iH  UdZddlZddlZddlZddlZddlZddlZddlZddlZddl Z ddl Z ddl Z ddl Z ddl Z ddlZddlZddlmZddlmZmZmZmZddlmZddlmZejeZddZddlm Z dd l!m"Z"m#Z#m$Z$dd l%m&Z&d Z'd Z(d Z)dZ*dZ+dZ,dZ-dZ.dZ/dZ0dZ1dZ2dZ3dZ4dZ5dZ6dZ7dZ8dZ9dZ:dZ;d Zd#Z?d$Z@d%ZAd&ZBd'ZCd(ZDd)ZEd$ZFd*ZGd+ZHd,eIfd-ZJeJZKd.d/d0d1d2d.d1d3d3d.d4 ZLeeIeMfeNd5<d.d.d1d1d1d1d6d7d8ZOeeIeMfeNd9<d2ZPePZQ dd:eeId;eeeIefd,eMfd<ZRd,eeIeffd=ZSd;eeIefd,eIfd>ZTeUhd?ZVd@ZWdAZXeUhdBZYd.ZZd;eeIefdCeId,eeIeffdDZ[d;eeIefdCeId,eeIeffdEZ\dFeeIefd,e]fdGZ^d:eId;eeIefd,eeeIeffdHZ_d;eeIeffdIZ`dFeeIefd,eafdJZb ddFeeIefdKeeId,eIfdLZcdFeeIefd,e]fdMZddNeIdOeMd,eeIfdPZedQeIdReeId,eIfdSZfdNeIdTeeIeIfd,eIfdUZgdVe jhd,dfdWZidXeIdYead,e jjfdZZkd[edFeeIefd,efd\Zld]eIdKeId^eIdFeeIefd;eeIefd,eIf d_Zmdd;eeeIefd,e]fd`Znd,e]fdaZodbeId,eeIfdcZpd]eIdKeId;eeIefd,eIfddZqd]eIdKeId;eeIefd,eIfdeZrd]eIdKeId;eeIefd,eIfdfZsd]eIdKeId;eeIefd,eIfdgZtd]eIdKeId;eeIefd,eIfdhZud]eIdKeId;eeIefd,eIfdiZveFeGeHfdjewdkeMdleMdmeMd,ewf dnZxd]eIdKeId;eeIefd,eIfdoZyd,e]fdpZzd,e]fdqZ{d,eIfdrZ|d,eIfdsZ}d]eIdKeId;eeIefd,eIfdtZ~iaeeIefeNdu<d,e]fdvZd,efdwZdxeIdyed,eIfdzZd]eIdKeId;eeIefd,eIfd{ZiaeeIefeNd|<d]eIdKeId;eeIefd,eIfd}Z dd]eIdKeeId,eIfd~Zd,e]fdZd,eeIeIffdZd,e]fdZejdZejdZejdZejdZejdZejdZejdZejdejZejdejZejdZejdZd]eId,eIfdZ ddejdejdejdeeeIgdffdZedkr-ededdZededee'drdndedee(drdndededrdndedee)drdndede$rdndededrdndederdndedeordndedeKeSZeTeZededdlmZmZddddddddedddd]gddZejddededdS)a Text-to-Speech Tool Module Built-in TTS providers: - Edge TTS (default, free, no API key): Microsoft Edge neural voices - ElevenLabs (premium): High-quality voices, needs ELEVENLABS_API_KEY - OpenAI TTS: Good quality, needs OPENAI_API_KEY - MiniMax TTS: High-quality with voice cloning, needs MINIMAX_API_KEY - Mistral (Voxtral TTS): Multilingual, native Opus, needs MISTRAL_API_KEY - Google Gemini TTS: Controllable, 30 prebuilt voices, needs GEMINI_API_KEY - xAI TTS: Grok voices, needs XAI_API_KEY - NeuTTS (local, free, no API key): On-device TTS via neutts - KittenTTS (local, free, no API key): On-device 25MB model - Piper (local, free, no API key): OHF-Voice/piper1-gpl neural VITS, 44 languages Custom command providers: - Users can declare any number of named providers with ``type: command`` under ``tts.providers.`` in ``~/.hermes/config.yaml``. Hermes writes the input text to a temp file and runs the configured shell command, which must produce the audio file at the expected path. See the Local Command section of ``website/docs/user-guide/features/tts.md``. Output formats: - Opus (.ogg) for Telegram voice bubbles (requires ffmpeg for Edge TTS) - MP3 (.mp3) for everything else (CLI, Discord, WhatsApp) Configuration is loaded from ~/.hermes/config.yaml under the 'tts:' key. The user chooses the provider and voice; the model just sends text. Usage: from tools.tts_tool import text_to_speech_tool, check_tts_requirements result = text_to_speech_tool(text="Hello world") N)Path)CallableDictAnyOptional)urljoin)display_hermes_homec~ ddlm}n%#t$rtj||cYSwxYw||}||n|S)aRead env values through the live config module. Tests may monkeypatch and later restore ``hermes_cli.config.get_env_value`` before this module is imported. Resolve the helper at call time so TTS does not keep a stale imported function for the rest of the test process. r) get_env_value)hermes_cli.configr ImportErrorosgetenv)namedefault_get_env_valuevalues 3/home/ubuntu/.hermes/hermes-agent/tools/tts_tool.pyr r ;sl(EEEEEEE (((yw'''''( N4 Em77.s  ++)resolve_managed_tool_gateway)managed_nous_tools_enabledprefers_gatewayresolve_openai_audio_api_key)hermes_xai_user_agentcddl}|S)z?Lazy import edge_tts. Returns the module or raises ImportError.rNedge_ttsrs r_import_edge_ttsrQsOOO Ocddlm}|S)zGLazy import ElevenLabs client. Returns the class or raises ImportError.r ElevenLabs)elevenlabs.clientr!r s r_import_elevenlabsr#Vs,,,,,, rcddlm}|S)zCLazy import OpenAI client. Returns the class or raises ImportError.r)OpenAI)openair%) OpenAIClients r_import_openai_clientr([s------ rcddlm}|S)zDLazy import Mistral client. Returns the class or raises ImportError.rMistral)mistralai.clientr+r*s r_import_mistral_clientr-`s(((((( Nrcddl}|S)zJLazy import sounddevice. Returns the module or raises ImportError/OSError.rN) sounddevice)sds r_import_sounddevicer1es Ircddlm}|S)z?Lazy import KittenTTS. Returns the class or raises ImportError.r KittenTTS) kittenttsr4r3s r_import_kittenttsr6ks###### rcddlm}|S)atLazy import Piper. Returns the PiperVoice class or raises ImportError. Piper is an optional, fully-local neural TTS engine (Home Assistant / Open Home Foundation). ``pip install piper-tts`` provides cross-platform wheels (Linux / macOS / Windows, x86_64 + ARM64) with embedded espeak-ng. Voice models (.onnx + .onnx.json) are downloaded on first use. r PiperVoice)piperr9r8s r _import_piperr;qs!     redgezen-US-AriaNeuralpNInz6obpgDQGcFmaJgBeleven_multilingual_v2eleven_flash_v2_5zgpt-4o-mini-ttsz!KittenML/kitten-tts-nano-0.8-int8Jasperzen_US-lessac-mediumalloyzhttps://api.openai.com/v1z speech-2.8-hdEnglish_Graceful_Ladyz https://api.minimax.io/v1/t2a_v2zvoxtral-mini-tts-2603z$c69964a6-ab8b-4f8a-9465-ec0925096ec8eveen]zhttps://api.x.ai/v1zgemini-2.5-flash-preview-ttsKorez0https://generativelanguage.googleapis.com/v1betareturnc@ddlm}t|ddS)Nrget_hermes_dirz cache/audio audio_cache)hermes_constantsrMstrrLs r_get_default_output_dirrQs.////// ~~m];; < <eleven_multilingual_v1eleven_english_sts_v2eleven_english_sts_v1eleven_flash_v2r? ELEVENLABS_MODEL_MAX_TEXT_LENGTHprovider tts_configcl|stS|}|pi}t||t r||ni}|r|dnd}t|t rd}t|tr|dkr|S|dkr[|pidpt}tt|}|r|S|tvr t|S|tvrot||}t|rP|d} t| t rd} t| tr| dkr| StStS)aReturn the input-character cap for *provider*. Resolution order: 1. ``tts..max_text_length`` (user override in config.yaml) 2. ``tts.providers..max_text_length`` for user-declared command providers 3. ElevenLabs model-aware table (keyed on configured ``model_id``) 4. ``PROVIDER_MAX_TEXT_LENGTH`` default 5. ``DEFAULT_COMMAND_TTS_MAX_TEXT_LENGTH`` when the provider is a command-type user provider without an explicit cap 6. ``FALLBACK_MAX_TEXT_LENGTH`` (4000) Non-positive or non-integer overrides fall through to the default so a broken config can't accidentally disable truncation entirely. max_text_lengthNrrVmodel_id)FALLBACK_MAX_TEXT_LENGTHlowerstrip isinstancegetdictboolintDEFAULT_ELEVENLABS_MODEL_IDr_rPrXBUILTIN_TTS_PROVIDERS_get_named_provider_config_is_command_provider_config#DEFAULT_COMMAND_TTS_MAX_TEXT_LENGTH) r`rakeycfgprov_cfgoverriderdmappednamednamed_overrides r_resolve_max_text_lengthrys& ('' ..   " "C  C *#''#,,==Eswws|||2H2:Dx||-...H(D!!(C  X\\ lN'' 33R7R155c(mm6I6I6K6KLL  M &&&',, '''*344 &u - - 7"YY'899N.$// &!%.#.. &>A3E3E%%6 6 ##rc  ddlm}|}|diS#t$rtdicYSt $r)}td|dicYd}~Sd}~wwxYw) z Load TTS configuration from ~/.hermes/config.yaml. Returns a dict with provider settings. Falls back to defaults for any missing fields. r) load_configttsz9hermes_cli.config not available, using default TTS configzFailed to load TTS config: %sTexc_infoN)r r{rir loggerdebug Exceptionwarning)r{configes r_load_tts_configrs 111111zz%$$$  PQQQ 6DIII s!%(&B BA=7B=Bc|dptS)z%Get the configured TTS provider name.r`)riDEFAULT_PROVIDERrfrg)ras r _get_providerrs2 NN: & & :*: A A C C I I K KKr> rRr<r:rUrWr&rSrTr5rVxmp3>roggwavflacrct|tsiS||}t|tr|niS)zBReturn a provider config block if it's a dict, else an empty dict.)rhrjri)rarsections r_get_provider_sectionrOsC j$ ' ' nnT""G $// 777R7rc t|d}t|tr||nd}t|tr|S|t vrt||}|r|SiS)aReturn the config dict for a user-declared provider. Looks up ``tts.providers.`` first (the canonical location), and falls back to ``tts.`` so users who followed the built-in layout still work. Returns an empty dict when the provider is not declared. providersN)rrhrjrirfrn)rarrrlegacys rroroWs&j+>>I%/ 4%@%@JimmD!!!dG'4   zz||000&z488  M Irrcjt|tsdSt|dpd}|r|dkrdS|d}t|to t |S)z;Return True when *config* declares a command-type provider.Ftypecommand)rhrjrPrirgrfrk)rptypers rrprpos fd # #u  6""(b ) ) / / 1 1 7 7 9 9E )##ujj##G gs # # =W]]__(=(==rc|sdS|}|tvrdSt||}t |r|SdS)zReturn the provider config if *provider* resolves to a command type. Built-in provider names are rejected (they have native handlers). Returns None when the name is a built-in, unknown, or not a command type. N)rfrgrnrorp)r`rarrrs r _resolve_command_provider_configrzsf t ..   " "C ###t ' C 8 8F"6** 4rc#Kt|tsdSt|d}|piD]J\}}t|tr0|t vrt|r||fVKdS)zDYield (name, config) pairs for every declared command-type provider.Nr)rhrjritemsrPrfrnrp)rarrrss r_iter_command_providersrs j$ ' '%j+>>Io2,,..  c dC  TZZ\\9N%N%N*3// Ci  rc|d|dt} t|}n+#ttf$rttcYSwxYw|dkrttS|S)z5Return timeout in seconds, falling back when invalid.timeouttimeout_secondsr)ri#DEFAULT_COMMAND_TTS_TIMEOUT_SECONDSfloat TypeError ValueError)rrawrs r_get_command_tts_timeoutrs **Y +<>a b b c cC:c  z ":::899999: zz8999 LsA%A)(A) output_pathc|rVt|jd}|t vr|S|dp|dpt}t|d}|t vr|ntS)z6Return the validated output format (mp3/wav/ogg/flac)..format output_format) rsuffixrfrglstripCOMMAND_TTS_OUTPUT_FORMATSri!DEFAULT_COMMAND_TTS_OUTPUT_FORMATrP)rrrrfmts r_get_command_tts_output_formatrs k"")//117799@@EE / / /M 8 - ::o & & - , c((..   " " ) )# . .C333339ZZrc|dd}t|tr(|dvSt |S)zEReturn True only when the user explicitly opted in to voice delivery.voice_compatibleF>1onyestrue)rirhrPrgrfrk)rrs r _is_command_tts_voice_compatiblersU JJ)5 1 1E%C{{}}""$$(BBB ;;rcommand_templatepositioncd}d}d}||kr\||}|dkr |dkrd}n:|dkr|rd}n/|dkrd}n&|dkrd}n|dkrd}n|dkrd}n |dkr|dz }|dz }||k\|S) zReturn the shell quote character active right before *position*. Returns ``"'"`` / ``'"'`` when inside a single- / double-quoted region of the template, ``None`` for bare context. NFr'"\TrH)rrquoteescapedichars r_shell_quote_contextrs  EG A h,," C<<s{{ c\\ s{{Q Q' h,,( Lrr quote_contextc\|dkr|ddS|dkrR|ddddddd d Stjd krtj|gSt j|S) zGQuote a placeholder value for its position in a shell command template.rz'\''rrz\\z\"$z\$`z\`nt)replacerr subprocess list2cmdlineshlexr)rrs r_quote_command_tts_placeholderrs}}S'***  WT6 " " WS% WS% WS%   w$&w/// ;u  r placeholderscddD}tjd|d|d}gdtjtdtffd }||}|d d d d }D]\}}|||}|S) z@Replace supported placeholders while preserving ``{{`` / ``}}``.|c3>K|]}tj|VdSN)reescape).0rs r z/_render_command_tts_template..s*>>RYt__>>>>>>rz(?z)\}\}|\{(?Pz)\})matchrJc |dp|d}dtd}|t|t |f|S)Ndoublesingle__HERMES_TTS_PLACEHOLDER___)grouplenappendrrstart)rrtokenrr replacementss r replace_matchz3_render_command_tts_template..replace_matchs{{8$$= H(=(=AC ,=,=AAA  *T"$%5u{{}}EE       rz{{{z}}})joinrcompileMatchrPsubr) rrnamespatternrrenderedrrrs `` @r_render_command_tts_templaters HH>>>>> > >EjPuPPEPPPG+-L RXc] s        {{=*:;;Hc**224==H$22 u##E511 Orprocc |dStjdkrk tjddddt |jgtjtjdn$#t$r| YnwxYwdS tj |jtj n0#t$rYdSt$r|YnwxYw |d dS#tj$rYnwxYw tj |jtjdS#t$rYdSt$r| YdSwxYw) zCBest-effort termination of a shell process and all of its children.Nrtaskkillz/Fz/Tz/PID)stdoutstderrrrIr)pollrrrrunrPpidDEVNULLrkillkillpgsignalSIGTERMProcessLookupError terminatewaitTimeoutExpiredSIGKILL)rs r#_terminate_command_tts_process_treers yy{{ w$  NT4TX?!)!)           IIKKKKK  $(FN++++     !   $      $(FN+++++   sTAA,,B  B $B88 C%C%$C%)DDD$D== E+ E+*E+rrcdtjtjdd}tjdkrt tdd|d<nd|d<tj|fi|} ||\}}n#tj$rz}t| |d \}}n2#t$r%t |d d }t |d d }YnwxYwtj|||| |d }~wwxYw|j rtj |j ||| tj ||j ||S)zGRun a command-provider shell command with process-tree timeout cleanup.T)shellrrtextrCREATE_NEW_PROCESS_GROUPr creationflagsstart_new_sessionrrHoutputNr)r r) rPIPErrgetattrPopen communicaterrr returncodeCalledProcessErrorCompletedProcess)rr popen_kwargsrrrexcs r_run_command_ttsr5s// $$L  w$(/ ..%+..   /!B ==?? fkkmm3q88 L] L LF L L    v;;sDB:KG&%K&K5H  KB#KKKKKcT|t}t|D]\}}dSdS)z=Return True when any command-type TTS provider is configured.NTF)rr)ra_name_cfgs r_has_any_command_tts_providerr9s:%'' .z:: ttt 5rc.tjdduS)z+Check if ffmpeg is available on the system.ffmpegN)shutilwhichrrr _has_ffmpegr>s < ! ! --rmp3_pathctsdS|ddddz} tjdd|dd d d d d dd|dg dd}|jdkrEt d|j|jdddddStj |r%tj |dkr|Sn#tj $rt dYnXt$rt dYn3t$r'}t d|dYd}~nd}~wwxYwdS)z Convert an MP3 file to OGG Opus format for Telegram voice bubbles. Args: mp3_path: Path to the input MP3 file. Returns: Path to the .ogg file, or None if conversion fails. NrrHr.oggr;-i-acodeclibopus-acr-b:a64k-vbroff-yTcapture_outputrz0ffmpeg conversion failed with return code %d: %srignoreerrorsz)ffmpeg OGG conversion timed out after 30szffmpeg not found in PATHz ffmpeg OGG conversion failed: %sr})r>rsplitrrrrrrdecoderrr*getsizerFileNotFoundErrorr)r?ogg_pathresultrs r_convert_to_opusrXs ==tsA&&q)F2HM tXy) Cx G      ! ! NNM +V]-A-A'RZ-A-[-[\`]`\`-a c c c4 7>>( # # (A(AA(E(EO  $DDDBCCCCC 333122222 MMM91tLLLLLLLLM 4s+A2C*%AC**)E*#E*: E*E%%E*c Kt}|di}|dt}t|d|dd}d|i}|dkrt |dz dz}|dd|d<|j|fi|} | |d {V|S) z Generate audio using Edge TTS. Args: text: Text to convert. output_path: Where to save the MP3 file. tts_config: TTS config dict. Returns: Path to the saved audio file. r<r!r?dz+d%rateN)rriDEFAULT_EDGE_VOICErround Communicatesave) rrra _edge_tts edge_configr!rkwargspctrs r_generate_edge_ttsrfs!""I..,,K OOG%7 8 8E +//':>>'3+G+GHH I IEu F ||US[C'((v')'7777K   ; ' '''''''' rc tdpd}|std|di}|dt}|dt}|drd}nd }t }|| } | j|||| } t|d 5} | D]} | |  d d d n #1swxYwY|S)z Generate audio using ElevenLabs. Args: text: Text to convert. output_path: Where to save the audio file. tts_config: TTS config dict. Returns: Path to the saved audio file. ELEVENLABS_API_KEYrz=ELEVENLABS_API_KEY not set. Get one at https://elevenlabs.io/rVvoice_idrdrA opus_48000_64 mp3_44100_128api_keyrrirdrwbN) r rriDEFAULT_ELEVENLABS_VOICE_IDrmendswithr#text_to_speechconvertopenwrite) rrrarm el_configrirdrr!clientaudio_generatorfchunks r_generate_elevenlabsr{se1228bG ZXYYY|R00I}}Z)DEEH}}Z)DEEHF##(' ' #%%J Z ( ( (F+33 # 4O k4 A$  E GGENNNN  sC99C=C=c t\}}|di}|dt}|dt}|d|}t |d|dd}|drd} nd } t } | || } t|||| d ttj i } |dkr!td td|| d<| j jjdi| } | ||t#| dd}t%|r |SS#t#| dd}t%|r |wwxYw)z Generate audio using OpenAI TTS. Args: text: Text to convert. output_path: Where to save the audio file. tts_config: TTS config dict. Returns: Path to the saved audio file. r&r"r!base_urlrrZrAopusr)rmr}zx-idempotency-key)r"r!inputresponse_format extra_headersg?g@closeNr)#_resolve_openai_audio_client_configriDEFAULT_OPENAI_MODELDEFAULT_OPENAI_VOICErrqr(rjrPuuiduuid4maxminaudiospeechcreatestream_to_filer callable)rrrarmr} oai_configr"r!rrr'rw create_kwargsresponsers r_generate_openai_ttsr*s<==GX"--J NN7$8 9 9E NN7$8 9 9E~~j(33H *..*..#*F*FGG H HEF##  (**L \'H = = =F+.DJLL0A0AB     C<<%(s3%?%?M' "-6<&->> >> ,,,.. E??  EGGGG .. E??  EGGGG s !BF-Gcddl}tdpd}|std|di}t |dt pt }t |dtpt}t|d t}t|d t} t |d ptd pt d } | drdnd} |||d} | dks|tks| dkr(| tkrd| i} |r|| d <| dkr| r| | d <| | d<|| dd|dtd| d}|t#|d5}||jdddn #1swxYwY|S)z Generate audio using xAI TTS. xAI exposes a dedicated /v1/tts endpoint instead of the OpenAI audio.speech API shape, so this is implemented as a separate backend. rN XAI_API_KEYrz5XAI_API_KEY not set. Get one at https://console.x.ai/rRrilanguage sample_ratebit_rater} XAI_BASE_URL/.wavrr)rrircodecrz/ttsBearer application/json) Authorization Content-Typez User-Agent<)headersjsonrro)requestsr rgrrirPDEFAULT_XAI_VOICE_IDDEFAULT_XAI_LANGUAGErlDEFAULT_XAI_SAMPLE_RATEDEFAULT_XAI_BIT_RATEDEFAULT_XAI_BASE_URLrstriprqpostrraise_for_statusrtrucontent)rrrarrm xai_configrirrrr}rpayloadrrrys r_generate_xai_ttsr]sOOO]++1r88::G RPQQQr**J:>>*.BCCDDJJLLdPdH:>>*.BCCDDJJLLdPdHjnn]4KLLMMK:>>*.BCCDDHz""  ( (  eggffSkk  !))&11 >h>(0M* %#0 }} 0w00./11    H  k4 "A  !!!""""""""""""""" s/IIIc ddl}tdpd}|std|di}|dt}|dt }|d |d d }|d d } |d d} |d t } |drd} n|drd} nd} ||d||| | ddd| d dd} dd|d}|| | |d}| | }|di}|dd }|dkr+|d!d"}td#|d$||d%id&d}|std't |}t|d(5}||dddn #1swxYwY|S))a Generate audio using MiniMax TTS API. MiniMax returns hex-encoded audio data. Supports streaming (SSE) and non-streaming modes. This implementation uses non-streaming for simplicity. Args: text: Text to convert (max 10,000 characters). output_path: Where to save the audio file. tts_config: TTS config dict. Returns: Path to the saved audio file. rNMINIMAX_API_KEYrz@MINIMAX_API_KEY not set. Get one at https://platform.minimax.io/rSr"rirrHvolpitchr}rr.flacrrF)rirrri}rF)rbitraterchannel)r"rstream voice_setting audio_settingrr)rrr)rrr base_resp status_code status_msg unknown errorzMiniMax TTS API error (code ): datarz%MiniMax TTS returned empty audio dataro)rr rriDEFAULT_MINIMAX_MODELDEFAULT_MINIMAX_VOICE_IDDEFAULT_MINIMAX_BASE_URLrqrrrr/bytesfromhexrtru)rrrarrm mm_configr"rirrrr} audio_formatrrrrWrrr hex_audio audio_bytesrys r_generate_minimax_ttsrsOOO.//52G ][\\\y"--I MM'#8 9 9E}}Z)ABBH MM':>>'1#=#= > >E --q ! !C MM'1 % %E}}Z)ABBHF##   g & &      !"   G&+,7,,G }}XGWb}QQH  ]]__F ;++I-- r22Ka]]<AA V+VV*VVWWW 62&&**7B77I DBCCC-- **K k4 A   s3IIIctdpd}|std|di}|dt}|dpt}|drd}n2|d rd }n|d rd }nd }t } ||5} | jj ||||} tj | j } dddn #1swxYwYn^#t$rt$rG} td| dt!dt#| j| d} ~ wwxYwt'|d5} | | dddn #1swxYwY|S)zGenerate audio using Mistral Voxtral TTS API. The API returns base64-encoded audio; this function decodes it and writes the raw bytes to *output_path*. Supports native Opus output for Telegram voice bubbles. MISTRAL_API_KEYrz?MISTRAL_API_KEY not set. Get one at https://console.mistral.ai/rTr"rirAr~rrrrrrl)r"rrirNzMistral TTS failed: %sTr}zMistral TTS failed: ro)r rriDEFAULT_MISTRAL_TTS_MODELDEFAULT_MISTRAL_TTS_VOICE_IDrqr-rrcompletebase64 b64decode audio_datarrerrorr/r__name__rtru)rrrarm mi_configr"rirr+rwrrrrys r_generate_mistral_ttsrs.//52G \Z[[[y"--I MM'#< = =E}}Z((H,HHF##     f % %    g & &  $&&G M WW % % % @|*33! / 4H !*8+>??K @ @ @ @ @ @ @ @ @ @ @ @ @ @ @  MMM -q4 @@@D$q''2BDDEE1LM k4 A   sO D*=D D*D""D*%D"&D**F>AFFF::F>F> pcm_bytesrchannels sample_widthc @ddl}||z|z}||z}t|}|dddd|||||dz }|dd |} d t|zt| z|z} |d d | d } | |z| z|zS)aWrap raw signed-little-endian PCM with a standard WAV RIFF header. Gemini TTS returns audio/L16;codec=pcm;rate=24000 -- raw PCM samples with no container. We add a minimal WAV header so the file is playable and ffmpeg can re-encode it to MP3/Opus downstream. rNz <4sIHHIIHHsfmt rHz<4sIsdataz<4sI4ssRIFFsWAVE)structrpack) rrrrr byte_rate block_align data_size fmt_chunkdata_chunk_header riff_size riff_headers r_wrap_pcm_as_wavr&sMMMh&5I\)KII   q  I FGY??C NN"S):%;%;;iGI++hGDDK  "%6 6 BBrc ddl}tdptdpd}|std|di}t |dt pt }t |d tpt}t |d ptd pt d }d d|igigdgddd|iiidd} |d|d} | | d|iddi| d} | j dkr | di} | dp| j dd} n#t$r| j dd} YnwxYwtd | j d!|  | }|d"dd#d }t!d$|Dd}|td%|d&p|d'pi}|d(d}n2#t"t$t&f$r}td)||d}~wwxYw|std*t)j|}t-|}|d+r?t3|d,5}||dddn #1swxYwY|St7jd+d-.5}|||j}dddn #1swxYwY t=jd/}|r|d0r|d1|d2d3d4d5d6d7d8d9d:d;d|g}n |d1|d:d;d|g}tAj!|d}|j"dkr6|j#$d?d@Add}tdB|n0tJ&dC|t=j'|| tQj)|n:#tT$rYn.wxYw# tQj)|w#tT$rYwwxYwxYw|S)DaGenerate audio using Google Gemini TTS. Gemini's generateContent endpoint with responseModalities=["AUDIO"] returns raw 24kHz mono 16-bit PCM (L16) as base64. We wrap it with a WAV RIFF header to produce a playable file, then ffmpeg-convert to MP3 / Opus if the caller requested those formats (same pattern as NeuTTS). Args: text: Text to convert (prompt-style; supports inline direction like "Say cheerfully:" and audio tags like [whispers]). output_path: Where to save the audio file (.wav, .mp3, or .ogg). tts_config: TTS config dict. Returns: Path to the saved audio file. rNGEMINI_API_KEYGOOGLE_API_KEYrzIGEMINI_API_KEY not set. Get one at https://aistudio.google.com/app/apikeyrUr"r!r}GEMINI_BASE_URLrpartsrAUDIO voiceConfigprebuiltVoiceConfig voiceName)responseModalities speechConfig)contentsgenerationConfigz/models/z:generateContentrrrrr)paramsrrrrQrmessage,zGemini TTS API error (HTTP r candidatesrc3*K|]}d|vsd|v |VdS) inlineData inline_dataNr)rps rrz'_generate_gemini_tts..s7WW|q/@/@MUVDVDV1DVDVDVDVWWrz+Gemini TTS response contained no audio datarrrz#Gemini TTS response was malformed: z$Gemini TTS returned empty audio datarroFrdeleter;rArBrCrDrErrFrGrHrIrJ -loglevelTrKrLrrNrOzffmpeg conversion failed: zEffmpeg not found; writing raw WAV to %s (extension may be misleading))+rr rgrrirPDEFAULT_GEMINI_TTS_MODELDEFAULT_GEMINI_TTS_VOICEDEFAULT_GEMINI_TTS_BASE_URLrrrrrrr/nextKeyError IndexErrorrrrrrfrqrtrur,NamedTemporaryFilerr<r=rrrrrSrrcopyfilerremoveOSError)rrrarrm gemini_configr"r!r}rendpointrerrr4rr audio_partinline audio_b64rr wav_bytesrytmpwav_pathr;cmdrWrs r_generate_gemini_ttsrHs&"OOO-..W-@P2Q2QWUW^^``G   W   NN8R00M  !!'+CDD E E K K M M iQiE  !!'+CDD E E K K M M iQiE*%% ' * + + ' & eggffSkk  /01#*))K+?     G;;E;;;H}}w!34  Hs"" )--//%%gr22CWWY''>8=#+>FF ) ) )]4C4(FFF ) K(*> K K6 K K    M}}\"1%i09WWeWWWY]^^  LMM M --T 1N1NTRTJJvr** j) ,MMMDDDEE1LM CABBB ++I ++I##F++ +t $ $  GGI                    $F5 A A AS )8h''  3  ""++F33 XD(y%E65+w tXt[';W^CbIIIF A%%--gh-GGM"#H#H#HIII& NNW    OHk 2 2 2  Ih        D   Ih        D  sA GG-,G- BJ''K>KKM&&M*-M* N22N69N6>C!S R55 SSS,SS, S)&S,(S))S,cf ddl}|jdduS#t$rYdSwxYw)z=Check if the neutts engine is importable (installed locally).rNrWFimportlib.utilutil find_specr importlibs r_check_neutts_availablersP~''11== uu " 00cf ddl}|jdduS#t$rYdSwxYw)z@Check if the kittentts engine is importable (installed locally).rNr5Frrs r_check_kittentts_availablersP~'' 44D@@ uurcZtttjdz dz S)z9Return path to the bundled default voice reference audio.neutts_sampleszjo.wavrPr__file__r(rrr_default_neutts_ref_audior$$ tH~~$'77(B C CCrcZtttjdz dz S)z>Return path to the bundled default voice reference transcript.r!zjo.txtr"rrr_default_neutts_ref_textr'r%rcddl}|di}|ddp t}|ddp t}|dd}|d d }|} |d s|d d dd z} t ttj dz } |j | d|d| d|d|d|d|g} tj | ddd} | j dkrk| j} d| D}t#dt%d|pd| |kr`t)jd}|r5|d| ddd |g}tj |dd!"t-j| nt-j| ||S)#aGenerate speech using the local NeuTTS engine. Runs synthesis in a subprocess via tools/neutts_synth.py to keep the ~500MB model in a separate process that exits after synthesis. Outputs WAV; the caller handles conversion for Telegram if needed. rNrW ref_audiorref_textr"zneuphonic/neutts-air-q4-ggufdevicecpurrrHzneutts_synth.pyz--textz--outz --ref-audioz --ref-textz--modelz--deviceTrrMrrc<g|]}|d|S)zOK:) startswith)rls r z$_generate_neutts..s)QQQQQ\\%=P=PQqQQQrzNeuTTS synthesis failed: rr;rBrJrrrKcheckr)sysrir$r'rqrRrPrr#r( executablerrrrrg splitlinesr/chrrr<r=rrrename)rrrar5 neutts_configr)r*r"r+r synth_scriptrrWr error_linesr;conv_cmds r_generate_neuttsr>sJJJNN8R00M!!+r22Q6O6Q6QI  R00N4L4N4NH   g'E F FE   x / /FH    ' ':%%c1--a069tH~~,/@@AAL  $yh5F C^C4 M M MF A$$&&QQ&"3"3"5"5QQQ es2ww||K7P7P7cTceefff;h''  -hk7KXH N84 < < < < Ih     Ih , , , r_piper_voice_cachecf ddl}|jdduS#t$rYdSwxYw)z2Check whether the piper-tts package is importable.rNr:Frrs r_check_piper_availablerA*sP~''00<< uurcrddlm}t|dd}|dd|S)zReturn the directory where Hermes caches Piper voice models. Resolves to ``~/.hermes/cache/piper-voices/`` under the active HERMES_HOME so voice downloads follow profile boundaries. rrLzcache/piper-voicespiper_voices_cacheTr)rOrMrr))rMroots r_get_piper_voices_dirrE3sL 0///// 35IJJ K KDJJtdJ+++ Krr! download_dirc P|st}t|}|jdkr#|rt |S||dz }|r)||dz rt |Sddl}t d|| tj |j dd|dt |gd d d }n+#tj $r}td |d |d}~wwxYw|jdkr:|jpdpd}td|d|dd|std|dt |S)a}Resolve *voice* (a model name or path) to a concrete .onnx file path. Accepts any of: - Absolute / expanded path to an .onnx file the user already has - A voice *name* like ``en_US-lessac-medium`` (downloads to ``download_dir`` on first use via ``python -m piper.download_voices``) Raises RuntimeError if the model can't be located or downloaded. z.onnxz .onnx.jsonrNz0[Piper] Downloading voice '%s' to %s (first use)z-mzpiper.download_voicesz--download-dirTrr-z/Piper voice download timed out after 300s for 'rrzno stderr outputz!Piper voice download failed for 'z': iz#Piper voice download completed but uh is missing — check voice name (see: https://github.com/OHF-Voice/piper1-gpl/blob/main/docs/VOICES.md))DEFAULT_PIPER_VOICErr'rrfr*rPr5rinforrr6rr/rrrg)r!rF candidatecached_sysrWrrs r_resolve_piper_voice_pathrM?s $#U &&((I7**y/?/?/A/A*9~~uOOO +F }}Le+?+?+??GGII6{{ KKBE<XXX  _d$;U s<00 2dC     $ Fe F F F    A-%2,,..D2D H H H&#, H H    ==??  )& ) ) )   v;;s0DD*D%%D*c t}ddl}t|tr|dinidpt }t dp t}| ddtdd }t||}|d |} | tvrTtd |||| t| <td t| } d} t!fddD} | r ddlm} | t'ddt'ddt'ddt'ddtdd} n*#t($rtdYnwxYw|}|ds|ddddz}||d5}| | ||| n| ||dddn #1swxYwY||krqt5jd}|rF|d |d!d"d#|g}t9j|dd$% t=j|n%#t@$rYnwxYwt=j!|||S)&zGenerate speech using the local Piper engine. Loads the voice model once per process (cached by absolute path) and writes a WAV file. Caller is responsible for converting to MP3/Opus via ffmpeg when a different output format is required. rNr:r! voices_dirTruse_cudaFz::cuda=z[Piper] Loading voice: %s)rPz[Piper] Voice loadedc3 K|]}|vV dSrr)rk piper_configs rrz&_generate_piper_tts..s<  \r) length_scale noise_scale noise_w_scalevolumenormalize_audio)SynthesisConfigrTrZrUgMbX?rVg?rWrXuZ[Piper] SynthesisConfig not available in this piper-tts version — advanced knobs ignoredrrrHro) syn_configr;rBrJrrrKr3)"r;waverhrjrirHrrEr'r)rkrMr?rrIloadanyr:rYrr rrqrRrtsynthesize_wavr<r=rrrrr r9)rrrar9r[ voice_namerFrP model_path cache_keyr!rZ has_advancedrYrwav_filer;r=rSs @r_generate_piper_ttsrdssJKKK2>'2...TVL!!'**A.AJ ((66Q:O:Q:QRR]]__Ltd333L$$Z7788H*:|DDJ00h00I*** /<<<(2 X(V(V9% *+++ y )E J^L  - - - - - -("<#3#3NC#H#HII!,"2"2=%"H"HII#L$4$4_c$J$JKK\--h<<== $\%5%56G%N%N O O JJ    NN5      H    ' ':%%c1--a069 8T " "1h  !  xJ G G G G  x 0 0 0 111111111111111;h''  -hk7KXH N84 < < < <  (####     Ih , , , s7:B;H66$II,2K**K.1K.2M MM_kittentts_model_cachecnt}|di}|dt}|dt}|dd}|dd}|tvrHt d|||t|<t d t|} | |||| } d d l} |} | d s| ddd d z} | | | d| |kr`tj d} | r5| d| ddd|g}tj|ddt!j| nt!j| ||S)atGenerate speech using KittenTTS local ONNX model. KittenTTS is a lightweight TTS engine (25-80MB models) that runs entirely on CPU without requiring a GPU or API key. Args: text: Text to convert to speech. output_path: Where to save the audio file. tts_config: TTS config dict. Returns: Path to the saved audio file. r5r"r!rrZ clean_textTz[KittenTTS] Loading model: %sz%[KittenTTS] Model loaded successfully)r!rrgrNrrrHrEr;rBrJrrrKr3)r6riDEFAULT_KITTENTTS_MODELDEFAULT_KITTENTTS_VOICErerrIgenerate soundfilerqrRrur<r=rrrrr9)rrrar4 kt_config model_namer!rrgr"rsfrr;r=s r_generate_kittenttsros"##I{B//Iw(?@@J MM'#: ; ;E MM'3 ' 'E|T22J/// 3Z@@@-6Yz-B-Bz* ;<<< ": .E NN4uEjN Q QEH    ' ':%%c1--a069HHXue$$$;h''  -hk7KXH N84 < < < < Ih     Ih , , , rcrstddStt}t |}t |}t |kr4td|t |d|ddl m }|dd  }|d k}|r4t| }|t||}ntjd } tt$} | d d |t)|} | d| d| z }n|r|dvr | d| dz }n | d| dz }|jd d t-| |0td|t1||n>|dkrf t3n)#t4$rt7jddddcYSwxYwtdt;n|dkrf t=n)#t4$rt7jddddcYSwxYwtdt?nf|dkr-tdtAn3|dkr-tdtCn|d krf tEn)#t4$rt7jdd!ddcYSwxYwtd"tGn|d#kr-td$tIna|d%krTtKst7jdd&ddStd'tMn|d(krf tOn)#t4$rt7jdd)ddcYSwxYwtd*tQn|d+krf tSn)#t4$rt7jdd,ddcYSwxYwtd-tUn/d } tWn#t4$rd} YnwxYw| rtd. ddl,} | j-.d/05}|/fd10d23dddn #1swxYwYn#tb$r&tej3tiYnYwxYwtKr.td4d%}tMnt7jdd5ddStjj67r#tjj68dkrt7jdd6|d7ddSd}|Mts|r=:dstw}|r|:d}nH|d8vr+:dstw}|r|d }n|d9vr:d}tjj68}td:|d;|d<}|rd=|}t7jd |||d>dS#tx$r>}d?|d@|}t=dA|t|dcYd}~Sd}~wt|$r@}dB|d@|}t=dA|d Ct|dcYd}~Sd}~wt~$r@}dD|d@|}t=dA|d Ct|dcYd}~Sd}~wwxYw)Eac Convert text to speech audio. Reads provider/voice config from ~/.hermes/config.yaml (tts: section). The model sends text; the user configures voice and provider. On messaging platforms, the returned MEDIA: tag is intercepted by the send pipeline and delivered as a native voice message. In CLI mode, the file is saved to ~/voice-memos/. Args: text: The text to convert to speech. output_path: Optional custom save path. Defaults to ~/voice-memos/.mp3 Returns: str: JSON result with success, file_path, and optionally MEDIA tag. zText is requiredF)successz>TTS text too long for provider %s (%d chars), truncating to %dNr)get_session_envHERMES_SESSION_PLATFORMrtelegramz %Y%m%d_%H%M%STrtts_r)r&rVrTrUrAz.mp3z3Generating speech with command TTS provider '%s'...rVz`ElevenLabs provider selected but 'elevenlabs' package not installed. Run: pip install elevenlabs)rqr) ensure_asciiz$Generating speech with ElevenLabs...r&zz%text_to_speech_tool..sGK0B4S]0^0^$_$_rrrz9Edge TTS not available, falling back to NeuTTS (local)...zhNo TTS provider available. Install edge-tts (pip install edge-tts) or set up NeuTTS for local synthesis.z-TTS generation produced no output (provider: ))r<rWrSrRr5r:)rVr&rTrUz,TTS audio saved: %s (%s bytes, provider: %s),zMEDIA:z[[audio_as_voice]] )rq file_path media_tagr`rzTTS configuration error (rz%szTTS dependency missing (r}zTTS generation failed ()@rg tool_errorrrrryrrrgateway.session_contextrrrfrr'rdatetimenowstrftimeDEFAULT_OUTPUT_DIRr)rr(rPrIr5r#r rdumpsr{r(rrrr-rrrr>r6ror;rdrconcurrent.futuresfuturesThreadPoolExecutorsubmitrWr/ryrrfrrr*rTrrqrXrrrUr)rrr`command_provider_configmax_lenrrplatform want_opusr~ timestampout_dirredge_available concurrentpoolr opus_path file_sizerr error_msgrzras` @@rtext_to_speech_toolrs1 * =tzz||=,e<<<<!##JZ((H ?xTT'x< ? ? ? x < < < <  ! ! '%'''' ' ' 'z$[##!&'''''' ' KK> ? ? ? x < < < <  " " KK? @ @ @ !$* = = = =    KK; < < < dHj 9 9 9 9  " " '&(((( ' ' 'z$H##!& '''''' ' KKG H H H !$* = = = =  ! ! KKE F F F x < < < <  ! !*,, 'z$F##!& '''' KKB C C C T8Z 8 8 8 8  $ $ '!#### ' ' 'z$J## !& '''''' ' KKL M M M h ; ; ; ;   ' ' ' 'z$5## !& '''''' ' KKA B B B h ; ; ; ;"N ' """" ' ' '!& ' ' @AAAP----#+>>1>MM-QU ______ &&,,,---------------$PPPK 24: N NOOOOOP(** ' WXXX# x<<<<z$E##!& ''''w~~h'' #27??8+D+D+I+I: TTTT"### #! " .00GHH =((00- 0 : :I -#,#+#4#4V#<#< S S S\d\m\mnt\u\u S(22I ($#' F F F'0088 GOOH--  BHQZN^N^`hiii(X''  ;:y::Iz!" 0      444@@@Q@@  T9%%%)U333333333 444?x??A??  T9t 444)U333333333 444>h>>1>>  T9t 444)U333333333 4sy 8^H^#H;8^:H;;6^2J^#J'$^&J''B^M^#M96^8M99B^ 3^=Q ^ #Q2/^1Q226^)R87^8#S^S2^T ^ T/,^.T//^V5./V) V5)V--V50V-1V54^5-W%"^$W%%A^=A^C>^ a3%3_a3 a3+5`& a3& a335a.(a3.a3cptrdS tdS#t$rYnwxYw tt drdSn#t$rYnwxYw t t rdSn#t$rYnwxYwt drdSt drdSt dst drdS tt drdSn#t$rYnwxYwtrdStrdStrdSdS) a, Check if at least one TTS provider is available. Edge TTS needs no API key and is the default, so if the package is installed, TTS is available. A user-declared command provider also satisfies the requirement. Returns: bool: True if at least one provider can work. TrhrrrrrF) r9rr r#r r(_has_openai_audio_backendr-rrrArrrcheck_tts_requirementsr s%&&t t        - . . 4         $ & & 4       &''t]##t%&&-8H*I*It    * + + 4         t!##tt 5sA" //A A A $B BBC66 DDc t}|rtds |tfStd}|$d}t r|dz }t ||jt|j dddfS)zReturn direct OpenAI audio config or a managed gateway fallback. When ``tts.use_gateway`` is set in config, the Tool Gateway is preferred even if direct OpenAI credentials are present. r| openai-audioNz8Neither VOICE_TOOLS_OPENAI_KEY nor OPENAI_API_KEY is setz5, and the managed OpenAI audio gateway is unavailablerv1) rrDEFAULT_OPENAI_BASE_URLrrrnous_user_tokenrgateway_originr)direct_api_keymanaged_gatewayrs rrr?s 233N7oe4476662>BBOL % ' ' O N NG!!!  *G  ) 0 0 5 5888$-- rcVttptdS)zPReturn True when OpenAI audio can use direct credentials or the managed gateway.r)rkrrrrrrrUs% ,..^2N~2^2^ _ __rz(?<=[.!?])(?:\s|\n)|(?:\n\n)z```[\s\S]*?```z\[([^\]]+)\]\([^)]+\)z https?://\S+z \*\*(.+?)\*\*z \*(.+?)\*z`(.+?)`z^#+\s*flagsz ^\s*[-*]\s+z---+z\n{3,}cFtd|}td|}td|}td|}t d|}t d|}td|}td|}td|}td|}| S)z:Remove markdown formatting that shouldn't be spoken aloud. z\1rz ) _MD_CODE_BLOCKr_MD_LINK_MD_URL_MD_BOLD _MD_ITALIC_MD_INLINE_CODE _MD_HEADER _MD_LIST_ITEM_MD_HR _MD_EXCESS_NLrg)rs r_strip_markdown_for_ttsrms   c4 ( (D <<t $ $D ;;r4 D <<t $ $D >>% & &D   ud + +D >>"d # #D   R & &D ::b$  D   VT * *D ::<<r text_queue stop_eventtts_done_eventdisplay_callbackc | ddttt}|di}|d|d|dt di|di|diit dpd}|stdn t}|| n*#t$rtd YnwxYw t}| d d d  nj#ttf$r'} td| dYd} ~ n7d} ~ wt $r'} td| dYd} ~ nd} ~ wwxYwd} d} d} d} gt#jdt"j}dt(ff d }ds; || }n5#t,j$r#t1| | kr || d} Y^wxYw|6|d| } | r || n| |z } |d| } d| vrd| vr t6| }|n_|}| d|}| |d} t1|| kr|| z} n |||; |n#t,j$rYnwxYw,n2#t $r%} td| Yd} ~ nd} ~ wwxYw:  n#t $rYnwxYw|!dS#:  n#t $rYnwxYw|!wxYw)aConsume text deltas from *text_queue*, buffer them into sentences, and stream each sentence through ElevenLabs TTS to the speaker in real-time. Protocol: * The producer puts ``str`` deltas onto *text_queue*. * A ``None`` sentinel signals end-of-text (flush remaining buffer). * *stop_event* can be set to abort early (e.g. user interrupt). * *tts_done_event* is **set** in the ``finally`` block so callers waiting on it (continuous voice mode) know playback is finished. NrVristreaming_model_idrdrhrz8ELEVENLABS_API_KEY not set; streaming TTS audio disabledrlz8elevenlabs package not installed; streaming TTS disabledrErHint16) sampleraterdtypezsounddevice not available: %sz#sounddevice OutputStream failed: %sr[g?z].*?rsentencecH rdSt|}|sdS|d} D]0}|d|krdS1 |  | dSt |kr |d} j| d}h|D]a}rnLddl }| ||j } | ddbdSdS |dS#t$r&}td |Yd}~dSd}~wwxYw) z6Display sentence and optionally generate + play audio.Nz.!, pcm_24000rnr)rrrHz!Streaming TTS sentence failed: %s)is_setrrgrfrrrrrrsnumpy frombufferrrureshaperrr)rcleaned cleaned_lowerprev audio_iterrz_np audio_arrayr_play_via_tempfile_spoken_sentencesrwrrd output_streamrstream_max_lenris r_speak_sentencez.stream_tts_to_speaker.._speak_sentences  "" -h77==??G #MMOO22599M)  ::<<&&u-->>FF?  $ $W - - -+  ***~7||n,,!/>/2 I#2:: %%"- ; !,!+HH%,,.."!E++++&)nnU#)n&L&L %++K,?,?A,F,FGGGG HH!E '&z:>>>>> I I IBCHHHHHHHHH IsBE1# E11 F!;FF!cd} ddl}tjdd}|j}||d5}|d|d|d |D]-}|rn| |.dddn #1swxYwYdd l m }||n2#t$r%}td |Yd}~nd}~wwxYw|r( tj|dS#t"$rYdSwxYwdS#|r& tj|w#t"$rYwwxYwwxYw) z0Write PCM chunks to a temp WAV file and play it.NrrFrrorHrIrE)play_audio_filez!Temp-file TTS fallback failed: %s)r[r,rrrt setnchannels setsampwidth setframerater writeframestools.voice_moderrrrrr+r ) rstop_evttmp_pathr[rwfrzrrs rrz1stream_tts_to_speaker.._play_via_tempfilesH  1NNN8YYsD)).ROOA&&&OOA&&&OOE***!+..#??,,"!Eu----...............=<<<<<)))) I I IBCHHHHHHHH I (+++++"8 (++++"s7CA0B7+ C7B;;C>B;?CD2 DC>9D2>DD2 D D-,D-2E6E  E EEEErzTz Streaming TTS pipeline error: %s)"clearrp%DEFAULT_ELEVENLABS_STREAMING_MODEL_IDrriryr rrr#r r1 OutputStreamrr rrrrDOTALLrPrqueueEmptyrrrg_SENTENCE_BOUNDARY_REsearchend get_nowaitstoprset)rrrrrarvrmr!r0r sentence_bufmin_sentence_lenlong_flush_len queue_timeout_think_block_rerdeltamend_posrrrrwrdrrris ` ` @@@@@@@rstream_tts_to_speakerr|s"{ .8%'' NN<44 ==X66==!5!*z8!D!DFF 2  Mz M<)LI)Lz8)L)L M M   !!566<" ) NNU V V V V [/11 #G444 [ [ [YZZZZZ [ ! ),..B$&OO#(1G%4%%M"''))))#W-)))LL!@#FFF$(MMMMMM )))NN#H#NNN$(MMMMMM)  ')*%>biPPP( Ic( I( I( I( I( I( I( I( I( I( I( I( I( I( IT   4##%%* * "}==;   |$$~55#OL111#%L  }.222|DD %%''2#OL111 E !L +..r<@@L<''Jl,J,J *)00>>9%%'''1+GHH5 x~~''((+;;;#+l#:L))) *?##%%* *Z  %%'''';      @@@93????????@  $ ""$$$##%%%%      $ ""$$$##%%%%    sCN C;:N ;$D"N !D""N (:E#"N #G 4FN  G #GN G  AN %H<;N =Q#> Q Q# Q  Q#__main__u🔊 Text-to-Speech Tool Modulez2==================================================c> |dS#t$rYdSwxYw)NTF)r )importerlabels r_checkrTs9  HJJJ4   55 s  z Provider availability:z Edge TTS: installedz$not installed (pip install edge-tts)z ElevenLabs: elz¬ installed (pip install elevenlabs)z API Key: rhrznot setz OpenAI: oaiz not installedz2not set (VOICE_TOOLS_OPENAI_KEY or OPENAI_API_KEY)z MiniMax: rz API key setznot set (MINIMAX_API_KEY)z Piper: z%not installed (pip install piper-tts)z ffmpeg: u ✅ foundu(❌ not found (needed for Telegram Opus)z Output dir: z Configured provider: )registryrrraConvert text to speech audio. Returns a MEDIA: path that the platform delivers as native audio. Compatible providers render as a voice bubble on Telegram; otherwise audio is sent as a regular attachment. In CLI mode, saves to ~/voice-memos/. Voice and provider are user-configured (built-in providers like edge/openai or custom command providers under tts.providers.), not model-selected.objectstringzThe text to convert to speech. Provider-specific character caps apply and are enforced automatically (OpenAI 4096, xAI 15000, MiniMax 10000, ElevenLabs 5k-40k depending on model); over-long input is truncated.)r descriptionz9Optional custom file path to save the audio. Defaults to z/audio_cache/.mp3rr)r propertiesrequired)rr parametersr|c rt|dd|dS)Nrrrr)rri)argskws rr{r{s62 XXfb ! !HH]++ - - -ru🔊)rtoolsetschemahandlercheck_fnemojir)__doc__ryrrrloggingrrrrr<rrr, threadingrpathlibrtypingrrrr urllib.parserrOr getLoggerrrr tools.managed_tool_gatewayrtools.tool_backend_helpersrrrtools.xai_httprrr#r(r-r1r6r;rr^rprmrrrhrirHrrrrrrrrrrrrrrrGEMINI_TTS_SAMPLE_RATEGEMINI_TTS_CHANNELSGEMINI_TTS_SAMPLE_WIDTHrPrQrrXrl__annotations__r_reMAX_TEXT_LENGTHryrr frozensetrnrrrrqrrorkrprrrrrrrrrr rrrrr5r9r>rXrfr{rrrrrrrrrr$r'r>r?rArErMrdrerorrtuplerrrrrrrrrr MULTILINErrrrrQueueEventrprintrrr`tools.registryrr TTS_SCHEMAregisterrrrrsz!!!F    000000000000 000000  8 $ $ / / / /DCCCCCpppppppppp000000        '46(;%(="+5'2=3E,9!P=====-,..    , ,$sCx.   ##"" 4 4 $sCx.    + ,05$5$sm5$c3h(5$ 5$5$5$5$v$sCx.&Ld38nLLLLLH" # # #  '*#$)!&Y'D'D'DEE&*#8d38n8C8DcN8888S#X  #s(^0>S#X>4>>>>S#Xd38n* S#X     T#s(^     "&[[ cN[#[ [[[[$T#s(^3#(3-@#hsmPS"sCx. <#j.>#4####L%Qc%QE%Qj6Q%Q%Q%Q%QP'd'DcN't'''' A AAA cN A S#X A  AAAAHhtCH~.FRV.T.... s x}    L3Sd3PS8nY\<(s(($sCx.(UX((((\-s--$sCx.-UX----f;C;c;tCH~;RU;;;;BQQ#Q4S>QVYQQQQn++#+4S>+VY++++f.'/ CCCCC C  CCCCD@s@@$sCx.@UX@@@@NDD3DDDD D#DDDD 232S2d38n2QT2222x&(DcN''' t    1S111111hKcKKc3hKTWKKKKf*,S#X+++4c44c3h4TW4444x"&B4B4 B4#B4 B4B4B4B4P11111hU38_,`4````# #BCC-.. 2:. / / "*_ % % 2:& ' ' RZ % % "*Z(( RZ  6 6 6  >>>> G   9%%  # #    &9= NN NNONxt 45 NNNNh z E +,,, E(OOO E $%%% E x&&1A6*J*Jv;;Pv x xyyy E z&&1CT*J*Jx;;Px z z{{{ E XMM2F$G$GV55Y X XYYY E e&&1F*N*Nc;;Tc e efff E o0022 l558l o o E oMM:K,L,Lm==Rm o oppp E q*@*@*B*Bo;;Ho q qrrr E i++--g;;=g i ijjj E 1/ 1 1222    F}V$$H E .H . ./// 0/////// `! s ! O[n[n[p[p O O O   H   &    - -$       r