Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Improve performance by not extracting compressed image data if retainImageContent was set to false #590

Merged
merged 12 commits into from
Apr 13, 2023
Merged

Improve performance by not extracting compressed image data if retainImageContent was set to false #590

Changes from 6 commits
Commits
Show all changes
12 commits
Select commit Hold shift + click to select a range
1709b82
skip image streams if retainImageContent is false
se-ti Mar 14, 2023
afe2a66
minor refactoring
se-ti Mar 26, 2023
2b9648d
use /Length field to help preg_match parse stream
se-ti Mar 26, 2023
d1a1173
check nullness explicitly
se-ti Mar 27, 2023
04a966c
fix doc style and braces
se-ti Mar 27, 2023
29a4a80
Fixed coding style issues; made getHeaderValue private
k00ni Mar 30, 2023
b8de42e
fix code style issues
se-ti Mar 30, 2023
6037b9c
try fix cs-fixer complains
se-ti Mar 30, 2023
450fdfd
micro: fix comparison style
se-ti Mar 31, 2023
0001a58
added further comments to explain new code parts
k00ni Apr 11, 2023
b0ba56f
Update RawDataParser.php
k00ni Apr 11, 2023
084a01f
fixed further coding style issues
k00ni Apr 11, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
54 changes: 47 additions & 7 deletions src/Smalot/PdfParser/RawData/RawDataParser.php
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -553,16 +553,18 @@ 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 +607,12 @@ 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 +761,20 @@ 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]);

$streamLen = (int) $this->getHeaderValue($headerDic, 'Length', 'numeric', 0);
k00ni marked this conversation as resolved.
Show resolved Hide resolved
$skip = !$this->config->getRetainImageContent() && 'XObject' == $this->getHeaderValue($headerDic, 'Type', '/') && 'Image' == $this->getHeaderValue($headerDic, 'Subtype', '/');
k00ni marked this conversation as resolved.
Show resolved Hide resolved

$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 +804,38 @@ protected function getRawObject(string $pdfData, int $offset = 0): array
return [$objtype, $objval, $offset];
}

/**
* Get value of an object header's section
*
* @param string $key header's section name
* @param string $type type of the section (i.e. 'numeric', '/', '<<', etc.)
* @param string|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, $default = '')
{
if (!\is_array($headerDic)) {
k00ni marked this conversation as resolved.
Show resolved Hide resolved
return $default;
}

foreach ($headerDic as $i => $val) {
if (
\is_array($val)
&& 3 == \count($val)
&& '/' == $val[0]
&& $val[1] == $key
&& isset($headerDic[$i + 1])
) {
return \is_array($headerDic[$i + 1]) && 1 < \count($headerDic[$i + 1]) && $type == $headerDic[$i + 1][0]
? $headerDic[$i + 1][1]
: $default;
}
k00ni marked this conversation as resolved.
Show resolved Hide resolved
}

return $default;
}

/**
* Get Cross-Reference (xref) table and trailer data from PDF document data.
*
Expand Down