DocParser::assertValidUTF8() - Code Metrics - Inspection of "Merge pull request #409 from thephpleague/gfm-exte..." - thephpleague/commonmark - Measure and Improve Code Quality continuously with Scrutinizer

Completed

Push — master ( 83739a...a13ee4 )

by Colin

created 2020-02-08 22:57 UTC

DocParser::assertValidUTF8() A

↳ Parent: DocParser

Complexity

Conditions	2
Paths	2

Size

Total Lines

Duplication

Lines	0
Ratio	0 %

Code Coverage

Tests	4
CRAP Score	2

Importance

Changes

Metric	Value
dl	0
loc	6
ccs	4
cts	4
cp	1
rs	10
c	0
b	0
f	0
cc	2
nc	2
nop	1
crap	2

<?php

/*
 * This file is part of the league/commonmark package.
 *
 * (c) Colin O'Dell <[email protected]>
 *
 * Original code based on the CommonMark JS reference parser (https://bitly.com/commonmark-js)
 *  - (c) John MacFarlane
 *
 * For the full copyright and license information, please view the LICENSE
 * file that was distributed with this source code.
 */

namespace League\CommonMark;

use League\CommonMark\Block\Element\AbstractBlock;
use League\CommonMark\Block\Element\AbstractStringContainerBlock;
use League\CommonMark\Block\Element\Document;
use League\CommonMark\Block\Element\Paragraph;
use League\CommonMark\Block\Element\StringContainerInterface;
use League\CommonMark\Event\DocumentParsedEvent;
use League\CommonMark\Exception\UnexpectedEncodingException;

final class DocParser implements DocParserInterface
{
    /**
     * @var EnvironmentInterface
     */
    private $environment;

    /**
     * @var InlineParserEngine
     */
    private $inlineParserEngine;

    /**
     * @var int|float
     */
    private $maxNestingLevel;

    /**
     * @param EnvironmentInterface $environment
     */
    public function __construct(EnvironmentInterface $environment)
    {
        $this->environment = $environment;
        $this->inlineParserEngine = new InlineParserEngine($environment);
        $this->maxNestingLevel = $environment->getConfig('max_nesting_level', \INF);
    }

    /**
     * @param string $input
     *
     * @return string[]
     */
    private function preProcessInput(string $input): array
    {
        /** @var string[] $lines */
        $lines = \preg_split('/\r\n|\n|\r/', $input);

        // Remove any newline which appears at the very end of the string.
        // We've already split the document by newlines, so we can simply drop
        // any empty element which appears on the end.
        if (\end($lines) === '') {
            \array_pop($lines);
        }

        return $lines;
    }

    /**
     * @param string $input
     *
     * @throws \RuntimeException
     *
     * @return Document
     */
    public function parse(string $input): Document
    {
        $document = new Document();
        $context = new Context($document, $this->environment);

        $this->assertValidUTF8($input);
        $lines = $this->preProcessInput($input);
        foreach ($lines as $line) {
            $context->setNextLine($line);
            $this->incorporateLine($context);
        }

        $lineCount = \count($lines);
        while ($tip = $context->getTip()) {
            $tip->finalize($context, $lineCount);
        }

        $this->processInlines($context);

        $this->environment->dispatch(new DocumentParsedEvent($document));

        return $document;
    }

    private function incorporateLine(ContextInterface $context)
    {
        $context->getBlockCloser()->resetTip();
        $context->setBlocksParsed(false);

        $cursor = new Cursor($context->getLine());

        $this->resetContainer($context, $cursor);
        $context->getBlockCloser()->setLastMatchedContainer($context->getContainer());

        $this->parseBlocks($context, $cursor);

        // What remains at the offset is a text line.  Add the text to the appropriate container.
        // First check for a lazy paragraph continuation:
        if ($this->handleLazyParagraphContinuation($context, $cursor)) {
            return;
        }

        // not a lazy continuation
        // finalize any blocks not matched
        $context->getBlockCloser()->closeUnmatchedBlocks();

        // Determine whether the last line is blank, updating parents as needed
        $this->setAndPropagateLastLineBlank($context, $cursor);

        // Handle any remaining cursor contents
        if ($context->getContainer() instanceof StringContainerInterface) {
            $context->getContainer()->handleRemainingContents($context, $cursor);
        } elseif (!$cursor->isBlank()) {
            // Create paragraph container for line
            $p = new Paragraph();
            $context->addBlock($p);
            $cursor->advanceToNextNonSpaceOrTab();
            $p->addLine($cursor->getRemainder());
        }
    }

    private function processInlines(ContextInterface $context)
    {
        $walker = $context->getDocument()->walker();

        while ($event = $walker->next()) {
            if (!$event->isEntering()) {
                continue;
            }

            $node = $event->getNode();
            if ($node instanceof AbstractStringContainerBlock) {
                $this->inlineParserEngine->parse($node, $context->getDocument()->getReferenceMap());
            }
        }
    }

    /**
     * Sets the container to the last open child (or its parent)
     *
     * @param ContextInterface $context
     * @param Cursor           $cursor
     */
    private function resetContainer(ContextInterface $context, Cursor $cursor)
    {
        $container = $context->getDocument();

        while ($lastChild = $container->lastChild()) {
            if (!($lastChild instanceof AbstractBlock)) {
                break;
            }

            if (!$lastChild->isOpen()) {
                break;
            }

            $container = $lastChild;
            if (!$container->matchesNextLine($cursor)) {
                $container = $container->parent(); // back up to the last matching block
                break;
            }
        }

        $context->setContainer($container);
    }

    /**
     * Parse blocks
     *
     * @param ContextInterface $context
     * @param Cursor           $cursor
     */
    private function parseBlocks(ContextInterface $context, Cursor $cursor)
    {
        while (!$context->getContainer()->isCode() && !$context->getBlocksParsed()) {
            $parsed = false;
            foreach ($this->environment->getBlockParsers() as $parser) {
                if ($parser->parse($context, $cursor)) {
                    $parsed = true;
                    break;
                }
            }

            if (!$parsed || $context->getContainer() instanceof StringContainerInterface || (($tip = $context->getTip()) && $tip->getDepth() >= $this->maxNestingLevel)) {
                $context->setBlocksParsed(true);
                break;
            }
        }
    }

    /**
     * @param ContextInterface $context
     * @param Cursor           $cursor
     *
     * @return bool
     */
    private function handleLazyParagraphContinuation(ContextInterface $context, Cursor $cursor): bool
    {
        $tip = $context->getTip();

        if ($tip instanceof Paragraph &&
            !$context->getBlockCloser()->areAllClosed() &&
            !$cursor->isBlank() &&
            \count($tip->getStrings()) > 0) {

            // lazy paragraph continuation
            $tip->addLine($cursor->getRemainder());

            return true;
        }

        return false;
    }

    /**
     * @param ContextInterface $context
     * @param Cursor           $cursor
     */
    private function setAndPropagateLastLineBlank(ContextInterface $context, Cursor $cursor)
    {
        $container = $context->getContainer();

        if ($cursor->isBlank() && $lastChild = $container->lastChild()) {
            if ($lastChild instanceof AbstractBlock) {
                $lastChild->setLastLineBlank(true);
            }
        }

        $lastLineBlank = $container->shouldLastLineBeBlank($cursor, $context->getLineNumber());

        // Propagate lastLineBlank up through parents:
        while ($container instanceof AbstractBlock && $container->endsWithBlankLine() !== $lastLineBlank) {
            $container->setLastLineBlank($lastLineBlank);
            $container = $container->parent();
        }
    }

    private function assertValidUTF8(string $input)
    {
        if (!\mb_check_encoding($input, 'UTF-8')) {
            throw new UnexpectedEncodingException('Unexpected encoding - UTF-8 or ASCII was expected');
        }
    }
}


1		<?php
2
3		/*
4		* This file is part of the league/commonmark package.
5		*
6		* (c) Colin O'Dell <[email protected]>
7		*
8		* Original code based on the CommonMark JS reference parser (https://bitly.com/commonmark-js)
9		* - (c) John MacFarlane
10		*
11		* For the full copyright and license information, please view the LICENSE
12		* file that was distributed with this source code.
13		*/
14
15		namespace League\CommonMark;
16
17		use League\CommonMark\Block\Element\AbstractBlock;
18		use League\CommonMark\Block\Element\AbstractStringContainerBlock;
19		use League\CommonMark\Block\Element\Document;
20		use League\CommonMark\Block\Element\Paragraph;
21		use League\CommonMark\Block\Element\StringContainerInterface;
22		use League\CommonMark\Event\DocumentParsedEvent;
23		use League\CommonMark\Exception\UnexpectedEncodingException;
24
25		final class DocParser implements DocParserInterface
26		{
27		/**
28		* @var EnvironmentInterface
29		*/
30		private $environment;
31
32		/**
33		* @var InlineParserEngine
34		*/
35		private $inlineParserEngine;
36
37		/**
38		* @var int\|float
39		*/
40		private $maxNestingLevel;
41
42		/**
43		* @param EnvironmentInterface $environment
44		*/
45	2442	public function __construct(EnvironmentInterface $environment)
46		{
47	2442	$this->environment = $environment;
48	2442	$this->inlineParserEngine = new InlineParserEngine($environment);
49	2442	$this->maxNestingLevel = $environment->getConfig('max_nesting_level', \INF);
50	2442	}
51
52		/**
53		* @param string $input
54		*
55		* @return string[]
56		*/
57	2427	private function preProcessInput(string $input): array
58		{
59		/** @var string[] $lines */
60	2427	$lines = \preg_split('/\r\n\|\n\|\r/', $input);
61
62		// Remove any newline which appears at the very end of the string.
63		// We've already split the document by newlines, so we can simply drop
64		// any empty element which appears on the end.
65	2427	if (\end($lines) === '') {
66	2187	\array_pop($lines);
67		}
68
69	2427	return $lines;
70		}
71
72		/**
73		* @param string $input
74		*
75		* @throws \RuntimeException
76		*
77		* @return Document
78		*/
79	2433	public function parse(string $input): Document
80		{
81	2433	$document = new Document();
82	2433	$context = new Context($document, $this->environment);
83
84	2433	$this->assertValidUTF8($input);
85	2427	$lines = $this->preProcessInput($input);
86	2427	foreach ($lines as $line) {
87	2427	$context->setNextLine($line);
88	2427	$this->incorporateLine($context);
89		}
90
91	2427	$lineCount = \count($lines);
92	2427	while ($tip = $context->getTip()) {
93	2427	$tip->finalize($context, $lineCount);
94		}
95
96	2427	$this->processInlines($context);
97
98	2427	$this->environment->dispatch(new DocumentParsedEvent($document));
99
100	2427	return $document;
101		}
102
103	2427	private function incorporateLine(ContextInterface $context)
104		{
105	2427	$context->getBlockCloser()->resetTip();
106	2427	$context->setBlocksParsed(false);
107
108	2427	$cursor = new Cursor($context->getLine());
109
110	2427	$this->resetContainer($context, $cursor);
111	2427	$context->getBlockCloser()->setLastMatchedContainer($context->getContainer());
112
113	2427	$this->parseBlocks($context, $cursor);
114
115		// What remains at the offset is a text line. Add the text to the appropriate container.
116		// First check for a lazy paragraph continuation:
117	2427	if ($this->handleLazyParagraphContinuation($context, $cursor)) {
118	36	return;
119		}
120
121		// not a lazy continuation
122		// finalize any blocks not matched
123	2427	$context->getBlockCloser()->closeUnmatchedBlocks();
124
125		// Determine whether the last line is blank, updating parents as needed
126	2427	$this->setAndPropagateLastLineBlank($context, $cursor);
127
128		// Handle any remaining cursor contents
129	2427	if ($context->getContainer() instanceof StringContainerInterface) {
130	864	$context->getContainer()->handleRemainingContents($context, $cursor);
131	2175	} elseif (!$cursor->isBlank()) {
132		// Create paragraph container for line
133	2094	$p = new Paragraph();
134	2094	$context->addBlock($p);
135	2094	$cursor->advanceToNextNonSpaceOrTab();
136	2094	$p->addLine($cursor->getRemainder());
137		}
138	2427	}
139
140	2427	private function processInlines(ContextInterface $context)
141		{
142	2427	$walker = $context->getDocument()->walker();
143
144	2427	while ($event = $walker->next()) {
145	2427	if (!$event->isEntering()) {
146	2427	continue;
147		}
148
149	2427	$node = $event->getNode();
150	2427	if ($node instanceof AbstractStringContainerBlock) {
151	2385	$this->inlineParserEngine->parse($node, $context->getDocument()->getReferenceMap());
152		}
153		}
154	2427	}
155
156		/**
157		* Sets the container to the last open child (or its parent)
158		*
159		* @param ContextInterface $context
160		* @param Cursor $cursor
161		*/
162	2427	private function resetContainer(ContextInterface $context, Cursor $cursor)
163		{
164	2427	$container = $context->getDocument();
165
166	2427	while ($lastChild = $container->lastChild()) {
167	1263	if (!($lastChild instanceof AbstractBlock)) {
168		break;
169		}
170
171	1263	if (!$lastChild->isOpen()) {
172	513	break;
173		}
174
175	1254	$container = $lastChild;
176	1254	if (!$container->matchesNextLine($cursor)) {
177	822	$container = $container->parent(); // back up to the last matching block
178	822	break;
179		}
180		}
181
182	2427	$context->setContainer($container);
183	2427	}
184
185		/**
186		* Parse blocks
187		*
188		* @param ContextInterface $context
189		* @param Cursor $cursor
190		*/
191	2427	private function parseBlocks(ContextInterface $context, Cursor $cursor)
192		{
193	2427	while (!$context->getContainer()->isCode() && !$context->getBlocksParsed()) {
194	2427	$parsed = false;
195	2427	foreach ($this->environment->getBlockParsers() as $parser) {
196	2427	if ($parser->parse($context, $cursor)) {
197	918	$parsed = true;
198	918	break;
199		}
200		}
201
202	2427	if (!$parsed \|\| $context->getContainer() instanceof StringContainerInterface \|\| (($tip = $context->getTip()) && $tip->getDepth() >= $this->maxNestingLevel)) {
203	2400	$context->setBlocksParsed(true);
204	2400	break;
205		}
206		}
207	2427	}
208
209		/**
210		* @param ContextInterface $context
211		* @param Cursor $cursor
212		*
213		* @return bool
214		*/
215	2427	private function handleLazyParagraphContinuation(ContextInterface $context, Cursor $cursor): bool
216		{
217	2427	$tip = $context->getTip();
218
219	2427	if ($tip instanceof Paragraph &&
220	2427	!$context->getBlockCloser()->areAllClosed() &&
221	2427	!$cursor->isBlank() &&
222	2427	\count($tip->getStrings()) > 0) {
223
224		// lazy paragraph continuation
225	36	$tip->addLine($cursor->getRemainder());
226
227	36	return true;
228		}
229
230	2427	return false;
231		}
232
233		/**
234		* @param ContextInterface $context
235		* @param Cursor $cursor
236		*/
237	2427	private function setAndPropagateLastLineBlank(ContextInterface $context, Cursor $cursor)
238		{
239	2427	$container = $context->getContainer();
240
241	2427	if ($cursor->isBlank() && $lastChild = $container->lastChild()) {
242	579	if ($lastChild instanceof AbstractBlock) {
243	579	$lastChild->setLastLineBlank(true);
244		}
245		}
246
247	2427	$lastLineBlank = $container->shouldLastLineBeBlank($cursor, $context->getLineNumber());
248
249		// Propagate lastLineBlank up through parents:
250	2427	while ($container instanceof AbstractBlock && $container->endsWithBlankLine() !== $lastLineBlank) {
251	729	$container->setLastLineBlank($lastLineBlank);
252	729	$container = $container->parent();
253		}
254	2427	}
255
256	2433	private function assertValidUTF8(string $input)
257		{
258	2433	if (!\mb_check_encoding($input, 'UTF-8')) {
259	6	throw new UnexpectedEncodingException('Unexpected encoding - UTF-8 or ASCII was expected');
260		}
261	2427	}
262		}
263

thephpleague / commonmark

Push — master ( 83739a...a13ee4 )

DocParser::assertValidUTF8() A

Complexity

Size

Duplication

Code Coverage

Importance

Duplication Side-by-Side

Filter issues like