o c۷i4@sddlZddlmZddlmZddlmZddlmZddl m Z m Z m Z m Z ddlmZddlmZmZdd lmZdd lmZGd d d eeZGd ddeZGdddeZGdddeZGdddeZGdddeZGdddeZGdddeeZGdddeZ GdddeZ!Gdd d eZ"Gd!d"d"eZ#ee"eBeBe!Be Be#Be d#d$fZ$ee"eBeBe!Be Be d#d$fZ%d%e&eee&eefBfd&e$fd'd(Z'dS))N)Enum)Path)Literal)urlparse) ConfigDictFieldValidationErrorfield_validator) Annotated)EXPECTED_FORMAT_VALUESAudio) MistralBase)SerializableImagec@s(eZdZdZdZdZdZdZdZdZ dS) ChunkTypesaEnum for the types of chunks that can be sent to the model. Attributes: text: A text chunk. image: An image chunk. image_url: An image url chunk. input_audio: An input audio chunk. audio_url: An audio url chunk. Examples: >>> from mistral_common.protocol.instruct.chunk import ChunkTypes >>> chunk_type = ChunkTypes.text textimage image_url input_audio audio_urlthinkingN) __name__ __module__ __qualname____doc__rrrrrrrr\/home/ubuntu/vllm_env/lib/python3.10/site-packages/mistral_common/protocol/instruct/chunk.pyrsrc@seZdZUdZeejejejej ej ej fe d<de eee eefBffddZede eee eefBfddfddZd S) BaseContentChunkzBase class for all content chunks. Content chunks are used to send different types of content to the model. Attributes: type: The type of the chunk. typereturncCstdt|j)z_Converts the chunk to the OpenAI format. Should be implemented by subclasses. z%to_openai method not implemented for )NotImplementedErrorrrselfrrr to_openai8szBaseContentChunk.to_openai openai_chunkcCstd|j)zgConverts the OpenAI chunk to the Mistral format. Should be implemented by subclasses. z'from_openai method not implemented for )rrclsr#rrr from_openai?szBaseContentChunk.from_openaiN)rrrrrrrrrrrr__annotations__dictstrr" classmethodr&rrrrr&s " ,rc@seZdZUdZejZeejed<e ed<e ddZ de e e e e e fBffddZed e e e e e e fBfddfd d Zd S) ImageChunkzImage chunk. Attributes: image: The image to be sent to the model. Examples: >>> from PIL import Image >>> image_chunk = ImageChunk(image=Image.new('RGB', (200, 200), color='blue')) rrTarbitrary_types_allowedrcCs&|jdhddidd}dd|idS)(Converts the chunk to the OpenAI format.radd_format_prefixT)includecontextrurlrr model_dump)r! base64_imagerrrr"WszImageChunk.to_openair#cCsp|ddks J||d}t|trd|vsJ|td|dr/|ddd|d<|d|diS)0Converts the OpenAI chunk to the Mistral format.rrr2z^data:image/\w+;base64,,r)get isinstancer(rematchsplitmodel_validate)r%r#image_url_dictrrrr&\s zImageChunk.from_openaiN)rrrrrrrrr'rr model_configr(r)r"r*r&rrrrr+Hs   ",r+c@s*eZdZUdZeed<dZedBed<dS)ImageURLzImage URL or a base64 encoded image. Attributes: url: The URL of the image. detail: The detail of the image. Examples: >>> image_url = ImageURL(url="https://example.com/image.png") r2Ndetail)rrrrr)r'rCrrrrrBjs  rBc@seZdZUdZejZeejed<e e Bed<e ddZ de fddZ dee e ee e fBffd d Zed ee e ee e fBfddfd d ZdS) ImageURLChunkzImage URL chunk. Attributes: image_url: The URL of the image or a base64 encoded image to be sent to the model. Examples: >>> image_url_chunk = ImageURLChunk(image_url="data:image/png;base64,iVBORw0") rrTr,rcCt|jtr |jjS|jSN)r;rrBr2r rrrget_urls zImageURLChunk.get_urlcCs>d|i}t|jtr|jjdur|jj|d<d|d}|S)r.r2NrCrr3)rGr;rrBrC)r!r@out_dictrrrr"s  zImageURLChunk.to_openair#cCs|d|diS)r7rr?r$rrrr&szImageURLChunk.from_openaiN)rrrrrrrrr'rBr)rrArGr(r"r*r&rrrrrDys   " ,rDc@sVeZdZUdZeeBed<eed<ededdfddZ e ddedefd d Z d S) RawAudioaKBase64 encoded audio data. This class represents raw audio data encoded in base64 format. Attributes: data: The base64 encoded audio data, which can be a string or bytes. format: The format of the audio data. Examples: >>> audio = RawAudio(data="base64_encoded_audio_data", format="mp3") dataformataudiorcCs|j}||d}|||dS)zCreates a RawAudio instance from an Audio object. Args: audio: An Audio object containing audio data, format, and duration. Returns: An AudioChunk instance initialized with the audio data. FrKrL)rL to_base64)r%rMrLrKrrr from_audios  zRawAudio.from_audiovcCs"|tvrtdtd|d|S)Nz`format` should be one of z. Got: `)r rr%rQrrrshould_not_be_emptyszRawAudio.should_not_be_emptyN) rrrrr)bytesr'r*r rPr rTrrrrrJs  rJc@seZdZUdZeed<dS)AudioURLzHAudio URL. Attributes: url: The URL of the audio file. r2N)rrrrr)r'rrrrrVs  rVc@s eZdZdZdZdZdZdZdS) AudioURLTypezEnum for the types of audio URLs. Attributes: url: A URL. base64: A base64 encoded audio. Can be prefixed with `data:audio/;base64,`. file: A file path. file_uri: A file URI (eg. `file:///path/to/file`). r2base64filefile_uriN)rrrrr2rXrYrZrrrrrWs  rWc@seZdZUdZejZeejed<e e Bed<e de fddZ de fddZdee e ee e fBffd d Zed ee e ee e fBfddfd d ZdS) AudioURLChunkzAudio URL chunk. Attributes: type: The type of the chunk, which is always `ChunkTypes.audio_url`. audio_url: The URL of the audio file. rrrcCrErF)r;rrVr2r rrrr2s zAudioURLChunk.urlcCstt|jj}|dvr tjS|dkrtjS|dkrtjSz t|j}|}Wn ty1d}Ynw|r7tj StjS)aReturns the type of the audio URL. Note: URLs should be either: - a valid URL (http:// or https://) - a valid file path (e.g. /path/to/file) - a valid file URI (e.g. file:///path/to/file) - a base64 encoded audio. It is assumed to be base64 encoded if it is not a valid URL or file path. Returns: The type of the audio URL. >httphttpsrKrYF) rr2schemerWrXrZrexistsOSErrorrY)r! url_schemeurl_path exist_pathrrr get_url_types    zAudioURLChunk.get_url_typecCs&t|jtr |S|jd|jidS)r.r2)rr)r;rrVr5rr rrrr"s zAudioURLChunk.to_openair#cC ||Sr7rIr$rrrr& zAudioURLChunk.from_openaiN)rrrrrrrrr'r)rVpropertyr2rWrdr(r"r*r&rrrrr[s  " ,r[c@seZdZUdZejZeejed<e ed<e dde de fddZ e de ddfd d ZdeeeeeefBffd d Ze d eeeeeefBfddfddZdS) AudioChunkaAudio chunk containing raw audio data. This class represents a chunk of audio data that can be used as input. Attributes: type: The type of the chunk, which is always ChunkTypes.input_audio. input_audio: The RawAudio object containing the audio data. Examples: >>> audio_chunk = AudioChunk(input_audio=RawAudio(data="base64_encoded_audio_data", format="mp3")) rrrQrcCs|js td|d|S)Nz'`InputAudio` should not be empty. Got: rR)rKstriprrSrrrrT.s zAudioChunk.should_not_be_emptyrMcCs|t|dS)zCreates an AudioChunk instance from an Audio object. Args: audio: An Audio object containing audio data. Returns: An AudioChunk instance initialized with the audio data. )r)rJrP)r%rMrrrrP5s zAudioChunk.from_audiocCs@t|jjtr|jjdn|jj}|jt||jjddS)zConverts the chunk to the OpenAI format. Returns: A dictionary representing the audio chunk in the OpenAI format. zutf-8rN)rr) r;rrKrUdecoderrJrLr5)r!contentrrrr"As "zAudioChunk.to_openair#cCre)aConverts the OpenAI chunk to the Mistral format. Args: openai_chunk: A dictionary representing the audio chunk in the OpenAI format. Returns: An AudioChunk instance initialized with the data from the OpenAI chunk. rIr$rrrr&Os zAudioChunk.from_openaiN)rrrrrrrrr'rJr rTr*r rPr(r)r"r&rrrrris  " ,ric@szeZdZUdZejZeejed<e ed<de e e e e e fBffddZ e de e e e e e fBfddfdd Z d S) TextChunkzText chunk. Attributes: text: The text to be sent to the model. Examples: >>> text_chunk = TextChunk(text="Hello, how can I help you?") rrrcC|Sr.r4r rrrr"izTextChunk.to_openair#cCrerfrIr$rrrr&mrgzTextChunk.from_openaiN)rrrrrrrrr'r)r(r"r*r&rrrrrm\s  ",rmc@seZdZUdZejZeejed<e ed<e dddZ e ed<de e e e e e fBffd d Zed e e e e e e fBfddfd d ZdS) ThinkChunkzThinking chunk. Attributes: type: The type of the chunk, which is always ChunkTypes.thinking. thinking: The list of text chunks of the thinking. closed: Whether the thinking chunk is closed or not. rrTz,Whether the thinking chunk is closed or not.)default descriptionclosedrcCrnror4r rrrr"rpzThinkChunk.to_openair#cCrerfrIr$rrrr&rgzThinkChunk.from_openaiN)rrrrrrrrr'r)rrtboolr(r"r*r&rrrrrqss ",rqr) discriminatoropenai_content_chunksrcCs|d}|dur tdt|}|tjkrt|S|tjkr%t|S|tjkr/t |S|tj kr9t |S|tj krCt |S|tjkrMt|Std|)Nrz%Content chunk must have a type field.zUnknown content chunk type: )r: ValueErrorrrrmr&rrDrr+rrirr[rrq)rwcontent_type_str content_typerrr_convert_openai_content_chunkss"             r{)(r<enumrpathlibrtypingr urllib.parserpydanticrrrr typing_extensionsr mistral_common.audior r mistral_common.baser mistral_common.imagerr)rrr+rBrDrJrVrWr[rirmrq ContentChunkUserContentChunkr(r{rrrrs:       ""&' >> *