* deepgram gather: if both endpointing and utterance_end_ms are set (but not continous asr) return either when we get speech_final or UtteranceEnd. This is the belt-and-suspenders apprach deepgram is recommending
* include verb id in action hook if one was provided in the verb set
* minor
* update to fsmrf with fix
* changes to support elevenlabs tts streaming
* say: add vendor data to span
* bug: tts spans must include cached property
* add env for JAMBONES_USE_FREESWITCH_TIMER_FD
* fix bug in prev commit
* wip
* linting
* wip - caching files generating by streaming tts
* wip caching
* cleanup some logs
* handle tts streaming failure, write alert
* update node version dependency
* set timerfd on outbound call scenarios
* default model to nova-2-phonecall when using deepgram
---------
Co-authored-by: Dave Horton <daveh@beachdognet.com>
* verbhook on ws connection should be ended in next redirect command
* wip
* wip
* minor change for readability
---------
Co-authored-by: Dave Horton <daveh@beachdognet.com>
* fix to be more precise about removing custom event handlers so that when we stop a gather we dont also inadvertently stop a background transcribe as well
* test fixes
* fix: endpointing=false was being ignored for Deepgram
* Update say task and add possibility to use elevenlabs options from synthesizer
* revert ms change
* fix contdition for alerting
---------
Co-authored-by: Markus Frindt <m.frindt@cognigy.com>
* fix release freeswitch media properly
* if a leg is opus, modify b leg offer opus first
* if a leg is opus, modify b leg offer opus first
* wip
* wip
* fix review comments
* fix review comments
* fix review comments
* feat Audult call session should have its own requestor
* fix eslint
* fix eslint
* if user authenticate from http server instead of databse
* wip
* fix custom STT
* fix custom STT
* fix custom STT
* deepgram: rework continuous asr, and resolve on speech_final not is_final (wip)
* wip
* deepgram: empty final transcript should trigger resolve with speech if we have buffered transcripts
* wip
* fixes for deepgram compiling multiple transcripts
* test deepgram utteranceEndMs
* more handling of utteranceEndMs
* wip
* better handling of digit strings collected over multiple deepgram responses
* wip
* add support for deepgramOptions.shortUtterance which triggers off of is_final instead of speech_final
* apply deepgram fixes to transcribe
* cleanup continnuous asr
* more continuous asr fixes for deepgram
* update to verb-specifications for handling SttTask properties
* set log level for tests back to error
