o +wi @sddlmZddlmZddlZddlZddlZddlZddlZejddkr,ddl Z nddl Z ddl m Z ddl mZmZddd ZgZed d ejd d ejD7Zejddkrheejdg7Zejdredddejejdejejddejejdddg7Znegd7Zede jg7ZddZd,ddZGdd d eZGd!d"d"ee Z!d#d$Z"d%d&Z#Gd'd(d(eZ$Gd)d*d*eZ%gd+Z&dS)-)absolute_import)unicode_literalsN)utils)DATA_DIRmake_dirz?A serialized Python object, stored using the ``pickle`` module.z)The raw (byte string) contents of a file.)picklerawcCsg|]}|r|qSr ).0dr r H/home/ubuntu/sommelier/.venv/lib/python3.10/site-packages/konlpy/data.py "sr KONLPY_DATAz~/z ~/konlpy_datawinzC:\konlpy_datazD:\konlpy_datazE:\konlpy_data konlpy_datalibAPPDATAzC:\)z/usr/share/konlpy_dataz/usr/local/share/konlpy_dataz/usr/lib/konlpy_dataz/usr/local/lib/konlpy_dataz%s/datacCsxtD]}tj||}tj|rt|Sqd}tjd|dddd}|dddd tD7}td |||f) aFind the path of a given resource URL by searching through directories in ``konlpy.data.path``. If the given resource is not found, raise a ``LookupError``, whose message gives a pointer to the installation instructions for ``konlpy.download()``. :type resource_url: str :param resource_url: The URL of the resource to search for. URLs are posix-style relative path names, such as ``corpora/kolaw``. In particular, directory names should always be separated by the forward slash character (i.e., '/'), which will be automatically converted to a platform-appropriate path separator by KoNLPy. zF**********************************************************************zjResource %s not found. Please use the KoNLPy data downloader to obtain the resource: >>> konlpy.download()z B)initial_indentsubsequent_indentwidthz Searched in:rcss|]}d|VqdS)z - %sNr )r pr r r Tszfind..z %s %s %s)pathosjoinexistsFileSystemPathPointertextwrapfill LookupError) resource_urlrfsepmsgr r r find:s  r'autocCsh|dkrtj|dd}|dkrtt|}|S|dkr(t|}|S|tvs.Jt d|)aLoad a given resource from the KoNLPy data package. If no format is specified, ``load()`` will attempt to determine a format based on the resource name's file extension. If that fails, ``load()`` will raise a ``ValueError`` exception. :type resource_url: str :param resource_url: A URL specifying where the resource should be loaded from. :param format: Format type of resource. r(.rr zUnknown format type: %s) rrsplitextstriprloadr'openFORMATS ValueError)r#format resource_valr r r r-Xs   r-c@s"eZdZdZdddZddZdS) PathPointerzAn abstract base class for path pointers. One subclass exists: 1. ``FileSystemPathPointer``: Identifies a file by an absolute path. utf-8cCtdNzAbstract base classNotImplementedErrorselfencodingr r r r.uzPathPointer.opencCr5r6r7r:r r r file_sizexr<zPathPointer.file_sizeNr4)__name__ __module__ __qualname____doc__r.r>r r r r r3qs  r3c@s*eZdZdZddZd ddZddZd S) rz:A path pointer that identifies a file by an absolute path.cCs.tj|}tj|std|||_dS)NzNo such file or directory: %s)rrabspathrIOError)r:rr r r __init__s    zFileSystemPathPointer.__init__r4cCs t|jSN)rload_txtrr9r r r r.s zFileSystemPathPointer.opencCst|jjSrG)rstatrst_sizer=r r r r>szFileSystemPathPointer.file_sizeNr?)r@rArBrCrFr.r>r r r r r|s   rcCs&ttd}|D]}t|q dS)z~clear the konlpy output data directory .. code-block:: python >>> import konlpy >>> konlpy.clear() *N)rlistdirrremove)itemsitemr r r clears  rPcCstttdS)z}list konlpy default data directory. .. code-block:: python >>> import konlpy >>> konlpy.listdir() N)rpprintrrLrr r r r rLs rLc@seZdZdddZddZdS) CorpusReader.txtcCsttd||_i|_dS)aCorpusReader reads corpuses in konlpy data directory. extension (str, optional): Defaults to '.txt'. extension of corpus to load. .. code-block:: python >>> from konlpy.data import CorpusReader >>> reader = CorpusReader() >>> reader.read() >>> reader.corpus {...} >>> reader.items = ["data/specific_corpus.txt"] >>> reader.read() >>> reader.corpus['specific_corpus.txt'] content of corpus rKN)globrrNcorpus)r: extensionr r r rFs zCorpusReader.__init__cCs6|jD]}tj|ddd}||jtj|<qdS)zgread method reads all files included in items attr and save it into corpus dictionary. zr+r4moder;N)rNior.readrUrrbasename)r:filenamereaderr r r rZs zCorpusReader.readN)rS)r@rArBrFrZr r r r rRs  rRc@seZdZddZddZdS) StringWritercCs ttjt|ddd|_dS)Nar4rW)rrYr.rwriter)r:r\r r r rFszStringWriter.__init__cCs|j||jddS)N )r`write)r:stringr r r rbs zStringWriter.writeN)r@rArBrFrbr r r r r^s r^) r'r-rLrPrrr3rRr^)r()' __future__rrrrYsysrTr version_inforcPicklekonlpyrkonlpy.constantsrrr/renvirongetsplitpathsep expanduserplatform startswithrprefix installpathr'r-objectr3strrrPrLrRr^__all__r r r r sJ    $