ia UdZddlZddlZddlZddlZddlZddlZddlZddlZddl Z ddl Z ddl Z ddl m Z mZmZmZejeZdZdefdZddlmZdefdZdeefd Zdefd Zdefd Zdefd Z d Z!dZ"dZ#dZ$dZ%dZ&ej'(ej)dZ*d/de+de,de+ddfdZ-GddZ.GddZ/de/e.zfdZ0hdZ1ej2d ej3!Z4d"edefd#Z5d0d$ed%eedeee ffd&Z6da7eej8e9d'<e j:Z;d1d(Zd2d-e+de+fd.Z?dS)3aVVoice Mode -- Push-to-talk audio recording and playback for the CLI. Provides audio capture via sounddevice, WAV encoding via stdlib wave, STT dispatch via tools.transcription_tools, and TTS playback via sounddevice or system audio players. Dependencies (optional): pip install sounddevice numpy or: pip install hermes-agent[voice] N)AnyDictListOptionalcddl}ddl}||fS)zLazy-import sounddevice and numpy. Returns (sd, np). Raises ImportError or OSError if the libraries are not available (e.g. PortAudio missing on headless servers). rN) sounddevicenumpy)sdnps 5/home/ubuntu/.hermes/hermes-agent/tools/voice_mode.py _import_audior s%  r6MreturncT tdS#ttf$rYdSwxYw)z/Return True if audio libraries can be imported.TF)r ImportErrorOSErrorrr _audio_availabler+s<t  !uus '') is_termuxc&trdSdS)NzGpkg install python-numpy portaudio && python -m pip install sounddevicezpip install sounddevice numpy)_is_termux_environmentrrr _voice_capture_install_hintr7sYXX * *rcJtsdStjdS)Nztermux-microphone-record)rshutilwhichrrr _termux_microphone_commandr=s& ! # #t <2 3 33rctsdS tjgddddd}d|jpdvS#t$rYdSwxYw)NF)pmlistpackageszcom.termux.apiTcapture_outputtexttimeoutcheckzpackage:com.termux.api)r subprocessrunstdout Exception)results r _termux_api_app_installedr-Ds{ ! # #u  8 8 8    (FM,?R@@ uus%8 AAc>tduo tSN)rr-rrr _termux_voice_capture_availabler0Ts % ' 't 3 S8Q8S8SSrcg}g}t}t}t|o|}tddDr|dddlm}|r|d tdd5}d | vrJtj d r|d n|d d d d n #1swxYwYn#tttf$rYnwxYw t!\}} |} | s-|r|dn|dnr#t$$retj d r|dn-|r|dn|dYnwxYwn#t&$rZ|r|dn?|r|s|dn%|dt)dYnzt$rn|r|dnS|r|s|dn9t+r|dn|dYnwxYw| ||dS)zDetect if the current environment supports audio I/O. Returns dict with 'available' (bool), 'warnings' (list of hard-fail reasons that block voice mode), and 'notices' (list of informational messages that do NOT block voice mode). c3TK|]#}tj|V$dSr/)osenvironget).0vs r z+detect_audio_environment..fs0 R R2:>>!   R R R R R Rr) SSH_CLIENTSSH_TTYSSH_CONNECTIONz.Running over SSH -- no audio devices availabler) is_containerz3Running inside Docker container -- no audio devicesz /proc/versionr microsoft PULSE_SERVERz%Running in WSL with PulseAudio bridgezRunning in WSL -- audio requires PulseAudio bridge. 1. Set PULSE_SERVER=unix:/mnt/wslg/PulseServer 2. Create ~/.asoundrc pointing ALSA at PulseAudio 3. Verify with: arecord -d 3 /tmp/test.wav && aplay /tmp/test.wavNzMNo PortAudio devices detected, but Termux:API microphone capture is availablez&No audio input/output devices detectedz?Audio device query failed but PULSE_SERVER is set -- continuingzMPortAudio device query failed, but Termux:API microphone capture is availablez6Audio subsystem error (PortAudio cannot query devices)zDTermux:API microphone recording available (sounddevice not required)zkTermux:API Android app is not installed. Install/update the Termux:API app to use termux-microphone-record.zAudio libraries not installed ()zBTermux:API microphone recording available (PortAudio not required)zmPortAudio system library not found -- install it first: Termux: pkg install portaudio Then retry /voice on.zPortAudio system library not found -- install it first: Linux: sudo apt-get install libportaudio2 macOS: brew install portaudio Then retry /voice on.) availablewarningsnotices)rr-boolanyappendhermes_constantsr<openreadlowerr3r4r5FileNotFoundErrorPermissionErrorrr query_devicesr+rrr) rBrCtermux_mic_cmdtermux_app_installedtermux_capturer<fr _devicess r detect_audio_environmentrTXsHG/11N466.A-ABBN R R&Q R R RRRJHIII.-----|~~OMNNN  /3 ' ' 1affhhnn....:>>.11NN#JKKKKOO^                 8     .A Z&&((G N!NNN#rssssOO$LMMM Z Z Zz~~n-- Z`aaaa Znoooo XYYY Z ```  ` NNa b b b b  `$8 ` OO}     OO^>Y>[>[^^^ _ _ _    NN_ ` ` ` `  $8  OO}    $ % %  OO(     OO(   ,"\  swD!A3D D!DD!DD!!D;:D;?HAFHA,HHHHA!K"+A4K"!K"i>int16@ hermes_voicepQ? frequencydurationcountc z t\}}n#ttf$rYdSwxYw d}tt|z}tt|z}g}t |D]!} |d||d} |d|jz|z| z} tttdz|dz} | d| xx|dd | zcc<| | dxx|d d| zcc<| | d zd z |j | |d z kr/| | ||j #||} || t t!jdz}|r|jrot!j|krXt!jd|r0|jrt!j|kX|dS#t,$r&}t.d|Yd}~dSd}~wwxYw)zPlay a short beep tone using numpy + sounddevice. Args: frequency: Tone frequency in Hz (default 880 = A5). duration: Duration of each beep in seconds. count: Number of beeps to play (with short gap between). NgQ?rF)endpointrW{Gz?rU333333?idtype samplerate@zBeep playback failed: %s)r rrint SAMPLE_RATErangelinspacesinpiminrFastyperVzeros concatenateplaytime monotonic get_streamactivesleepstopr+loggerdebug)r]r^r_r r gapsamples_per_beepsamples_per_gappartsittonefade_lenaudiodeadlinees r play_beeprsBB  !4{X566kC/00u H HA Ax)9E JJA66!be)i/!344D3{T1224D4IJJH (OOOr{{1a:: :OOO (    Aq( ; ; ;    LL$*u,44RX>> ? ? ?519}} RXXoRXXFFGGGu%% +...>##c)mmoo "--//"8 T^=M=MPX=X=X Jt   mmoo "--//"8 T^=M=MPX=X=X  444 /3333333334s$))IJ J:J55J:ceZdZdZdZddZedefdZede fdZ ede fdZ ddd Z dd Zdeefd Zdd Zdd ZdS)TermuxAudioRecorderzBRecorder backend that uses Termux:API microphone capture commands.FrNcntj|_d|_d|_d|_d|_dS)NFr) threadingLock_lock _recording _start_time_recording_path _current_rmsselfs r __init__zTermuxAudioRecorder.__init__s6^%% .2rc|jSr/rrs r is_recordingz TermuxAudioRecorder.is_recordings rcJ|jsdStj|jz SNrrrurvrrs r elapsed_secondsz#TermuxAudioRecorder.elapsed_seconds' 3~$"222rc|jSr/rrs r current_rmszTermuxAudioRecorder.current_rms s   rc ~t}|stdtstd|j5|jr ddddSt jtdtj d}t j td|d|_ dddn #1swxYwY|d|j d d d d d ttdttg } t!j|ddddny#t j$rG}|jp|jpt|}td||d}~wt,$r}td||d}~wwxYw|j5tj|_d|_d|_dddn #1swxYwYt4ddS)NzTermux voice capture requires the termux-api package and app. Install with: pkg install termux-api Then install/update the Termux:API Android app.zrTermux voice capture requires the Termux:API Android app. Install/update the Termux:API app, then retry /voice on.Texist_ok %Y%m%d_%H%M%S recording_z.aacz-fz-l0z-eaacz-rz-cr"z Termux microphone start failed: rzTermux voice recording started)r RuntimeErrorr-rrr3makedirs _TEMP_DIRrustrftimepathjoinrstrrkCHANNELSr(r)CalledProcessErrorstderrr*stripr+rvrrr{info)ron_silence_stopmic_cmd timestampcommandrdetailss r startzTermuxAudioRecorder.starts ,.. B  )** K  Z Y Y  Y Y Y Y Y Y Y Y K D 1 1 1 1 o66I#%7<< ;W ;W;W;W#X#XD  Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y  $& # % #k"" #h--   N N74dBVZ [ [ [ [ [, T T Tx5185s1vv<<>>GK'KKLLRS S N N NE!EEFFA M NZ " "#~//D "DO !D  " " " " " " " " " " " " " " "  455555sP CACC CD F/AE11 F>FF 'GGGc`t}|sdStj|dgdddddS)N-qTrFr")rr(r))rrs r _stop_termux_recordingz*TermuxAudioRecorder._stop_termux_recording;sA,..  Ft$PRZ_``````rcr|j5|js ddddSd|_|j}d|_|j}d|_dddn #1swxYwY||rt j|sdStj |z dkr( t j |n#t$rYnwxYwdSt j |dkr( t j |n#t$rYnwxYwdStd||S)NFrrdz"Termux voice recording stopped: %s)rrrrrrr3risfilerurvunlinkrgetsizer{r)rr started_ats r rzzTermuxAudioRecorder.stopAs Z " "?  " " " " " " " "$DO'D#'D )J !D  " " " " " " " " " " " " " " " ##%%% 27>>$// 4 >  j (3 . .  $    4 7??4 A % %  $    4 8$??? s: A #A  AA*B?? C  C 5D DDc|j5|j}d|_d|_d|_dddn #1swxYwY |n#t $rYnwxYw|rEt j|r& t j |n#t$rYnwxYwt ddS)NFrz Termux voice recording cancelled) rrrrrr+r3rrrrr{r)rrs r cancelzTermuxAudioRecorder.cancel]s" Z " "'D#DO#'D !D   " " " " " " " " " " " " " " "    ' ' ) ) ) )    D   BGNN4((   $     677777s/155A AAB B&%B&c.|dSr/)rrs r shutdownzTermuxAudioRecorder.shutdownns rrNr/)__name__ __module__ __qualname____doc__supports_silence_autostoprpropertyrDrfloatrrjrrrrrrzrrrrr rrs"LL %dX3333X3 !S!!!X!*6*6*6*6*6Xaaaa hsm88888"rrceZdZdZdZddZedefdZede fdZ ede fdZ dd Z ddd Zdd eddfd ZdeefdZddZddZedefdZdS) AudioRecorderaThread-safe audio recorder using sounddevice.InputStream. Usage:: recorder = AudioRecorder() recorder.start(on_silence_stop=my_callback) # ... user speaks ... wav_path = recorder.stop() # returns path to WAV file # or recorder.cancel() # discard without saving If ``on_silence_stop`` is provided, recording automatically stops when the user is silent for ``silence_duration`` seconds and calls the callback. TrNcFtj|_d|_g|_d|_d|_d|_d|_d|_ d|_ d|_ d|_ d|_ d|_d|_t |_t$|_d|_d|_d|_dS)NFrrdg.@r)rrr_stream_framesrr _has_spoken _speech_start _dip_start_min_speech_duration_max_dip_tolerance_silence_start _resume_start_resume_dip_start_on_silence_stopSILENCE_RMS_THRESHOLD_silence_thresholdSILENCE_DURATION_SECONDS_silence_duration _max_wait _peak_rmsrrs r rzAudioRecorder.__init__s^%%   "$ "% $'!$+.!),%($'(+ $'<(@ $!"rcJ|jsdStj|jz Srrrs r rzAudioRecorder.elapsed_secondsrrc|jS)zBCurrent audio input RMS level (0-32767). Updated each audio chunk.rrs r rzAudioRecorder.current_rmss   rc|jS)z,Whether audio recording is currently active.rrs r rzAudioRecorder.is_recordings rctjdSt\}fd}d} |ttt |}|nN#t$rA}|& |n#t$rYnwxYwtd|d|d}~wwxYw|_dS)a_Create the audio InputStream once and keep it alive. The stream stays open for the lifetime of the recorder. Between recordings the callback simply discards audio chunks (``_recording`` is ``False``). This avoids the CoreAudio bug where closing and re-opening an ``InputStream`` hangs indefinitely on macOS. Nc n |rtd| jsdS j|t  | j dz}| _ | j kr| _ j ktj}| jz }| jkrd _ jdkr| _nD js=| jz jkr*d _td| jz  jsd _nd _ jdkr| _n| jz jkrd _d _n jr@ jdkr4 jdkr| _n| jz jkrd _d _nb jdkrW jdkr| _nD| jz jkr1td| jz d _d _d} jrT| jkrI jdkr| _nj| jz jkr"td jd}n4 js-| jkr"td jd}|r` j5 j d _ dddn #1swxYwY r3 fd }t=j|d  dSdSdSdS) Nzsounddevice status: %srWrTz(Speech confirmed (%.2fs above threshold)rz'Speech attempt reset (dip lasted %.2fs)Fz'Silence detected (%.1fs), auto-stoppingz%No speech within %.0fs, auto-stoppingc dS#t$r(}td|dYd}~dSd}~wwxYw)NzSilence callback failed: %sT)exc_info)r+r{error)rcbs r _safe_cbzAAudioRecorder._ensure_stream.._callback.._safe_cbsg^ "#,^^^ & -JAX\ ] ] ] ] ] ] ] ] ]^s  A<Atargetdaemon)!r{r|rrrFcopyrjsqrtmeanrqfloat64rrrrurvrrrrrrrrrrrrrrrThreadr) indataframes time_infostatusrmsnowelapsed should_firerrr rs @r _callbackz/AudioRecorder._ensure_stream.._callbacks ? 5v>>>?  L   . . .bggbggfmmBJ&?&?1&DEEFFGGC #D T^##!$$0n&& 00000&)DO)S00-0**!-?#8J2JdNg2g2g+/( %O%(4+=%=???  + 5.1++ 25.-4414D.. 4#559RRR25D/14D.%.)A--1S8858D22 4#99T=TTT14D.58D2'!++#--*-t.$2III %N%(4?%:<<<-0**- $ # 't/F(F(F*c11.1++t22d6LLL $M$($:<<<&* )'g.G.GKK G $000"&K O55!204-555555555555555O^^^^^ "(FFFLLNNNNN_10J O OOOsK55K9<K9)rhchannelsrfcallbackz#Failed to open audio input stream: z6. Check that a microphone is connected and accessible.) rr InputStreamrkrDTYPErr+closer)rr rstreamrr s` @r _ensure_streamzAudioRecorder._ensure_streams# < # FB^ O^ O^ O^ O^ O^ OB ^^&!" $F LLNNNN   !LLNNNN DGaGGG   s;._do_closelsH       s (- ;;Trrug?r%u:Audio stream close timed out after %.1fs — forcing ahead) rrrr __import__rvis_aliverr{warning)rr%r rrrs @r _close_stream_with_timeoutz(AudioRecorder._close_stream_with_timeoutds$ <  F        Id ; ; ;  f%%//11G;jjll z&11;;==HH FF3F   jjll z&11;;==HH ::<< b NNWY` a a a a a b brc|j5|js ddddSd|_d|_|js ddddSt \}}||jd}g|_t j|jz }t d|t|ttdz}t||kr6tdt| ddddS|jt kr4t d|jt  ddddS||cdddS#1swxYwYdS) uStop recording and write captured audio to a WAV file. The underlying stream is kept alive for reuse — only frame collection is stopped. Returns: Path to the WAV file, or ``None`` if no audio was captured. NFr)axisz+Voice recording stopped (%.1fs, %d samples)rdz,Recording too short (%d samples), discardingz2Recording too quiet (peak RMS=%d < %d), discarding)rrrrr rsrurvrr{rlenrjrkr|rr _write_wav)rrRr audio_datar min_sampless r rzzAudioRecorder.stop|s.Z / /?  / / / / / / / /$DO !D <  / / / / / / / /"OOEAr 1==JDLn&&)99G KKEwPST^P_P_ ` ` `kC/00K:,, KSQ[__]]]/ / / / / / / / /6~ 555 P N,ACCC= / / / / / / / /@??:..A / / / / / / / / / / / / / / / / / /s) E>E>C E>7E>E>>FFc|j5d|_g|_d|_d|_dddn #1swxYwYt ddS)zoStop recording and discard all captured audio. The underlying stream is kept alive for reuse. FNrzVoice recording cancelled)rrrrrr{rrs r rzAudioRecorder.cancels Z " "#DODL$(D ! !D   " " " " " " " " " " " " " " "  /00000s 155c|j5d|_g|_d|_dddn #1swxYwY|t ddS)z%продолжение следует(продолжение следует...*ご視聴ありがとうございました,sottotitoli creati dalla comunità amara.org5sous-titres réalisés par la communauté d'amara.orgbye.the endthe end. amara.org thank you thank you. sous-titres www.mooji.orgplease subscribeplease subscribe.like and subscribelike and subscribe.thanks for watchingthanks for watching.thank you for watchingsubscribe to my channelthank you for watching.subscribe to my channel.untertitel von stephanie geigesbyeyouz9^(?:thank you|thanks|bye|you|ok|okay|the end|\.|\s|,|!)+$)flags transcriptc|}|sdS|dtvs |tvrdSt|rdSdS)zBCheck if a transcript is a known Whisper hallucination on silence.Tz.!F)rrJrstripWHISPER_HALLUCINATIONS_HALLUCINATION_REPEAT_REmatch)rDcleaneds r is_whisper_hallucinationrKst  &&((G t~~d555DZ9Z9Zt%%g..t 5rr#modelcddlm}|||}|drJt|ddr'td|ddddd S|S) aTranscribe a WAV recording using the existing Whisper pipeline. Delegates to ``tools.transcription_tools.transcribe_audio()``. Filters out known Whisper hallucinations on silent audio. Args: wav_path: Path to the WAV file. model: Whisper model name (default: from config or ``whisper-1``). Returns: Dict with ``success``, ``transcript``, and optionally ``error``. r)transcribe_audio)rLsuccessrDr'z"Filtered Whisper hallucination: %rT)rOrDfiltered)tools.transcription_toolsrNr5rKr{r)r#rLrNr,s r transcribe_recordingrRs;:::::  he 4 4 4Fzz)E!9&**\SU:V:V!W!WE 8&:NOOOrtDDD Mr_active_playbackcpt5t}dadddn #1swxYwY|rT|@ |tdn#t $rYnwxYw t\}}|dS#t $rYdSwxYw)z/Interrupt the currently playing audio (if any).NzAudio playback interrupted) _playback_lockrSpoll terminater{rr+r rz)procr rRs r stop_playbackrY7s                    #  NN    KK4 5 5 5 5    D  A       s0 "".A// A<;A<%B'' B54B5 file_pathcTtj|std|dS|dr t \}}tj|d5}| | }| ||j }| }dddn #1swxYwY|||t||z }t!j|zdz}|r|jrot!j|krXt!jd |r0|jrt!j|kX|d S#t,t.f$rYn1t0$r%} td | Yd} ~ nd} ~ wwxYwt5j} g} | d kr| d |g| ddddd|g| dkr| dd|g| D]}} t;j| d} | r] t?j | t>j!t>j!}tD5|a#dddn #1swxYwY|$dtD5da#dddn #1swxYwYd S#t>j%$rntd| d|&|$tD5da#dddn #1swxYwYY#t0$rN} td| d| tD5da#dddn #1swxYwYYd} ~ ud} ~ wwxYwtd|dS)ayPlay an audio file through the default output device. Strategy: 1. WAV files via ``sounddevice.play()`` when available. 2. System commands: ``afplay`` (macOS), ``ffplay`` (cross-platform), ``aplay`` (Linux ALSA). Playback can be interrupted by calling ``stop_playback()``. Returns: ``True`` if playback succeeded, ``False`` otherwise. zAudio file not found: %sFrrbreNrgrirbTzsounddevice playback failed: %sDarwinafplayffplayz-nodispz -autoexitz -loglevelquietLinuxaplayrr)r*ri,r z+System player %s timed out, killing processzSystem player %s failed: %sz No audio player available for %s)'r3rrr{rendswithr rrH readframes getnframes frombufferrV getframeratertrrurvrwrxryrzrrr+r|platformsystemrFrrr(PopenDEVNULLrUrSwaitTimeoutExpiredkill)rZr r r$rr sample_rate duration_secsrrriplayerscmdexerXs r play_audio_filertKs> 7>>) $ $19===u&!!? ?"__FB9d++ 0rr}}77]]6]BB  oo//  0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 GGJ;G 7 7 7 OOk9M~''-7#=H--// !bmmoo&< !AQAQT\A\A\ 4   --// !bmmoo&< !AQAQT\A\A\ GGIII4W%    D ? ? ? LL:A > > > > > > > > ?_  FG ),--- NNHik7IVWWW y1222,,l3q6""  , ,!'J4FzOabbb#,,'+$,,,,,,,,,,,,,,, # &&&#,,'+$,,,,,,,,,,,,,,,tt, , , ,LcRSfUUU  #,,'+$,,,,,,,,,,,,,,, , , , :CFAFFF#,,'+$,,,,,,,,,,,,,,, , ,( NN5yAAA 5s&G:AC GC""G%C"&C5GH0 H9HH/2L5!K0$ L50K4 4L57K4 8 L5L' L5'L+ +L5.L+ /L55APN# P#N' 'P*N' +P1 P:)P#O2& P2O6 6P9O6 :PPcddlm}m}m}|}||}||}|o|dk}g}t }t p|} | s|ddgt} | o |o| d} g} |r| dn=| r| dn%| d td |s| d ni|d kr| d nM|dkr| dn1|dkr| dn| d| dD]} | d| | dgD]}| d|| | ||d | | dS)zCheck if all voice mode requirements are met. Returns: Dict with ``available``, ``audio_available``, ``stt_available``, ``missing_packages``, and ``details``. r) _get_provider_load_stt_configis_stt_enablednonerr rAz)Audio capture: OK (Termux:API microphone)zAudio capture: OKzAudio capture: MISSING (r@z5STT provider: DISABLED in config (stt.enabled: false)localz'STT provider: OK (local faster-whisper)groqzSTT provider: OK (Groq)openaizSTT provider: OK (OpenAI)z`STT provider: MISSING (pip install faster-whisper, or set GROQ_API_KEY / VOICE_TOOLS_OPENAI_KEY)rBz Environment: rC )rAaudio_available stt_availablemissing_packagesr environment) rQrvrwrxr0rextendrTrFrr5r)rvrwrx stt_config stt_enabled stt_providerrmissingrP has_audio env_checkrA details_partsrnotices r check_voice_requirementsrstZYYYYYYYYY!!##J .,,K =,,L:LF$:MG466N ""4nI 1 w/000)**IFmF +0FIMZHIIII Z01111X8S8U8UXXXYYY   TUUUU  FGGGG   67777  ! !89999 <   Z(886W667777-- 2..775V556666$&#99]++   rmax_age_secondsc"tjtsdSd}t j}tjtD]}|r|jdrl|j drR || j z }||krtj |j|dz }#t$rYwxYw|rtd||S)zRemove old temporary voice recording files. Args: max_age_seconds: Delete files older than this (default: 1 hour). Returns: Number of files deleted. rrrrUz"Cleaned up %d old voice recordings)r3risdirrruscandiris_filename startswithrcstatst_mtimerrr{r|)rdeletedrentryages r cleanup_temp_recordingsrs 7== # #qG )++CI&& ==?? uz44\BB uzGZGZ[aGbGb  EJJLL11((Iej)))qLG    D 97CCC Ns AC!! C.-C.)r[r\rUr/r)r)@rloggingr3rhrerr(rtempfilerrurtypingrrrr getLoggerrr{r rDrrGrrrrrr-r0dictrTrkrrrrrrr gettempdirrrjrrrrr(rGcompile IGNORECASErHrKrRrSrj__annotations__rrUrYrtrrrrr rs       ,,,,,,,,,,,,  8 $ $$A@@@@@+S++++ 4HSM4444 4     TTTTT\$\\\\B   GLL,,.. ? ? &4&4&4e&43&4t&4&4&4&4Xzzzzzzzz@\\\\\\\\~ }/BB>&2:@ -      "3x}SRUX<04(:+,333!!    (KsKtKKKKb<$sCx.<<<<DSCr