ޠj:tSrSSKJr SSKrSSKrSSKJrJrJrJ r J r SSK J r J r SSKrSSKrSSKJr SSKJrJr SSKJr SS KJrJrJr SS KJr SS KJrJrJ r SS K!J"r"J#r#J$r$J%r% SS K&J'r' SSK(J)r) SSK*J+r+ \RX"\-5r.SSjr/SSjr0"SS\5r1"SS\15r2g)zQBase classes for OpenAI large language models. Chat models are in `chat_models/`.) annotationsN) AsyncIteratorCallable CollectionIteratorMapping)AnyLiteral) deprecated)AsyncCallbackManagerForLLMRunCallbackManagerForLLMRun)BaseLLM) GenerationGenerationChunk LLMResult)get_pydantic_field_names)_build_model_kwargsfrom_envsecret_from_env) ConfigDictField SecretStrmodel_validator)Self) __version__) _PROFILEScURUS5nUH&nXB;a USUX$'MX$==USU- ss'M( g)zUpdate token usage.usageN) intersection)keysresponse token_usage _keys_to_use_keys T/var/www/html/gaurav/venv/lib/python3.13/site-packages/langchain_openai/llms/base.py_update_token_usager&sS$$Xg%67L  " ( 1$ 7K   '!24!8 8  cUS(d [SS9$[USSS=(d SUSSRSS5USSRSS5S .S 9$) z0Convert a stream response to a generation chunk.choices)textrr+ finish_reasonNlogprobsr,r-r+generation_info)rget)stream_responses r%$_stream_response_to_generation_chunkr3*su 9 %B''  Y ' *6 2 8b,Y7:>>PTU' 21599*dK  r'cl^\rSrSr%Sr\"SSS9rS\S'\"SSS9rS\S'\"S S S 9r S \S 'Sr S\S'Sr S\S'Sr S\S'Sr S\S'SrS\S'SrS\S'SrS\S'\"\S9rS\S'\"S\"SSS 9S!9rS"\S#'\"S$\"S%SS 9S!9rS&\S''\"S(\"S)S*/SS 9S!9rS&\S+'\"\"S,SS 9S9rS&\S-'S.rS\S/'\"SS0S 9rS1\S2'SrS3\S4'S5rS\S6'SrS7\S8'SrS7\S9'S:rS;\S<'\ "5r!S=\S>'S?r"S@\SA'Sr#S&\SB'Sr$SC\SD'Sr%SE\SF'Sr&SG\SH'Sr'SG\SI'Sr(SJ\SK'\)"SSL9r*\+"SMSN9\,SgSOj55r-\+"SPSN9ShSQj5r.\+"SPSN9ShSRj5r/\0SiSSj5r1SjSkSTjjr2SjSlSUjjr3SjSmSVjjr4SjSnSWjjr5SoSpSXjjr6SSY.SqSZjjr7\0SiS[j5r8\0SrS\j5r9\0SsS]j5r:StU4S^jjr;\<\="S_S`SaSb9SuScj55r>\0SvSdj5r?SwSejr@SfrAU=rB$)x BaseOpenAI9u3 Base OpenAI large language model class. Setup: Install `langchain-openai` and set environment variable `OPENAI_API_KEY`. ```bash pip install -U langchain-openai export OPENAI_API_KEY="your-api-key" ``` Key init args — completion params: model_name: Name of OpenAI model to use. temperature: Sampling temperature. max_tokens: Max number of tokens to generate. top_p: Total probability mass of tokens to consider at each step. frequency_penalty: Penalizes repeated tokens according to frequency. presence_penalty: Penalizes repeated tokens. n: How many completions to generate for each prompt. best_of: Generates best_of completions server-side and returns the "best". logit_bias: Adjust the probability of specific tokens being generated. seed: Seed for generation. logprobs: Include the log probabilities on the logprobs most likely output tokens. streaming: Whether to stream the results or not. Key init args — client params: openai_api_key: OpenAI API key. If not passed in will be read from env var `OPENAI_API_KEY`. openai_api_base: Base URL path for API requests, leave blank if not using a proxy or service emulator. Falls back to env var `OPENAI_API_BASE`, then to `OPENAI_BASE_URL` (read by the underlying SDK client). openai_organization: OpenAI organization ID. If not passed in will be read from env var `OPENAI_ORG_ID`. request_timeout: Timeout for requests to OpenAI completion API. max_retries: Maximum number of retries to make when generating. batch_size: Batch size to use when passing multiple documents to generate. See full list of supported init args and their descriptions in the params section. Instantiate: ```python from langchain_openai.llms.base import BaseOpenAI model = BaseOpenAI( model_name="gpt-3.5-turbo-instruct", temperature=0.7, max_tokens=256, top_p=1, frequency_penalty=0, presence_penalty=0, # openai_api_key="...", # openai_api_base="...", # openai_organization="...", # other params... ) ``` Invoke: ```python input_text = "The meaning of life is " response = model.invoke(input_text) print(response) ``` ```txt "a philosophical question that has been debated by thinkers and scholars for centuries." ``` Stream: ```python for chunk in model.stream(input_text): print(chunk, end="") ``` ```txt a philosophical question that has been debated by thinkers and scholars for centuries. ``` Async: ```python response = await model.ainvoke(input_text) # stream: # async for chunk in model.astream(input_text): # print(chunk, end="") # batch: # await model.abatch([input_text]) ``` ``` "a philosophical question that has been debated by thinkers and scholars for centuries." ``` NT)defaultexcluder client async_clientgpt-3.5-turbo-instructmodel)r7aliasstr model_namegffffff?float temperatureint max_tokenstop_prfrequency_penaltypresence_penaltynbest_of)default_factorydict[str, Any] model_kwargsapi_keyOPENAI_API_KEY)r7)r=rKz$SecretStr | None | Callable[[], str]openai_api_keybase_urlOPENAI_API_BASE str | Noneopenai_api_base organization OPENAI_ORG_IDOPENAI_ORGANIZATIONopenai_organization OPENAI_PROXY openai_proxy batch_sizetimeoutz(float | tuple[float, float] | Any | Nonerequest_timeoutzdict[str, float] | None logit_bias max_retriesz int | Noneseedr-Fbool streamingzLiteral['all'] | set[str]allowed_specialallz Literal['all'] | Collection[str]disallowed_specialtiktoken_model_namezMapping[str, str] | Nonedefault_headerszMapping[str, object] | None default_queryz Any | None http_clienthttp_async_clientzMapping[str, Any] | None extra_body)populate_by_namebefore)modec.[U5n[X5$)z>Build extra kwargs from additional params that were passed in.)rr)clsvaluesall_required_field_namess r% build_extraBaseOpenAI.build_extra3s$ 1.z'Cannot stream results when best_of > 1.N)rNrUrQr]rarirjrk)rI ValueErrorrdrJrP isinstancerget_secret_valuecallablerXrTr^rarirjr9rkopenaiOpenAI completionsr:rl AsyncOpenAI)r{msg api_key_value client_params sync_specificasync_specifics r%validate_environmentBaseOpenAI.validate_environment@s{ 66A:)CS/ ! >>dffqj5CS/ ! >>dllQ.;CS/ !9=    *$--y99 $ 3 3 D D F $--.. $ 3 3 % 44,,++++#33!// {{*D,<,<=M --I-I=IUUDK  +T-C-CDN & 2 2!! !k    r'cURURURURURUR UR S.nURbURUS'URbURUS'URbURUS'URS:aURUS'0UEURE$)z2Get the default parameters for calling OpenAI API.)rArFrGrHrIrbr-r_rDrmrErJ) rArFrGrHrIrbr-r_rDrmrJrM)r{ normal_paramss r%_default_paramsBaseOpenAI._default_paramsjs ++ZZ!%!7!7 $ 5 5II )  ?? &*.//M, ' ?? &*.//M, ' ?? &*.//M, ' < total_tokens prompt_tokenscompletion_tokensNrE,Cannot stream results with multiple prompts.r$Generation is empty after streaming.r,r-r+r,r-rerrorr)system_fingerprintrr)rrrdlenrrappendr+r0r1r9rrrrextendr&create_llm_resultr{promptsrrrr sub_promptsr)r"_keysr_promptsr generationrr!s r% _generateBaseOpenAI._generates.((%F%f%**6DA &( G)-#H~~x=1$HC$S/)59 !\\(1+tSFSE!)%* "+ T %@C$S/) * *99'66::?K!% *99'66:::F!%   ;;--HXHH!(D11 (224H <<(($X\\'%:;;x 23#E[A)))16J)K&_$`%% fFX&  r'cF# URn0UEUEnURXQU5n/n0n1Skn Sn UHn UR(aK[U 5S:a Sn [ U 5eSn UR "U SX#40UD6ShvN nU cUn MX- n MURR"S S U 0UD6IShvN n[U[5(dUR5nURUS 5 [!XU5 M UR#XqXXU S 9$N U c Sn [ U 5eUR U RU R(aU RRS5OSU R(aU RRS5OSS .5 GMlN7f)z:Call out to OpenAI's endpoint async with k unique prompts.>rrrNrErrrr,r-rrr)rr)rrrdrrrrr+r0r1r:rrrrrr&rrs r% _agenerateBaseOpenAI._agenerates((%F%f%**6DA &( G)-#H~~x=1$HC$S/)59 #'==QK$6<$,%")%* "+ ("&!2!2!9!9!T!TV!TT!(D11'224Hx 23#E[AK$L%% fFX&  ?,$%@C$S/) * *99'66::?K!% *99'66:::F!%  Us8A:F!<DDD/F!3F4AF!DB F!cUbX1S'USS:Xa3[U5S:wa Sn[U5eURUS5US'[S[U5UR5Vs/sHnX%XPR-PM sn$s snf)z!Get the sub prompts for llm call.rrDrEz7max_tokens set to -1 not supported for multiple inputs.r)rrmax_tokens_for_promptranger\)r{rrrris r%rBaseOpenAI.get_sub_promptsNs  !6N , 2 %7|q O o%#'#=#=gaj#IF< 1c'lDOO< < OO+ ,<   s&Brc h/nURSUR5n[U5H_upXU-US-U-n URU V s/sH1n [ U SU RS5U RS5S.S9PM3 sn 5 Ma X@R S.n U(aX\S '[ XlS 9$s sn f) z2Create the LLMResult from the choices and prompts.rIrEr+r,r-r.r/)r"r?r) generations llm_output)r1rI enumeraterrr?r) r{r)rrr"rrrIr_ sub_choiceschoicers r%rBaseOpenAI.create_llm_resultas JJsDFF #g&DA!a%1q5A+6K   #. #.#F^-3ZZ-H(. :(>)#.  '&1P /A+ ,[HH s 7B/ cUR$)z,Get the parameters used to invoke the model.)rrzs r%rBaseOpenAI._invocation_paramss###r'c6SUR0URE$)zGet the identifying parameters.r?)r?rrzs r%_identifying_paramsBaseOpenAI._identifying_paramssdooF1E1EFFr'cg)zReturn type of llm.rrrzs r% _llm_typeBaseOpenAI._llm_typesr'c>URbURU5$[RSS:a[TU]U5$UR =(d UR n[R"U5nURUURURS9$![a [R"S5nNGf=f)z-Get the token IDs using the tiktoken package.rE cl100k_base)rerg)custom_get_token_idssys version_infosuperget_num_tokensrhr?tiktokenencoding_for_modelKeyError get_encodingencodererg)r{r+r?enc __class__s r% get_token_idsBaseOpenAI.get_token_idss  $ $ 0,,T2 2   A  "7)$/ /--@  7--j9Czz  00#66   7'' 6C 7s&B!! CCz1.2z2.0zgthe model profile's `max_input_tokens` field (e.g. `ChatOpenAI(model=...).profile['max_input_tokens']`))sinceremoval alternativec "0SS_SS_SS_SS_SS_SS _S S _S S _S S_SS_SS _SS_SS_SS_SS_SS_SS_SSSSSSSS.EnSU;aURSSS 9S!n[R"U5nU(aURS"5OS#nUbUc[R S$U5 UcURU5nUcS[ 1[R "5kUR 5k5nS%US&3S'RU5-n[U5eU$)(aLReturn the maximum input context size for a model. Prefers the model's profile (`max_input_tokens`) and falls back to a mapping of legacy models that have no profile. !!! warning "Changed in 1.2" Now returns `max_input_tokens` from the model profile, which is the input context window. Earlier releases returned a hand-maintained number that for some newer models (e.g. `gpt-5`) reflected the *total* context (input + output). Callers using the result as an input-token budget are unaffected; callers using it as a combined input+output budget should switch to the profile fields directly. Args: modelname: The modelname we want to know the context size for. Returns: The maximum input context size. Example: ```python max_tokens = openai.modelname_to_contextsize("gpt-3.5-turbo-instruct") ``` z gpt-4-0314i z gpt-4-0613z gpt-4-32kizgpt-4-32k-0314zgpt-4-32k-0613zgpt-4o-2024-05-13izgpt-3.5-turbo-0301izgpt-3.5-turbo-0613zgpt-3.5-turbo-16ki@zgpt-3.5-turbo-16k-0613r;z text-ada-001iadaztext-babbage-001ibabbageztext-curie-001curieiiAi)davinciztext-davinci-003ztext-davinci-002zcode-davinci-002zcode-davinci-001zcode-cushman-002zcode-cushman-001zft-:rE)maxsplitrmax_input_tokensNzSProfile for model %s is missing `max_input_tokens`; falling back to legacy mapping.zUnknown model: zz. Please provide a valid OpenAI model name, or read `max_input_tokens` from the model profile directly. Known models are: z, ) splitrr1loggerwarningsortedr joinr) modelnamelegacy_token_mappingprofile context_sizeknownrs r%modelname_to_contextsize#BaseOpenAI.modelname_to_contextsizesH $ $    e  e     !$  !$    %e  %d  D  4    t d! " T# $ $ $ $ $ $ $1 8 I !a8;I-- *:Aw{{#56t  <#7 NN2   /33I>L  LY^^-L0D0I0I0KLME!)-%%'+yy'78  S/ !r'c8URUR5$)z$Get max context size for this model.)rr?rzs r%max_context_sizeBaseOpenAI.max_context_sizes,,T__==r'cBURU5nURU- $)a&Calculate the maximum number of tokens possible to generate for a prompt. Args: prompt: The prompt to pass into the model. Returns: The maximum number of tokens to generate for a prompt. Example: ```python max_tokens = openai.max_tokens_for_prompt("Tell me a joke.") ``` )rr)r{r num_tokenss r%r BaseOpenAI.max_tokens_for_prompts%((0 $$z11r')r:r9)rsrLreturnr )rrrrL)NN) rr>rlist[str] | NonerCallbackManagerForLLMRun | Nonerr rzIterator[GenerationChunk]) rr>rr r$AsyncCallbackManagerForLLMRun | Nonerr rzAsyncIterator[GenerationChunk]) r list[str]rr rr rr rr) rr rr rr rr rr)N)rrLrr rr rzlist[list[str]]) r)r rr rrLr"zdict[str, int]rrSrr)rzMapping[str, Any])rr>)r+r>rz list[int])rr>rrC)rrC)rr>rrC)C__name__ __module__ __qualname____firstlineno____doc__rr9__annotations__r:r?rArDrFrGrHrIrJrrMrrPrrTrXrZr\r^r_rarbr-rdsetrergrhrirjrkrlrmr model_configr classmethodrur|rpropertyrrrrrrrrrrr staticmethodr rrr__static_attributes__ __classcell__rs@r%r5r59spdd3FC3dD9L#9$L.>V;@9ISW)X<N8P"'(3Dd*S#OZ', 3 4d ' O % > L*JH@EIAO= +/J'.DK<D*Hj"It/14O.63;@8@7&** 15O-415M.5#K" %)z( ,0J(/@t4L(#E$E '"# '"'#'R66>"&7; 5    # <"&<@ :    ( @"&7; R R R 5 R  R  R n"&<@ 9 9 9 : 9  9  9 ~"&        4*.III I $ I'I I>$$GG ( I NN`>>22r'r5c^\rSrSrSr\S Sj5r\S Sj5r\S U4Sjj5r \S Sj5r \S Sj5r Sr U=r $) riuJ OpenAI completion model integration. Setup: Install `langchain-openai` and set environment variable `OPENAI_API_KEY`. ```bash pip install -U langchain-openai export OPENAI_API_KEY="your-api-key" ``` Key init args — completion params: model: Name of OpenAI model to use. temperature: Sampling temperature. max_tokens: Max number of tokens to generate. logprobs: Whether to return logprobs. stream_options: Configure streaming outputs, like whether to return token usage when streaming (`{"include_usage": True}`). Key init args — client params: timeout: Timeout for requests. max_retries: Max number of retries. api_key: OpenAI API key. If not passed in will be read from env var `OPENAI_API_KEY`. base_url: Base URL for API requests. Only specify if using a proxy or service emulator. organization: OpenAI organization ID. If not passed in will be read from env var `OPENAI_ORG_ID`. See full list of supported init args and their descriptions in the params section. Instantiate: ```python from langchain_openai import OpenAI model = OpenAI( model="gpt-3.5-turbo-instruct", temperature=0, max_retries=2, # api_key="...", # base_url="...", # organization="...", # other params... ) ``` Invoke: ```python input_text = "The meaning of life is " model.invoke(input_text) ``` ```txt "a philosophical question that has been debated by thinkers and scholars for centuries." ``` Stream: ```python for chunk in model.stream(input_text): print(chunk, end="|") ``` ```txt a| philosophical| question| that| has| been| debated| by| thinkers| and| scholars| for| centuries|. ``` ```python "".join(model.stream(input_text)) ``` ```txt "a philosophical question that has been debated by thinkers and scholars for centuries." ``` Async: ```python await model.ainvoke(input_text) # stream: # async for chunk in (await model.astream(input_text)): # print(chunk) # batch: # await model.abatch([input_text]) ``` ```txt "a philosophical question that has been debated by thinkers and scholars for centuries." ``` c /SQ$)z[Get the namespace of the LangChain object. Returns: `["langchain", "llms", "openai"]` ) langchainllmsrrrrs r%get_lc_namespaceOpenAI.get_lc_namespacess /.r'cg)z9Return whether this model can be serialized by LangChain.Trrs r%is_lc_serializableOpenAI.is_lc_serializable|sr'c4>SUR0[TU]E$)Nr<)r?rr)r{rs r%rOpenAI._invocation_paramssGEG,FGGr'c SS0$)z0Mapping of secret keys to environment variables.rPrOrrzs r% lc_secretsOpenAI.lc_secretss!"233r'c0nUR(aURUS'UR(aURUS'UR(aURUS'U$)z$LangChain attributes for this class.rTrXrZ)rTrXrZ)r{ attributess r% lc_attributesOpenAI.lc_attributess\&(   ,0,@,@J( )  # #040H0HJ, -   )-):):J~ &r'r)rr )rrcr)rzdict[str, str])r rrrrrr r#rrr(r,rrrs@r%rrsp]~//HH44  r'r)r zset[str]r!rLr"rLrNone)r2rLrr)3r __future__rloggingrcollections.abcrrrrrtypingr r rrlangchain_core._api.deprecationr langchain_core.callbacksr r #langchain_core.language_models.llmsrlangchain_core.outputsrrrlangchain_core.utilsrlangchain_core.utils.utilsrrrpydanticrrrrtyping_extensionsrlangchain_openai._versionrlangchain_openai.data._profilesr getLoggerr rr&r3r5rrr'r%r>sW" RR 68II9UUBB"15   8 $ 9  9, 9;I 9  9 #  W 2W 2tDZDr'