Ports the intelligent-caching branch concept to a plain Python in-memory
implementation — no SQLAlchemy, no extra DB table, no new dependencies.
_LiveCache (threading.Lock + dicts) caches:
- device info: indefinite, invalidated by POST /device/config
- events: keyed by (conn_key, device_event_count); count-probe fast path
(~2s poll+count_events) avoids full downloads when nothing is new
- monitor status: 30-second TTL, invalidated by monitor start/stop
- waveforms: permanent per (conn_key, event_index)
All four cached endpoints accept ?force=true to bypass the cache.
Removes sfm/cache.py (SQLAlchemy experiment, now superseded).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
sfm/database.py (new)
- SeismoDb class: three tables keyed by unit serial number
- ach_sessions: one row per ACH call-home
- events: one row per triggered event, deduped by (serial, waveform_key)
- monitor_log: one row per monitoring interval, deduped by (serial, waveform_key)
- WAL mode, per-request connections, silent dedup via UNIQUE constraint
- Query helpers: query_events(), query_monitor_log(), get_sessions(), query_units()
- false_trigger flag on events for future review UI / report filtering
bridges/ach_server.py
- Import SeismoDb; create shared instance at startup pointed at
bridges/captures/seismo_relay.db
- After each call-home: insert_events() + insert_monitor_log() + insert_ach_session()
- DB failures logged as warnings, never abort the session
sfm/server.py
- Import SeismoDb; lazy singleton via _get_db()
- New DB read endpoints: GET /db/units, /db/events, /db/monitor_log, /db/sessions
- PATCH /db/events/{id}/false_trigger for manual review flagging
CLAUDE.md / CHANGELOG.md
- Document DB schema, SFM DB endpoints, architecture decision (unit-keyed only)
- Version bump to v0.11.0
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Introduces sfm/cache.py — a SQLite-backed cache (via SQLAlchemy) that
sits between the SFM REST endpoints and the device, eliminating redundant
cellular downloads for data that doesn't change.
Cache behaviour by data type:
- Device info / compliance config: cached until a config write occurs;
POST /device/config now calls mark_config_dirty() to force a fresh read
on the next /device/info call.
- Event headers + peak values: cached permanently (append-only). On
subsequent calls to /device/events, the server does a fast count_events()
(~2s) instead of a full download (~10-30s); only new events are fetched
from the device and merged into the cache.
- Full waveforms (raw ADC samples): cached permanently — immutable once
recorded. Repeated requests for the same waveform return instantly with
zero device contact.
- Monitor status (battery, memory, is_monitoring): 30-second TTL; auto-
invalidated on start/stop monitoring commands.
All endpoints gain a ?force=true param to bypass the cache when needed.
New endpoints: GET /cache/stats, DELETE /cache/device.
Adds requirements.txt listing fastapi, uvicorn, sqlalchemy, pyserial.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
section[6] is the monitoring flag (was wrongly section[1] — section[1] is always
0x00 in both states). Battery and memory fields use relative-from-end offsets
(section[-11:-9], section[-9:-5], section[-5:-1]) instead of absolute positions,
which broke when the payload grew by 3 bytes in monitoring mode.
Confirmed from full byte diff of 142 0xE3 frames in 4-8-26/2ndtry capture.
SFM start_monitoring now polls /device/monitor/status every 5s for up to 60s
instead of a fixed 25s delay (unit runs ~40s on-device sensor check before
confirming monitoring state).
Also corrects stale 1C→6E response anomaly claim in protocol reference — no
exceptions to the 0xFF−SUB rule are known.
- Introduced new SUBs for monitoring status, start, and stop commands in protocol.py.
- Implemented read_monitor_status, start_monitoring, and stop_monitoring methods in MiniMateProtocol class.
- Added new API endpoints for monitoring status retrieval and control in server.py.
- Enhanced the web application with a monitoring panel, including battery and memory status display.
- Created a new Python script to parse SUB 0x1C response frames for monitoring status.
- Documented the monitoring status response format and field locations in markdown and text files.
Two improvements to eliminate the ~2-min-per-event wait and unnecessary
full-event-list download when only one event is requested:
1. protocol.py: pass timeout=10.0 to _recv_one in the 5A chunk loop.
Device responds within ~1s per chunk; 10s gives a safe 10x buffer.
End-of-stream detection (raw_bytes=1) now fires in 10s instead of 120s,
cutting ~110s of dead wait per event.
2. client.py: add stop_after_index parameter to get_events(). When set,
iteration stops immediately after the target event is collected — no
further 0A/1E/0C/5A/1F cycles for events the caller doesn't need.
3. server.py: pass stop_after_index=index to both /device/event/{idx}
and /device/event/{idx}/waveform endpoints so a single-event request
only downloads that one event.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
The 210-byte waveform record only stores "Project:" — client, operator,
sensor_location, and notes are device-level settings in SUB 1A, not
per-event fields. Backfill those into each event's project_info after
download, same pattern as the sample_rate backfill.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- DeviceInfo.event_count: Optional[int] = None (new field in models.py)
- connect() now calls proto.read_event_index() after compliance config and
stores the decoded count in device_info.event_count
- _serialise_device_info() exposes event_count in /device/info and /device/events
JSON responses
event_count is decoded from uint32 BE at offset +3 of the 88-byte F7 payload
(🔶 inferred — needs live device confirmation against a multi-event device).
Any ProtocolError from the index read is caught and logged; event_count stays
None rather than failing the whole connect().
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
sample_rate is a device-level setting stored in the compliance config,
not per-event in the waveform record. After downloading events, backfill
ev.sample_rate from info.compliance_config.sample_rate for any event
that didn't get it from the waveform record decode path.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Adds full support for reading device compliance configuration (2090-byte E5
response) containing record time, trigger/alarm levels, and project strings.
protocol.py:
- Implement read_compliance_config() two-step read (SUB 1A → E5)
- Fixed length 0x082A (2090 bytes)
models.py:
- Add ComplianceConfig dataclass with fields: record_time, sample_rate,
trigger_level_geo, alarm_level_geo, max_range_geo, project strings
- Add compliance_config field to DeviceInfo
client.py:
- Implement _decode_compliance_config_into() to extract:
* Record time float at offset +0x28 ✅
* Trigger/alarm levels per-channel (heuristic parsing) 🔶
* Project/setup strings from E5 payload
* Placeholder for sample_rate (location TBD ❓)
- Update connect() to read SUB 1A after SUB 01, cache in device_info
- Add ComplianceConfig to imports
sfm/server.py:
- Add _serialise_compliance_config() JSON encoder
- Include compliance_config in /device/info response
- Updated _serialise_device_info() to output compliance config
Both record_time (at fixed offset 0x28) and project strings are ✅ CONFIRMED
from protocol reference §7.6. Trigger/alarm extraction uses heuristics
pending more detailed field mapping from captured data.
Sample rate remains undiscovered in the E5 payload — likely in the
mystery flags at offset +0x12 or requires a "fast mode" capture.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Confirmed 2026-04-01 against Blastware event report for BE11529 thump
event ("00:28:12 April 1, 2026", PVS 3.906 in/s).
models.py:
- Timestamp.from_waveform_record(): decode 9-byte format from 0C record
bytes[0-8]: [day][sub_code][month][year:2BE][?][hour][min][sec]
- Timestamp: add hour/minute/second optional fields; __str__ includes
time when available
- PeakValues: add peak_vector_sum field (confirmed fixed offset 87)
client.py:
- _decode_waveform_record_into: add timestamp decode from bytes[0:9]
- _extract_record_type: decode byte[1] (sub_code), not ASCII string
search; 0x10 → "Waveform", histogram TBD
- _extract_peak_floats: add PVS from offset 87 (IEEE 754 BE float32)
= √(T²+V²+L²) at max instantaneous vector moment
sfm/server.py:
- _serialise_timestamp: add hour/minute/second/day fields to JSON
- _serialise_peak_values: add peak_vector_sum to JSON
docs: update §7.7.5 and §8 with confirmed 9-byte timestamp layout,
PVS field, and byte[1] record type encoding; update command table;
close resolved open questions.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- _decode_serial_number: read from data[11:] not data[:8] — was returning
the LENGTH_ECHO byte (0x0A = '\n') instead of the serial string
- _extract_peak_floats: search for channel label strings ("Tran" etc) and
read float at label+6; old step-4 aligned scan was reading trigger levels
instead of PPV values
- get_events: add debug=False param; stashes raw 210-byte record on
Event._raw_record when True for field-layout inspection
- server /device/events: add ?debug=true query param; includes
raw_record_hex + raw_record_len in response when set
- models: add Event._raw_record optional bytes field
- bridges/tcp_serial_bridge.py: increase default boot_delay 2s → 8s to
cover MiniMate Plus cold-start time (unit wakes from RS-232 line
assertion but takes 5-10s to be ready for POLL_PROBE).
- sfm/server.py: add _run_with_retry() — on TCP connections only, retries
once on ProtocolError. Serial timeouts are not retried (usually a real
fault). Confirmed behaviour: unit wakes purely from RS-232 line voltage,
no software wake-up frame needed.
- minimateplus/transport.py: add TcpTransport — stdlib socket-based transport
with same interface as SerialTransport. Overrides read_until_idle() with
idle_gap=1.5s to absorb the modem's 1-second serial data forwarding buffer.
- minimateplus/client.py: make `port` param optional (default "") so
MiniMateClient works cleanly when a pre-built transport is injected.
- minimateplus/__init__.py: export SerialTransport and TcpTransport.
- sfm/server.py: add `host` / `tcp_port` query params to all device endpoints.
New _build_client() helper selects TCP or serial transport automatically.
OSError (connection refused, timeout) now returns HTTP 502.
- docs/instantel_protocol_reference.md: add changelog entry and full §14
(TCP/Modem Transport) documenting confirmed transparent passthrough, no ENQ
on connect, modem forwarding delay, call-up vs ACH modes, and hardware note
deprecating Raven X in favour of RV55/RX55.
Usage: GET /device/info?host=<modem_ip>&tcp_port=12345
claude
serversdown