10 Commits

Author	SHA1	Message	Date
you	84f33aef7b	M3: Restructure in-memory store around transmissions ... - load() reads from transmissions JOIN observations (with legacy fallback) - byHash now maps hash → single transmission object (1:1) - byNode maps pubkey → [transmissions] (deduped, no inflated observations) - byTransmission is the primary data structure - byId maps observation IDs for backward-compat packet detail links - byObserver still maps observer_id → [observations] - getSiblings() returns observations from transmission - findPacketsForNode() returns unique transmissions - query()/queryGrouped() work with transmission-centric model - All returned objects maintain backward-compatible fields - SQLite-only fallback path (NO_MEMORY_STORE=1) unchanged - Tested: 11.6K transmissions from 37.5K observations (3.2x dedup)	2026-03-20 20:44:32 +00:00
you	baa60cac0f	M2: Dual-write ingest to transmissions/observations tables ... - 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	2026-03-20 20:29:03 +00:00
you	fa40ede9e7	fix: findPacketsForNode always resolves name, even when called with pubkey	2026-03-20 08:31:07 +00:00
you	8ce2262813	refactor: single findPacketsForNode() replaces 4 duplicate node lookups ... 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).	2026-03-20 08:29:45 +00:00
you	87bbd93d12	fix: node-only search path also resolves name→pubkey + text search ... The fast-path for single node filter was bypassing the name resolution, using raw name string on pubkey-indexed byNode map	2026-03-20 08:20:43 +00:00
you	e837dba000	fix: node search combines index + text search for name AND pubkey ... Previous fix only used index OR text, missing packets that reference nodes by name in decoded_json	2026-03-20 08:13:27 +00:00
you	fa72e6242d	fix: node name search returns all packet types, not just adverts ... Resolves node name → pubkey, then searches byNode index and paths table instead of only matching decoded_json text	2026-03-20 08:10:24 +00:00
you	c2bc07bb4a	feat: live A/B benchmark — launches SQLite-only vs in-memory servers ... NO_MEMORY_STORE=1 env var makes packet-store fall through to SQLite for all reads. Benchmark spins up both servers on temp ports and compares: SQLite cold, Memory cold, Memory cached. Results on 27K packets (ARM64): Subpaths 5-8: SQLite 4.7s → cached 1.1ms (4,273×) Bulk health: SQLite 1.8s → cached 1.7ms (1,059×) Topology: SQLite 1.1s → cached 3.0ms (367×) Channels: SQLite 617ms → cached 1.9ms (325×) RF Analytics: SQLite 448ms → cached 1.6ms (280×)	2026-03-20 04:47:31 +00:00
you	1be6b4f4ad	perf: ALL packet reads from RAM — analytics, channels, topology, subpaths, RF, observers ... 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.	2026-03-20 03:43:23 +00:00
you	d8d0572abb	perf: in-memory packet store — all reads from RAM, SQLite write-only ... - 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	2026-03-20 03:38:37 +00:00