CompletionContext::getTokenValue() - Code Metrics - Inspection of "WIP: Rewrite command-line splitting to tokenize qu..." - stecman/symfony-console-completion - Measure and Improve Code Quality continuously with Scrutinizer

Completed

Pull Request — master (#82)

by Stephen

created 2019-01-07 03:03 UTC

CompletionContext::getTokenValue() A

↳ Parent: CompletionContext

Complexity

Conditions	2
Paths	2

Size

Total Lines

Duplication

Lines	0
Ratio	0 %

Importance

Changes

Metric	Value
dl	0
loc	14
rs	9.7998
c	0
b	0
f	0
cc	2
nc	2
nop	1

<?php


namespace Stecman\Component\Symfony\Console\BashCompletion;

/**
 * Command line context for completion
 *
 * Represents the current state of the command line that is being completed
 */
class CompletionContext
{
    /**
     * The current contents of the command line as a single string
     *
     * Bash equivalent: COMP_LINE
     *
     * @var string
     */
    protected $commandLine;

    /**
     * The index of the user's cursor relative to the start of the command line.
     *
     * If the current cursor position is at the end of the current command,
     * the value of this variable is equal to the length of $this->commandLine
     *
     * Bash equivalent: COMP_POINT
     *
     * @var int
     */
    protected $charIndex = 0;

    /**
     * An array containing the individual words in the current command line.
     *
     * This is not set until $this->splitCommand() is called, when it is populated by
     * $commandLine exploded by $wordBreaks
     *
     * Bash equivalent: COMP_WORDS
     *
     * @var array|null
     */
    protected $words = null;

    /**
     * The index in $this->words containing the word at the current cursor position.
     *
     * This is not set until $this->splitCommand() is called.
     *
     * Bash equivalent: COMP_CWORD
     *
     * @var int|null
     */
    protected $wordIndex = null;

    /**
     * Characters that $this->commandLine should be split on to get a list of individual words
     *
     * Bash equivalent: COMP_WORDBREAKS
     *
     * @var string
     */
    protected $wordBreaks = "= \t\n";

    /**
     * Set the whole contents of the command line as a string
     *
     * @param string $commandLine
     */
    public function setCommandLine($commandLine)
    {
        $this->commandLine = $commandLine;
        $this->reset();
    }

    /**
     * Return the current command line verbatim as a string
     *
     * @return string
     */
    public function getCommandLine()
    {
        return $this->commandLine;
    }

    /**
     * Return the word from the command line that the cursor is currently in
     *
     * Most of the time this will be a partial word. If the cursor has a space before it,
     * this will return an empty string, indicating a new word.
     *
     * @return string
     */
    public function getCurrentWord()
    {
        if (isset($this->words[$this->wordIndex])) {
            return $this->words[$this->wordIndex];
        }

        return '';
    }

    /**
     * Return a word by index from the command line
     *
     * @see $words, $wordBreaks
     * @param int $index
     * @return string
     */
    public function getWordAtIndex($index)
    {
        if (isset($this->words[$index])) {
            return $this->words[$index];
        }

        return '';
    }

    /**
     * Get the contents of the command line, exploded into words based on the configured word break characters
     *
     * @see $wordBreaks, setWordBreaks
     * @return array
     */
    public function getWords()
    {
        if ($this->words === null) {
            $this->splitCommand();
        }

        return $this->words;
    }

    /**
     * Get the index of the word the cursor is currently in
     *
     * @see getWords, getCurrentWord
     * @return int
     */
    public function getWordIndex()
    {
        if ($this->wordIndex === null) {
            $this->splitCommand();
        }

        return $this->wordIndex;
    }

    /**
     * Get the character index of the user's cursor on the command line
     *
     * This is in the context of the full command line string, so includes word break characters.
     * Note that some shells can only provide an approximation for character index. Under ZSH for
     * example, this will always be the character at the start of the current word.
     *
     * @return int
     */
    public function getCharIndex()
    {
        return $this->charIndex;
    }

    /**
     * Set the cursor position as a character index relative to the start of the command line
     *
     * @param int $index
     */
    public function setCharIndex($index)
    {
        $this->charIndex = $index;
        $this->reset();
    }

    /**
     * Set characters to use as split points when breaking the command line into words
     *
     * This defaults to a sane value based on BASH's word break characters and shouldn't
     * need to be changed unless your completions contain the default word break characters.
     *
     * @deprecated This is becoming an internal setting that doesn't make sense to expose publicly.
     *
     * @see wordBreaks
     * @param string $charList - a single string containing all of the characters to break words on
     */
    public function setWordBreaks($charList)
    {
        // Drop quotes from break characters - strings are handled separately to word breaks now
        $this->wordBreaks = str_replace(array('"', '\''), '', $charList);;
        $this->reset();
    }

    /**
     * Split the command line into words using the configured word break characters
     *
     * @return string[]
     */
    protected function splitCommand()
    {
        $tokens = $this->tokenizeString($this->commandLine);

        foreach ($tokens as $token) {
            if ($token['type'] != 'break') {
                $this->words[] = $this->getTokenValue($token);
            }

            // Determine which word index the cursor is inside once we reach it's offset
            if ($this->wordIndex === null && $this->charIndex <= $token['offsetEnd']) {
                $this->wordIndex = count($this->words) - 1;

                if ($token['type'] == 'break') {
                    // Cursor is in the break-space after a word
                    // Push an empty word at the cursor to allow completion of new terms at the cursor, ignoring words ahead
                    $this->wordIndex++;
                    $this->words[] = '';
                    continue;
                }

                if ($this->charIndex < $token['offsetEnd']) {
                    // Cursor is inside the current word - truncate the word at the cursor
                    // (This emulates normal BASH completion behaviour I've observed, though I'm not entirely sure if it's useful)
                    $relativeOffset = $this->charIndex - $token['offset'];
                    $truncated = substr($token['value'], 0, $relativeOffset);

                    $this->words[$this->wordIndex] = $truncated;
                }
            }
        }

        // Cursor position is past the end of the command line string - consider it a new word
        if ($this->wordIndex === null) {
            $this->wordIndex = count($this->words);
            $this->words[] = '';
        }
    }

    /**
     * Return a token's value with escaping and quotes removed
     *
     * @see self::tokenizeString()
     * @param array $token
     * @return string
     */
    protected function getTokenValue($token)
    {
        $value = $token['value'];

        // Remove outer quote characters (or first quote if unclosed)
        if ($token['type'] == 'quoted') {
            $value = preg_replace('/^(?:[\'"])(.*?)(?:[\'"])?$/', '$1', $value);
        }

        // Remove escape characters
        $value = preg_replace('/\\\\(.)/', '$1', $value);

        return $value;
    }

    /**
     * Break a string into words, quoted strings and non-words (breaks)
     *
     * Returns an array of unmodified segments of $string with offset and type information.
     *
     * @param string $string
     * @return array as [ [type => string, value => string, offset => int], ... ]
     */
    protected function tokenizeString($string)
    {
        // Map capture groups to returned token type
        $typeMap = array(
            'double_quote_string' => 'quoted',
            'single_quote_string' => 'quoted',
            'word' => 'word',
            'break' => 'break',
        );

        // Escape every word break character including whitespace
        // preg_quote won't work here as it doesn't understand the ignore whitespace flag ("x")
        $breaks = preg_replace('/(.)/', '\\\$1', $this->wordBreaks);

        $pattern = <<<"REGEX"
            /(?:
                (?P<double_quote_string>
                    "(\\\\.|[^\"\\\\])*(?:"|$)
                ) |
                (?P<single_quote_string>
                    '(\\\\.|[^'\\\\])*(?:'|$)
                ) |
                (?P<word>
                    (?:\\\\.|[^$breaks])+
                ) |
                (?P<break>
                     [$breaks]+
                )
            )/x
REGEX;

        $tokens = array();

        if (!preg_match_all($pattern, $string, $matches, PREG_OFFSET_CAPTURE | PREG_SET_ORDER)) {
            return $tokens;
        }

        foreach ($matches as $set) {
            foreach ($set as $groupName => $match) {

                // Ignore integer indices preg_match outputs (duplicates of named groups)
                if (is_integer($groupName)) {
                    continue;
                }

                // Skip if the offset indicates this group didn't match
                if ($match[1] === -1) {

                    continue;
                }

                $tokens[] = array(
                    'type' => $typeMap[$groupName],
                    'value' => $match[0],
                    'offset' => $match[1],
                    'offsetEnd' => $match[1] + strlen($match[0])
                );

                // Move to the next set (only one group should match per set)
                continue;
            }
        }

        return $tokens;
    }

    /**
     * Reset the computed words so that $this->splitWords is forced to run again
     */
    protected function reset()
    {
        $this->words = null;
        $this->wordIndex = null;
    }
}


1			<?php
2
3
4			namespace Stecman\Component\Symfony\Console\BashCompletion;
5
6			/**
7			* Command line context for completion
8			*
9			* Represents the current state of the command line that is being completed
10			*/
11			class CompletionContext
12			{
13			/**
14			* The current contents of the command line as a single string
15			*
16			* Bash equivalent: COMP_LINE
17			*
18			* @var string
19			*/
20			protected $commandLine;
21
22			/**
23			* The index of the user's cursor relative to the start of the command line.
24			*
25			* If the current cursor position is at the end of the current command,
26			* the value of this variable is equal to the length of $this->commandLine
27			*
28			* Bash equivalent: COMP_POINT
29			*
30			* @var int
31			*/
32			protected $charIndex = 0;
33
34			/**
35			* An array containing the individual words in the current command line.
36			*
37			* This is not set until $this->splitCommand() is called, when it is populated by
38			* $commandLine exploded by $wordBreaks
39			*
40			* Bash equivalent: COMP_WORDS
41			*
42			* @var array\|null
43			*/
44			protected $words = null;
45
46			/**
47			* The index in $this->words containing the word at the current cursor position.
48			*
49			* This is not set until $this->splitCommand() is called.
50			*
51			* Bash equivalent: COMP_CWORD
52			*
53			* @var int\|null
54			*/
55			protected $wordIndex = null;
56
57			/**
58			* Characters that $this->commandLine should be split on to get a list of individual words
59			*
60			* Bash equivalent: COMP_WORDBREAKS
61			*
62			* @var string
63			*/
64			protected $wordBreaks = "= \t\n";
65
66			/**
67			* Set the whole contents of the command line as a string
68			*
69			* @param string $commandLine
70			*/
71			public function setCommandLine($commandLine)
72			{
73			$this->commandLine = $commandLine;
74			$this->reset();
75			}
76
77			/**
78			* Return the current command line verbatim as a string
79			*
80			* @return string
81			*/
82			public function getCommandLine()
83			{
84			return $this->commandLine;
85			}
86
87			/**
88			* Return the word from the command line that the cursor is currently in
89			*
90			* Most of the time this will be a partial word. If the cursor has a space before it,
91			* this will return an empty string, indicating a new word.
92			*
93			* @return string
94			*/
95			public function getCurrentWord()
96			{
97			if (isset($this->words[$this->wordIndex])) {
98			return $this->words[$this->wordIndex];
99			}
100
101			return '';
102			}
103
104			/**
105			* Return a word by index from the command line
106			*
107			* @see $words, $wordBreaks
108			* @param int $index
109			* @return string
110			*/
111			public function getWordAtIndex($index)
112			{
113			if (isset($this->words[$index])) {
114			return $this->words[$index];
115			}
116
117			return '';
118			}
119
120			/**
121			* Get the contents of the command line, exploded into words based on the configured word break characters
122			*
123			* @see $wordBreaks, setWordBreaks
124			* @return array
125			*/
126			public function getWords()
127			{
128			if ($this->words === null) {
129			$this->splitCommand();
130			}
131
132			return $this->words;
133			}
134
135			/**
136			* Get the index of the word the cursor is currently in
137			*
138			* @see getWords, getCurrentWord
139			* @return int
140			*/
141			public function getWordIndex()
142			{
143			if ($this->wordIndex === null) {
144			$this->splitCommand();
145			}
146
147			return $this->wordIndex;
148			}
149
150			/**
151			* Get the character index of the user's cursor on the command line
152			*
153			* This is in the context of the full command line string, so includes word break characters.
154			* Note that some shells can only provide an approximation for character index. Under ZSH for
155			* example, this will always be the character at the start of the current word.
156			*
157			* @return int
158			*/
159			public function getCharIndex()
160			{
161			return $this->charIndex;
162			}
163
164			/**
165			* Set the cursor position as a character index relative to the start of the command line
166			*
167			* @param int $index
168			*/
169			public function setCharIndex($index)
170			{
171			$this->charIndex = $index;
172			$this->reset();
173			}
174
175			/**
176			* Set characters to use as split points when breaking the command line into words
177			*
178			* This defaults to a sane value based on BASH's word break characters and shouldn't
179			* need to be changed unless your completions contain the default word break characters.
180			*
181			* @deprecated This is becoming an internal setting that doesn't make sense to expose publicly.
182			*
183			* @see wordBreaks
184			* @param string $charList - a single string containing all of the characters to break words on
185			*/
186			public function setWordBreaks($charList)
187			{
188			// Drop quotes from break characters - strings are handled separately to word breaks now
189			$this->wordBreaks = str_replace(array('"', '\''), '', $charList);;
190			$this->reset();
191			}
192
193			/**
194			* Split the command line into words using the configured word break characters
195			*
196			* @return string[]
197			*/
198			protected function splitCommand()
199			{
200			$tokens = $this->tokenizeString($this->commandLine);
201
202			foreach ($tokens as $token) {
203			if ($token['type'] != 'break') {
204			$this->words[] = $this->getTokenValue($token);
205			}
206
207			// Determine which word index the cursor is inside once we reach it's offset
208			if ($this->wordIndex === null && $this->charIndex <= $token['offsetEnd']) {
209			$this->wordIndex = count($this->words) - 1;
210
211			if ($token['type'] == 'break') {
212			// Cursor is in the break-space after a word
213			// Push an empty word at the cursor to allow completion of new terms at the cursor, ignoring words ahead
214			$this->wordIndex++;
215			$this->words[] = '';
216			continue;
217			}
218
219			if ($this->charIndex < $token['offsetEnd']) {
220			// Cursor is inside the current word - truncate the word at the cursor
221			// (This emulates normal BASH completion behaviour I've observed, though I'm not entirely sure if it's useful)
222			$relativeOffset = $this->charIndex - $token['offset'];
223			$truncated = substr($token['value'], 0, $relativeOffset);
224
225			$this->words[$this->wordIndex] = $truncated;
226			}
227			}
228			}
229
230			// Cursor position is past the end of the command line string - consider it a new word
231			if ($this->wordIndex === null) {
232			$this->wordIndex = count($this->words);
233			$this->words[] = '';
234			}
235			}
236
237			/**
238			* Return a token's value with escaping and quotes removed
239			*
240			* @see self::tokenizeString()
241			* @param array $token
242			* @return string
243			*/
244			protected function getTokenValue($token)
245			{
246			$value = $token['value'];
247
248			// Remove outer quote characters (or first quote if unclosed)
249			if ($token['type'] == 'quoted') {
250			$value = preg_replace('/^(?:[\'"])(.*?)(?:[\'"])?$/', '$1', $value);
251			}
252
253			// Remove escape characters
254			$value = preg_replace('/\\\\(.)/', '$1', $value);
255
256			return $value;
257			}
258
259			/**
260			* Break a string into words, quoted strings and non-words (breaks)
261			*
262			* Returns an array of unmodified segments of $string with offset and type information.
263			*
264			* @param string $string
265			* @return array as [ [type => string, value => string, offset => int], ... ]
266			*/
267			protected function tokenizeString($string)
268			{
269			// Map capture groups to returned token type
270			$typeMap = array(
271			'double_quote_string' => 'quoted',
272			'single_quote_string' => 'quoted',
273			'word' => 'word',
274			'break' => 'break',
275			);
276
277			// Escape every word break character including whitespace
278			// preg_quote won't work here as it doesn't understand the ignore whitespace flag ("x")
279			$breaks = preg_replace('/(.)/', '\\\$1', $this->wordBreaks);
280
281			$pattern = <<<"REGEX"
282			/(?:
283			(?P<double_quote_string>
284			"(\\\\.\|[^\"\\\\])*(?:"\|$)
285			) \|
286			(?P<single_quote_string>
287			'(\\\\.\|[^'\\\\])*(?:'\|$)
288			) \|
289			(?P<word>
290			(?:\\\\.\|[^$breaks])+
291			) \|
292			(?P<break>
293			[$breaks]+
294			)
295			)/x
296			REGEX;
297
298			$tokens = array();
299
300			if (!preg_match_all($pattern, $string, $matches, PREG_OFFSET_CAPTURE \| PREG_SET_ORDER)) {
301			return $tokens;
302			}
303
304			foreach ($matches as $set) {
305			foreach ($set as $groupName => $match) {
306
307			// Ignore integer indices preg_match outputs (duplicates of named groups)
308			if (is_integer($groupName)) {
309			continue;
310			}
311
312			// Skip if the offset indicates this group didn't match
313			if ($match[1] === -1) {
			0 ignored issues – show Unused Code Bug introduced 2018-02-12 13:13 UTC by Report Bug Copy Issue Report The strict comparison `===` seems to always evaluate to `false` as the types of `$match[1]` (`string`) and `-1` (`integer`) can never be identical. Maybe you want to use a loose comparison `==` instead? Loading history...
314			continue;
315			}
316
317			$tokens[] = array(
318			'type' => $typeMap[$groupName],
319			'value' => $match[0],
320			'offset' => $match[1],
321			'offsetEnd' => $match[1] + strlen($match[0])
322			);
323
324			// Move to the next set (only one group should match per set)
325			continue;
326			}
327			}
328
329			return $tokens;
330			}
331
332			/**
333			* Reset the computed words so that $this->splitWords is forced to run again
334			*/
335			protected function reset()
336			{
337			$this->words = null;
338			$this->wordIndex = null;
339			}
340			}
341

stecman / symfony-console-completion

Pull Request — master (#82)

CompletionContext::getTokenValue() A

Complexity

Size

Duplication

Importance

Duplication Side-by-Side

Filter issues like