o i] @sddlZddlZddlmZddlmZmZmZddlm Z m Z m Z m Z m Z mZmZmZmZmZmZddlmZmZmZmZmZmZmZmZdZdZdZgd Z d d e DZ!d d e DZ"Gdddej#Z$Gdddej#Z%Gdddej#Z&Gdddej#Z'Gdddej#Z(Gdddej#Z)ej*jdddddej+j,dd ej*jd!d"d#ddej+j,d$d ej-jd%d&d'd(ej*jd)d#d#d*dej+j,d+d gZ.Gd,d-d-ej#Z/d.e0d/e1fd0d1Z2Gd2d3d3ej#Z3Gd4d5d5eZ4d/e4fd6d7Z5dS)8N)override)IOComfyExtensionInput) AddVoiceRequestAddVoiceResponse DialogueInputDialogueSettingsSpeechToSpeechRequestSpeechToTextRequestSpeechToTextResponseTextToDialogueRequestTextToSoundEffectsRequestTextToSpeechRequestTextToSpeechVoiceSettings) ApiEndpointaudio_bytes_to_audio_inputaudio_ndarray_to_bytesio"audio_tensor_to_contiguous_ndarraysync_op sync_op_rawupload_audio_to_comfyapivalidate_stringELEVENLABS_MUSIC_SECTIONSELEVENLABS_COMPOSITION_PLANELEVENLABS_VOICE))CwhRBWXzGAHq8TQ4Fs17Rogermaleamerican)EXAVITQu4vr4xnSDxMaLSarahfemaler)FGY2WhTYpPnrIDTdsKH5Laurar"r)IKne3meq5aSn9XLyUdCDCharlier australian)JBFqnCBsd6RMkjVDRZzbGeorgerbritish)N2lVS1w4EtoT3dr4eOWOCallumrr)SAz9YHcvj6GT2YYXdXwwRiverneutralr)SOYHLrjzK2X1ezoPC6crHarryrr)TX3LPaxmHKxFdv7VOQHJLiamrr)Xb7hH8MSUJpSbSDYk0k2Alicer"r*)XrExE9yKIg1WjnnlVkGXMatildar"r)bIHbv24MWmeRgasZH58oWillrr)cgSgspJ2msm6clMCkdW9Jessicar"r)cjVigY5qzO86Huf0OWalEricrr)hpp4J3VqNfWAUOO0d1UsBellar"r)iP95p4xoKVk53GoZ742BChrisrr)nPczCjzI2devNBz1zQrbBrianrr)onwK4e9ZLuTAKqWW03F9Danielrr*)pFZP5JQG7iQjIQuC4BkuLilyr"r*)pNInz6obpgDQGcFmaJgBAdamrr)pqHfZKP75CvOlQylNhV4BillrrcCs*g|]\}}}}|d|d|dqSz (z, )).0_namegenderaccentrNrN>/mnt/c/Users/fbmor/ComfyUI/comfy_api_nodes/nodes_elevenlabs.py <s*rUcCs,i|]\}}}}|d|d|d|qSrLrN)rOvoice_idrQrRrSrNrNrT =s rWc @sJeZdZedejfddZedejde de de de dej f d d Z d S) ElevenLabsSpeechToTextreturncCs0tjddddtjjdddtjjdtjd tjjd d d d tjjdd dd tjjdddddtjj ddtjjdddddtjj ddtj jdgddddggd d!tj jd"d#d$d tj jd%d&d&d'tjj d(d)tj jd*d+d&d,d-d.gtj j d/d0tj j d"d0tj j d1d0gtjjtjjtjjgd2tjd3d4d5 S)6NrXzElevenLabs Speech to Textapi node/audio/ElevenLabsznTranscribe audio to text. Supports automatic language detection, speaker diarization, and audio event tagging.audiozAudio to transcribe.tooltipmodel scribe_v2tag_audio_eventsFz=Annotate sounds like (laughter), (music), etc. in transcript.defaultr]diarizez"Annotate which speaker is talking.diarization_thresholdg)\(?皙?g?{Gz?zSSpeaker separation sensitivity. Lower values are more sensitive to speaker changes.rbminmaxstep display_moder] temperatureg@zNRandomness control. 0.0 uses model default. Higher values increase randomness.timestamps_granularity)word characternoneroz&Timing precision for transcript words.)optionsrbr]zModel to use for transcription.rrr] language_codedISO-639-1 or ISO-639-3 language code (e.g., 'en', 'es', 'fra'). Leave empty for automatic detection. num_speakersr zHMaximum number of speakers to predict. Set to 0 for automatic detection.)rbrhrirkr]seed6Seed for reproducibility (determinism not guaranteed).rbrhrir]text display_name words_jsonTzL{"type":"usd","usd":0.0073,"format":{"approximate":true,"suffix":"/minute"}}expr node_idrcategory descriptioninputsoutputshidden is_api_node price_badge)rSchemaAudior DynamicComboOptionBooleanFloat NumberDisplaysliderComboStringIntOutputHiddenauth_token_comfy_orgapi_key_comfy_org unique_id PriceBadgeclsrNrNrT define_schemaCs  +/ K  z$ElevenLabsSpeechToText.define_schemar[r^rtrwryc s|dr |r tdt|dt||ddddIdH|r |nd|d|d kr*|nd|d |d|dr9|d nd||d d }t|tdddt|ddIdH}tj|j r^dd|j Dngdd}t |j |j |S)NrczxNumber of speakers cannot be specified when diarization is enabled. Either disable diarization or set num_speakers to 0.r^mp4aac audio/mp4)container_format codec_name mime_typer`rrnrdrl) model_idcloud_storage_urlrtr`rwrnrcrdryrlz#/proxy/elevenlabs/v1/speech-to-textPOSTpathmethodmultipart/form-data)response_modeldata content_typecSsg|]}|jddqS)T exclude_none) model_dump)rOwrNrNrTrUsz2ElevenLabsSpeechToText.execute..)indent) ValueErrorr rstriprrr jsondumpswordsr NodeOutputr~rt) rr[r^rtrwryrequestresponserrNrNrTexecutes>     zElevenLabsSpeechToText.executeN)__name__ __module__ __qualname__ classmethodrrrrrdictstrintrrrNrNrNrTrXBs"arXc@s8eZdZedejfddZededejfddZ dS)ElevenLabsVoiceSelectorrYc Cs8tjddddtjjdtddgttjddgd d S) NrzElevenLabs Voice SelectorrZzCSelect a predefined ElevenLabs voice for text-to-speech generation.voicez5Choose a voice from the predefined ElevenLabs voices.rsrF)rrrrrrr)rrrrELEVENLABS_VOICE_OPTIONSCustomrrrrNrNrTrsz%ElevenLabsVoiceSelector.define_schemarcCs&t|}|std|t|S)NzUnknown voice: )ELEVENLABS_VOICE_MAPgetrrr)rrrVrNrNrTrs  zElevenLabsVoiceSelector.executeN) rrrrrrrrrrrNrNrNrTrs rc@sTeZdZedejfddZededededede d ed e d edej fd d Z dS)ElevenLabsTextToSpeechrYcCstjddddttjdddtjjdd d d d tjjd ddddtjjddtj jdgdddtj jdtj dtjjdddddtjjddtjjdddddtjjd dtj jd!d"d#d$tjjd%ddd&dtjjd'dgtj d(tjjdddddtjjddtjjdddddtjjd dggd)dtjjd*d d+d$tj jd,d-d.d/d0d1tj jd2d3d4gd5dgtjgtjjtjjtjjgd tjd6d7d8 S)9NrzElevenLabs Text to SpeechrZzConvert text to speech.rzVVoice to use for speech synthesis. Connect from Voice Selector or Instant Voice Clone.r\r~TruzThe text to convert to speech. multilinerbr] stability?rm?rfVoice stability. Lower values give broader emotional range, higher values produce more consistent but potentially monotonous speech.rgapply_text_normalizationautoonoffjText normalization mode. 'auto' lets the system decide, 'on' always applies normalization, 'off' skips it.rsr^eleven_multilingual_v2speedffffff??6Speech speed. 1.0 is normal, <1.0 slower, >1.0 faster.similarity_boost?LSimilarity boost. Higher values make the voice more similar to the original.use_speaker_boostF/Boost similarity to the original speaker voice.rastyle皙?YStyle exaggeration. Higher values increase stylistic expression but may reduce stability. eleven_v3z Model to use for text-to-speech.rtrvryrzrr{r|r} output_format mp3_44100_192opus_48000_192Audio output format.L{"type":"usd","usd":0.24,"format":{"approximate":true,"suffix":"/1K chars"}}rr)rrrrrrrrrrrrrrrrrrrrrrrNrNrTrs    & ?Crz$ElevenLabsTextToSpeech.define_schemarr~rrr^rtryrc st|ddt||d|r|ndt||d|d|dd|ddd||d } t|td |d d |id | ddIdH} tt | S)Nrz min_lengthr^rrrr)rrrrr)r~rrtvoice_settingsryrz$/proxy/elevenlabs/v1/text-to-speech/rrrr query_paramsTr as_binary) rrrrrrrrrr) rrr~rrr^rtryrrrrNrNrTrss6     zElevenLabsTextToSpeech.executeN) rrrrrrrrfloatrrrrrNrNrNrTrs0  rc@s:eZdZedejfddZedejdej fddZ dS)ElevenLabsAudioIsolationrYc CsLtjddddtjjdddgtjgtjjtjjtjjgdtj d d d S) NrzElevenLabs Voice IsolationrZz?Remove background noise from audio, isolating vocals or speech.r[z.Audio to process for background noise removal.r\TJ{"type":"usd","usd":0.24,"format":{"approximate":true,"suffix":"/minute"}}rr) rrrrrrrrrrrrNrNrTrs*z&ElevenLabsAudioIsolation.define_schemar[csXt|d}t||ddd}t|tddddd |d fid d d IdH}tt|S)Nwaveform sample_raterrz$/proxy/elevenlabs/v1/audio-isolationrrr[ audio.mp4rrT)filesrr)rrrrrrr)rr[ audio_data_npaudio_bytes_iorrNrNrTrs    z ElevenLabsAudioIsolation.executeN) rrrrrrrrrrrrNrNrNrTrsrc @s@eZdZedejfddZededededej fddZ d S) ElevenLabsTextToSoundEffectsrYcCstjddddtjjddddd tjjd tjd tjjd d dddtjjddtj jddddtjjdddddtjjddggddtj jdd d!gd"dgtj gtj jtj jtj jgdtjd#d$d% S)&Nrz ElevenLabs Text to Sound EffectsrZz.Generate sound effects from text descriptions.r~Truz1Text description of the sound effect to generate.rr^ eleven_sfx_v2durationg@rg>@rez'Duration of generated sound in seconds.rgloopFz'Create a smoothly looping sound effect.raprompt_influenceg333333?rmrrfzeHow closely generation follows the prompt. Higher values make the sound follow the text more closely.z)Model to use for sound effect generation.rsrrrrzJ{"type":"usd","usd":0.14,"format":{"approximate":true,"suffix":"/minute"}}rr)rrrrrrrrrrrrrrrrrrrrNrNrTrsx #1z*ElevenLabsTextToSoundEffects.define_schemar~r^rc s\t|ddt|tddd|idt||d|d|d dd d d IdH}tt|S) Nrzrz%/proxy/elevenlabs/v1/sound-generationrrrrrr)r~duration_secondsrrTr)rrrrrrrr)rr~r^rrrNrNrTrs$   z$ElevenLabsTextToSoundEffects.executeN) rrrrrrrrrrrrNrNrNrTrsDrc@s@eZdZedejfddZedejjde dej fddZ dS) ElevenLabsInstantVoiceClonerYc Cs~tjddddtjjdtjjtjddddd d d tjjd d ddgttj ddgtj j tj j tj j gdtjddd S)NrzElevenLabs Instant Voice ClonerZz]Create a cloned voice from audio samples. Provide 1-8 audio recordings of the voice to clone.rr[rz)prefixrhriz#Audio recordings for voice cloning.)templater]remove_background_noiseFzARemove background noise from voice samples using audio isolation.rarrTz{"type":"usd","usd":0.15}rr)rrAutogrowrTemplatePrefixrrrrrrrrrrrrNrNrTr)s<   z)ElevenLabsInstantVoiceClone.define_schemarrc sg}|D](}||}|d}|d}t|}t||dd} |d|d| dffqt|tdd d tttt |d |d d IdH} t | j S)Nrrrrrz.mp4rz/proxy/elevenlabs/v1/voices/addrr)rQrr)rrrr)rrappendgetvaluerrrrruuiduuid4rrrV) rrr file_tupleskeyr[rrrrrrNrNrTrNs*    z#ElevenLabsInstantVoiceClone.executeN) rrrrrrrrTypeboolrrrNrNrNrTr(s$rrrrrrfrrgrrrmrrFrrarrrc@sReZdZedejfddZededej de de ded e d e dejfd d Zd S)ElevenLabsSpeechToSpeechrYcCstjddddttjdddtjjdd dtjjd d d d dtjjddtj jdtj dt tj dt gddtj jdddgddtj jddddddtjjdd d!d"gtjgtjjtjjtjjgd#tjd$d%d& S)'NrzElevenLabs Speech to SpeechrZz]Transform speech from one voice to another while preserving the original content and emotion.rzXTarget voice for the transformation. Connect from Voice Selector or Instant Voice Clone.r\r[zSource audio to transform.rrrmrrfrrgr^eleven_multilingual_sts_v2eleven_english_sts_v2z1Model to use for speech-to-speech transformation.rsrrrrryrSeed for reproducibility.r}rFz?Remove background noise from input audio using audio isolation.raTrrr)rrrrrrrrrrrELEVENLABS_STS_VOICE_SETTINGSrrrrrrrrrrrNrNrTrs   5z&ElevenLabsSpeechToSpeech.define_schemarr[rr^rryrc st|d}t||ddd} t||d|d|d|dd } t|td |d d |id t|d| jdd||ddd| dfidddIdH} t t | S)Nrrrrrrrr)rrrrrz&/proxy/elevenlabs/v1/speech-to-speech/rrrr^Tr)rrryrr[rrr)rrrr) rrrrrr model_dump_jsonr rrr) rrr[rr^rryrrrrrrNrNrTrs8   z ElevenLabsSpeechToSpeech.executeN)rrrrrrrrrrrrrrrrrNrNrNrTrs*H rcountrYc Csfg}td|dD]'}|tjjd|ddd|ddttjd|d |d d gq |S) z>Generate input widgets for a given number of dialogue entries.rzr~Truz Text content for dialogue entry .rrzVoice for dialogue entry z5. Connect from Voice Selector or Instant Voice Clone.r\)rangeextendrrrrr)rrirNrNrT_generate_dialogue_inputss    rc@sPeZdZedejfddZedededede ded e d edej fd d Z d S)ElevenLabsTextToDialoguerYcCshtjddddtjjdddddtjjd d tjjd gd d dtjjddgddtjjdtjdt dtjdt dtjdt dtjdt dtjdt dtjdt dtjdt d tjd!t d"tjd#t d$tjd%t d&g d'dtj jd(d)d*d+tj jd,dd-d.d/d0tjjd1d2d3gd4dgtj gtjjtjjtjjgd5tjd6d7d8 S)9NrzElevenLabs Text to DialoguerZzZGenerate multi-speaker dialogue from text. Each dialogue entry has its own text and voice.rrrmrrrgrrrrsr^rz%Model to use for dialogue generation.r1rz2r345678r9 10 zNumber of dialogue entries.rtrurvraryrrrr}rrrrTrrr)rrrrrrrrrrrrrrrrrrrrrNrNrTrs  :z&ElevenLabsTextToDialogue.define_schemarrr^rrtryrcst|d}g} td|dD]} |d| } |d| } t| dd| t| | dqt| ||r9|ndt|d||d} t|t d d d |id | d dIdH}t t |S)Nrrzr~rr)r~rV)r)rrrtsettingsryrz%/proxy/elevenlabs/v1/text-to-dialoguerrrTr) rrrr rr rr rrrrr)rrrr^rrtryr num_entriesdialogue_inputsrr~rVrrrNrNrTres6    z ElevenLabsTextToDialogue.executeN) rrrrrrrrrrrrrrNrNrNrTrs*M rc@s(eZdZedeeejfddZdS)ElevenLabsExtensionrYcsttttttttgSN)rXrrrrrrr)selfrNrNrT get_node_listsz!ElevenLabsExtension.get_node_listN) rrrrlisttyper ComfyNoder7rNrNrNrTr4sr4cstSr5)r4rNrNrNrTcomfy_entrypointsr;)6rr typing_extensionsrcomfy_api.latestrrrcomfy_api_nodes.apis.elevenlabsrrrr r r r r rrrcomfy_api_nodes.utilrrrrrrrrrrrELEVENLABS_VOICESrrr:rXrrrrrrrrrrrrr8rrr4r;rNrNrNrTsx 4(  0-`D  $rv