Csv::getCsv() - Code Metrics - Inspection of "Php 8.4 Will Deprecate fgetcsv Parameter" - PHPOffice/PhpSpreadsheet - Measure and Improve Code Quality continuously with Scrutinizer

Passed

Pull Request — master (#4162)

by Owen

created 2024-09-07 14:36 UTC

Csv::getCsv() A

↳ Parent: Csv

Complexity

Conditions	3
Paths	2

Size

Total Lines	13
Code Lines	4

Duplication

Lines	0
Ratio	0 %

Code Coverage

Tests	0
CRAP Score	12

Importance

Changes	2
Bugs	0	Features	0

Metric	Value
eloc	4
c	2
b	0
f	0
dl	0
loc	13
ccs	0
cts	0
cp	0
rs	10
cc	3
nc	2
nop	5
crap	12

<?php

namespace PhpOffice\PhpSpreadsheet\Reader;

use PhpOffice\PhpSpreadsheet\Calculation\Calculation;
use PhpOffice\PhpSpreadsheet\Cell\Cell;
use PhpOffice\PhpSpreadsheet\Cell\Coordinate;
use PhpOffice\PhpSpreadsheet\Reader\Csv\Delimiter;
use PhpOffice\PhpSpreadsheet\Reader\Exception as ReaderException;
use PhpOffice\PhpSpreadsheet\Shared\StringHelper;
use PhpOffice\PhpSpreadsheet\Spreadsheet;
use PhpOffice\PhpSpreadsheet\Worksheet\Worksheet;

class Csv extends BaseReader
{
    const DEFAULT_FALLBACK_ENCODING = 'CP1252';
    const GUESS_ENCODING = 'guess';
    const UTF8_BOM = "\xEF\xBB\xBF";
    const UTF8_BOM_LEN = 3;
    const UTF16BE_BOM = "\xfe\xff";
    const UTF16BE_BOM_LEN = 2;
    const UTF16BE_LF = "\x00\x0a";
    const UTF16LE_BOM = "\xff\xfe";
    const UTF16LE_BOM_LEN = 2;
    const UTF16LE_LF = "\x0a\x00";
    const UTF32BE_BOM = "\x00\x00\xfe\xff";
    const UTF32BE_BOM_LEN = 4;
    const UTF32BE_LF = "\x00\x00\x00\x0a";
    const UTF32LE_BOM = "\xff\xfe\x00\x00";
    const UTF32LE_BOM_LEN = 4;
    const UTF32LE_LF = "\x0a\x00\x00\x00";

    /**
     * Input encoding.
     */
    private string $inputEncoding = 'UTF-8';

    /**
     * Fallback encoding if guess strikes out.
     */
    private string $fallbackEncoding = self::DEFAULT_FALLBACK_ENCODING;

    /**
     * Delimiter.
     */
    private ?string $delimiter = null;

    /**
     * Enclosure.
     */
    private string $enclosure = '"';

    /**
     * Sheet index to read.
     */
    private int $sheetIndex = 0;

    /**
     * Load rows contiguously.
     */
    private bool $contiguous = false;

    /**
     * The character that can escape the enclosure.
     * This will probably become unsupported in Php 9.
     * Not yet ready to mark deprecated in order to give users
     * a migration path.
     */
    private ?string $escapeCharacter = null;

    /**
     * The character that will be supplied to fgetcsv
     * when escapeCharacter is null.
     * It is anticipated that it will conditionally be set
     * to null-string for Php9 and above.
     */
    private static string $defaultEscapeCharacter = '\\';

    /**
     * Callback for setting defaults in construction.
     *
     * @var ?callable
     */
    private static $constructorCallback;

    /**
     * Attempt autodetect line endings (deprecated after PHP8.1)?
     */
    private bool $testAutodetect = true;

    protected bool $castFormattedNumberToNumeric = false;

    protected bool $preserveNumericFormatting = false;

    private bool $preserveNullString = false;

    private bool $sheetNameIsFileName = false;

    private string $getTrue = 'true';

    private string $getFalse = 'false';

    private string $thousandsSeparator = ',';

    private string $decimalSeparator = '.';

    /**
     * Create a new CSV Reader instance.
     */
    public function __construct()
    {
        parent::__construct();
        $callback = self::$constructorCallback;
        if ($callback !== null) {
            $callback($this);
        }
    }

    /**
     * Set a callback to change the defaults.
     *
     * The callback must accept the Csv Reader object as the first parameter,
     * and it should return void.
     */
    public static function setConstructorCallback(?callable $callback): void
    {
        self::$constructorCallback = $callback;
    }

    public static function getConstructorCallback(): ?callable
    {
        return self::$constructorCallback;
    }

    public function setInputEncoding(string $encoding): self
    {
        $this->inputEncoding = $encoding;

        return $this;
    }

    public function getInputEncoding(): string
    {
        return $this->inputEncoding;
    }

    public function setFallbackEncoding(string $fallbackEncoding): self
    {
        $this->fallbackEncoding = $fallbackEncoding;

        return $this;
    }

    public function getFallbackEncoding(): string
    {
        return $this->fallbackEncoding;
    }

    /**
     * Move filepointer past any BOM marker.
     */
    protected function skipBOM(): void
    {
        rewind($this->fileHandle);

        if (fgets($this->fileHandle, self::UTF8_BOM_LEN + 1) !== self::UTF8_BOM) {
            rewind($this->fileHandle);
        }
    }

    /**
     * Identify any separator that is explicitly set in the file.
     */
    protected function checkSeparator(): void
    {
        $line = fgets($this->fileHandle);
        if ($line === false) {
            return;
        }

        if ((strlen(trim($line, "\r\n")) == 5) && (stripos($line, 'sep=') === 0)) {
            $this->delimiter = substr($line, 4, 1);

            return;
        }

        $this->skipBOM();
    }

    /**
     * Infer the separator if it isn't explicitly set in the file or specified by the user.
     */
    protected function inferSeparator(): void
    {
        if ($this->delimiter !== null) {
            return;
        }

        $inferenceEngine = new Delimiter($this->fileHandle, $this->escapeCharacter ?? self::$defaultEscapeCharacter, $this->enclosure);

        // If number of lines is 0, nothing to infer : fall back to the default
        if ($inferenceEngine->linesCounted() === 0) {
            $this->delimiter = $inferenceEngine->getDefaultDelimiter();
            $this->skipBOM();

            return;
        }

        $this->delimiter = $inferenceEngine->infer();

        // If no delimiter could be detected, fall back to the default
        if ($this->delimiter === null) {
            $this->delimiter = $inferenceEngine->getDefaultDelimiter();
        }

        $this->skipBOM();
    }

    /**
     * Return worksheet info (Name, Last Column Letter, Last Column Index, Total Rows, Total Columns).
     */
    public function listWorksheetInfo(string $filename): array
    {
        // Open file
        $this->openFileOrMemory($filename);
        $fileHandle = $this->fileHandle;

        // Skip BOM, if any
        $this->skipBOM();
        $this->checkSeparator();
        $this->inferSeparator();

        $worksheetInfo = [];
        $worksheetInfo[0]['worksheetName'] = 'Worksheet';
        $worksheetInfo[0]['lastColumnLetter'] = 'A';
        $worksheetInfo[0]['lastColumnIndex'] = 0;
        $worksheetInfo[0]['totalRows'] = 0;
        $worksheetInfo[0]['totalColumns'] = 0;
        $delimiter = $this->delimiter ?? '';

        // Loop through each line of the file in turn
        $rowData = self::getCsv($fileHandle, 0, $delimiter, $this->enclosure, $this->escapeCharacter);
        while (is_array($rowData)) {
            ++$worksheetInfo[0]['totalRows'];
            $worksheetInfo[0]['lastColumnIndex'] = max($worksheetInfo[0]['lastColumnIndex'], count($rowData) - 1);
            $rowData = self::getCsv($fileHandle, 0, $delimiter, $this->enclosure, $this->escapeCharacter);
        }

        $worksheetInfo[0]['lastColumnLetter'] = Coordinate::stringFromColumnIndex($worksheetInfo[0]['lastColumnIndex'] + 1);
        $worksheetInfo[0]['totalColumns'] = $worksheetInfo[0]['lastColumnIndex'] + 1;

        // Close file
        fclose($fileHandle);

        return $worksheetInfo;
    }

    /**
     * Loads Spreadsheet from file.
     */
    protected function loadSpreadsheetFromFile(string $filename): Spreadsheet
    {
        // Create new Spreadsheet
        $spreadsheet = new Spreadsheet();

        // Load into this instance
        return $this->loadIntoExisting($filename, $spreadsheet);
    }

    /**
     * Loads Spreadsheet from string.
     */
    public function loadSpreadsheetFromString(string $contents): Spreadsheet
    {
        // Create new Spreadsheet
        $spreadsheet = new Spreadsheet();

        // Load into this instance
        return $this->loadStringOrFile('data://text/plain,' . urlencode($contents), $spreadsheet, true);
    }

    private function openFileOrMemory(string $filename): void
    {
        // Open file
        $fhandle = $this->canRead($filename);
        if (!$fhandle) {
            throw new ReaderException($filename . ' is an Invalid Spreadsheet file.');
        }
        if ($this->inputEncoding === self::GUESS_ENCODING) {
            $this->inputEncoding = self::guessEncoding($filename, $this->fallbackEncoding);
        }
        $this->openFile($filename);
        if ($this->inputEncoding !== 'UTF-8') {
            fclose($this->fileHandle);
            $entireFile = file_get_contents($filename);
            $fileHandle = fopen('php://memory', 'r+b');
            if ($fileHandle !== false && $entireFile !== false) {
                $this->fileHandle = $fileHandle;
                $data = StringHelper::convertEncoding($entireFile, 'UTF-8', $this->inputEncoding);
                fwrite($this->fileHandle, $data);
                $this->skipBOM();
            }
        }
    }

    public function setTestAutoDetect(bool $value): self
    {
        $this->testAutodetect = $value;

        return $this;
    }

    private function setAutoDetect(?string $value): ?string
    {
        $retVal = null;
        if ($value !== null && $this->testAutodetect) {
            $retVal2 = @ini_set('auto_detect_line_endings', $value);
            if (is_string($retVal2)) {
                $retVal = $retVal2;
            }
        }

        return $retVal;
    }

    public function castFormattedNumberToNumeric(
        bool $castFormattedNumberToNumeric,
        bool $preserveNumericFormatting = false
    ): void {
        $this->castFormattedNumberToNumeric = $castFormattedNumberToNumeric;
        $this->preserveNumericFormatting = $preserveNumericFormatting;
    }

    /**
     * Open data uri for reading.
     */
    private function openDataUri(string $filename): void
    {
        $fileHandle = fopen($filename, 'rb');
        if ($fileHandle === false) {
            // @codeCoverageIgnoreStart
            throw new ReaderException('Could not open file ' . $filename . ' for reading.');
            // @codeCoverageIgnoreEnd
        }

        $this->fileHandle = $fileHandle;
    }

    /**
     * Loads PhpSpreadsheet from file into PhpSpreadsheet instance.
     */
    public function loadIntoExisting(string $filename, Spreadsheet $spreadsheet): Spreadsheet
    {
        return $this->loadStringOrFile($filename, $spreadsheet, false);
    }

    /**
     * Loads PhpSpreadsheet from file into PhpSpreadsheet instance.
     */
    private function loadStringOrFile(string $filename, Spreadsheet $spreadsheet, bool $dataUri): Spreadsheet
    {
        // Deprecated in Php8.1
        $iniset = $this->setAutoDetect('1');

        // Open file
        if ($dataUri) {
            $this->openDataUri($filename);
        } else {
            $this->openFileOrMemory($filename);
        }
        $fileHandle = $this->fileHandle;

        // Skip BOM, if any
        $this->skipBOM();
        $this->checkSeparator();
        $this->inferSeparator();

        // Create new PhpSpreadsheet object
        while ($spreadsheet->getSheetCount() <= $this->sheetIndex) {
            $spreadsheet->createSheet();
        }
        $sheet = $spreadsheet->setActiveSheetIndex($this->sheetIndex);
        if ($this->sheetNameIsFileName) {
            $sheet->setTitle(substr(basename($filename, '.csv'), 0, Worksheet::SHEET_TITLE_MAXIMUM_LENGTH));
        }

        // Set our starting row based on whether we're in contiguous mode or not
        $currentRow = 1;
        $outRow = 0;

        // Loop through each line of the file in turn
        $delimiter = $this->delimiter ?? '';
        $rowData = self::getCsv($fileHandle, 0, $delimiter, $this->enclosure, $this->escapeCharacter);
        $valueBinder = Cell::getValueBinder();
        $preserveBooleanString = method_exists($valueBinder, 'getBooleanConversion') && $valueBinder->getBooleanConversion();
        $this->getTrue = Calculation::getTRUE();
        $this->getFalse = Calculation::getFALSE();
        $this->thousandsSeparator = StringHelper::getThousandsSeparator();
        $this->decimalSeparator = StringHelper::getDecimalSeparator();
        while (is_array($rowData)) {
            $noOutputYet = true;
            $columnLetter = 'A';
            foreach ($rowData as $rowDatum) {
                if ($preserveBooleanString) {
                    $rowDatum = $rowDatum ?? '';
                } else {
                    $this->convertBoolean($rowDatum);
                }
                $numberFormatMask = $this->castFormattedNumberToNumeric ? $this->convertFormattedNumber($rowDatum) : '';
                if (($rowDatum !== '' || $this->preserveNullString) && $this->readFilter->readCell($columnLetter, $currentRow)) {
                    if ($this->contiguous) {
                        if ($noOutputYet) {
                            $noOutputYet = false;
                            ++$outRow;
                        }
                    } else {
                        $outRow = $currentRow;
                    }
                    // Set basic styling for the value (Note that this could be overloaded by styling in a value binder)
                    if ($numberFormatMask !== '') {
                        $sheet->getStyle($columnLetter . $outRow)
                            ->getNumberFormat()
                            ->setFormatCode($numberFormatMask);
                    }
                    // Set cell value
                    $sheet->getCell($columnLetter . $outRow)->setValue($rowDatum);
                }
                ++$columnLetter;
            }
            $rowData = self::getCsv($fileHandle, 0, $delimiter, $this->enclosure, $this->escapeCharacter);
            ++$currentRow;
        }

        // Close file
        fclose($fileHandle);

        $this->setAutoDetect($iniset);

        // Return
        return $spreadsheet;
    }

    /**
     * Convert string true/false to boolean, and null to null-string.
     */
    private function convertBoolean(mixed &$rowDatum): void
    {
        if (is_string($rowDatum)) {
            if (strcasecmp($this->getTrue, $rowDatum) === 0 || strcasecmp('true', $rowDatum) === 0) {
                $rowDatum = true;
            } elseif (strcasecmp($this->getFalse, $rowDatum) === 0 || strcasecmp('false', $rowDatum) === 0) {
                $rowDatum = false;
            }
        } else {
            $rowDatum = $rowDatum ?? '';
        }
    }

    /**
     * Convert numeric strings to int or float values.
     */
    private function convertFormattedNumber(mixed &$rowDatum): string
    {
        $numberFormatMask = '';
        if ($this->castFormattedNumberToNumeric === true && is_string($rowDatum)) {
            $numeric = str_replace(
                [$this->thousandsSeparator, $this->decimalSeparator],
                ['', '.'],
                $rowDatum
            );

            if (is_numeric($numeric)) {
                $decimalPos = strpos($rowDatum, $this->decimalSeparator);
                if ($this->preserveNumericFormatting === true) {
                    $numberFormatMask = (str_contains($rowDatum, $this->thousandsSeparator))
                        ? '#,##0' : '0';
                    if ($decimalPos !== false) {
                        $decimals = strlen($rowDatum) - $decimalPos - 1;
                        $numberFormatMask .= '.' . str_repeat('0', min($decimals, 6));
                    }
                }

                $rowDatum = ($decimalPos !== false) ? (float) $numeric : (int) $numeric;
            }
        }

        return $numberFormatMask;
    }

    public function getDelimiter(): ?string
    {
        return $this->delimiter;
    }

    public function setDelimiter(?string $delimiter): self
    {
        $this->delimiter = $delimiter;

        return $this;
    }

    public function getEnclosure(): string
    {
        return $this->enclosure;
    }

    public function setEnclosure(string $enclosure): self
    {
        if ($enclosure == '') {
            $enclosure = '"';
        }
        $this->enclosure = $enclosure;

        return $this;
    }

    public function getSheetIndex(): int
    {
        return $this->sheetIndex;
    }

    public function setSheetIndex(int $indexValue): self
    {
        $this->sheetIndex = $indexValue;

        return $this;
    }

    public function setContiguous(bool $contiguous): self
    {
        $this->contiguous = $contiguous;

        return $this;
    }

    public function getContiguous(): bool
    {
        return $this->contiguous;
    }

    /**
     * Php9 intends to drop support for this parameter in fgetcsv.
     * Not yet ready to mark deprecated in order to give users
     * a migration path.
     */
    public function setEscapeCharacter(string $escapeCharacter): self
    {
        $this->escapeCharacter = $escapeCharacter;

        return $this;
    }

    public function getEscapeCharacter(): string
    {
        return $this->escapeCharacter ?? self::$defaultEscapeCharacter;
    }

    /**
     * Can the current IReader read the file?
     */
    public function canRead(string $filename): bool
    {
        // Check if file exists
        try {
            $this->openFile($filename);
        } catch (ReaderException) {
            return false;
        }

        fclose($this->fileHandle);

        // Trust file extension if any
        $extension = strtolower(pathinfo($filename, PATHINFO_EXTENSION));
        if (in_array($extension, ['csv', 'tsv'])) {
            return true;
        }

        // Attempt to guess mimetype
        $type = mime_content_type($filename);
        $supportedTypes = [
            'application/csv',
            'text/csv',
            'text/plain',
            'inode/x-empty',
            'text/html',
        ];

        return in_array($type, $supportedTypes, true);
    }

    private static function guessEncodingTestNoBom(string &$encoding, string &$contents, string $compare, string $setEncoding): void
    {
        if ($encoding === '') {
            $pos = strpos($contents, $compare);
            if ($pos !== false && $pos % strlen($compare) === 0) {
                $encoding = $setEncoding;
            }
        }
    }

    private static function guessEncodingNoBom(string $filename): string
    {
        $encoding = '';
        $contents = file_get_contents($filename);
        self::guessEncodingTestNoBom($encoding, $contents, self::UTF32BE_LF, 'UTF-32BE');
        self::guessEncodingTestNoBom($encoding, $contents, self::UTF32LE_LF, 'UTF-32LE');
        self::guessEncodingTestNoBom($encoding, $contents, self::UTF16BE_LF, 'UTF-16BE');
        self::guessEncodingTestNoBom($encoding, $contents, self::UTF16LE_LF, 'UTF-16LE');
        if ($encoding === '' && preg_match('//u', $contents) === 1) {
            $encoding = 'UTF-8';
        }

        return $encoding;
    }

    private static function guessEncodingTestBom(string &$encoding, string $first4, string $compare, string $setEncoding): void
    {
        if ($encoding === '') {
            if (str_starts_with($first4, $compare)) {
                $encoding = $setEncoding;
            }
        }
    }

    private static function guessEncodingBom(string $filename): string
    {
        $encoding = '';
        $first4 = file_get_contents($filename, false, null, 0, 4);
        if ($first4 !== false) {
            self::guessEncodingTestBom($encoding, $first4, self::UTF8_BOM, 'UTF-8');
            self::guessEncodingTestBom($encoding, $first4, self::UTF16BE_BOM, 'UTF-16BE');
            self::guessEncodingTestBom($encoding, $first4, self::UTF32BE_BOM, 'UTF-32BE');
            self::guessEncodingTestBom($encoding, $first4, self::UTF32LE_BOM, 'UTF-32LE');
            self::guessEncodingTestBom($encoding, $first4, self::UTF16LE_BOM, 'UTF-16LE');
        }

        return $encoding;
    }

    public static function guessEncoding(string $filename, string $dflt = self::DEFAULT_FALLBACK_ENCODING): string
    {
        $encoding = self::guessEncodingBom($filename);
        if ($encoding === '') {
            $encoding = self::guessEncodingNoBom($filename);
        }

        return ($encoding === '') ? $dflt : $encoding;
    }

    public function setPreserveNullString(bool $value): self
    {
        $this->preserveNullString = $value;

        return $this;
    }

    public function getPreserveNullString(): bool
    {
        return $this->preserveNullString;
    }

    public function setSheetNameIsFileName(bool $sheetNameIsFileName): self
    {
        $this->sheetNameIsFileName = $sheetNameIsFileName;

        return $this;
    }

    /**
     * Php8.4 deprecates use of anything other than null string
     * as escape Character.
     *
     * @param resource $stream
     * @param null|int<0, max> $length
     *
     * @return array<int,?string>|false
     */
    private static function getCsv(
        $stream,
        ?int $length = null,
        string $separator = ',',
        string $enclosure = '"',
        ?string $escape = null
    ): array|false {
        $escape = $escape ?? self::$defaultEscapeCharacter;
        if (PHP_VERSION_ID >= 80400 && $escape !== '') {
            return @fgetcsv($stream, $length, $separator, $enclosure, $escape);
        }

        return fgetcsv($stream, $length, $separator, $enclosure, $escape);
    }
}


1		<?php
2
3		namespace PhpOffice\PhpSpreadsheet\Reader;
4
5		use PhpOffice\PhpSpreadsheet\Calculation\Calculation;
6		use PhpOffice\PhpSpreadsheet\Cell\Cell;
7		use PhpOffice\PhpSpreadsheet\Cell\Coordinate;
8		use PhpOffice\PhpSpreadsheet\Reader\Csv\Delimiter;
9		use PhpOffice\PhpSpreadsheet\Reader\Exception as ReaderException;
10		use PhpOffice\PhpSpreadsheet\Shared\StringHelper;
11		use PhpOffice\PhpSpreadsheet\Spreadsheet;
12		use PhpOffice\PhpSpreadsheet\Worksheet\Worksheet;
13
14		class Csv extends BaseReader
15		{
16		const DEFAULT_FALLBACK_ENCODING = 'CP1252';
17		const GUESS_ENCODING = 'guess';
18		const UTF8_BOM = "\xEF\xBB\xBF";
19		const UTF8_BOM_LEN = 3;
20		const UTF16BE_BOM = "\xfe\xff";
21		const UTF16BE_BOM_LEN = 2;
22		const UTF16BE_LF = "\x00\x0a";
23		const UTF16LE_BOM = "\xff\xfe";
24		const UTF16LE_BOM_LEN = 2;
25		const UTF16LE_LF = "\x0a\x00";
26		const UTF32BE_BOM = "\x00\x00\xfe\xff";
27		const UTF32BE_BOM_LEN = 4;
28		const UTF32BE_LF = "\x00\x00\x00\x0a";
29		const UTF32LE_BOM = "\xff\xfe\x00\x00";
30		const UTF32LE_BOM_LEN = 4;
31		const UTF32LE_LF = "\x0a\x00\x00\x00";
32
33		/**
34		* Input encoding.
35		*/
36		private string $inputEncoding = 'UTF-8';
37
38		/**
39		* Fallback encoding if guess strikes out.
40		*/
41		private string $fallbackEncoding = self::DEFAULT_FALLBACK_ENCODING;
42
43		/**
44		* Delimiter.
45		*/
46		private ?string $delimiter = null;
47
48		/**
49		* Enclosure.
50		*/
51		private string $enclosure = '"';
52
53		/**
54		* Sheet index to read.
55		*/
56		private int $sheetIndex = 0;
57
58		/**
59		* Load rows contiguously.
60		*/
61		private bool $contiguous = false;
62
63		/**
64		* The character that can escape the enclosure.
65		* This will probably become unsupported in Php 9.
66		* Not yet ready to mark deprecated in order to give users
67		* a migration path.
68		*/
69		private ?string $escapeCharacter = null;
70
71		/**
72		* The character that will be supplied to fgetcsv
73		* when escapeCharacter is null.
74		* It is anticipated that it will conditionally be set
75		* to null-string for Php9 and above.
76		*/
77		private static string $defaultEscapeCharacter = '\\';
78
79		/**
80		* Callback for setting defaults in construction.
81		*
82		* @var ?callable
83		*/
84		private static $constructorCallback;
85
86		/**
87		* Attempt autodetect line endings (deprecated after PHP8.1)?
88		*/
89		private bool $testAutodetect = true;
90
91		protected bool $castFormattedNumberToNumeric = false;
92
93		protected bool $preserveNumericFormatting = false;
94
95		private bool $preserveNullString = false;
96
97		private bool $sheetNameIsFileName = false;
98
99	145	private string $getTrue = 'true';
100
101	145	private string $getFalse = 'false';
102	145
103	145	private string $thousandsSeparator = ',';
104	5
105		private string $decimalSeparator = '.';
106
107		/**
108		* Create a new CSV Reader instance.
109		*/
110		public function __construct()
111		{
112		parent::__construct();
113		$callback = self::$constructorCallback;
114	6	if ($callback !== null) {
115		$callback($this);
116	6	}
117		}
118
119	1	/**
120		* Set a callback to change the defaults.
121	1	*
122		* The callback must accept the Csv Reader object as the first parameter,
123		* and it should return void.
124	46	*/
125		public static function setConstructorCallback(?callable $callback): void
126	46	{
127		self::$constructorCallback = $callback;
128	46	}
129
130		public static function getConstructorCallback(): ?callable
131	1	{
132		return self::$constructorCallback;
133	1	}
134
135		public function setInputEncoding(string $encoding): self
136	5	{
137		$this->inputEncoding = $encoding;
138	5
139		return $this;
140	5	}
141
142		public function getInputEncoding(): string
143	1	{
144		return $this->inputEncoding;
145	1	}
146
147		public function setFallbackEncoding(string $fallbackEncoding): self
148		{
149		$this->fallbackEncoding = $fallbackEncoding;
150
151	123	return $this;
152		}
153	123
154		public function getFallbackEncoding(): string
155	123	{
156	108	return $this->fallbackEncoding;
157		}
158
159		/**
160		* Move filepointer past any BOM marker.
161		*/
162		protected function skipBOM(): void
163	123	{
164		rewind($this->fileHandle);
165	123
166	123	if (fgets($this->fileHandle, self::UTF8_BOM_LEN + 1) !== self::UTF8_BOM) {
167	1	rewind($this->fileHandle);
168		}
169		}
170	122
171	2	/**
172		* Identify any separator that is explicitly set in the file.
173	2	*/
174		protected function checkSeparator(): void
175		{
176	120	$line = fgets($this->fileHandle);
177		if ($line === false) {
178		return;
179		}
180
181		if ((strlen(trim($line, "\r\n")) == 5) && (stripos($line, 'sep=') === 0)) {
182	123	$this->delimiter = substr($line, 4, 1);
183
184	123	return;
185	22	}
186
187		$this->skipBOM();
188	112	}
189
190		/**
191	112	* Infer the separator if it isn't explicitly set in the file or specified by the user.
192	1	*/
193	1	protected function inferSeparator(): void
194		{
195	1	if ($this->delimiter !== null) {
196		return;
197		}
198	111
199		$inferenceEngine = new Delimiter($this->fileHandle, $this->escapeCharacter ?? self::$defaultEscapeCharacter, $this->enclosure);
200
201	111	// If number of lines is 0, nothing to infer : fall back to the default
202	9	if ($inferenceEngine->linesCounted() === 0) {
203		$this->delimiter = $inferenceEngine->getDefaultDelimiter();
204		$this->skipBOM();
205	111
206		return;
207		}
208
209		$this->delimiter = $inferenceEngine->infer();
210
211	12	// If no delimiter could be detected, fall back to the default
212		if ($this->delimiter === null) {
213		$this->delimiter = $inferenceEngine->getDefaultDelimiter();
214	12	}
215	11
216		$this->skipBOM();
217		}
218	11
219	11	/**
220	11	* Return worksheet info (Name, Last Column Letter, Last Column Index, Total Rows, Total Columns).
221		*/
222	11	public function listWorksheetInfo(string $filename): array
223	11	{
224	11	// Open file
225	11	$this->openFileOrMemory($filename);
226	11	$fileHandle = $this->fileHandle;
227	11
228	11	// Skip BOM, if any
229		$this->skipBOM();
230		$this->checkSeparator();
231	11	$this->inferSeparator();
232	11
233	11	$worksheetInfo = [];
234	11	$worksheetInfo[0]['worksheetName'] = 'Worksheet';
235	11	$worksheetInfo[0]['lastColumnLetter'] = 'A';
236		$worksheetInfo[0]['lastColumnIndex'] = 0;
237		$worksheetInfo[0]['totalRows'] = 0;
238	11	$worksheetInfo[0]['totalColumns'] = 0;
239	11	$delimiter = $this->delimiter ?? '';
240
241		// Loop through each line of the file in turn
242	11	$rowData = self::getCsv($fileHandle, 0, $delimiter, $this->enclosure, $this->escapeCharacter);
243		while (is_array($rowData)) {
244	11	++$worksheetInfo[0]['totalRows'];
245		$worksheetInfo[0]['lastColumnIndex'] = max($worksheetInfo[0]['lastColumnIndex'], count($rowData) - 1);
246		$rowData = self::getCsv($fileHandle, 0, $delimiter, $this->enclosure, $this->escapeCharacter);
247		}
248
249		$worksheetInfo[0]['lastColumnLetter'] = Coordinate::stringFromColumnIndex($worksheetInfo[0]['lastColumnIndex'] + 1);
250	109	$worksheetInfo[0]['totalColumns'] = $worksheetInfo[0]['lastColumnIndex'] + 1;
251
252		// Close file
253	109	fclose($fileHandle);
254
255		return $worksheetInfo;
256	109	}
257
258		/**
259		* Loads Spreadsheet from file.
260		*/
261		protected function loadSpreadsheetFromFile(string $filename): Spreadsheet
262	3	{
263		// Create new Spreadsheet
264		$spreadsheet = new Spreadsheet();
265	3
266		// Load into this instance
267		return $this->loadIntoExisting($filename, $spreadsheet);
268	3	}
269
270		/**
271	123	* Loads Spreadsheet from string.
272		*/
273		public function loadSpreadsheetFromString(string $contents): Spreadsheet
274	123	{
275	123	// Create new Spreadsheet
276	3	$spreadsheet = new Spreadsheet();
277
278	120	// Load into this instance
279	17	return $this->loadStringOrFile('data://text/plain,' . urlencode($contents), $spreadsheet, true);
280		}
281	120
282	120	private function openFileOrMemory(string $filename): void
283	37	{
284	37	// Open file
285	37	$fhandle = $this->canRead($filename);
286	37	if (!$fhandle) {
287	37	throw new ReaderException($filename . ' is an Invalid Spreadsheet file.');
288	37	}
289	37	if ($this->inputEncoding === self::GUESS_ENCODING) {
290	37	$this->inputEncoding = self::guessEncoding($filename, $this->fallbackEncoding);
291		}
292		$this->openFile($filename);
293		if ($this->inputEncoding !== 'UTF-8') {
294		fclose($this->fileHandle);
295	3	$entireFile = file_get_contents($filename);
296		$fileHandle = fopen('php://memory', 'r+b');
297	3	if ($fileHandle !== false && $entireFile !== false) {
298		$this->fileHandle = $fileHandle;
299	3	$data = StringHelper::convertEncoding($entireFile, 'UTF-8', $this->inputEncoding);
300		fwrite($this->fileHandle, $data);
301		$this->skipBOM();
302	115	}
303		}
304	115	}
305	115
306	112	public function setTestAutoDetect(bool $value): self
307	112	{
308	112	$this->testAutodetect = $value;
309
310		return $this;
311		}
312	115
313		private function setAutoDetect(?string $value): ?string
314		{
315	14	$retVal = null;
316		if ($value !== null && $this->testAutodetect) {
317		$retVal2 = @ini_set('auto_detect_line_endings', $value);
318		if (is_string($retVal2)) {
319	14	$retVal = $retVal2;
320	14	}
321		}
322
323		return $retVal;
324		}
325
326	3	public function castFormattedNumberToNumeric(
327		bool $castFormattedNumberToNumeric,
328	3	bool $preserveNumericFormatting = false
329	3	): void {
330		$this->castFormattedNumberToNumeric = $castFormattedNumberToNumeric;
331		$this->preserveNumericFormatting = $preserveNumericFormatting;
332		}
333
334		/**
335	3	* Open data uri for reading.
336		*/
337		private function openDataUri(string $filename): void
338		{
339		$fileHandle = fopen($filename, 'rb');
340		if ($fileHandle === false) {
341	112	// @codeCoverageIgnoreStart
342		throw new ReaderException('Could not open file ' . $filename . ' for reading.');
343	112	// @codeCoverageIgnoreEnd
344		}
345
346		$this->fileHandle = $fileHandle;
347		}
348
349	115	/**
350		* Loads PhpSpreadsheet from file into PhpSpreadsheet instance.
351		*/
352	115	public function loadIntoExisting(string $filename, Spreadsheet $spreadsheet): Spreadsheet
353		{
354		return $this->loadStringOrFile($filename, $spreadsheet, false);
355	115	}
356	3
357		/**
358	112	* Loads PhpSpreadsheet from file into PhpSpreadsheet instance.
359		*/
360	113	private function loadStringOrFile(string $filename, Spreadsheet $spreadsheet, bool $dataUri): Spreadsheet
361		{
362		// Deprecated in Php8.1
363	113	$iniset = $this->setAutoDetect('1');
364	113
365	113	// Open file
366		if ($dataUri) {
367		$this->openDataUri($filename);
368	113	} else {
369	4	$this->openFileOrMemory($filename);
370		}
371	113	$fileHandle = $this->fileHandle;
372	113
373	4	// Skip BOM, if any
374		$this->skipBOM();
375		$this->checkSeparator();
376		$this->inferSeparator();
377	113
378	113	// Create new PhpSpreadsheet object
379		while ($spreadsheet->getSheetCount() <= $this->sheetIndex) {
380		$spreadsheet->createSheet();
381	113	}
382	113	$sheet = $spreadsheet->setActiveSheetIndex($this->sheetIndex);
383	113	if ($this->sheetNameIsFileName) {
384	113	$sheet->setTitle(substr(basename($filename, '.csv'), 0, Worksheet::SHEET_TITLE_MAXIMUM_LENGTH));
385	113	}
386	113
387	113	// Set our starting row based on whether we're in contiguous mode or not
388	113	$currentRow = 1;
389	113	$outRow = 0;
390	112
391	112	// Loop through each line of the file in turn
392	112	$delimiter = $this->delimiter ?? '';
393	112	$rowData = self::getCsv($fileHandle, 0, $delimiter, $this->enclosure, $this->escapeCharacter);
394	4	$valueBinder = Cell::getValueBinder();
395		$preserveBooleanString = method_exists($valueBinder, 'getBooleanConversion') && $valueBinder->getBooleanConversion();
396	108	$this->getTrue = Calculation::getTRUE();
397		$this->getFalse = Calculation::getFALSE();
398	112	$this->thousandsSeparator = StringHelper::getThousandsSeparator();
399	112	$this->decimalSeparator = StringHelper::getDecimalSeparator();
400	112	while (is_array($rowData)) {
401	3	$noOutputYet = true;
402	3	$columnLetter = 'A';
403	3	foreach ($rowData as $rowDatum) {
404		if ($preserveBooleanString) {
405		$rowDatum = $rowDatum ?? '';
406	109	} else {
407		$this->convertBoolean($rowDatum);
408		}
409	112	$numberFormatMask = $this->castFormattedNumberToNumeric ? $this->convertFormattedNumber($rowDatum) : '';
410	7	if (($rowDatum !== '' \|\| $this->preserveNullString) && $this->readFilter->readCell($columnLetter, $currentRow)) {
411	7	if ($this->contiguous) {
412	7	if ($noOutputYet) {
413		$noOutputYet = false;
414		++$outRow;
415	112	}
416		} else {
417	112	$outRow = $currentRow;
418		}
419	112	// Set basic styling for the value (Note that this could be overloaded by styling in a value binder)
420	112	if ($numberFormatMask !== '') {
421		$sheet->getStyle($columnLetter . $outRow)
422		->getNumberFormat()
423		->setFormatCode($numberFormatMask);
424	113	}
425		// Set cell value
426	113	$sheet->getCell($columnLetter . $outRow)->setValue($rowDatum);
427		}
428		++$columnLetter;
429	113	}
430		$rowData = self::getCsv($fileHandle, 0, $delimiter, $this->enclosure, $this->escapeCharacter);
431		++$currentRow;
432		}
433
434		// Close file
435	108	fclose($fileHandle);
436
437	108	$this->setAutoDetect($iniset);
438	108
439	11	// Return
440	108	return $spreadsheet;
441	11	}
442
443		/**
444		* Convert string true/false to boolean, and null to null-string.
445		*/
446		private function convertBoolean(mixed &$rowDatum): void
447		{
448		if (is_string($rowDatum)) {
449		if (strcasecmp($this->getTrue, $rowDatum) === 0 \|\| strcasecmp('true', $rowDatum) === 0) {
450		$rowDatum = true;
451	14	} elseif (strcasecmp($this->getFalse, $rowDatum) === 0 \|\| strcasecmp('false', $rowDatum) === 0) {
452		$rowDatum = false;
453	14	}
454	14	} else {
455	14	$rowDatum = $rowDatum ?? '';
456	14	}
457	14	}
458	14
459	14	/**
460		* Convert numeric strings to int or float values.
461	14	*/
462	14	private function convertFormattedNumber(mixed &$rowDatum): string
463	14	{
464	7	$numberFormatMask = '';
465	7	if ($this->castFormattedNumberToNumeric === true && is_string($rowDatum)) {
466	7	$numeric = str_replace(
467	7	[$this->thousandsSeparator, $this->decimalSeparator],
468	7	['', '.'],
469		$rowDatum
470		);
471
472	14	if (is_numeric($numeric)) {
473		$decimalPos = strpos($rowDatum, $this->decimalSeparator);
474		if ($this->preserveNumericFormatting === true) {
475		$numberFormatMask = (str_contains($rowDatum, $this->thousandsSeparator))
476	14	? '#,##0' : '0';
477		if ($decimalPos !== false) {
478		$decimals = strlen($rowDatum) - $decimalPos - 1;
479	14	$numberFormatMask .= '.' . str_repeat('0', min($decimals, 6));
480		}
481	14	}
482
483		$rowDatum = ($decimalPos !== false) ? (float) $numeric : (int) $numeric;
484	10	}
485		}
486	10
487		return $numberFormatMask;
488	10	}
489
490		public function getDelimiter(): ?string
491	2	{
492		return $this->delimiter;
493	2	}
494
495		public function setDelimiter(?string $delimiter): self
496	9	{
497		$this->delimiter = $delimiter;
498	9
499	3	return $this;
500		}
501	9
502		public function getEnclosure(): string
503	9	{
504		return $this->enclosure;
505		}
506	1
507		public function setEnclosure(string $enclosure): self
508	1	{
509		if ($enclosure == '') {
510		$enclosure = '"';
511	5	}
512		$this->enclosure = $enclosure;
513	5
514		return $this;
515	5	}
516
517		public function getSheetIndex(): int
518	3	{
519		return $this->sheetIndex;
520	3	}
521
522	3	public function setSheetIndex(int $indexValue): self
523		{
524		$this->sheetIndex = $indexValue;
525	1
526		return $this;
527	1	}
528
529		public function setContiguous(bool $contiguous): self
530	8	{
531		$this->contiguous = $contiguous;
532	8
533		return $this;
534	8	}
535
536		public function getContiguous(): bool
537	1	{
538		return $this->contiguous;
539	1	}
540
541		/**
542		* Php9 intends to drop support for this parameter in fgetcsv.
543		* Not yet ready to mark deprecated in order to give users
544		* a migration path.
545	139	*/
546		public function setEscapeCharacter(string $escapeCharacter): self
547		{
548		$this->escapeCharacter = $escapeCharacter;
549	139
550	3	return $this;
551	3	}
552
553		public function getEscapeCharacter(): string
554	136	{
555		return $this->escapeCharacter ?? self::$defaultEscapeCharacter;
556		}
557	136
558	136	/**
559	105	* Can the current IReader read the file?
560		*/
561		public function canRead(string $filename): bool
562		{
563	31	// Check if file exists
564	31	try {
565	31	$this->openFile($filename);
566	31	} catch (ReaderException) {
567	31	return false;
568	31	}
569	31
570	31	fclose($this->fileHandle);
571
572	31	// Trust file extension if any
573		$extension = strtolower(pathinfo($filename, PATHINFO_EXTENSION));
574		if (in_array($extension, ['csv', 'tsv'])) {
575	20	return true;
576		}
577	20
578	20	// Attempt to guess mimetype
579	20	$type = mime_content_type($filename);
580	10	$supportedTypes = [
581		'application/csv',
582		'text/csv',
583		'text/plain',
584		'inode/x-empty',
585	20	'text/html',
586		];
587	20
588	20	return in_array($type, $supportedTypes, true);
589	20	}
590	20
591	20	private static function guessEncodingTestNoBom(string &$encoding, string &$contents, string $compare, string $setEncoding): void
592	20	{
593	20	if ($encoding === '') {
594	3	$pos = strpos($contents, $compare);
595		if ($pos !== false && $pos % strlen($compare) === 0) {
596		$encoding = $setEncoding;
597	20	}
598		}
599		}
600	30
601		private static function guessEncodingNoBom(string $filename): string
602	30	{
603	30	$encoding = '';
604	10	$contents = file_get_contents($filename);
605		self::guessEncodingTestNoBom($encoding, $contents, self::UTF32BE_LF, 'UTF-32BE');
606		self::guessEncodingTestNoBom($encoding, $contents, self::UTF32LE_LF, 'UTF-32LE');
607		self::guessEncodingTestNoBom($encoding, $contents, self::UTF16BE_LF, 'UTF-16BE');
608		self::guessEncodingTestNoBom($encoding, $contents, self::UTF16LE_LF, 'UTF-16LE');
609	30	if ($encoding === '' && preg_match('//u', $contents) === 1) {
610		$encoding = 'UTF-8';
611	30	}
612	30
613	30	return $encoding;
614	30	}
615	30
616	30	private static function guessEncodingTestBom(string &$encoding, string $first4, string $compare, string $setEncoding): void
617	30	{
618	30	if ($encoding === '') {
619		if (str_starts_with($first4, $compare)) {
620		$encoding = $setEncoding;
621	30	}
622		}
623		}
624	30
625		private static function guessEncodingBom(string $filename): string
626	30	{
627	30	$encoding = '';
628	20	$first4 = file_get_contents($filename, false, null, 0, 4);
629		if ($first4 !== false) {
630		self::guessEncodingTestBom($encoding, $first4, self::UTF8_BOM, 'UTF-8');
631	30	self::guessEncodingTestBom($encoding, $first4, self::UTF16BE_BOM, 'UTF-16BE');
632		self::guessEncodingTestBom($encoding, $first4, self::UTF32BE_BOM, 'UTF-32BE');
633		self::guessEncodingTestBom($encoding, $first4, self::UTF32LE_BOM, 'UTF-32LE');
634	1	self::guessEncodingTestBom($encoding, $first4, self::UTF16LE_BOM, 'UTF-16LE');
635		}
636	1
637		return $encoding;
638	1	}
639
640		public static function guessEncoding(string $filename, string $dflt = self::DEFAULT_FALLBACK_ENCODING): string
641	1	{
642		$encoding = self::guessEncodingBom($filename);
643	1	if ($encoding === '') {
644		$encoding = self::guessEncodingNoBom($filename);
645		}
646	4
647		return ($encoding === '') ? $dflt : $encoding;
648	4	}
649
650	4	public function setPreserveNullString(bool $value): self
651		{
652		$this->preserveNullString = $value;
653
654		return $this;
655		}
656
657		public function getPreserveNullString(): bool
658		{
659		return $this->preserveNullString;
660		}
661
662		public function setSheetNameIsFileName(bool $sheetNameIsFileName): self
663		{
664		$this->sheetNameIsFileName = $sheetNameIsFileName;
665
666		return $this;
667		}
668
669		/**
670		* Php8.4 deprecates use of anything other than null string
671		* as escape Character.
672		*
673		* @param resource $stream
674		* @param null\|int<0, max> $length
675		*
676		* @return array<int,?string>\|false
677		*/
678		private static function getCsv(
679		$stream,
680		?int $length = null,
681		string $separator = ',',
682		string $enclosure = '"',
683		?string $escape = null
684		): array\|false {
685		$escape = $escape ?? self::$defaultEscapeCharacter;
686		if (PHP_VERSION_ID >= 80400 && $escape !== '') {
687		return @fgetcsv($stream, $length, $separator, $enclosure, $escape);
688		}
689
690		return fgetcsv($stream, $length, $separator, $enclosure, $escape);
691		}
692		}
693

PHPOffice / PhpSpreadsheet

Pull Request — master (#4162)

Csv::getCsv() A

Complexity

Size

Duplication

Code Coverage

Importance

Duplication Side-by-Side

Filter issues like