ArrayDataTransformer::processPostgresValue() - Code Metrics - Inspection of "fix: bring back support of unquoted strings that c..." - martin-georgiev/postgresql-for-doctrine - Measure and Improve Code Quality continuously with Scrutinizer

Test Failed

Push — issue-329 ( 5fdddd )

by Martin

created 2025-04-02 23:57 UTC

ArrayDataTransformer::processPostgresValue() A

↳ Parent: ArrayDataTransformer

Complexity

Conditions	5
Paths	5

Size

Total Lines	22
Code Lines	10

Duplication

Lines	0
Ratio	0 %

Code Coverage

Tests	7
CRAP Score	6.2017

Importance

Changes

Metric	Value
cc	5
eloc	10
nc	5
nop	1
dl	0
loc	22
ccs	7
cts	11
cp	0.6364
crap	6.2017
rs	9.6111
c	0
b	0
f	0

<?php

declare(strict_types=1);

namespace MartinGeorgiev\Utils;

use MartinGeorgiev\Utils\Exception\InvalidArrayFormatException;

/**
 * @since 3.0
 *
 * @author Martin Georgiev <[email protected]>
 */
class ArrayDataTransformer
{
    private const POSTGRESQL_EMPTY_ARRAY = '{}';

    private const POSTGRESQL_NULL_VALUE = 'null';

    /**
     * This method supports only single-dimensioned text arrays and
     * relays on the default escaping strategy in PostgreSQL (double quotes).
     *
     * @throws InvalidArrayFormatException when the input is a multi-dimensional array or has invalid format
     */
    public static function transformPostgresTextArrayToPHPArray(string $postgresArray): array
    {
        $trimmed = \trim($postgresArray);

        if ($trimmed === '' || \strtolower($trimmed) === self::POSTGRESQL_NULL_VALUE) {
            return [];
        }

        if (\str_contains($trimmed, '},{') || \str_starts_with($trimmed, '{{')) {
            throw InvalidArrayFormatException::multiDimensionalArrayNotSupported();
        }

        if ($trimmed === self::POSTGRESQL_EMPTY_ARRAY) {
            return [];
        }

        // Check for malformed nesting - this is a more specific check than the one above
        // But we need to exclude cases where curly braces are part of quoted strings
        $content = \trim($trimmed, '{}');
        $inQuotes = false;
        $escaping = false;

        for ($i = 0, $len = \strlen($content); $i < $len; $i++) {
            $char = $content[$i];

            if ($escaping) {
                $escaping = false;

                continue;
            }

            if ($char === '\\' && $inQuotes) {
                $escaping = true;

                continue;
            }

            if ($char === '"') {
                $inQuotes = !$inQuotes;

            } elseif (($char === '{' || $char === '}') && !$inQuotes) {
                throw InvalidArrayFormatException::invalidFormat('Malformed array nesting detected');
            }
        }

        // Check for unclosed quotes
        if ($inQuotes) {

            throw InvalidArrayFormatException::invalidFormat('Unclosed quotes in array');
        }

        // First try with json_decode for properly quoted values
        $jsonArray = '['.\trim($trimmed, '{}').']';

        /** @var array<int, mixed>|null $decoded */
        $decoded = \json_decode($jsonArray, true, 512, JSON_BIGINT_AS_STRING);

        // If json_decode fails, try manual parsing for unquoted strings
        if ($decoded === null && \json_last_error() !== JSON_ERROR_NONE) {
            return self::parsePostgresArrayManually($content);
        }

        return \array_map(
            static fn (mixed $value): mixed => \is_string($value) ? self::unescapeString($value) : $value,
            (array) $decoded
        );
    }

    /**
     * Manually parse a PostgreSQL array content string.
     */
    private static function parsePostgresArrayManually(string $content): array
    {
        if ($content === '') {
            return [];
        }

        // Parse the array manually, handling quoted and unquoted values
        $result = [];
        $inQuotes = false;
        $currentValue = '';
        $escaping = false;

        for ($i = 0, $len = \strlen($content); $i < $len; $i++) {
            $char = $content[$i];

            // Handle escaping within quotes
            if ($escaping) {
                $currentValue .= $char;
                $escaping = false;

                continue;
            }

            if ($char === '\\' && $inQuotes) {
                $escaping = true;
                $currentValue .= $char;

                continue;
            }

            if ($char === '"') {
                $inQuotes = !$inQuotes;

                // For quoted values, we include the quotes for later processing
                $currentValue .= $char;
            } elseif ($char === ',' && !$inQuotes) {
                // End of value
                $result[] = self::processPostgresValue($currentValue);
                $currentValue = '';
            } else {
                $currentValue .= $char;
            }
        }

        // Add the last value
        if ($currentValue !== '') {
            $result[] = self::processPostgresValue($currentValue);
        }

        return $result;
    }

    /**
     * Process a single value from a PostgreSQL array.
     */
    private static function processPostgresValue(string $value): mixed
    {
        $value = \trim($value);

        if (self::isNullValue($value)) {
            return null;
        }

        if (self::isBooleanValue($value)) {
            return self::processBooleanValue($value);
        }

        if (self::isQuotedString($value)) {
            return self::processQuotedString($value);
        }

        if (self::isNumericValue($value)) {
            return self::processNumericValue($value);
        }

        // For unquoted strings, return as is
        return $value;
    }

    /**
     * Check if the value is a NULL value.
     */
    private static function isNullValue(string $value): bool
    {
        return $value === 'NULL' || $value === 'null';
    }

    /**
     * Check if the value is a boolean value.
     */
    private static function isBooleanValue(string $value): bool
    {
        return \in_array($value, ['true', 't', 'false', 'f'], true);
    }

    /**
     * Process a boolean value.
     */
    private static function processBooleanValue(string $value): bool
    {
        return $value === 'true' || $value === 't';
    }

    /**
     * Check if the value is a quoted string.
     */
    private static function isQuotedString(string $value): bool
    {
        return \strlen($value) >= 2 && $value[0] === '"' && $value[\strlen($value) - 1] === '"';
    }

    /**
     * Process a quoted string.
     */
    private static function processQuotedString(string $value): string
    {
        // Remove the quotes and unescape the string
        $unquoted = \substr($value, 1, -1);

        return self::unescapeString($unquoted);
    }

    /**
     * Check if the value is a numeric value.
     */
    private static function isNumericValue(string $value): bool
    {
        return \is_numeric($value);
    }

    /**
     * Process a numeric value.
     */
    private static function processNumericValue(string $value): float|int
    {
        // Convert to int or float as appropriate
        if (\str_contains($value, '.') || \stripos($value, 'e') !== false) {
            return (float) $value;
        }

        return (int) $value;
    }

    /**
     * This method supports only single-dimensioned PHP arrays.
     * This method relays on the default escaping strategy in PostgreSQL (double quotes).
     *
     * @throws InvalidArrayFormatException when the input is a multi-dimensional array or has invalid format
     */
    public static function transformPHPArrayToPostgresTextArray(array $phpArray): string
    {
        if ($phpArray === []) {
            return self::POSTGRESQL_EMPTY_ARRAY;
        }

        if (\array_filter($phpArray, 'is_array')) {
            throw InvalidArrayFormatException::multiDimensionalArrayNotSupported();
        }

        /** @var array<int|string, string> */
        $processed = \array_map(
            static fn (mixed $value): string => self::formatValue($value),
            $phpArray
        );

        return '{'.\implode(',', $processed).'}';
    }

    /**
     * Formats a single value for PostgreSQL array.
     */
    private static function formatValue(mixed $value): string
    {
        // Handle null
        if ($value === null) {
            return 'NULL';
        }

        // Handle actual numbers
        if (\is_int($value) || \is_float($value)) {
            return (string) $value;
        }

        // Handle booleans
        if (\is_bool($value)) {
            return $value ? 'true' : 'false';
        }

        if (\is_object($value)) {
            if (\method_exists($value, '__toString')) {
                $stringValue = $value->__toString();
            } else {
                // For objects without __toString, use a default representation
                $stringValue = $value::class;
            }
        } elseif (\is_resource($value)) {
            $stringValue = '(resource)';
        } else {
            $valueType = \get_debug_type($value);

            if ($valueType === 'string') {
                $stringValue = $value;
            } elseif (\in_array($valueType, ['int', 'float', 'bool'], true)) {
                /** @var bool|float|int $value */
                $stringValue = (string) $value;
            } else {
                $stringValue = $valueType;
            }
        }

        \assert(\is_string($stringValue));

        // Handle empty string
        if ($stringValue === '') {
            return '""';
        }

        // Always quote strings to match the test expectations
        // Double the backslashes and escape quotes
        $escaped = \str_replace(
            ['\\', '"'],
            ['\\\\', '\"'],
            $stringValue
        );

        return '"'.$escaped.'"';
    }

    private static function unescapeString(string $value): string
    {
        // First handle escaped quotes
        $value = \str_replace('\"', '___QUOTE___', $value);

        // Handle double backslashes
        $value = \str_replace('\\\\', '___DBLBACK___', $value);

        // Restore double backslashes
        $value = \str_replace('___DBLBACK___', '\\', $value);

        // Finally restore quotes
        return \str_replace('___QUOTE___', '"', $value);
    }
}


1		<?php
2
3		declare(strict_types=1);
4
5		namespace MartinGeorgiev\Utils;
6
7		use MartinGeorgiev\Utils\Exception\InvalidArrayFormatException;
8
9		/**
10		* @since 3.0
11		*
12		* @author Martin Georgiev <[email protected]>
13		*/
14		class ArrayDataTransformer
15		{
16		private const POSTGRESQL_EMPTY_ARRAY = '{}';
17
18		private const POSTGRESQL_NULL_VALUE = 'null';
19
20		/**
21		* This method supports only single-dimensioned text arrays and
22		* relays on the default escaping strategy in PostgreSQL (double quotes).
23		*
24		* @throws InvalidArrayFormatException when the input is a multi-dimensional array or has invalid format
25		*/
26	30	public static function transformPostgresTextArrayToPHPArray(string $postgresArray): array
27		{
28	30	$trimmed = \trim($postgresArray);
29
30	30	if ($trimmed === '' \|\| \strtolower($trimmed) === self::POSTGRESQL_NULL_VALUE) {
31	2	return [];
32		}
33
34	28	if (\str_contains($trimmed, '},{') \|\| \str_starts_with($trimmed, '{{')) {
35	1	throw InvalidArrayFormatException::multiDimensionalArrayNotSupported();
36		}
37
38	27	if ($trimmed === self::POSTGRESQL_EMPTY_ARRAY) {
39	1	return [];
40		}
41
42		// Check for malformed nesting - this is a more specific check than the one above
43		// But we need to exclude cases where curly braces are part of quoted strings
44	26	$content = \trim($trimmed, '{}');
45	26	$inQuotes = false;
46	26	$escaping = false;
47
48	26	for ($i = 0, $len = \strlen($content); $i < $len; $i++) {
49	26	$char = $content[$i];
50
51	26	if ($escaping) {
52	6	$escaping = false;
53
54	6	continue;
55		}
56
57	26	if ($char === '\\' && $inQuotes) {
58	6	$escaping = true;
59
60	6	continue;
61		}
62
63	26	if ($char === '"') {
64	21	$inQuotes = !$inQuotes;
		0 ignored issues – show introduced 2025-04-03 00:08 UTC by Report Bug Copy Issue Report The condition `$inQuotes` is always `false`. Loading history...
65	26	} elseif (($char === '{' \|\| $char === '}') && !$inQuotes) {
66	1	throw InvalidArrayFormatException::invalidFormat('Malformed array nesting detected');
67		}
68		}
69
70		// Check for unclosed quotes
71	25	if ($inQuotes) {
		0 ignored issues – show introduced 2025-04-03 00:08 UTC by Report Bug Copy Issue Report The condition `$inQuotes` is always `false`. Loading history...
72	2	throw InvalidArrayFormatException::invalidFormat('Unclosed quotes in array');
73		}
74
75		// First try with json_decode for properly quoted values
76	23	$jsonArray = '['.\trim($trimmed, '{}').']';
77
78		/** @var array<int, mixed>\|null $decoded */
79	23	$decoded = \json_decode($jsonArray, true, 512, JSON_BIGINT_AS_STRING);
80
81		// If json_decode fails, try manual parsing for unquoted strings
82	23	if ($decoded === null && \json_last_error() !== JSON_ERROR_NONE) {
83	1	return self::parsePostgresArrayManually($content);
84		}
85
86	22	return \array_map(
87	22	static fn (mixed $value): mixed => \is_string($value) ? self::unescapeString($value) : $value,
88	22	(array) $decoded
89	22	);
90		}
91
92		/**
93		* Manually parse a PostgreSQL array content string.
94		*/
95	1	private static function parsePostgresArrayManually(string $content): array
96		{
97	1	if ($content === '') {
98		return [];
99		}
100
101		// Parse the array manually, handling quoted and unquoted values
102	1	$result = [];
103	1	$inQuotes = false;
104	1	$currentValue = '';
105	1	$escaping = false;
106
107	1	for ($i = 0, $len = \strlen($content); $i < $len; $i++) {
108	1	$char = $content[$i];
109
110		// Handle escaping within quotes
111	1	if ($escaping) {
112		$currentValue .= $char;
113		$escaping = false;
114
115		continue;
116		}
117
118	1	if ($char === '\\' && $inQuotes) {
119		$escaping = true;
120		$currentValue .= $char;
121
122		continue;
123		}
124
125	1	if ($char === '"') {
126		$inQuotes = !$inQuotes;
		0 ignored issues – show introduced 2025-04-03 00:08 UTC by Report Bug Copy Issue Report The condition `$inQuotes` is always `false`. Loading history...
127		// For quoted values, we include the quotes for later processing
128		$currentValue .= $char;
129	1	} elseif ($char === ',' && !$inQuotes) {
130		// End of value
131	1	$result[] = self::processPostgresValue($currentValue);
132	1	$currentValue = '';
133		} else {
134	1	$currentValue .= $char;
135		}
136		}
137
138		// Add the last value
139	1	if ($currentValue !== '') {
140	1	$result[] = self::processPostgresValue($currentValue);
141		}
142
143	1	return $result;
144		}
145
146		/**
147		* Process a single value from a PostgreSQL array.
148		*/
149	1	private static function processPostgresValue(string $value): mixed
150		{
151	1	$value = \trim($value);
152
153	1	if (self::isNullValue($value)) {
154		return null;
155		}
156
157	1	if (self::isBooleanValue($value)) {
158		return self::processBooleanValue($value);
159		}
160
161	1	if (self::isQuotedString($value)) {
162		return self::processQuotedString($value);
163		}
164
165	1	if (self::isNumericValue($value)) {
166		return self::processNumericValue($value);
167		}
168
169		// For unquoted strings, return as is
170	1	return $value;
171		}
172
173		/**
174		* Check if the value is a NULL value.
175		*/
176	1	private static function isNullValue(string $value): bool
177		{
178	1	return $value === 'NULL' \|\| $value === 'null';
179		}
180
181		/**
182		* Check if the value is a boolean value.
183		*/
184	1	private static function isBooleanValue(string $value): bool
185		{
186	1	return \in_array($value, ['true', 't', 'false', 'f'], true);
187		}
188
189		/**
190		* Process a boolean value.
191		*/
192		private static function processBooleanValue(string $value): bool
193		{
194		return $value === 'true' \|\| $value === 't';
195		}
196
197		/**
198		* Check if the value is a quoted string.
199		*/
200	1	private static function isQuotedString(string $value): bool
201		{
202	1	return \strlen($value) >= 2 && $value[0] === '"' && $value[\strlen($value) - 1] === '"';
203		}
204
205		/**
206		* Process a quoted string.
207		*/
208		private static function processQuotedString(string $value): string
209		{
210		// Remove the quotes and unescape the string
211		$unquoted = \substr($value, 1, -1);
212
213		return self::unescapeString($unquoted);
214		}
215
216		/**
217		* Check if the value is a numeric value.
218		*/
219	1	private static function isNumericValue(string $value): bool
220		{
221	1	return \is_numeric($value);
222		}
223
224		/**
225		* Process a numeric value.
226		*/
227		private static function processNumericValue(string $value): float\|int
228		{
229		// Convert to int or float as appropriate
230		if (\str_contains($value, '.') \|\| \stripos($value, 'e') !== false) {
231		return (float) $value;
232		}
233
234		return (int) $value;
235		}
236
237		/**
238		* This method supports only single-dimensioned PHP arrays.
239		* This method relays on the default escaping strategy in PostgreSQL (double quotes).
240		*
241		* @throws InvalidArrayFormatException when the input is a multi-dimensional array or has invalid format
242		*/
243	28	public static function transformPHPArrayToPostgresTextArray(array $phpArray): string
244		{
245	28	if ($phpArray === []) {
246	1	return self::POSTGRESQL_EMPTY_ARRAY;
247		}
248
249	27	if (\array_filter($phpArray, 'is_array')) {
250	5	throw InvalidArrayFormatException::multiDimensionalArrayNotSupported();
251		}
252
253		/** @var array<int\|string, string> */
254	22	$processed = \array_map(
255	22	static fn (mixed $value): string => self::formatValue($value),
256	22	$phpArray
257	22	);
258
259	22	return '{'.\implode(',', $processed).'}';
260		}
261
262		/**
263		* Formats a single value for PostgreSQL array.
264		*/
265	22	private static function formatValue(mixed $value): string
266		{
267		// Handle null
268	22	if ($value === null) {
269		return 'NULL';
270		}
271
272		// Handle actual numbers
273	22	if (\is_int($value) \|\| \is_float($value)) {
274	6	return (string) $value;
275		}
276
277		// Handle booleans
278	20	if (\is_bool($value)) {
279	1	return $value ? 'true' : 'false';
280		}
281
282	19	if (\is_object($value)) {
283	1	if (\method_exists($value, '__toString')) {
284	1	$stringValue = $value->__toString();
285		} else {
286		// For objects without __toString, use a default representation
287		$stringValue = $value::class;
288		}
289	18	} elseif (\is_resource($value)) {
290	1	$stringValue = '(resource)';
291		} else {
292	17	$valueType = \get_debug_type($value);
293
294	17	if ($valueType === 'string') {
295	17	$stringValue = $value;
296		} elseif (\in_array($valueType, ['int', 'float', 'bool'], true)) {
297		/** @var bool\|float\|int $value */
298		$stringValue = (string) $value;
299		} else {
300		$stringValue = $valueType;
301		}
302		}
303
304	19	\assert(\is_string($stringValue));
305
306		// Handle empty string
307	19	if ($stringValue === '') {
308	1	return '""';
309		}
310
311		// Always quote strings to match the test expectations
312		// Double the backslashes and escape quotes
313	18	$escaped = \str_replace(
314	18	['\\', '"'],
315	18	['\\\\', '\"'],
316	18	$stringValue
317	18	);
318
319	18	return '"'.$escaped.'"';
320		}
321
322	19	private static function unescapeString(string $value): string
323		{
324		// First handle escaped quotes
325	19	$value = \str_replace('\"', '___QUOTE___', $value);
326
327		// Handle double backslashes
328	19	$value = \str_replace('\\\\', '___DBLBACK___', $value);
329
330		// Restore double backslashes
331	19	$value = \str_replace('___DBLBACK___', '\\', $value);
332
333		// Finally restore quotes
334	19	return \str_replace('___QUOTE___', '"', $value);
335		}
336		}
337

martin-georgiev / postgresql-for-doctrine

Push — issue-329 ( 5fdddd )

ArrayDataTransformer::processPostgresValue() A

Complexity

Size

Duplication

Code Coverage

Importance

Duplication Side-by-Side

Filter issues like