Changelog

All notable changes to Seismo Fleet Manager will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

0.2.1 - 2025-12-03

/settings roster manager page with CSV export/import, live stats, and danger-zone reset controls.
/api/settings router that exposes export-csv, stats, roster-units, import-csv-replace, and the clear-* endpoints backing the UI.
Dedicated HTMX partials/tabs for deployed, benched, retired, and ignored units plus new ignored-table UI to unignore or delete entries.

Roster and unit detail templates now display device-type specific metadata (calibration windows, modem pairings, IP/phone fields) alongside inline actions.
Base navigation highlights the new settings workflow and routes retired/ignored buckets through dedicated endpoints + partials.

Snapshot summary counts only consider deployed units, preventing dashboard alerts from including benched hardware.
Snapshot payloads now include address/coordinate metadata so map widgets and CSV exports stay accurate.

Device-type aware roster schema (seismographs vs modems) with new metadata columns plus backend/migrate_add_device_types.py for upgrading existing SQLite files.
create_test_db.py helper that generates a ready-to-use demo database with sample seismographs, modems, and emitter rows.
Ignore list persistence/API so noisy legacy emitters can be quarantined via /api/roster/ignore and surfaced in the UI.
Roster page enhancements: Add Unit modal, CSV import modal, and HTMX-powered table fed by /partials/roster-table.
Unit detail view rewritten to fetch data via API, expose deployment status, and allow edits to all metadata.

Snapshot service now merges roster + emitter data into active/benched/retired/unknown buckets and includes device-specific metadata in each record.
Roster edit endpoints parse date fields, manage modem/seismograph specific attributes, and guarantee records exist when toggling deployed/retired states.
Dashboard partial endpoints are grouped under /dashboard/* so HTMX tabs stay in sync with the consolidated snapshot payload.

Toggling deployed/retired flags no longer fails when a unit does not exist because the router now auto-creates placeholder roster rows.
CSV import applies address/coordinate updates instead of silently dropping unknown columns.

Roster Editing API: Full CRUD operations for roster management
- POST /api/roster/add - Add new units to roster
- POST /api/roster/set-deployed/{unit_id} - Toggle deployment status
- POST /api/roster/set-retired/{unit_id} - Toggle retired status
- POST /api/roster/set-note/{unit_id} - Update unit notes
CSV Import: Bulk roster import functionality
- POST /api/roster/import-csv - Import units from CSV file
- Support for all roster fields: unit_id, unit_type, deployed, retired, note, project_id, location
- Optional update_existing parameter to control duplicate handling
- Detailed import summary with added/updated/skipped/error counts
Enhanced Database Models:
- Added project_id field to RosterUnit model
- Added location field to RosterUnit model
- Added last_updated timestamp tracking
Dashboard Enhancements:
- Separate views for Active, Benched, and Retired units
- New endpoints: /dashboard/active and /dashboard/benched

Database session management bug in emit_status_snapshot()
- Added get_db_session() helper function for direct session access
- Implemented proper session cleanup with try/finally blocks
Database schema synchronization issues
- Database now properly recreates when model changes are detected