270ms to <30ms on cold calls.
- PacketStore loads all packets into memory on startup (~11MB for 27K packets)
- Indexed by id, hash, observer, and node pubkey for fast lookups
- /api/packets, /api/packets/timestamps, /api/packets/:id all served from RAM
- MQTT ingest writes to both RAM + SQLite
- Configurable maxMemoryMB (default 1024MB) in config.json packetStore section
- groupByHash queries computed in-memory
- Packet store stats exposed in /api/perf
- Expected: /api/packets goes from 77ms to <1ms
