qiadZddlZddlZddlZddlmZddlmZmZm Z ddl Z ddl m Z ddl mZddlmZmZmZmZmZmZmZddlmZdd lmZmZdd lmZmZmZdd l m!Z!m"Z"dd l#m$Z$ ddl%Z%dd l&m'Z(ddl)m*Z*de!de e/fdZ0eGddeZ1GddeZ2GddeZ3y#e+$r7Z,e jZde,e jZde.de,dZ,[,wwxYw)z-Async text-to-speech service implementations.N) dataclass)AnyAsyncGeneratorOptional)logger) BaseModel) CancelFrameEndFrame ErrorFrameFrame StartFrameTTSAudioRawFrameTTSStoppedFrame)FrameDirection) TTSSettings_warn_deprecated_param)TextAggregationMode TTSServiceWebsocketTTSService)Languageresolve_language) traced_tts)connect)Statez Exception: zEIn order to use Async, you need to `pip install pipecat-ai[asyncai]`.zMissing module: languagereturncBitjdtjdtjdtjdtj dtj dtjdtjdtjd tjd tjd tjd tjd tjdtjdtj d}t#||dS)zConvert a Language enum to Async language code. Args: language: The Language enum value to convert. Returns: The corresponding Async language code, or None if not supported. enfresdeitptnlarrurojahehytrhizhT) use_base_code)rENFRESDEITPTNLARRUROJAHEHYTRHIZHr)r LANGUAGE_MAPs N/opt/pipecat/venv/lib/python3.12/site-packages/pipecat/services/asyncai/tts.pylanguage_to_async_languagerA,s T T  T  T   T   T   T  T  T  T  T  T  T  T  T  T!L& Hl$ GGceZdZdZy)AsyncAITTSSettingsz9Settings for AsyncAITTSService and AsyncAIHttpTTSService.N)__name__ __module__ __qualname____doc__rBr@rDrDKsCrBrDceZdZUdZeZeed<GddeZddddddd ddddd d e d e e d e de de e de e de de de ede ede e de e ffdZdedee efffd Zde fdZdede e fdZd4de de d e de fd!Zd"effd# Zd"effd$ Zd"effd% Zfd&Zfd'Zd(Zd)Z d*Z!d5de e fd+Z"e#jHfd"e%d,e#ffd- Z&d.Z'd/Z(de fd0Z)de fd1Z*de fd2Z+e,de de de-e%dffd3Z.xZ/S)6AsyncAITTSServiceztAsync TTS service with WebSocket streaming. Provides text-to-speech using Async's streaming WebSocket API. _settingsc&eZdZUdZdZeeed<y)AsyncAITTSService.InputParamszInput parameters for Async TTS configuration. .. deprecated:: 0.0.105 Use ``AsyncAITTSSettings`` directly via the ``settings`` parameter instead. Parameters: language: Language to use for synthesis. NrrErFrGrHrrr__annotations__rIrBr@ InputParamsrN[ (,(8$+rBrQNv1z/wss://api.async.com/text_to_speech/websocket/ws pcm_s16leraw) voice_idversionurlmodel sample_rateencoding containerparamssettingsaggregate_sentencestext_aggregation_modeapi_keyrVrWrXrYrZr[r\r]r^r_r`c tddd}|tdtd||_|tdtd||_| @tdt| s.| jr|j | jnd|_| |j | t| d | | d|dd|d | ||_ ||_ ||_ ||_ ||_ d |_d|_d|_y) aInitialize the Async TTS service. Args: api_key: Async API key. voice_id: UUID of the voice to use for synthesis. See docs for a full list: https://docs.async.com/list-voices-16699698e0 .. deprecated:: 0.0.105 Use ``settings=AsyncAITTSSettings(voice=...)`` instead. version: Async API version. url: WebSocket URL for Async TTS API. model: TTS model to use (e.g., "async_flash_v1.0"). .. deprecated:: 0.0.105 Use ``settings=AsyncAITTSSettings(model=...)`` instead. sample_rate: Audio sample rate. encoding: Audio encoding format. container: Audio container format. params: Additional input parameters for voice customization. .. deprecated:: 0.0.105 Use ``settings=AsyncAITTSSettings(...)`` instead. settings: Runtime-updatable settings. When provided alongside deprecated parameters, ``settings`` values take precedence. aggregate_sentences: Deprecated. Use text_aggregation_mode instead. .. deprecated:: 0.0.104 Use ``text_aggregation_mode`` instead. text_aggregation_mode: How to aggregate text before synthesis. **kwargs: Additional arguments passed to the parent service. async_flash_v1.0NrYvoicerrVrerYr]T)r_r`pause_frame_processingrZpush_start_framepush_stop_framesr^rrI)rDrrerYrlanguage_to_service_language apply_updatesuper__init___api_key _api_version_url_output_container_output_encoding_output_sample_rate _receive_task_keepalive_task)selfrarVrWrXrYrZr[r\r]r^r_r`kwargsdefault_settings __class__s r@rlzAsyncAITTSService.__init__gsj.$    ":/A7 K%-  "   "7,> H%*  "   "8-? @JP//D55fooF_c!)    ) )( 3  3"7#'#!!%    # "+ (#$ !#rBdeltarclKt||d{}|s|S|j||S7w)zgApply a settings delta. Settings are stored but not applied to the active connection. N)rk_update_settings _warn_unhandled_updated_settings)rurychangedrxs r@r{z"AsyncAITTSService._update_settingss; 077N --g68s 424cy)zCheck if this service can generate processing metrics. Returns: True, as Async service supports metrics generation. TrIrus r@can_generate_metricsz&AsyncAITTSService.can_generate_metrics rBrct|SzConvert a Language enum to Async language format. Args: language: The language to convert. Returns: The Async-specific language code, or None if not supported. rArurs r@riz.AsyncAITTSService.language_to_service_language*(33rBtext context_idforcec8|||d}tj|S)N) transcriptrr)jsondumps)rurrrmsgs r@ _build_msgzAsyncAITTSService._build_msgs!eLzz#rBframecKt||d{|j|_|j d{y7.7w)z}Start the Async TTS service. Args: frame: The start frame containing initialization parameters. N)rkstartrZrr_connectrurrxs r@rzAsyncAITTSService.startsA gmE"""#'#3#3 mmo #s A A(A AA A ctKt||d{|jd{y77w)zUStop the Async TTS service. Args: frame: The end frame. N)rkstop _disconnectrs r@rzAsyncAITTSService.stops6 gl5!!!    " 848688ctKt||d{|jd{y77w)zZCancel the Async TTS service. Args: frame: The cancel frame. N)rkcancelrrs r@rzAsyncAITTSService.cancels6 gnU###    $ rc|Kt|d{|jd{|jr;|js/|j |j |j|_|jr2|js%|j |j|_yyy77wN) rkr_connect_websocket _websocketrs create_task_receive_task_handler _report_errorrt_keepalive_task_handlerrurxs r@rzAsyncAITTSService._connectsg   %%''' ??4#5#5!%!1!1$2L2LTM_M_2`!aD  ??4#7#7#'#3#3D4P4P4R#SD $8? !'s B<B8B<B:B B<:B<cRKt|d{|jr*|j|jd{d|_|jr*|j|jd{d|_|j d{y77\7(7 wr)rkrrs cancel_taskrt_disconnect_websocketrs r@rzAsyncAITTSService._disconnectsg!###   ""4#5#56 6 6!%D    ""4#7#78 8 8#'D ((*** $ 7 9 +sEB'B.B'B!5B':B#;B'B%B'!B'#B'%B'cLK |jr'|jjtjuryt j dt |jd|jd|jd{|_|jjd|jjd|j|j|jd|jj d}|j#j%t'j(|d{|j+dd{y77 7 #t,$rL}|j/d || d{7d|_|j+d |d{7Yd}~yd}~wwxYww) NzConnecting to Asyncz ?api_key=z &version=idmoderr\r[rZ)model_idre output_formatr on_connectedUnknown error occurred:  error_msg exceptionon_connection_error)rstaterOPENrdebugwebsocket_connectrormrnrLrYrerprqrrr_get_websocketsendrr_call_event_handler Exception push_error)ruinit_msges r@rz$AsyncAITTSService._connect_websocket&s_ J4??#8#8EJJ#F LL. /$599+Yt}}oYt?P?P>QR%DO!NN00"&dnn.B.BC!%!7!7 $ 5 5#'#;#;" !NN33 H%%',,TZZ-AB B B**>: : :! C : J//.Fqc,JVW/X X X"DO**+@QCI I I JsF$2E F$A E EB&E 'E(E E E F$E E E F!F.E1/"FFFF$F!!F$cK |jd{|jrtjd|j r8|jj t jddid{|jjd{tjdd|_|jd{|jdd{y77v7V#t$r)}|jd||d{7Yd}~ld}~wwxYw7Y7B#d|_|jd{7|jdd{7wxYww)NzDisconnecting from Async terminateTzDisconnected from Asyncrron_disconnected) stop_all_metricsrrrhas_active_audio_contextrrrcloserrremove_active_audio_contextr)rurs r@rz'AsyncAITTSService._disconnect_websocketAs1 >'') ) ) 78002//..tzz;:M/NOOOoo++--- 67#DO224 4 4**+<= = = * P- Y//.Fqc,JVW/X X X Y 5 =#DO224 4 4**+<= = =sE3C;C5A(C;C7!C;"C9#C;<E3D0E3/D20E35C;7C;9C;; D-D(D D(#D4(D--D40E32E34E0EE0)E,*E00E3cH|jr |jStd)NzWebsocket not connected)rrrs r@rz AsyncAITTSService._get_websocketSs ???? "122rBcK|xs|j}|r |jsytj|d|j d|d}|jj |d{y7w)zFlush any pending audio. Args: context_id: The specific context to flush. If None, falls back to the currently active context. Nz: flushing audio T)rrr)get_active_audio_context_idrrtracerr)rurflush_idrs r@ flush_audiozAsyncAITTSService.flush_audioXsgC!A!A!Ct  v-./oo384oHoo""3'''sA.A80A61A8 directioncBKt|||d{y7w)zPush a frame downstream with special handling for stop conditions. Args: frame: The frame to push. direction: The direction to push the frame. N)rk push_frame)rurrrxs r@rzAsyncAITTSService.push_framefsg  222s cK|j23d{}tj|}|s!|jd}|jddurt j d|^|j |s^|j|k(r2t jd||j|d{nt jd||jds|jd{tj|d}t||jd| }|j||d{D7@77V7 6yw) NrfinalTz#Received final message for context z4Received a delayed message, recreating the context: z+Ignoring message from unavailable context: audior)rrloadsgetrraudio_context_availablerrcreate_audio_contextstop_ttfb_metricsbase64 b64decoderrZappend_to_audio_context)rumessagerreceived_ctx_idrrs r@_receive_messagesz#AsyncAITTSService._receive_messagesosT!002 K K'**W%C!ggl3Owww4' B?BSTU//@335HLLNN_`33ODDD LL#NN_!`awww,,...((W6(0@0@!P_`22?EJJJ? K&E/K?3scE"E EE B&E"E/E"2E"EA E"EE"E E"E"E" E"c&Kd} tj|d{ |jr|jjtj ur||j }|rd|d}tjdnddi}tjd|jjtj|d{77#tj$r$}tj|d|Yd}~yd}~wwxYww) zBSend periodic keepalive messages to maintain WebSocket connection. Nr)rrzSending keepalive messagerz!Sending keepalive without contextz keepalive error: )asynciosleeprrrrrrrrrr websocketsConnectionClosedwarning)ruKEEPALIVE_SLEEPrkeepalive_messagers r@rz)AsyncAITTSService._keepalive_task_handlers--0 0 0 ??t'<'< 'J!%!A!A!CJ!*-*4-) %@A .:3,?) %HI//..tzz:K/LMMM# 0 N.. $'9!=> sKDCDB)C CCDCD*D D DDc K|rI|jr< |jjtj|dddd{yyy7#t$r'}t j |d|d|Yd}~yd}~wwxYww)NT)r close_contextrz: Error closing context : )rrrrrrerror)rurrs r@_close_contextz AsyncAITTSService._close_contexts $// Qoo**JJj4_abc*: Q v%=j\A3OPP Qs@B5AAA BA BA<7B<BBcBK|j|d{y7w)z7Close the Async AI context when the bot is interrupted.Nrrurs r@on_audio_context_interruptedz.AsyncAITTSService.on_audio_context_interrupteds!!*--- cBK|j|d{y7w)aClose the Async AI context after all audio has been played. Async AI does not send a server-side signal when a context is exhausted, so Pipecat must explicitly close it with ``close_context: True`` to free server-side resources. Nrrs r@on_audio_context_completedz,AsyncAITTSService.on_audio_context_completeds!!*---rcVKtj|d|d |jr&|jjtj ur|j d{ |j|d|}|jj|d{|j|d{dy7_7%7#t$r+}td|t|Yd}~yd}~wwxYw#t$r}td|Yd}~yd}~wwxYww) a!Generate speech from text using Async API websocket endpoint. Args: text: The text to synthesize into speech. context_id: The context ID for tracking audio frames. Yields: Frame: Audio frames containing the synthesized speech. : Generating TTS []NT)rrrr)rr)rrrrrCLOSEDrrrrstart_tts_usage_metricsrr r)rurrrrs r@run_ttszAsyncAITTSService.run_ttss   v/vQ78 C??doo&;&;u||&Kmmo%% oo4t oS))+00555224888 J&68  )A!'EFF%<<   C%=aS#AB B B CsD)AD$C%D*6C C!C 9C:C >DD)DC C C>!C94D8D)9C>>D D& D!D)!D&&D))rrFr)0rErFrGrHrDSettingsrPrrQstrrintboolrrlrdictrr{rrrirr rr rr rrrrrrrr DOWNSTREAMr rrrrrrrrr __classcell__rxs@r@rKrKRs) "H!! ,i , #'D#%)#(,15.2?Cd$d$3- d$  d$  d$}d$c]d$d$d$%d$-.d$&d^d$ ((;<d$L K DcN d 4X 4(3- 4sSdWZ!!!+! T +J6>$3 (HSM (JXIbIb3e33 KD2 Qs Q.S..3.C#C3C>%QU+;VCCrBrKceZdZUdZeZeed<GddeZddddddd ddd d e d e e d e jde e de de de e de de de ede effdZdefdZdede e fdZdeffd Zede de deedffdZxZS) AsyncAIHttpTTSServicezHTTP-based Async TTS service. Provides text-to-speech using Async's HTTP streaming API for simpler, non-WebSocket integration. Suitable for use cases where streaming WebSocket connection is not required or desired. rLc&eZdZUdZdZeeed<y)!AsyncAIHttpTTSService.InputParamszInput parameters for Async API. .. deprecated:: 0.0.105 Use ``AsyncAITTSSettings`` directly via the ``settings`` parameter instead. Parameters: language: Language to use for synthesis. NrrOrIrBr@rQrrRrBrQNzhttps://api.async.comrSrTrU) rVrYrXrWrZr[r\r]r^rarVaiohttp_sessionrYrXrWrZr[r\r]r^c  tddd} |tdtd|| _|tdtd|| _| @tdt| s.| jr|j | jnd| _| | j | t| d |dd| d | ||_ ||_ ||_ | |_ ||_ d |_||_y) aInitialize the Async TTS service. Args: api_key: Async API key. voice_id: ID of the voice to use for synthesis. .. deprecated:: 0.0.105 Use ``settings=AsyncAITTSSettings(voice=...)`` instead. aiohttp_session: An aiohttp session for making HTTP requests. model: TTS model to use (e.g., "async_flash_v1.0"). .. deprecated:: 0.0.105 Use ``settings=AsyncAITTSSettings(model=...)`` instead. url: Base URL for Async API. version: API version string for Async API. sample_rate: Audio sample rate. encoding: Audio encoding format. container: Audio container format. params: Additional input parameters for voice customization. .. deprecated:: 0.0.105 Use ``settings=AsyncAITTSSettings(...)`` instead. settings: Runtime-updatable settings. When provided alongside deprecated parameters, ``settings`` values take precedence. **kwargs: Additional arguments passed to the parent TTSService. rcNrdrVrerYr]T)rZrgrhr^rrI)rDrrerYrrirjrkrlrm _base_urlrnrprqrr_session)rurarVrrYrXrWrZr[r\r]r^rvrwrxs r@rlzAsyncAIHttpTTSService.__init__s\.$    ":/A7 K%-  "   "7,> H%*  "   "8-? @JP//D55fooF_c!)    ) )( 3  #!!%      #"+ (#$ ' rBrcy)zCheck if this service can generate processing metrics. Returns: True, as Async HTTP service supports metrics generation. TrIrs r@rz*AsyncAIHttpTTSService.can_generate_metricsSrrBrct|Srrrs r@riz2AsyncAIHttpTTSService.language_to_service_language[rrBrcbKt||d{|j|_y7w)zStart the Async HTTP TTS service. Args: frame: The start frame containing initialization parameters. N)rkrrZrrrs r@rzAsyncAIHttpTTSService.startfs. gmE"""#'#3#3  #s /-/rrcKtj|d|d d|jjd}|jj|||j |j |jd|jjd}|j|jdd}|jd }|jj||| 4d {}|jd k7rP|jd {}|j!d |d {t#d|jd|t%} |j&j)d23d {} | s |j+d {| j-| 6777~7776t/| } d d d d {7n#1d {7swYnxYw|j1|d {7t3 |j4d|} | n5#t"$r)} |j!d| | d {7Yd } ~ nd } ~ wwxYw|j+d {7y #|j+d {7wxYww)aGenerate speech from text using Async's HTTP streaming API. Args: text: The text to synthesize into speech. context_id: The context ID for tracking audio frames. Yields: Frame: Audio frames containing the synthesized speech. rrrrr)rrrerrzapplication/json)rWz x-api-keyz Content-Typez/text_to_speech/streaming)rheadersNzAsync API error: )rzAsync API returned status rir)rrZ num_channelsrrr)rrrLrerYrprqrrrrnrmr r poststatusrrr bytearraycontent iter_chunkedrextendbytesrrrZ)rurr voice_configpayloadrrXresponse error_textbufferchunk audio_datarrs r@rzAsyncAIHttpTTSService.run_ttsosU  v/vQ783 +$(0D0DEL!NN00"%!%!7!7 $ 5 5#'#;#;" !NN33 G ,,!]] 2G ^^$$=>C}}))#GW)M + +QY??c)'/}}!6J//6G |4T/UUU#&@@QQST^S_$`aa##+#3#3#@#@#K))%  00222MM%( +!6U )3$L #6]  + + + + +..t4 4 4$  ,,% EK Y//.Fqc,JVW/X X X Y((* * *$((* * *sI7B6H FH #F<<F=F<FAF<F"F#F&F<=F>F<H F<F<FF< F<* H 5F86H <GG G H %G(&#H I H<H7,H/-H72I7H<<I?I7II7I4-I0.I44I7)rErFrGrHrDrrPrrQrraiohttp ClientSessionrrlrrrrir rrrr rrrs@r@rrs="H!! ,i , #'#*%)#(,15Y(Y(3- Y( !.. Y( } Y(Y(Y(c]Y(Y(Y(%Y(-.Y(vd 4X 4(3- 444?+#?+3?+>%QU+;V?+?+rBr)4rHrrr dataclassesrtypingrrrr!logururpydanticrpipecat.frames.framesr r r r r rr"pipecat.processors.frame_processorrpipecat.services.settingsrrpipecat.services.tts_servicerrrpipecat.transcriptions.languagerr(pipecat.utils.tracing.service_decoratorsrrwebsockets.asyncio.clientrrwebsockets.protocolrModuleNotFoundErrorrrrrrArDrKrrIrBr@r0s4 !00>I]]F?,F)HHhsmH>     LC+LC^ N+JN+w ,FLL;qc"#FLLXY &qc* ++,s(B**C&/2C!!C&