jddlZddlmZmZmZddlZddlmZddlm Z ddl m Z ddl m Z ddlmZddlmZmZ dd eed eed eeedeedededeedee fdZ ddeeedeeedeeedeeedededeeedeee fdZdS)N)IOListOptional)UnstructuredClient)shared)Element)logger) exactly_one)dict_to_elementselements_from_json.https://api.unstructured.io/general/v0/generalfilename content_typefile file_filenameapi_urlapi_keymetadata_filenamereturnc t|||r|rtd||}tjdd|vr |ddn|}t ||} |Qt |d5} t j| | } dddn #1swxYwYn)|'|td t j|| } t j dd | i|} | j | } | j d krt| jj Std| j d)aPartitions a document using the Unstructured REST API. This is equivalent to running the document through partition. See https://api.unstructured.io/general/docs for the hosted API documentation or https://github.com/Unstructured-IO/unstructured-api for instructions on how to run the API locally as a container. Parameters ---------- filename A string defining the target filename path. content_type A string defining the file content in MIME type file A file-like object using "rb" mode --> open(filename, "rb"). metadata_filename When file is not None, the filename (string) to store in element metadata. E.g. "foo.txt" api_url The URL for the Unstructured API. Defaults to the hosted Unstructured API. api_key The API key to pass to the Unstructured API. request_kwargs Additional parameters to pass to the data field of the request to the Unstructured API. For example the `strategy` parameter. )rrzOnly one of metadata_filename and file_filename is specified. metadata_filename is preferred. file_filename is marked for deprecation.NzuThe file_filename kwarg will be deprecated in a future version of unstructured. Please use metadata_filename instead.z/general/v0/generali) api_key_auth server_urlrb)content file_namezWIf file is specified in partition_via_api, metadata_filename must be specified as well.files)textReceive unexpected status code  from the API.)r ValueErrorr warnropenrFilesreadPartitionParametersgeneral partition status_coder raw_responser)rrrrrrrrequest_kwargsbase_urlsdkfrreqresponses d/lsinfo/ai/hellotax_ai/base_platform/venv/lib/python3.11/site-packages/unstructured/partition/api.pypartition_via_apir4sF---- ]  W    )  4   !6 @ @wtt}}gH 'h G G GC (D ! ! QL"E                   $?  '     $      C{$$S))Hs""!x'<'ABBBB Rh.B R R R   s0)B%%B),B) filenames content_typesrfile_filenamesmetadata_filenamesc d|d}|r|rtd||}tjd||r/t|t|krtdt j5fd|D}g} t |D]3\} } || } ||| nd} | d| | | ff4tj |||| }dddn #1swxYwYn||r/t|t|krtd |std t|t|krtd g} t |D]3\} }||| nd} || } | d| || ff4tj |||| }|j d kr]g}| }t|dts|g}|D]$}|t|%|Std|j d)aPartitions multiple documents using the Unstructured REST API by batching the documents into a single HTTP request. See https://api.unstructured.io/general/docs for the hosted API documentation or https://github.com/Unstructured-IO/unstructured-api for instructions on how to run the API locally as a container. Parameters ---------- filenames A list of strings defining the target filename paths. content_types A list of strings defining the file contents in MIME types. files A list of file-like object using "rb" mode --> open(filename, "rb"). metadata_filename When file is not None, the filename (string) to store in element metadata. E.g. "foo.txt" api_url The URL for the Unstructured API. Defaults to the hosted Unstructured API. api_key The API key to pass to the Unstructured API. request_kwargs Additional parameters to pass to the data field of the request to the Unstructured API. For example the `strategy` parameter. zapplication/json)ACCEPTzUNSTRUCTURED-API-KEYzOnly one of metadata_filenames and file_filenames is specified. metadata_filenames is preferred. file_filenames is marked for deprecation.NzwThe file_filenames kwarg will be deprecated in a future version of unstructured. Please use metadata_filenames instead.z6content_types and filenames must have the same length.cVg|]%}t|d&S)r) enter_contextr%).0r0stacks r3 z.partition_multiple_via_api..s/KKKAU((a77KKKr)headersdatarz2content_types and files must have the same length.z8metadata_filenames must be specified if files are passedz7metadata_filenames and files must have the same length.rrr r!)r#r r$len contextlib ExitStack enumerateappendrequestspostr+json isinstancelistr )r5r6rr7rrr8r-rA_filesirrrr2_file documents response_listdocumentr>s @r3partition_multiple_via_apirSis4H% 'G  n  Y   !+  5     WS//3y>>AAUVV V  ! # # uKKKKKKKEF$U++ I I4$Q<3@3L}Q//RV  w4(FGHHHH}# H                   SS//3u::==QRR R! XWXX X # $ $E 2 2VWW W!%(( F FHAu/r]s2 222222------333333&&&&&&555555JJJJJJJJ#"& $#'C'+T T smT 3-T  2e9 T C= T  T  T  }T  ']T T T T p&*)-!%*.C.2m m S "m DI&m  DI m T#Y' m  m  m !c+m  $w-m m m m m m r@