Warning: session_start(): Session cannot be started after headers have already been sent in /home/tvrreohg/public_html/manga.php on line 13
3
j+ @ s* d dl Z d dlZd dlZd dlZd dlmZ d dlmZ d dlmZ d dl m
Z
m Z m Z m
Z
mZmZmZmZmZ ddlmZmZ ddlmZmZ dd lmZmZm Z m Z m Z m Z m Z m!Z!m"Z"m#Z#m$Z$ dd
l%m&Z' dd l%m(Z) dd l%m*Z* dd
l%m+Z, ddl%m-Z. ddl%m/Z/m0Z0 ddl%m1Z2 ddl%m3Z4 ddl%m5Z6 ddl7m8Z8m9Z9m:Z:m;Z;mZ>m?Z?m@Z@mAZAmBZBmCZCmDZDmEZEmFZFmGZGmHZHmIZImJZJmKZKmLZLmMZM ddlNmOZOmPZP ddlQmRZR eSeTeeTee
df f dddZUeSeTeeTee
df f ddd ZVG d d d eBZWG d d d ZXG d!d" d"eXZYdS )# N)datetime)BytesIO)Path) AnyCallableDictIterableListOptionalTupleUnioncast )
Encryption PasswordType)
PageObject _VirtualList) StrByteType
StreamTypeb_deprecate_no_replacementdeprecate_with_replacementlogger_warningread_non_whitespaceread_previous_lineread_until_whitespaceskip_over_commentskip_over_whitespace)CatalogAttributes)CatalogDictionary) CheckboxRadioButtonAttributes)Core) DocumentInformationAttributes)FieldDictionaryAttributesGoToActionArguments)PageAttributes)PagesAttributes) TrailerKeys)EmptyFileErrorFileNotDecryptedError PdfReadErrorPdfStreamErrorWrongPasswordError) ArrayObject
ContentStreamDecodedStreamObject DestinationDictionaryObjectEncodedStreamObjectField FloatObjectIndirectObject
NameObject
NullObject NumberObject PdfObjectTextStringObject
TreeObject read_object) OutlineType PagemodeType)XmpInformation.)dsizereturnc C s4 |dkrt dd| } | dd } tjd| d S )N z invalid size in convert_to_ints z>qr i)r* structunpack)r@ rA rF 0/tmp/pip-build-vnmlutbd/PyPDF2/PyPDF2/_reader.pyconvert_to_intf s
rH c C s t dd t| |S )N convertToIntrH )r rH )r@ rA rF rF rG rI n s
rI c @ sz e Zd ZdZddddZeee dddZeee dd d
Ze ee dd d Z
e ee dd
dZ e ee dddZ e ee dddZ
e ee dddZe ee dddZe ee dddZe ee dddZe ee ddd Ze ee dd d Ze ee dd d Ze ee dd!d"Ze ee dd#d$Ze ee dd%d&ZdS )'DocumentInformationa
A class representing the basic document metadata provided in a PDF File.
This class is accessible through :py:class:`PdfReader.metadata`.
All text properties of the document metadata have
*two* properties, eg. author and author_raw. The non-raw property will
always return a ``TextStringObject``, making it ideal for a case where
the metadata is being displayed. The raw property can sometimes return
a ``ByteStringObject``, if PyPDF2 was unable to decode the string's
text encoding; this requires additional safety in the caller and
therefore is not as commonly accessed.
N)rB c C s t j| d S )N)r1 __init__)selfrF rF rG rK s z DocumentInformation.__init__)keyrB c C s | j |d }t|tr|S d S )N)get
isinstancer: )rL rM retvalrF rF rG _get_text s
z DocumentInformation._get_textc C s t d | j|S )z
The text value of the specified key or None.
.. deprecated:: 1.28.0
Use the attributes (e.g. :py:attr:`title` / :py:attr:`author`).
getText)r rQ )rL rM rF rF rG rR s zDocumentInformation.getTextc C s, | j tjr(| jtjp*| j tjj S dS )z
Read-only property accessing the document's **title**.
Returns a unicode string (``TextStringObject``) or ``None``
if the title is not specified.
N)rN DITITLErQ
get_object)rL rF rF rG title s
(zDocumentInformation.titlec C s | j tjS )z>The "raw" version of title; can return a ``ByteStringObject``.)rN rS rT )rL rF rF rG title_raw s z DocumentInformation.title_rawc C s | j tjS )z
Read-only property accessing the document's **author**.
Returns a unicode string (``TextStringObject``) or ``None``
if the author is not specified.
)rQ rS AUTHOR)rL rF rF rG author s zDocumentInformation.authorc C s | j tjS )z?The "raw" version of author; can return a ``ByteStringObject``.)rN rS rX )rL rF rF rG
author_raw s z DocumentInformation.author_rawc C s | j tjS )z
Read-only property accessing the document's **subject**.
Returns a unicode string (``TextStringObject``) or ``None``
if the subject is not specified.
)rQ rS SUBJECT)rL rF rF rG subject s zDocumentInformation.subjectc C s | j tjS )z@The "raw" version of subject; can return a ``ByteStringObject``.)rN rS r[ )rL rF rF rG subject_raw s z DocumentInformation.subject_rawc C s | j tjS )ax
Read-only property accessing the document's **creator**.
If the document was converted to PDF from another format, this is the
name of the application (e.g. OpenOffice) that created the original
document from which it was converted. Returns a unicode string
(``TextStringObject``) or ``None`` if the creator is not specified.
)rQ rS CREATOR)rL rF rF rG creator s
zDocumentInformation.creatorc C s | j tjS )z@The "raw" version of creator; can return a ``ByteStringObject``.)rN rS r^ )rL rF rF rG creator_raw s z DocumentInformation.creator_rawc C s | j tjS )a\
Read-only property accessing the document's **producer**.
If the document was converted to PDF from another format, this is
the name of the application (for example, OSX Quartz) that converted
it to PDF. Returns a unicode string (``TextStringObject``)
or ``None`` if the producer is not specified.
)rQ rS PRODUCER)rL rF rF rG producer s
z DocumentInformation.producerc C s | j tjS )zAThe "raw" version of producer; can return a ``ByteStringObject``.)rN rS ra )rL rF rF rG producer_raw s z DocumentInformation.producer_rawc C s, | j tj}|dkrdS tj|jdddS )zP
Read-only property accessing the document's **creation date**.
N' zD:%Y%m%d%H%M%S%z)rQ rS
CREATION_DATEr strptimereplace)rL textrF rF rG
creation_date s z!DocumentInformation.creation_datec C s | j tjS )z
The "raw" version of creation date; can return a ``ByteStringObject``.
Typically in the format D:YYYYMMDDhhmmss[+-]hh'mm where the suffix is the
offset from UTC.
)rN rS rf )rL rF rF rG creation_date_raw s z%DocumentInformation.creation_date_rawc C s, | j tj}|dkrdS tj|jdddS )z
Read-only property accessing the document's **modification date**.
The date and time the document was most recently modified.
Nrd re zD:%Y%m%d%H%M%S%z)rQ rS MOD_DATEr rg rh )rL ri rF rF rG modification_date s z%DocumentInformation.modification_datec C s | j tjS )z
The "raw" version of modification date; can return a ``ByteStringObject``.
Typically in the format D:YYYYMMDDhhmmss[+-]hh'mm where the suffix is the
offset from UTC.
)rN rS rl )rL rF rF rG modification_date_raw
s z)DocumentInformation.modification_date_raw)__name__
__module__ __qualname____doc__rK strr
rQ rR propertyrV rW rY rZ r\ r] r_ r` rb rc r rj rk rm rn rF rF rF rG rJ u s@
rJ c @ s. e Zd ZdZdeeef eedee f ddddZ
e eddd Z e e
e dd
d Ze
e dd d
Ze e
e dddZe e
e dddZe
e dddZe e
e dddZedddZedddZe edddZeed d d Zeed d d!Z e e ee f dd"d#Z e e ee f dd$d%Z de
e! e
e e e f e
e e
e ee f d&d'd(Z"de
e! e
e e e f e
e e
e ee f d&d)d*Z#ee!e$f e e e f e e dd+d,d-Z%ee!e$f e e dd&d.d/Z&e e e dd0d1d2Z'e ee f dd3d4Z(e ee f dd5d6Z)dee!df e
e e ee f d7d8d9Z*dee!df e
e e ee f d7d:d;Z+e e,ddd?Z.de
e$ e
e e,d@dAdBZ/de
e$ e
e e,d@dCdDZ0e e
e1 ddEdFZ2edee3e4f edGdHdIZ5eedJdKdLZ6eedJdMdNZ7e8edOdPdQZ9e8edOdRdSZ:ee
e;eee e;e ddZd[Z?e e
e dd\d]Z@e
e dd^d_ZAe e
e dd`daZBe e
eC ddbdcZDe
eC ddddeZEe e
eC ddfdgZFdede$ef e
e ee f e
e4 ddhdidjZGe4eeeHef dGdkdlZIeee
eH dmdndoZJeee4f e
eH dGdpdqZKe4e
eH drdsdtZLeMeNeef dudvdwZOeMeNeef dudxdyZPeee
eH dzd{d|ZQeee
eH dzd}d~ZReee
eH e
eH dddZSeee
eH e
eH dddZTeMdduddZUeMdduddZVeMdduddZWeMeduddZXeMdduddZYeMe
e eddddZZeMe
e duddZ[eMee
e dddZ\eMee]e^e_f duddZ`eaeMeedddZbeMdduddZce;e edegeeeNedf f f edeeeeNedf f gef ddddZee;e efeNeef dddZgdeMee dddZhdeMee dddZieee f ejdddZkee eef dddZle edddZmedddZne edddZoe e
e ee f dddZpdS ) PdfReadera
Initialize a PdfReader object.
This operation can take some time, as the PDF stream's cross-reference
tables are read into memory.
:param stream: A File object or an object that supports the standard read
and seek methods similar to a File object. Could also be a
string representing a path to a PDF file.
:param bool strict: Determines whether user should be warned of all
problems and also causes some correctable problems to be fatal.
Defaults to ``False``.
:param None/str/bytes password: Decrypt PDF file at initialization. If the
password is None, the file will not be decrypted.
Defaults to ``None``
FN)streamstrictpasswordrB c C s2 || _ d | _i | _d| _d | _t|dr`
instance that can be used to access XMP metadata from the document.
or ``None`` if no metadata was found on the document root.
TNF)r r r ROOT xmp_metadata)rL rF rF rG r s zPdfReader.xmp_metadatac C s t dd | jS )zh
.. deprecated:: 1.28.0
Use the attribute :py:attr:`xmp_metadata` instead.
getXmpMetadatar )r r )rL rF rF rG r s
zPdfReader.getXmpMetadatac C s t dd | jS )zh
.. deprecated:: 1.28.0
Use the attribute :py:attr:`xmp_metadata` instead.
xmpMetadatar )r r )rL rF rF rG r s
zPdfReader.xmpMetadatac C s: | j r| jtj d d S | jdkr,| j t| jS dS )z
Calculate the number of pages in this PDF file.
:return: number of pages
:raises PdfReadError: if file is encrypted and restrictions prevent
this action.
z/Pagesz/CountN)r r r r r} _flattenlen)rL rF rF rG _get_num_pages s
zPdfReader._get_num_pagesc C s t dd | j S )z\
.. deprecated:: 1.28.0
Use :code:`len(reader.pages)` instead.
zreader.getNumPageszlen(reader.pages))r r )rL rF rF rG getNumPages s
zPdfReader.getNumPagesc C s t dd | j S )z\
.. deprecated:: 1.28.0
Use :code:`len(reader.pages)` instead.
zreader.numPageszlen(reader.pages))r r )rL rF rF rG numPages s
zPdfReader.numPages)
pageNumberrB c C s t dd | j|S )zd
.. deprecated:: 1.28.0
Use :code:`reader.pages[page_number]` instead.
zreader.getPage(pageNumber)zreader.pages[page_number])r _get_page)rL r rF rF rG getPage s zPdfReader.getPage) page_numberrB c C s. | j dkr| j | j dk s$td| j | S )z
Retrieve a page by number from this PDF file.
:param int page_number: The page number to retrieve
(pages begin at zero)
:return: a :class:`PageObject` instance.
Nz
hint for mypy)r} r AssertionError)rL r rF rF rG r s
zPdfReader._get_pagec C s t dd | jS )z`
.. deprecated:: 1.28.0
Use :py:attr:`named_destinations` instead.
namedDestinationsnamed_destinations)r r )rL rF rF rG r s
zPdfReader.namedDestinationsc C s | j S )zv
A read-only dictionary which maps names to
:class:`Destinations`
)_get_named_destinations)rL rF rF rG r s z PdfReader.named_destinations)treerP fileobjrB c
C s t j }|jtj |dkrXi }tt| jtj }t j
|krTtt t |t j
}ndS |dkrd|S | j
||| x&|D ] }||krx| j|||| P qxW d|krtt|d }x$|D ] }|j } | j| ||| qW |S )a
Extract field data if this PDF contains interactive form fields.
The *tree* and *retval* parameters are for recursive use.
:param fileobj: A file object (usually a text file) to write
a report to on all interactive form fields found.
:return: A dictionary where each key is a field name, and each
value is a :class:`Field` object. By
default, the mapping name is used for keys.
``None`` if form data could not be located.
Nz/Fields)r# Zattributes_dictr r r
r1 r r r CDZ ACRO_FORMr
r; _check_kids _build_fieldr- rU )
rL r rP r field_attributescatalogattrfieldsffieldrF rF rG
get_fields s*
zPdfReader.get_fieldsc C s t dd | j|||S )zU
.. deprecated:: 1.28.0
Use :meth:`get_fields` instead.
getFieldsr )r r )rL r rP r rF rF rG r 0 s
zPdfReader.getFields)r rP r r rB c C s~ | j ||| y |d }W n6 tk
rP y |d }W n tk
rJ d S X Y nX |rn| j||| |jd t|||< d S )Nz/TMz/T
)r KeyError _write_fieldwriter3 )rL r rP r r rM rF rF rG r > s
zPdfReader._build_fieldc C s4 t j|kr0x$|t j D ]}| j|j || qW d S )N)PAKIDSr rU )rL r rP r kidrF rF rG r S s
zPdfReader._check_kids)r r r rB c C s t j }|tj }x|D ]}|t jt jfkr0q|| }y|t jkr~ddddd}|| |kr|j|d ||| d nv|t jkry|| t j }W n" t k
r || t j
}Y nX |j|d | d n |j|d t || d W q t k
r Y qX qW d S )NZButtonZTextZChoice Signature)z/Btnz/Txz/Chz/Sigz: r ) r#
attributesr ZKidsZAAZFTr ZParentZTMr Trs ) rL r r r Zfield_attributes_tupler attr_nametypesnamerF rF rG r [ s2
"zPdfReader._write_fieldc s& | j dkri S fdd D S )a6
Retrieve form fields from the document with textual data.
The key is the name of the form field, the value is the content of the
field.
If the document contains multiple form fields with the same name, the
second and following will get the suffix _2, _3, ...
Nc s6 i | ].} | j d dkr | j d | d qS )z/FTz/Txz/Vz/T)rN ).0r )
formfieldsrF rG
s z2PdfReader.get_form_text_fields..)r )rL rF )r rG get_form_text_fields s
z PdfReader.get_form_text_fieldsc C s t dd | j S )z_
.. deprecated:: 1.28.0
Use :meth:`get_form_text_fields` instead.
getFormTextFieldsr )r r )rL rF rF rG r s
zPdfReader.getFormTextFields)r rP rB c
C s |dkrni }t t| jtj }tj|kr:t t|tj }n4tj|krnt t|tj }tj|krnt t|tj }|dkrz|S t j
|krxt t |t j
D ]}| j |j
| qW ntj|krDt t|tj }xtdt|dD ]`}t t|| j
}||d j
}t|tr"d|kr"|d }| j||} | dk r| ||< qW n>x<|j D ]0\}
} | j
} | j|
| } | dk rN| ||
< qNW |S )z
Retrieve the named destinations present in the document.
:return: a dictionary which maps names to
:class:`Destinations`.
Nr r z/D)r
r1 r r r CAZDESTSr; ZNAMESr r r- r rU ranger rs rO _build_destinationitems)
rL r rP r namesr irM valuedestZk__Zv__valrF rF rG r s<
z!PdfReader._get_named_destinationsc C s t dd | j||S )z`
.. deprecated:: 1.28.0
Use :py:attr:`named_destinations` instead.
getNamedDestinationsr )r r )rL r rP rF rF rG r s
z PdfReader.getNamedDestinationsc C s | j S )z
Read-only property for the outline (i.e., a collection of 'outline items'
which are also known as 'bookmarks') present in the document.
:return: a nested list of :class:`Destinations`.
) _get_outline)rL rF rF rG outline s zPdfReader.outlinec C s t dd | jS )zT
.. deprecated:: 2.9.0
Use :py:attr:`outline` instead.
outlinesr )r r )rL rF rF rG r s
zPdfReader.outlines)noder rB c C s |d krng }t t| jtj }tj|krdt t|tj }t|trF|S |d k rdd|krdt t|d }| j | _
|d krz|S xd| j |}|r|j | d|krg }| j
t t|d | |r|j | d|krP t t|d }q|W |S )Nz/Firstz/Next)r
r1 r r r COZOUTLINESrO r7 r _namedDests_build_outline_itemappendr )rL r r r linesZ outline_objZ sub_outlinerF rF rG r s2
zPdfReader._get_outlinec C s t dd | j||S )zU
.. deprecated:: 1.28.0
Use :py:attr:`outline` instead.
getOutlinesr )r r )rL r r rF rF rG r s
zPdfReader.getOutlinesc C s4 t t| jtj }tj|kr,t d|tj S dS dS )u
Read-only property for the list of threads see §8.3.2 from PDF 1.7 spec
:return: an Array of Dictionnaries with "/F" and "/I" properties
or None if no articles.
r- N)r
r1 r r r r ZTHREADS)rL r rF rF rG threads# s
zPdfReader.threads)indirect_referencerB c C sp | j dkr dd t| jD | _ |dks2t|tr6dS t|trF|}n|j}| j dk s^td| j j|d}|S )zGenerate _page_id2numNc S s i | ]\}}||j jqS rF )r idnum)r r xrF rF rG r 6 s z:PdfReader._get_page_number_by_indirect..r z
hint for mypyr ) r enumeratepagesrO r7 intr r rN )rL r r retrF rF rG _get_page_number_by_indirect1 s
z&PdfReader._get_page_number_by_indirect)pagerB c C s | j |jS )a
Retrieve page number of a given PageObject
:param PageObject page: The page to get page number. Should be
an instance of :class:`PageObject`
:return: the page number or -1 if page not found
)r r )rL r rF rF rG get_page_numberD s zPdfReader.get_page_numberc C s t dd | j|S )zZ
.. deprecated:: 1.28.0
Use :meth:`get_page_number` instead.
getPageNumberr )r r )rL r rF rF rG r N s
zPdfReader.getPageNumber) destinationrB c C s | j |jS )z
Retrieve page number of a given Destination object.
:param Destination destination: The destination to get page number.
:return: the page number or -1 if page not found
)r r )rL r rF rF rG get_destination_page_numberW s z%PdfReader.get_destination_page_numberc C s t dd | j|S )zf
.. deprecated:: 1.28.0
Use :meth:`get_destination_page_number` instead.
getDestinationPageNumberr )r r )rL r rF rF rG r ` s z"PdfReader.getDestinationPageNumber)rV arrayrB c C s d\}}t |ttfs4t |tr,t|dks4|d krNt }td}t|||S |dd \}}|dd }yt|||f| S tk
r td| d| t | j
r | j d j }|d krt n|}t||tdS X d S )Nr z/Fitr zUnknown destination: )NN)
rO r7 rs r- r r: r0 r* r ro rw r r )rL rV r r typtmpr rF rF rG r m s& z PdfReader._build_destination)r rB c C s d\}}}yt d|d }W n, tk
rH | jr@td|d}Y nX d|krt t|d }t t|tj }|dkr|tj }n*d|kr|d }t |trd|kr|d }t |t
r| j ||}nt |t ry| j || j
| j}W n" tk
r | j |d }Y nX nN|d kr(| j ||}n6| jr@td |ntd
|d t | j |d }|rd |krt
d
d |d D |td < d|kr|d |td< d|kr|d |td< |S )Nrs z/Titlez(Outline Entry Missing /Title attribute: re z/Az/GoToz/Destz/DzUnexpected destination z Removed unexpected destination z from destinationz/Cc s s | ]}t |V qd S )N)r4 )r crF rF rG s z0PdfReader._build_outline_item..z/Fz/Count)NNN)r
r rw r* r1 r6 r$ SDrO r- r rs r Z
dest_arrayr ro )rL r r rV Z outline_itemactionZ action_typerF rF rG r sP
z PdfReader._build_outline_itemc C s t | j| jS )zWRead-only property that emulates a list of :py:class:`Page` objects.)r r r )rL rF rF rG r s zPdfReader.pagesc C s0 t t| jtj }tj|kr,t t|tj S dS )a
Get the page layout.
:return: Page layout currently being used.
.. list-table:: Valid ``layout`` values
:widths: 50 200
* - /NoLayout
- Layout explicitly not specified
* - /SinglePage
- Show one page at a time
* - /OneColumn
- Show one column at a time
* - /TwoColumnLeft
- Show pages in two columns, odd-numbered pages on the left
* - /TwoColumnRight
- Show pages in two columns, odd-numbered pages on the right
* - /TwoPageLeft
- Show two pages at a time, odd-numbered pages on the left
* - /TwoPageRight
- Show two pages at a time, odd-numbered pages on the right
N)r
r1 r r r r Z PAGE_LAYOUTr6 )rL r rF rF rG page_layout s
zPdfReader.page_layoutc C s t dd | jS )zY
.. deprecated:: 1.28.0
Use :py:attr:`page_layout` instead.
getPageLayoutr )r r )rL rF rF rG r s
zPdfReader.getPageLayoutc C s t dd | jS )zY
.. deprecated:: 1.28.0
Use :py:attr:`page_layout` instead.
pageLayoutr )r r )rL rF rF rG r s
zPdfReader.pageLayoutc C s* y| j tj d S tk
r$ dS X dS )aO
Get the page mode.
:return: Page mode currently being used.
.. list-table:: Valid ``mode`` values
:widths: 50 200
* - /UseNone
- Do not show outline or thumbnails panels
* - /UseOutlines
- Show outline (aka bookmarks) panel
* - /UseThumbs
- Show page thumbnails panel
* - /FullScreen
- Fullscreen view
* - /UseOC
- Show Optional Content Group (OCG) panel
* - /UseAttachments
- Show attachments panel
z /PageModeN)r r r r )rL rF rF rG page_mode s zPdfReader.page_modec C s t dd | jS )zW
.. deprecated:: 1.28.0
Use :py:attr:`page_mode` instead.
getPageModer )r r )rL rF rF rG r % s
zPdfReader.getPageModec C s t dd | jS )zW
.. deprecated:: 1.28.0
Use :py:attr:`page_mode` instead.
pageModer )r r )rL rF rF rG r . s
zPdfReader.pageMode)r inheritr rB c
C s4 t tjt tjt tjt tjf}|d kr0i }|d krZ| jtj j }|d j }g | _
d}t j |krr|t j }|dkrx |D ]}||kr|| ||< qW x|t j
D ].}i } t|tr|| d< | j|j |f| qW nT|dkr0x(t|j D ]\}
} |
|kr| ||
< qW t| |} | j| | j
j| d S )Nz/Pagesr z/Page)r6 PGZ RESOURCESZMEDIABOXZCROPBOXZROTATEr r r rU r} r ZTYPEr rO r5 r listr r r r )
rL r r r Zinheritable_page_attributesr tr r ZaddtZattr_inr Zpage_objrF rF rG r 8 s:
zPdfReader._flattenc C s | j |j \}}t|d| j }tt|d dks6t||d k sFttt|j }x,t
|d D ]}t | |j dd t
j|}t | |j dd t
j|}t | |j dd ||jkrqf| jr||krtd|j t|d | d t | |j dd yt|| } W nf tk
r~ }
zHtd| d |j d
|j d |
t | jrhtd |
t } W Y d d }
~
X nX | S W | jrtd
t S )Nr z/Typez/ObjStmz/Nr zObject is in wrong index.z/FirstzInvalid stream (index z) within object r z: zCan't read object stream: z%This is a fatal error in strict mode.r r r r ) xref_objStmr r5 rU r
rs r r r get_datar r r r8 Zread_from_streamrw r* r r< r+ r
generationro r7 ) rL r ZstmnumidxZobj_stm stream_datar Zobjnumoffsetr excrF rF rG _get_object_from_streamf sD
z!PdfReader._get_object_from_stream)numgenrB c C s t ||| j S )zr
used to ease development
equivalent to generic.IndirectObject(num,gen,self).get_object()
)r5 rU )rL r r rF rF rG _get_indirect_object s z PdfReader._get_indirect_objectc C sB t |trt|d| }| j|j|j}|d k r2|S |jdkrV|j| jkrV| j|}n|j| jkor|j| j|j kr| j j
|ji j
|jdrt S | j|j |j }| j j
|d y| j| j \}}W n tk
r t| j drt| j j }n2| j j }| j j
dd | j jd}| j j
|d tjd|j d|j dj |}|d k rtd|j d |j d
t |jdd | j|j |j< | j j
|jdd | j| j \}}nd}Y nX ||jkr| jr| j r>t d |j d |j d
| d | d n:||jkr>| j r>t d |j d |j d
| d | d | j rV||jksVt t | j | }| j r,| j!d k r,| j!j" st#dt$t%|}| j!j&||j|j}n~t| j drt| j j }n2| j j }| j j
dd | j jd}| j j
|d tjd|j d|j dj |}|d k r td|j d |j dt |j| jkr`i | j|j< |jdd | j|j |j< | j j
|j'dd t(| j | j j
dd t | j | }| j r,| j!d k r,| j!j" st#dt$t%|}| j!j&||j|j}n,td|j d |j dt | j r,t d| j)|j|j| |S )Nr F getbufferr z\sz\s+z\s+objz
Object ID ,z
ref repairedzExpected object ID (r z) does not match actual (z ); xref table not zero-indexed.z).zFile has not been decryptedzObject z foundz
not defined.zCould not find object.r r r r )*rO r r5 cache_get_indirect_objectr r r r
xrefxref_free_entryrN r7 rv r read_object_header Exceptionr bytesr r r researchencoder ro startr rw r* r r< r r Z is_decryptedr) r
r9 Zdecrypt_objectendr cache_indirect_object) rL r rP r r r bufpmrF rF rG rU s
&$
zPdfReader.get_object)indirectReferencerB c C s t dd | j|S )zU
.. deprecated:: 1.28.0
Use :meth:`get_object` instead.
getObjectrU )r rU )rL r rF rF rG r % s
zPdfReader.getObject)rv rB c C s d}t | |t|O }|jdd t|}|t|O }|jdd t|}|t|O }|jdd |jd}t| |jd d |r| jrtd| d| t t |t |fS )
NFr z.Superfluous whitespace found in object header r r r r r )
r r r r r r rw r ro r )rL rv extrar r _objrF rF rG r 0 s$
z PdfReader.read_object_headerc C s t dd | j|S )z]
.. deprecated:: 1.28.0
Use :meth:`read_object_header` instead.
readObjectHeaderr )r r )rL rv rF rF rG r$ L s
zPdfReader.readObjectHeader)r r rB c C s | j j||fS )N)r~ rN )rL r r rF rF rG r W s z#PdfReader.cache_get_indirect_objectc C s t dd | j||S )zd
.. deprecated:: 1.28.0
Use :meth:`cache_get_indirect_object` instead.
cacheGetIndirectObjectr )r r )rL r r rF rF rG r% \ s z PdfReader.cacheGetIndirectObject)r r r rB c C sH ||f| j kr6d| d| }| jr,t|t|t || j ||f< |S )NzOverwriting cache for r )r~ rw r* r ro )rL r r r msgrF rF rG r i s
z PdfReader.cache_indirect_objectc C s t dd | j|||S )z`
.. deprecated:: 1.28.0
Use :meth:`cache_indirect_object` instead.
cacheIndirectObjectr )r r )rL r r r rF rF rG r' t s
z PdfReader.cacheIndirectObjectc
C s4 | j | | j| | j|}| j||}|dkrV| jrD|rDtdtd| dt | j||| | j op| j r0|j
}x| j j D ]\}}|dkrqt
|j }xx|D ]p}|j|| d y| j|\} }
W n tk
r P Y nX | || j kr| j | | | j | | < | j | |= qW qW |j|d d S )Nr zBroken xref tablez incorrect startxref pointer()i )_basic_validation_find_eof_marker_find_startxref_pos_get_xref_issuesrw r* r ro _read_xref_tables_and_trailersr r r r sortedkeysr r
ValueError) rL rv startxref
xref_issue_nrr r Z
xref_entryZxref_kidpidZ_pgenrF rF rG r s4
zPdfReader.readc C sl |j dtj |j s td| jrh|j dtj |jd}|dkrZtd|j d d|j dtj d S )Nr zCannot read an empty file s %PDF-zPDF starts with 'utf8z', but '%PDF-' expected)
r osSEEK_ENDr r( rw SEEK_SETr r* r )rL rv Z header_byterF rF rG r) s
zPdfReader._basic_validationc C s> d}d}x0|d d dkr8|j |k r.tdt|}q
W d S )NrC r| r5 s %%EOFzEOF marker not found)r r* r )rL rv Zlast_mblinerF rF rG r* s zPdfReader._find_eof_markerc C s~ t |}y t|}W nD tk
rX |jds6tdt|dd j }tdt Y n"X t |}|dd dkrztd|S )z5Find startxref entry - the location of the xref tables startxrefzstartxref not found Nz startxref on same line as offset)r r r0
startswithr* stripr ro )rL rv r: r1 rF rF rG r+ s
z PdfReader._find_startxref_posc C s |j d}|d d dkr"tdt| |jdd d}xttt|| }|rp|dkrp|| _| jrpt dt
d }t| |jdd ttt|| }t| |jd d d}x||k r|j d
}x&|d d kr|jd d |j d
}qW |d d kr
|jd d y8|d d
j d\}} |dd }
t|t| } } W n t k
r t
|drnt|j }
n(|j }|jdd |j d!}
|j| tj| dj |
}|d krt d| dt
d} d"} n(t d| dt
t|jd} |j } Y nX | | jkr i | j| < i | j| < || j| kr2nj| | j| |< y|
dk| j| |< W n t k
rl Y nX y|
dk| jd |< W n t k
r Y nX |d7 }|d7 }qW t| |jd#d |j d}|dkr|jd$d q>P q>W d S )%N r! s refzxref table read errorr Tr zFXref table not zero-indexed. ID numbers for objects will be corrected.F s
s 0123456789t r z\s+(\d+)\s+objzentry z( in Xref table invalid; object not foundi z' in Xref table invalid but object found f s trailerr r r iir r r r r i)r r* r r r
r r< r rw r ro splitr r r r r r r r groupr r r )rL rv refZ firsttimer rA Zcntr: Zoffset_bZ generation_bZ entry_type_br r r r r Z
trailertagrF rF rG _read_standard_xref_table s
z#PdfReader._read_standard_xref_table)rv r1 r2 rB c
C s i | _ i | _i | _t | _x|d k r|j|d |jd}|dkrP|jd}|dkrd| j|}q |ry| j| P W n t k
r d}Y nX q |j
ry| j |}W nV t k
r } z8t j
| jkrtd|j t P ntd|j W Y d d }~X nX t j
t jt jt jf}x8|D ]0}||kr || jkr |j|| jt|< q W d|kr|j } |jtt|d d d | j | |j| d d|krtt|d }nP q | j||}q W d S ) Nr r s
xz!Previous trailer can not be read ztrailer can not be read z/XRefStmz/Prev)r r r r1 r r r
_read_xref_rebuild_xref_tabler isdigit_read_pdf15_xref_streamr r r argsro r* r r r Zraw_getr6 r r
r _read_xref_other_error)
rL rv r1 r2 r
xrefstreameZ trailer_keysrM r rF rF rG r- : sP
"
z(PdfReader._read_xref_tables_and_trailersc C s | j | t| |jdd ttttf t|| }x(|j D ] \}}|| j kr@|| j |< q@W d|kr|j
}|jtt |d d d y| j | W n* t
k
r td|d dt Y nX |j|d d|kr|d }|S d S d S )Nr z/XRefStmr zXRef object at z, can not be read, some object may be missingz/Prevr )rI r r r
r rs r r< r r r r rN r r ro )rL rv new_trailerrM r r r1 rF rF rG rK k s*
zPdfReader._read_xref)rv r1 rB c C s |dkr$| j rtdtdt d S |jdd |jd}|jd}|dkr\|d| 8 }|S |j|d x*td D ] }|jdj rr||7 }|S qrW d
| j kr| j rtd t y| j
| d S t k
r td Y nX td
d S )Nr z6/Prev=0 in the trailer (try opening with strict=False)zA/Prev=0 in the trailer - assuming there is no previous xref table r r? s xref
r5 z/Rootz"Invalid parent xref., rebuild xrefzcan not rebuild xrefz/Could not find xref table at specified locationir ) rw r* r ro r r findr rM r rL r )rL rv r1 r Zxref_loclookrF rF rG rP s6
z PdfReader._read_xref_other_errorc s |j dd j|\}}ttt|}tt|d dks@tj||| tt |j
|j dd|j dg}tt t
t
f |j d t dkstjrt dkrtd tttttd
f f d fd d
}tttttd
f f tdfdd }j||| |S )Nr z/Typez/XRefz/Indexr z/Sizez/Wr! zToo many entry sizes: .)r rB c s< | dkr(j | }t| | S | dkr4dS dS d S )Nr r )r rH )r r@ ) entry_sizesr rF rG get_entry s z4PdfReader._read_pdf15_xref_stream..get_entry)r r rB c s | j j|g kp| jkS )N)r rN r )r r )rL rF rG used_before s z6PdfReader._read_pdf15_xref_stream..used_beforer )r r r
r. r< rs r r r r r rN r r r rw r* r r r bool_read_xref_subsections)rL rv r r rQ idx_pairsrY rZ rF )rX rL r rG rN s &&z!PdfReader._read_pdf15_xref_streamc C s | j |d d | jd}|dkr&dS | jd}|dkr~d}x"|dkr^| jd}|dkr>dS q>W || jd7 }|j d kr~d
S dS ) zBReturn an int which indicates an issue. 0 means there is no issue.r r s
r> s xrefr| s 0123456789 r s objr! )r r lower)rv r1 r: rF rF rG r, s
zPdfReader._get_xref_issuesc C s i | _ |jdd |jd}xXtjd|D ]H}t|jd}t|jd}|| j kr^i | j |< |jd| j | |< q*W |jdd xbtjd|D ]R}|j|jdd tt t
t
f t || }x"t |j
D ]\}}|| j|< qW qW d S )Nr r s( [\r\n \t][ \t]*(\d+)[ \t]+(\d+)[ \t]+objr s$ [\r\n \t][ \t]*trailer[\r\n \t]*(<<)r )r r r r finditerr rG r r
r r r< r r r ) rL rv Zf_r r r rS rM r rF rF rG rL s
z PdfReader._rebuild_xref_table.)r] rY rZ rB c C s d}x| j |D ]\}}||ks$t|| }xt||| D ]}|d}|dkrb|d} |d}
q<|dkr|d} |d} | | jkri | j| < ||| s| | j| |< q<|dkr|d}
|d}d} ||| s|
|f| j|< q<| jr d}x4|| ||d fV |d7 }|d t |krP qW d S )Nr r r )r )rL r r rF rF rG r` & s zPdfReader._pairsr )rv limit_offsetrB c C s t ddd g }x|j dks*|j |kr2td|jd}|j dk rPtd|jdd |dkrd }xN|dkr|jd}|dkr|jdd d }|j dk rtd|jdd qjW |j|rdndd P q|j| qW |j d
j|S )z.. deprecated:: 2.1.0read_next_end_linez4.0.0)Z
removed_inr z!Could not read malformed PDF filer r zEOL marker not found
FTr| )rc rd )rc rd )rc rd r re )r r r* r r r reversejoin)rL rv ra Z
line_partsr ZcrlfrF rF rG rb . s2
z PdfReader.read_next_end_linec C s t d | j||S )z.. deprecated:: 1.28.0readNextEndLine)r rb )rL rv ra rF rF rG rh O s zPdfReader.readNextEndLine)rx rB c C s | j std| j j|S )ag
When using an encrypted / secured PDF file with the PDF Standard
encryption handler, this function will allow the file to be decrypted.
It checks the given password against the document's user password and
owner password, and then stores the resulting decryption key if either
password is correct.
It does not matter which password was matched. Both passwords provide
the correct decryption key that will allow the document to be used with
this library.
:param str password: The password to match.
:return: `PasswordType`.
zNot encrypted file)r r* r )rL rx rF rF rG decryptV s zPdfReader.decrypt)permissions_coderB c C s i }|d@ dk|d< |d@ dk|d< |d@ dk|d< |d@ dk|d
< |d @ dk|d < |d @ dk|d< |d @ dk|d< |d"@ dk|d< |S )#Nr r! r printr> modifyr5 copy annotationsr; ZformsrU Z
accessabilityrT Zassemble Zprint_high_qualityr r> r! rC r> r@ r5 rC r; i rU i rT i rF )rL rj permissionsrF rF rG decode_permissionsj s z PdfReader.decode_permissionsc C s t j| jkS )z
Read-only boolean property showing whether this PDF file is encrypted.
Note that this property, if true, will remain true even after the
:meth:`decrypt()` method is called.
)r r r )rL rF rF rG r y s zPdfReader.is_encryptedc C s t dd | jS )zZ
.. deprecated:: 1.28.0
Use :py:attr:`is_encrypted` instead.
getIsEncryptedr )r r )rL rF rF rG ru s
zPdfReader.getIsEncryptedc C s t dd | jS )zZ
.. deprecated:: 1.28.0
Use :py:attr:`is_encrypted` instead.
isEncryptedr )r r )rL rF rF rG rv s
zPdfReader.isEncryptedc
C s d }i }t t| jtj }d|ks,|d r0d S t t|d }d|krt t|d }t|}xL|D ]D}|}t|}t |t
rbt t t |j
}|rbtj|j} | ||< qbW |S )Nz /AcroFormz/XFA)r
r1 r r r r; r- iternextrO r5 r
r2 rU zlib
decompress_data)
rL r rP r r r r tagr esrF rF rG xfa s$
z
PdfReader.xfa)FN)NNN)NNN)NN)NN)NN)NN)NNN)r )r )qro rp rq rr r r r r[ rs r rK rt r r
rJ r r r r? r r r r r r r r r r r r r r r; r r r1 r r r r r r r r= r r r r r- r r7 r5 r r r r0 r r r r8 r r r r r r r> r r r r r9 r
r
rU r r r r r$ r r% r r' r r) r* r+ rI r- rK rP r. r2 r/ rN staticmethodr, rL r r\ r r` rb rh r ri rt r ru rv r~ rF rF rF rG ru s / *
% /
(
B "*4
&n
/(& ru c s& e Zd Zeedd fdd Z ZS )
PdfFileReaderN)rO kwargsrB c s8 t dd d|kr&t|dk r&d|d< t j|| d S )Nr ru rw r T)r r superrK )rL rO r ) __class__rF rG rK s
zPdfFileReader.__init__)ro rp rq r rK
__classcell__rF rF )r rG r s r )Zr7 r rD ry r ior Zpathlibr typingr r r r r r
r r r
r r r Z_pager r _utilsr r r r r r r r r r r constantsr r r r r r! r r" rS r# r$ r% r r& r r' r errorsr( r) r* r+ r, Zgenericr- r. r/ r0 r1 r2 r3 r4 r5 r6 r7 r8 r9 r: r; r< r r= r> xmpr? r r rH rI rJ ru r rF rF rF rG sV , 4
H " ! &