asterisk-voicemail

jp/asterisk-voicemail

Fork 0

Commit Graph

Author	SHA1	Message	Date
jp	21dc883e99	Fix contacts matcher: full-number (15-digit) matching, junk-name guard - Index by full E.164 digits instead of last-9 only; previously different people sharing a 9-digit suffix collided and lookups returned the wrong name. Resolve now tries longest match first. - match_digits default 9 -> 15 in code and example config. - Skip emoji/punctuation-only contact names so a junk card can't shadow a real one when a number is duplicated in the export. - Verified: 513 contacts self-consistent; 6 remaining collisions are genuine duplicate entries in the source VCF (two cards, same number), not a matcher error. Also: placed /home/jp/contacts.vcf at /var/lib/vm-transcribe/contacts.vcf (640 asterisk:asterisk), set backends=file, cleared stale cache, restarted vm-portal.	2026-08-13 11:13:12 +01:00
jp	857284abbf	Initial build: Asterisk voicemail transcription + portal - mailcmd replacement (vm_mailcmd.py): faster-whisper transcription (CPU int8), extractive summary + intent tags + spoken-digit number extraction, multipart/alternative HTML email, fail-safe relay of original message - Telegram DM delivery (vm_telegram.py) with per-mailbox routing - Caller-ID -> name (vm_contacts.py): file / google / carddav backends - SQLite store (vm_store.py) with content-addressed audio - FastAPI portal (vm_web.py): PIN login, list/play/delete, per-user settings, zero JS, loopback-only behind Apache TLS - Backfill importer (vm_import.py) for existing spool recordings - systemd unit, Apache vhost + certbot TLS, install.sh - Docs: INSTALL, CONFIGURATION, ARCHITECTURE, OPERATIONS, SECURITY, TESTING Verified end-to-end on mail.txt3.net: 157 historical messages backfilled, live voicemail -> transcribed -> stored -> visible at https://vm.txt3.net.	2026-08-13 09:31:41 +01:00

Author

SHA1

Message

Date

21dc883e99

Fix contacts matcher: full-number (15-digit) matching, junk-name guard

- Index by full E.164 digits instead of last-9 only; previously different
  people sharing a 9-digit suffix collided and lookups returned the wrong
  name. Resolve now tries longest match first.
- match_digits default 9 -> 15 in code and example config.
- Skip emoji/punctuation-only contact names so a junk card can't shadow a
  real one when a number is duplicated in the export.
- Verified: 513 contacts self-consistent; 6 remaining collisions are genuine
  duplicate entries in the source VCF (two cards, same number), not a matcher
  error.

Also: placed /home/jp/contacts.vcf at /var/lib/vm-transcribe/contacts.vcf
(640 asterisk:asterisk), set backends=file, cleared stale cache, restarted
vm-portal.

2026-08-13 11:13:12 +01:00

857284abbf

Initial build: Asterisk voicemail transcription + portal

- mailcmd replacement (vm_mailcmd.py): faster-whisper transcription (CPU int8),
  extractive summary + intent tags + spoken-digit number extraction,
  multipart/alternative HTML email, fail-safe relay of original message
- Telegram DM delivery (vm_telegram.py) with per-mailbox routing
- Caller-ID -> name (vm_contacts.py): file / google / carddav backends
- SQLite store (vm_store.py) with content-addressed audio
- FastAPI portal (vm_web.py): PIN login, list/play/delete, per-user settings,
  zero JS, loopback-only behind Apache TLS
- Backfill importer (vm_import.py) for existing spool recordings
- systemd unit, Apache vhost + certbot TLS, install.sh
- Docs: INSTALL, CONFIGURATION, ARCHITECTURE, OPERATIONS, SECURITY, TESTING

Verified end-to-end on mail.txt3.net: 157 historical messages backfilled,
live voicemail -> transcribed -> stored -> visible at https://vm.txt3.net.

2026-08-13 09:31:41 +01:00

2 Commits