Csv::guessEncodingTestBom() - 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 2024-09-06 08:13 UTC

Csv::guessEncodingTestBom() A

↳ Parent: Csv

Complexity

Conditions	3
Paths	3

Size

Total Lines	5
Code Lines	3

Duplication

Lines	0
Ratio	0 %

Code Coverage

Tests	4
CRAP Score	3

Importance

Changes

Metric	Value
cc	3
eloc	3
nc	3
nop	4
dl	0
loc	5
ccs	4
cts	4
cp	1
crap	3
rs	10
c	0
b	0
f	0

<?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.
     */
    private string $escapeCharacter = '\\';

    /**
     * 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, $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;
    }

    public function setEscapeCharacter(string $escapeCharacter): self
    {
        $this->escapeCharacter = $escapeCharacter;

        return $this;
    }

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

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

PHPOffice / PhpSpreadsheet

Pull Request — master (#4162)

Csv::guessEncodingTestBom() A

Complexity

Size

Duplication

Code Coverage

Importance

Duplication Side-by-Side

Filter issues like