o hK9@sddlZddlmZmZmZmZddlmZmZm Z m Z ddl m Z m Z er+ddlZe r2ddlZe eZGdddZee d Gd d d e ZdS) N)AnyDictListUnion)add_end_docstringsis_tf_availableis_torch_availablelogging)PIPELINE_INIT_ARGSPipelinec@seZdZdZ d*deeeeeeffdej fddZ ddZ d eeeffd d Z d+d ede fddZdefddZddZddZddZddZddZddZd d!Zed"d#Zed$d%Zed&d'Zed(d)ZdS), Conversationa Utility class containing a conversation and its history. This class is meant to be used as an input to the [`ConversationalPipeline`]. The conversation contains several utility functions to manage the addition of new user inputs and generated model responses. Arguments: messages (Union[str, List[Dict[str, str]]], *optional*): The initial messages to start the conversation, either a string, or a list of dicts containing "role" and "content" keys. If a string is passed, it is interpreted as a single message with the "user" role. conversation_id (`uuid.UUID`, *optional*): Unique identifier for the conversation. If not provided, a random UUID4 id will be assigned to the conversation. Usage: ```python conversation = Conversation("Going to the movies tonight - any suggestions?") conversation.add_message({"role": "assistant", "content": "The Big lebowski."}) conversation.add_message({"role": "user", "content": "Is it good?"}) ```Nmessagesconversation_idc Ks|st}|dur|dd}|durd|dg}ng}n t|tr)d|dg}d|_|dd}|dd}|durD|durDtd|durg}|durPg}ttt |t |gD]"}|t |krn| d||d|t |kr~| d||dq\||}||_||_ dS) Ntextuserrolecontentrgenerated_responsespast_user_inputsz>generated_responses cannot be passed without past_user_inputs! assistant) uuiduuid4pop isinstancestr_num_processed_user_inputs ValueErrorrangemaxlenappendr) selfrrdeprecated_kwargsrrrlegacy_messagesir([/var/www/html/ai/venv/lib/python3.10/site-packages/transformers/pipelines/conversational.py__init__(s6        zConversation.__init__cCs&t|tsdS|j|jkp|j|jkS)NF)rrrr)r$otherr(r(r)__eq__Ms zConversation.__eq__messagecCs@t|ddhkstd|ddvrtd|j|dS)Nrrz6Message should contain only 'role' and 'content' keys!)rrsystemzBOnly 'user', 'assistant' and 'system' roles are supported for now!)setkeysrrr#r$r-r(r(r) add_messageRs  zConversation.add_messageFr overwritecCst|dkr:|dddkr:|r(td|ddd|d||dd<d Std|ddd |d d S|jd|d d S) a3 Add a user input to the conversation for the next round. This is a legacy method that assumes that inputs must alternate user/assistant/user/assistant, and so will not add multiple user messages in succession. We recommend just using `add_message` with role "user" instead. rrrz8User input added while unprocessed input was existing: "rz" was overwritten with: "z".z" new input ignored: "z>". Set `overwrite` to True to overwrite unprocessed user inputrN)r"loggerwarningrr#)r$rr3r(r(r)add_user_inputYszConversation.add_user_inputresponsecCs|jd|ddS)zr This is a legacy method. We recommend just using `add_message` with an appropriate role instead. rrN)rr#)r$r8r(r(r)append_responsenszConversation.append_responsecCst|j|_dS)a) This is a legacy method, as the Conversation no longer distinguishes between processed and unprocessed user input. We set a counter here to keep behaviour mostly backward-compatible, but in general you should just read the messages directly when writing new code. N)r"_user_messagesrr$r(r(r)mark_processedtszConversation.mark_processedccs|jD]}|VqdSNrr1r(r(r)__iter__|s zConversation.__iter__cCs |j|Sr=r>)r$itemr(r(r) __getitem__ zConversation.__getitem__cCs||j|<dSr=r>)r$keyvaluer(r(r) __setitem__szConversation.__setitem__cCs t|jSr=)r"rr;r(r(r)__len__rBzConversation.__len__cCs:d|jd}|jD]}||dd|dd7}q |S)a Generates a string representation of the conversation. Returns: `str`: Example: Conversation id: 7d15686b-dc94-49f2-9c4b-c9eac6a1f114 user: Going to the movies tonight - any suggestions? bot: The Big Lebowski zConversation id:  rz: r)rr)r$outputr-r(r(r)__repr__s zConversation.__repr__ccs(|jD] }|ddk|dfVqdS)Nrrrr>r1r(r(r) iter_textss zConversation.iter_textscCdd|jDS)NcS g|] }|ddkr|dqS)rrrr(.0r-r(r(r)  z/Conversation._user_messages..r>r;r(r(r)r:zConversation._user_messagescCs@|jsgS|jdddks|jt|jkr|jS|jddS)Nr4rr)r:rrr"r;r(r(r)rs "zConversation.past_user_inputscCrK)NcSrL)rrrr(rMr(r(r)rOrPz4Conversation.generated_responses..r>r;r(r(r)rrQz Conversation.generated_responsescCs |jdS)Nr4)r:r;r(r(r)new_user_inputs zConversation.new_user_input)NN)F)__name__ __module__ __qualname____doc__rrrrrUUIDr*r,r2boolr7r9r<r?rArErFrIrJpropertyr:rrrRr(r(r(r)rs6 %   ra  min_length_for_response (`int`, *optional*, defaults to 32): The minimum length (in number of tokens) for a response. minimum_tokens (`int`, *optional*, defaults to 10): The minimum length of tokens to leave for a response. cseZdZdZfddZ dddZddeeee ee fffd d Z dd e d ee e ffddZ dddZdddZZS)ConversationalPipelinea Multi-turn conversational pipeline. Example: ```python >>> from transformers import pipeline, Conversation # Any model with a chat template can be used in a ConversationalPipeline. >>> chatbot = pipeline(model="facebook/blenderbot-400M-distill") >>> # Conversation objects initialized with a string will treat it as a user message >>> conversation = Conversation("I'm looking for a movie - what's your favourite one?") >>> conversation = chatbot(conversation) >>> conversation.messages[-1]["content"] ' I don't really have a favorite movie, but I do like action movies. What about you?' >>> conversation.add_message({"role": "user", "content": "That's interesting, why do you like action movies?"}) >>> conversation = chatbot(conversation) >>> conversation.messages[-1]["content"] ' I think it's just because they're so fast-paced and action-fantastic.' ``` Learn more about the basics of using a pipeline in the [pipeline tutorial](../pipeline_tutorial) This conversational pipeline can currently be loaded from [`pipeline`] using the following task identifier: `"conversational"`. This pipeline can be used with any model that has a [chat template](https://huggingface.co/docs/transformers/chat_templating) set. cs2tj|i||jjdur|jj|j_dSdSr=)superr* tokenizer pad_token_id eos_token pad_token)r$argskwargs __class__r(r)r*s zConversationalPipeline.__init__NcKshi}i}i}|dur||d<|dur||d<d|vr |d|d<|dur(||d<|r/|||||fS)Nmin_length_for_responseminimum_tokens max_lengthclean_up_tokenization_spaces)update)r$rdrerggenerate_kwargspreprocess_paramsforward_paramspostprocess_paramsr(r(r)_sanitize_parameterss   z+ConversationalPipeline._sanitize_parametersr conversationsc st|trt|dtrt|}nt|tr$t|dtr$dd|D}tj|fd|i|}t|tr?t|dkr?|dS|S)a/ Generate responses for the conversation(s) given as inputs. Args: conversations (a [`Conversation`] or a list of [`Conversation`]): Conversation to generate responses for. Inputs can also be passed as a list of dictionaries with `role` and `content` keys - in this case, they will be converted to `Conversation` objects automatically. Multiple conversations in either format may be passed as a list. clean_up_tokenization_spaces (`bool`, *optional*, defaults to `False`): Whether or not to clean up the potential extra spaces in the text output. generate_kwargs: Additional keyword arguments to pass along to the generate method of the model (see the generate method corresponding to your framework [here](./model#generative-models)). Returns: [`Conversation`] or a list of [`Conversation`]: Conversation(s) with updated generated responses for those containing a new user input. rcSsg|]}t|qSr()r)rNconvr(r(r)rO sz3ConversationalPipeline.__call__.. num_workersr )rlistdictrr[__call__r")r$rnrpraoutputsrbr(r)rss zConversationalPipeline.__call__ conversationreturncCsH|jj|dd}|jdkrt|g}n |jdkrt|g}||dS)NT)add_generation_promptpttf) input_idsrv)r\apply_chat_template frameworktorch LongTensorrzconstant)r$rvrdr{r(r(r) preprocess&s     z!ConversationalPipeline.preprocess cKst|djd}|d}d|vrd|vrd|d<|jjdi||}|jjjr+d}n|}|dd|df|dS) Nr{r rvrfmax_new_tokens) output_idsrvr()shapermodelgenerateconfigis_encoder_decoder)r$ model_inputsrerinrvrstart_positionr(r(r)_forward/s  zConversationalPipeline._forwardTcCs:|d}|jj|dd|d}|d}|d|d|S)NrrT)skip_special_tokensrgrvrr)r\decoder2)r$ model_outputsrgranswerrvr(r(r) postprocess;sz"ConversationalPipeline.postprocess)NNN)r)ru)r)T)rSrTrUrVr*rmrrrrrsrrrrr __classcell__r(r(rbr)rZs  &  rZ)rtypingrrrrutilsrrr r baser r tensorflowrzr~ get_loggerrSr5rrZr(r(r(r)s  1