math_cms/app/Support/BlankPlaceholderRenderer.php

<?php

namespace App\Support;

class BlankPlaceholderRenderer
{
    private const DEFAULT_BLANK_SPAN = '<span style="display:inline-block; min-width:80px; border-bottom:1.2px dashed #444; vertical-align:bottom;">&nbsp;</span>';

    /**
     * 将题干中的空括号/下划线/部分异常占位符统一替换为标准空位样式。
     *
     * @return array{0:string,1:bool} [renderedContent, replacedAnyPlaceholder]
     */
    public static function replaceToBlankSpan(
        string $content,
        ?string $blankSpan = null,
        bool $collapseAdjacentBlanks = false,
        bool $normalizeChineseTerminalPeriod = true
    ): array
    {
        $blankSpan = $blankSpan ?: self::DEFAULT_BLANK_SPAN;
        $renderedContent = $content;

        $latexPlaceholders = [];
        $counter = 0;
        $renderedContent = preg_replace_callback('/\$(?:[^\$]|\\\\.)*\$/u', function ($matches) use (&$latexPlaceholders, &$counter, $blankSpan) {
            $latexContent = $matches[0];
            $inner = mb_substr($latexContent, 1, mb_strlen($latexContent) - 2);

            // 数学环境内也可能包含填空占位符（如 $\\underline{\\qquad}$ / $\\angle A=\\underline{\\quad}$）
            $blankToken = '<<<BLANK_IN_MATH_'.$counter.'>>>';
            $innerWithBlanks = preg_replace(
                [
                    '/\\\\underline\{[^}]*\}/u',
                    '/\\\\qquad+/u',
                    '/\\\\quad+/u',
                    '/[（(](?:\s|&nbsp;|&#160;|　)*[）)]/u',
                    '/_{2,}/u',
                ],
                $blankToken,
                $inner,
                -1,
                $blankCount
            );
            if ($blankCount > 0) {
                $parts = explode($blankToken, $innerWithBlanks);
                $rebuilt = '';
                $lastIndex = count($parts) - 1;
                foreach ($parts as $index => $part) {
                    if ($part !== '') {
                        // 纯标点不再包进数学环境，避免生成 "$.$" 这类尾部格式。
                        if (preg_match('/^[\.．。]$/u', $part)) {
                            $rebuilt .= $part;
                        } else {
                            $rebuilt .= htmlspecialchars('$'.$part.'$', ENT_QUOTES | ENT_HTML5, 'UTF-8');
                        }
                    }
                    if ($index < $lastIndex) {
                        $rebuilt .= $blankSpan;
                    }
                }

                return $rebuilt === '' ? $blankSpan : $rebuilt;
            }

            $placeholder = '<<<LATEX_BLANK_'.$counter.'>>>';
            $latexPlaceholders[$placeholder] = $latexContent;
            $counter++;

            return $placeholder;
        }, $renderedContent);

        // 兼容常见空位写法：\underline{...}、\qquad、空括号（含 nbsp 等空白）、连续下划线、尾部 \\$
        $patterns = [
            '/\\\underline\{[^}]*\}/u',
            '/\\\qquad+/u',
            '/[（(](?:\s|&nbsp;|&#160;|　)*[）)]/u',
            '/_{2,}/u',
            '/\\\\+\$(?=\s*$)/u',
        ];
        $renderedContent = preg_replace($patterns, $blankSpan, $renderedContent);
        if ($collapseAdjacentBlanks) {
            $quotedBlankSpan = preg_quote($blankSpan, '/');
            $renderedContent = preg_replace('/(?:'.$quotedBlankSpan.'(?:\s|&nbsp;|&#160;|　)*){2,}/u', $blankSpan, $renderedContent);
        }

        foreach ($latexPlaceholders as $placeholder => $latexContent) {
            if (preg_match('/^\$(.*?)(\\\\+)\$$/u', $latexContent, $match)) {
                $inner = rtrim($match[1]);
                if ($inner === '' || preg_match('/[=:：]\s*$/u', $inner)) {
                    if ($inner === '') {
                        $replacement = $blankSpan;
                    } else {
                        $replacement = htmlspecialchars('$'.$inner.'$', ENT_QUOTES | ENT_HTML5, 'UTF-8').' '.$blankSpan;
                    }
                    $renderedContent = str_replace($placeholder, $replacement, $renderedContent);
                    continue;
                }
            }

            $encodedLatex = htmlspecialchars($latexContent, ENT_QUOTES | ENT_HTML5, 'UTF-8');
            $renderedContent = str_replace($placeholder, $encodedLatex, $renderedContent);
        }

        if ($normalizeChineseTerminalPeriod) {
            $renderedContent = self::normalizeChineseTerminalPeriod($renderedContent);
        }

        return [$renderedContent, $renderedContent !== $content];
    }

    public static function defaultBlankSpan(): string
    {
        return self::DEFAULT_BLANK_SPAN;
    }

    /**
     * 统一句尾标点（仅处理句尾，不影响中间小数/表达式）
     *
     * $mode:
     * - remove: 去掉句尾句号
     * - dot:    句尾统一为英文实心点 "."
     * - cn:     句尾统一为中文句号 "。"
     */
    public static function normalizeTerminalPunctuation(string $content, string $mode): string
    {
        $replacement = match ($mode) {
            'remove' => '',
            'dot' => '.',
            'cn' => '。',
            default => null,
        };
        if ($replacement === null) {
            return $content;
        }

        // 先处理数学片段尾点（如 "$.$" / "$。$" / "$．$"）。
        $content = preg_replace('/\$\s*[\.．。]\s*\$(?=\s*(?:(?:<\/[^>]+>|<[^>]+\/>)\s*)*$)/u', $replacement, $content);

        // 再处理普通句尾点。
        return preg_replace('/[\.．。](?=\s*(?:(?:<\/[^>]+>|<[^>]+\/>)\s*)*$)/u', $replacement, $content);
    }

    private static function normalizeChineseTerminalPeriod(string $content): string
    {
        // 仅在存在中文语境时，把句末英文句号统一为中文句号。
        if (! preg_match('/\p{Han}/u', $content)) {
            return $content;
        }

        return self::normalizeTerminalPunctuation($content, 'cn');
    }
}