Skip to content

Commit

Permalink
Improve performance by not extracting compressed image data if retain…
Browse files Browse the repository at this point in the history 
…ImageContent was set to false (#590)

* skip image streams if retainImageContent is false

* minor refactoring

* use /Length field to help preg_match parse stream

* check nullness explicitly

* fix doc style and braces

* Fixed coding style issues; made getHeaderValue private

Made getHeaderValue private, because it is only used internally. We don't wanna expand our API, if it can be avoided.

* fix code style issues

* try fix cs-fixer complains

* micro: fix comparison style

Co-authored-by: Konrad Abicht <hi@inspirito.de>

* added further comments to explain new code parts

* Update RawDataParser.php

* fixed further coding style issues

---------

Co-authored-by: Konrad Abicht <hi@inspirito.de>
  • Loading branch information
@se-ti @k00ni
se-ti and k00ni authored Apr 13, 2023
1 parent 68af1e3 commit 78cd9d4
Showing 1 changed file with 63 additions and 14 deletions.
77 changes: 63 additions & 14 deletions src/Smalot/PdfParser/RawData/RawDataParser.php
View file
Original file line number Diff line number Diff line change
Expand Up @@ -262,8 +262,7 @@ protected function decodeXrefStream(string $pdfData, int $startxref, array $xref
if (
('/' == $v[0])
&& ('Type' == $v[1])
&& (
isset($sarr[$k + 1])
&& (isset($sarr[$k + 1])
&& '/' == $sarr[$k + 1][0]
&& 'XRef' == $sarr[$k + 1][1]
)
Expand All @@ -289,17 +288,15 @@ protected function decodeXrefStream(string $pdfData, int $startxref, array $xref
if (
'/' == $vdc[0]
&& 'Columns' == $vdc[1]
&& (
isset($decpar[$kdc + 1])
&& (isset($decpar[$kdc + 1])
&& 'numeric' == $decpar[$kdc + 1][0]
)
) {
$columns = (int) $decpar[$kdc + 1][1];
} elseif (
'/' == $vdc[0]
&& 'Predictor' == $vdc[1]
&& (
isset($decpar[$kdc + 1])
&& (isset($decpar[$kdc + 1])
&& 'numeric' == $decpar[$kdc + 1][0]
)
) {
Expand Down Expand Up @@ -553,16 +550,18 @@ protected function getIndirectObject(string $pdfData, array $xref, string $objRe
$offset += $objHeaderLen;
$objContentArr = [];
$i = 0; // object main index
$header = null;
do {
$oldOffset = $offset;
// get element
$element = $this->getRawObject($pdfData, $offset);
$element = $this->getRawObject($pdfData, $offset, null != $header ? $header[1] : null);
$offset = $element[2];
// decode stream using stream's dictionary information
if ($decoding && ('stream' === $element[0]) && (isset($objContentArr[$i - 1][0])) && ('<<' === $objContentArr[$i - 1][0])) {
$element[3] = $this->decodeStream($pdfData, $xref, $objContentArr[$i - 1][1], $element[1]);
if ($decoding && ('stream' === $element[0]) && null != $header) {
$element[3] = $this->decodeStream($pdfData, $xref, $header[1], $element[1]);
}
$objContentArr[$i] = $element;
$header = isset($element[0]) && '<<' === $element[0] ? $element : null;
++$i;
} while (('endobj' !== $element[0]) && ($offset !== $oldOffset));
// remove closing delimiter
Expand Down Expand Up @@ -605,11 +604,12 @@ protected function getObjectVal(string $pdfData, $xref, array $obj): array
/**
* Get object type, raw value and offset to next object
*
* @param int $offset Object offset
* @param int $offset Object offset
* @param array|null $headerDic obj header's dictionary, parsed by getRawObject. Used for stream parsing optimization
*
* @return array containing object type, raw value and offset to next object
*/
protected function getRawObject(string $pdfData, int $offset = 0): array
protected function getRawObject(string $pdfData, int $offset = 0, ?array $headerDic = null): array
{
$objtype = ''; // object type to be returned
$objval = ''; // object value to be returned
Expand Down Expand Up @@ -758,15 +758,21 @@ protected function getRawObject(string $pdfData, int $offset = 0): array
$offset += 6;
if (1 == preg_match('/^([\r]?[\n])/isU', substr($pdfData, $offset, 4), $matches)) {
$offset += \strlen($matches[0]);

// we get stream length here to later help preg_match test less data
$streamLen = (int) $this->getHeaderValue($headerDic, 'Length', 'numeric', 0);
$skip = false === $this->config->getRetainImageContent() && 'XObject' == $this->getHeaderValue($headerDic, 'Type', '/') && 'Image' == $this->getHeaderValue($headerDic, 'Subtype', '/');

$pregResult = preg_match(
'/(endstream)[\x09\x0a\x0c\x0d\x20]/isU',
$pdfData,
$matches,
\PREG_OFFSET_CAPTURE,
$offset
$offset + $streamLen
);

if (1 == $pregResult) {
$objval = substr($pdfData, $offset, $matches[0][1] - $offset);
$objval = $skip ? '' : substr($pdfData, $offset, $matches[0][1] - $offset);
$offset = $matches[1][1];
}
}
Expand Down Expand Up @@ -796,6 +802,48 @@ protected function getRawObject(string $pdfData, int $offset = 0): array
return [$objtype, $objval, $offset];
}

/**
* Get value of an object header's section (obj << YYY >> part ).
*
* It is similar to Header::get('...')->getContent(), the only difference is it can be used during the parsing process,
* when no Smalot\PdfParser\Header objects are created yet.
*
* @param string $key header's section name
* @param string $type type of the section (i.e. 'numeric', '/', '<<', etc.)
* @param string|array|null $default default value for header's section
*
* @return string|array|null value of obj header's section, or default value if none found, or its type doesn't match $type param
*/
private function getHeaderValue(?array $headerDic, string $key, string $type, $default = '')
{
if (false === \is_array($headerDic)) {
return $default;
}

/*
* It recieves dictionary of header fields, as it is returned by RawDataParser::getRawObject,
* iterates over it, searching for section of type '/' whith requested key.
* If such a section is found, it tries to receive it's value (next object in dictionary),
* returning it, if it matches requested type, or default value otherwise.
*/
foreach ($headerDic as $i => $val) {
$isSectionName = \is_array($val) && 3 == \count($val) && '/' == $val[0];
if (
$isSectionName
&& $val[1] == $key
&& isset($headerDic[$i + 1])
) {
$isSectionValue = \is_array($headerDic[$i + 1]) && 1 < \count($headerDic[$i + 1]);

return $isSectionValue && $type == $headerDic[$i + 1][0]
? $headerDic[$i + 1][1]
: $default;
}
}

return $default;
}

/**
* Get Cross-Reference (xref) table and trailer data from PDF document data.
*
Expand All @@ -821,7 +869,8 @@ protected function getXrefData(string $pdfData, int $offset = 0, array $xref = [
// find last startxref
$pregResult = preg_match_all(
'/[\r\n]startxref[\s]*[\r\n]+([0-9]+)[\s]*[\r\n]+%%EOF/i',
$pdfData, $matches,
$pdfData,
$matches,
\PREG_SET_ORDER,
$offset
);
Expand Down

0 comments on commit 78cd9d4

Please sign in to comment.