On dual-nozzle printers (H2C, H2D Pro) with multiple AMS units, usage
tracking attributed filament consumption to the wrong spools. The MQTT
mapping field — a snow-encoded array mapping slicer slots to physical
AMS trays — was preserved in state but never parsed.
Add _decode_mqtt_mapping() to convert snow encoding (ams_hw_id * 256 +
slot) to bambuddy global tray IDs. Insert as priority 2 in the mapping
resolution chain, after print_cmd ams_mapping but before queue/tray_now
fallbacks. Works for all printer models and print sources.
AMS-HT trays (IDs 128-135) are single-slot units, but the snow_slot
computation used `snow_tray % 4 if snow_tray < 128 else -1`, yielding
-1 for AMS HT. This caused a misleading "slot mismatch" debug log on
every MQTT update despite producing the correct tray_now value. Now
returns slot 0 for AMS-HT IDs (128-135), matching the firmware's
tray_now=0 for these single-slot units.
P2S (and X1C) with multiple AMS units incorrectly entered the H2D
dual-nozzle tray_now disambiguation code path, causing warning log
spam. The check `tray_now <= 3` matched any printer loading from
AMS 0. Now uses a persistent `_is_dual_nozzle` flag detected from
device.extruder.info (>= 2 entries), which only dual-nozzle printers
(H2D, H2D Pro) report. Flag is set early in message processing so
it's available before AMS data is parsed.
Recent BambuStudio/OrcaSlicer updates require a bind/detect handshake on
port 3000 before connecting via MQTT/FTP. Without this, slicers cannot
discover or connect to the virtual printer in any mode.
- Add BindServer for server modes (immediate/review/print_queue)
- Add TCPProxy for raw TCP forwarding (proxy mode)
- Update Dockerfile (EXPOSE 3000) and docker-compose.yml (bridge port)
- Add 10 new tests for BindServer protocol and integration
Some printer MQTT brokers (e.g. P1S) reject subscriptions to the request
topic by killing the TCP connection, causing a disconnect loop. The request
topic subscription for ams_mapping capture is now fail-safe with two
detection paths:
- SUBACK rejection: on_subscribe callback detects failure code 0x80
- Broker kills connection: on_disconnect detects disconnect within 10s of
unconfirmed subscription attempt
After detection, the request topic is disabled for that printer and
reconnection succeeds with only the report topic. Printers that support
the request topic (e.g. X1E, H2D) get ams_mapping capture; printers that
don't (e.g. P1S) gracefully degrade to tray_now-based tracking.
Subscribe to the MQTT request topic (device/{serial}/request) to intercept
print commands from any source (BambuStudio, OrcaSlicer, Bambu Handy, or
Bambuddy itself). Captures ams_mapping from project_file commands and passes
it through print start/complete callbacks. Fixes wrong tray attribution on
H2D Pro when prints are started from external slicers.
Usage tracking works on X1E but fails on H2D Pro — the wrong tray
is identified, attributing usage to the wrong spool. The root cause
is unknown because we don't know what MQTT mapping data each printer
model reports.
Adds INFO-level logging at print start and completion:
- state.raw_data["mapping"] (unknown format, preserved but never parsed)
- tray_now, last_loaded_tray values
- All raw_data keys containing "map" (discovery)
- Per-AMS-tray summary: tray_type, tray_color, tray_now, tray_tar
This enables comparing MQTT behavior across printer models without
requiring DEBUG mode, so testers can send normal logs.
Three fixes for inventory spool usage not updating after prints:
1. Store ams_mapping from print command (reprint, library print, queue)
so the usage tracker maps 3MF slots to the actual physical trays
the user selected, not the default slicer mapping.
2. Track last_loaded_tray on printer state — the last valid tray_now
(0-253) seen during printing. On H2D printers, tray_now is always
255 in AMS data; the real tray resolves via snow field ~44s after
print start but reverts to "unloaded" at completion. The fallback
chain is: tray_now_at_start > current tray_now > last_loaded_tray.
3. Use color similarity (Euclidean RGB distance, threshold 50) instead
of exact hex match for auto-unlink fingerprint checks. RFID sensors
report slightly different shades across reads (e.g. distance ~43.6
for the same spool), causing false assignment unlinks after prints.
When a print was queued from the file manager (library_file_id path),
the scheduler never created an archive or called register_expected_print().
This forced on_print_start to re-download the 3MF via FTP, and if that
failed, a fallback archive with no file was created — breaking 3MF-based
filament usage tracking entirely. The queue item's archive_id also stayed
NULL, so the usage tracker couldn't find the AMS slot mapping.
Now the scheduler creates an archive from the library file before uploading,
sets item.archive_id, and the existing register_expected_print() call fires
since archive is no longer None.
Add a "Clear Errors" button to the HMS error modal that sends
clean_print_error via MQTT and locally clears hms_errors for
immediate UI feedback. Useful for dismissing stale print_error
values that persist after print cancellation or transient events.
Pushover and Discord already received print thumbnails, but Telegram
and ntfy only got text. Telegram now uses sendPhoto with the image as
a caption attachment. ntfy sends the image as a binary PUT with
Filename/Message headers. No config changes needed.
The duplicate badge on archive cards only matched by content hash,
so re-sliced prints of the same model (different GCODE, same name)
were not flagged. Now also matches by print name (case-insensitive),
consistent with the detail view's find_duplicates() logic.
The firmware update cache stored downloaded files under a mangled name
(e.g. X1C_01_09_00_10.bin) instead of the original filename from
Bambu Lab's CDN. On cache hit the wrong filename was uploaded to the
printer's SD card, so the printer couldn't find the firmware file.
Now caches with the original filename so it's always correct.
Add a chronological, table-based print log as a 4th view mode in the
Archives page. Log entries are stored in a separate print_log_entries
table — clearing the log never touches archives or queue items.
Backend:
- New PrintLogEntry model with independent table
- GET /print-log/ endpoint with search, printer, user, status, date filters
- DELETE /print-log/ clears only log entries
- Thumbnail serving endpoint for log entries
- write_log_entry() service called on print completion
- Auth: ARCHIVES_READ for viewing, ARCHIVES_DELETE_ALL for clearing
Frontend:
- New 'log' ViewMode with ClipboardList icon toggle
- Filterable/searchable table with pagination (10/25/50/100 rows)
- Colored status badges, duration formatting, filament color swatches
- Clear button with confirmation modal
- All filter state persisted to localStorage
- i18n: EN, DE, JA, FR, IT translations
Suppress SQL/aiosqlite debug noise (~90% log volume reduction), add
caller-traced PRINT COMMAND logging to start_print(), log scheduler
queue checks, tighten stale archive ilike match to exact match, and
warn on multiple queue items in "printing" status. Includes 18 new
unit tests.
Notify users when the print bed cools below a configurable threshold
(default 35°C) after a print finishes, so they know when to remove parts.
- Backend: DB migration, model, schemas, notification template, service
method, background cooldown monitor (polls every 15s, 30min timeout)
- Frontend: event toggle in provider card/modal, threshold setting in
Settings > Notifications, i18n keys for all 5 locales
- Tests: 4 backend + 4 frontend tests
- Docs: README, website, wiki updated
Downgrade 58 diagnostic logger.info calls to logger.debug in
bambu_mqtt.py — payload dumps, detector state changes, field
discovery, H2D disambiguation, and periodic status updates no longer
flood logs at the default INFO level. User-initiated actions (print,
stop, calibration, AMS load/unload) remain at INFO. Also suppress
paho-mqtt library INFO messages in production mode.
Backend:
- Add dual external spool support for H2D (vt_tray as list: Ext-L/Ext-R)
- Add cloud filament ID map endpoint (/cloud/filament-id-map)
- Fix RFID spool data erased by periodic AMS updates (skip tag matcher
for RFID-tagged trays)
- Fix AMS slot config overwrites RFID spool state
- Fix K-profile selection corrupts existing profiles on X1C/P1S
- Resolve K-profiles filament name via cloud filament ID map
- Update print scheduler and usage tracker for dual external spools
Frontend:
- Add printer model filtering to ConfigureAmsSlotModal (cloud/local/builtin
presets filtered by @BBL model suffix and compatible_printers)
- Add pre-population for configured slots (preset, color, K-profile)
- Add K-Profiles view with accurate filament name resolution
- Internationalize all ConfigureAmsSlotModal strings (en/de/fr/it/ja — 21 keys)
- Add 5 new ConfigureAmsSlotModal tests (model filtering, pre-selection,
color pre-population, i18n)
- Update PrintersPage for dual external spool rendering
Docs:
- Update CHANGELOG, README, website features, and wiki AMS docs
filament_details now shows which AMS unit and tray each filament came
from (e.g. "AMS-A T1 PLA: 12.4g | AMS-A T3 PETG: 2.8g"). External
spool holders show "Ext". Falls back to type-only format when usage
tracking data is unavailable (e.g. Spoolman mode).
- usage_tracker: include spool material in result dicts
- main.py: pass usage_results to notification flow via archive_data
- notification_service: build filament_details with AMS labels when
usage_results available, fall back to existing slot-based format
- Update sample data and changelog
3MF slicer estimates are now the primary tracking source for ALL spools
(BL and non-BL), with AMS remain% delta as fallback. Per-layer G-code
analysis provides accurate partial usage for failed/cancelled prints.
Fix wrong-spool tracking bug: 3MF slot_id was mapped directly to AMS
tray position, but the printer remaps slicer slots at print time. Now
uses queue ams_mapping for queue prints and tray_now from printer state
for single-filament non-queue prints.
Add filament_grams, filament_details, and progress template variables
to print_complete, print_failed, and print_stopped notification events.
Webhook payloads include per-slot filament breakdown. Per-slot data
stored in archive extra_data during 3MF parsing.
Status summary bar on the Printers page now shows availability count
("X available") and a "Next available" indicator with printer name,
progress bar, percentage, and remaining time for the printer finishing
soonest. Always visible when printers are printing; hidden otherwise.
Fix race condition in bulk archive deletion where files were removed
from disk before the database commit. Concurrent SQLite writes could
cause a lock timeout, rolling back the DB delete while files were
already gone — leaving orphaned records with broken thumbnails. Now
deletes the DB record first and only removes files after a successful
commit.
Dual-nozzle H2D/H2D Pro: filament matching now respects nozzle assignments
from the 3MF file. Each AMS unit feeds a specific nozzle (L/R), and the
scheduler/frontend constrain matching to only trays on the correct nozzle.
Falls back to unfiltered matching when no trays exist on the target nozzle.
L/R badges shown in the filament mapping UI. Translated in en/de/ja/it.
Fix AMS slot config overwritten on startup: on_ams_change unconditionally
unlinked BL spool assignments on every MQTT pushall, then re-assigned them
sending ams_filament_setting without setting_id — clearing the printer's
filament preset. Now compares spool RFID identifiers before unlinking.
Fix BL spool detection false positives: removed tray_info_idx from detection
logic in both backend is_bambu_lab_spool() and frontend isBambuLabSpool().
Third-party spools using Bambu generic presets had GF-prefixed tray_info_idx
values, causing misidentification. Now uses only tray_uuid and tag_uid.
SQLite WAL mode with 5s busy timeout reduces "database is locked" errors.
Existing code did not allow the MQTT relay to be shut down gracefully.
Moved _client.loop_stop() after _client.disconnect() call to allow
Paho to be notified to disconnect and shut down without reconnecting.
Added _client._disconnection_event to help prevent a race condition where
a disconnection is requested but the loop is immediately stopped.
Modify _client.disconnect() to take a timeout parameter (default: 0).
Wait in that routine for the disconnection event with the given timeout.
Apply same logic to MQTTSmartPlugService and add to shutdown sequence.
Shutdown the MQTT relay last in lifespan() with a hard-coded 2 sec timeout.
The disconnect timeout is also available to printer MQTT clients.
It is left to the default of 0 at this time. If later enabled,
it should be a configuration parameters, especially for large farms.
Fixes: #327
Signed-off-by: Jeff Kletsky <git-commits@allycomm.com>
Attach printer camera snapshots to Discord webhook notifications using
multipart form-data with embed image references. Images are included
for print start, complete, progress milestones, and error events.
When no image is available, notifications send as text-only as before.
When the same file was queued multiple times from the archive, all jobs
after the first were automatically skipped with "already printed X hours
ago". The scheduler's 4-hour safety check compared the shared archive's
completed_at timestamp, incorrectly treating legitimate repeat prints as
phantom reprints from crash recovery. This also affected single queue
items created from recently completed archives.
Remove the check entirely — the crash scenario it guarded against is
already prevented by committing item.status="printing" before sending
the print command.
The print scheduler previously treated FINISH/FAILED as idle states
and would auto-start the next queued print before the user cleared
the build plate. Now requires explicit user confirmation via a
"Clear Plate & Start Next" button on the printer card. Uses an
in-memory plate-cleared flag — no MQTT command needed since the
scheduler's start_print overrides the printer state.
Underscores in job names or error codes caused parse_mode to be
disabled entirely, rendering *bold* markers as literal asterisks.
Now escapes underscores in the body with \_ so Markdown stays enabled.
The snapshot endpoint always used the internal printer camera even when
an external camera was configured. Now checks for external camera first,
matching the stream endpoint pattern. Also added retry logic (3 attempts,
2s delay) to MJPEG and RTSP stream generators so they reconnect on
timeout instead of silently ending the stream.
There were several instances where print() debugging was used instead of
or in addition to formal logging. These messages were tagged `[EXT-CAM]`.
Where there was already a logging call, the print statement was removed.
Where there was not, logger.debug() was used as a replacement.
Signed-off-by: Jeff Kletsky <git-commits@allycomm.com>
Three fixes for the H2C nozzle rack reported by testers:
1. Empty slot position: rack nozzles were mapped by sequential array
index instead of by actual nozzle ID (16-21), so a missing slot
always appeared at position 6 instead of its real position. Now
maps each nozzle to its correct rack position via id offset.
2. Missing filament colors/materials: the H2C MQTT payload uses
different field names than H2D — color_m (not filament_colour),
fila_id (not filament_id), sn (not serial_number), tm (not
max_temp). Added fallback field name resolution in the backend
MQTT parser.
3. Medium card layout: the temperature row with nozzle rack overflowed
on medium-sized cards. Added flex-wrap and a 190px flex-basis on
the rack card so it wraps to a second line when space is tight.
AMS-HT global tray ID was calculated as ams_id * 4 (= 512 for unit 128)
but AMS-HT uses the raw ams_id directly since it has a single tray.
The backend misidentified 512 as an external spool, producing wrong
ams_mapping2. Fixed in 4 locations: frontend getGlobalTrayId(), backend
start_print() ams_mapping2 builder, print scheduler, and Spoolman tracking.
- Show all 6 nozzles including mounted: backend includes all nozzle_info
entries (removes id >= 2 filter), frontend filters to non-empty
- Translate nozzle type codes to full names: HS→Hardened Steel, 00/01/05
mapped to Stainless Steel/Hardened Steel/Tungsten Carbide
- Add flow type to hover card: HH→High Flow, HS→Standard
- Show filament material type in hover card (PLA, PETG etc.) from MQTT
tray_type/filament_type field
- Add filament_type to NozzleRackSlot schema (backend + frontend)
- Add translations (en, de, ja, it): nozzleTungstenCarbide, nozzleFlow,
nozzleHighFlow, nozzleStandardFlow
The remote_interface_ip setting only worked in proxy mode but was
completely ignored in server modes (immediate/review/print_queue).
Users with multiple NICs (LAN + Tailscale, Docker bridges) got wrong
auto-detected IP in SSDP broadcasts and TLS certificates.
H2C was mapped to the "h2d" firmware API key, causing it to check
the H2D firmware page and offer H2D firmware for download. H2C has
its own separate firmware track (01.01.x.x vs H2D's 01.02.x.x).
Changed H2C to use its own "h2c" API key and added the reverse
mapping. Also added missing H2C (O1C) and H2S (O1S) entries to
the printer model ID and 3MF model name maps.
The nozzle_info MQTT array contains L/R nozzle heads (IDs 0, 1) and
rack slots (IDs 16-21). Backend was dumping all entries into nozzle_rack,
and frontend slice(0,6) grabbed L, R, plus only 4 rack slots — cutting
off the last position (e.g. the 0.6mm nozzle) and showing the mounted
nozzle as docked.
Backend: filter nozzle_rack to id >= 2 (rack-only), sort by ID.
Frontend: compact single-row layout with bottom accent bars for
mounted/docked status, wider rack card (flex-[2]), vertically
centered temp cards.
maziggy