* gather/transcribe: support for azure custom speech models (endpoint id)
* allow azure stt custom speech endpoint id to be passed as property in recognizer
* fix to add custom stt endpoint to session speech credentials object
* typo for media bug name in azure and punctuation fix
* say: split very long text intelligently
* more fixes from testing
* update to latest synthAudio
* feat: update time-series 0.11.12
* feat: support play verb url in plain text or array
* fix: review comment
Co-authored-by: Dave Horton <daveh@beachdognet.com>
* fixes from testing
* modify Task#exec to take resources as an object rather than argument list
* pass 2 endpoints to Transcribe when invoked in a SipRec call session
* logging
* change siprec invite to sendrecv just so freeswitch does not try to reinvite (TODO: block outgoing media at rtpengine)
* Config: when enabling recording, block until siprec dialog is established
* missed play verb in commit 031c79d
* linting
* bugfix: get final transcript in siprec call
* initial changes for amd
* wip
* fix bug where transcripts were discarded
* a bit of refactoring, and adding support for avmd in config verb
* bug fixes
* initial changes to use undici for http client and connection pooling
* use body.json() mixin
* logging
* add pipelining env var
* implement socket close
* initial changes to support siprec recording
* include additional params on SIP INFO to start recording
* add support for maniupulating recording via REST API
* fixes from testing pause/resume recording
* bugfix: background gather for speech-only should still kill audio on dtmf entry when dtmfBargein is true
* initial changes for continuous asr
* move properties under recognizer
* update drachtio-srf@4.5.1
* catch exception on destroy
* issue# 107: on gather timeout, if minDigits are collected, resolve with dtmf-num-digits
* gather timeout: use conditional instead of if/else
Co-authored-by: Prashanth Gujjeti <prashanth@minervacq.com>
