Teren-app/app/Console/Commands/TemplateScanCommand.php

<?php

namespace App\Console\Commands;

use App\Models\DocumentTemplate;
use App\Services\Documents\TokenScanner;
use Illuminate\Console\Command;
use Illuminate\Support\Facades\Storage;
use ZipArchive;

class TemplateScanCommand extends Command
{
    protected $signature = 'template:scan {slug : Template slug} {--tpl-version= : Specific template version number} {--parts : Show per-part tokens}';

    protected $description = 'Scan a stored DOCX template by slug/version and dump detected tokens directly from storage.';

    public function handle(TokenScanner $scanner): int
    {
        $slug = (string) $this->argument('slug');
        $version = $this->option('tpl-version');

        /** @var DocumentTemplate|null $template */
        $query = DocumentTemplate::query()->where('slug', $slug);
        if (! empty($version)) {
            $query->where('version', (int) $version);
        } else {
            $query->orderByDesc('version');
        }
        $template = $query->first();
        if (! $template) {
            $this->error("Template not found for slug '{$slug}'".($version ? " v{$version}" : ''));

            return self::FAILURE;
        }

        $disk = 'public';
        $path = $template->file_path;
        if (! $path || ! Storage::disk($disk)->exists($path)) {
            $this->error('Template file not found on disk: '.$path);

            return self::FAILURE;
        }

        $bytes = Storage::disk($disk)->get($path);
        $tmp = tempnam(sys_get_temp_dir(), 'tmpl');
        file_put_contents($tmp, $bytes);

        $zip = new ZipArchive;
        if ($zip->open($tmp) !== true) {
            $this->error('Unable to open DOCX (zip).');

            return self::FAILURE;
        }

        // Collect parts: main + headers/footers + notes/comments
        $parts = [];
        $doc = $zip->getFromName('word/document.xml');
        if ($doc !== false) {
            $parts['word/document.xml'] = $doc;
        }
        for ($i = 0; $i < $zip->numFiles; $i++) {
            $name = $zip->getNameIndex($i);
            if (! is_string($name)) {
                continue;
            }
            if (preg_match('#^word/(header\d*|footer\d*|footnotes|endnotes|comments)\.xml$#i', $name)) {
                $xml = $zip->getFromName($name);
                if ($xml !== false) {
                    $parts[$name] = $xml;
                }
            }
        }

        // Normalize and scan
        $all = [];
        $perPart = [];
        foreach ($parts as $name => $xml) {
            $norm = $this->normalizeRunsForTokens($xml);
            $found = $scanner->scan($norm);
            $perPart[$name] = $found;
            if ($found) {
                $all = array_merge($all, $found);
            }
        }
        $union = array_values(array_unique($all));

        $this->info("Template: {$template->name} (slug={$template->slug}, v{$template->version})");
        $this->line('File: '.$path);
        $this->line('Tokens found (union): '.count($union));
        foreach ($union as $t) {
            $this->line(' - '.$t);
        }

        if ($this->option('parts')) {
            $this->line('');
            $this->info('Per-part details:');
            foreach ($perPart as $n => $list) {
                $this->line("[{$n}] (".count($list).')');
                foreach ($list as $t) {
                    $this->line('   - '.$t);
                }
            }
        }

        $zip->close();
        @unlink($tmp);

        return self::SUCCESS;
    }

    private function normalizeRunsForTokens(string $xml): string
    {
        // Remove proofing error markers
        $xml = preg_replace('#<w:proofErr[^>]*/>#i', '', $xml) ?? $xml;
        // Collapse boundaries between runs and inside runs (include tabs/line breaks)
        $patterns = [
            '#</w:t>\s*</w:r>\s*(?:<(?:w:proofErr|w:tab|w:br)[^>]*/>\s*)*(?:<w:r[^>]*>\s*(?:<w:rPr>.*?</w:rPr>\s*)*)?<w:t[^>]*>#is',
            '#</w:t>\s*(?:<(?:w:proofErr|w:tab|w:br)[^>]*/>\s*)*<w:t[^>]*>#is',
        ];
        $prev = null;
        while ($prev !== $xml) {
            $prev = $xml;
            foreach ($patterns as $pat) {
                $xml = preg_replace($pat, '', $xml) ?? $xml;
            }
        }
        // Clean inside {{ ... }}
        $xml = preg_replace_callback('/\{\{.*?\}\}/s', function (array $m) {
            $inner = substr($m[0], 2, -2);
            $inner = preg_replace('/<[^>]+>/', '', $inner) ?? $inner;
            $inner = preg_replace('/\s+/', '', $inner) ?? $inner;

            return '{{'.$inner.'}}';
        }, $xml) ?? $xml;
        // Clean inside { ... } if it looks like a token
        $xml = preg_replace_callback('/\{[^{}]*\}/s', function (array $m) {
            $raw = $m[0];
            $inner = substr($raw, 1, -1);
            $clean = preg_replace('/<[^>]+>/', '', $inner) ?? $inner;
            $clean = preg_replace('/\s+/', '', $clean) ?? $clean;
            if (preg_match('/^[a-zA-Z0-9_]+(?:\.[a-zA-Z0-9_]+)*\.[a-zA-Z0-9_.-]+$/', $clean)) {
                return '{'.$clean.'}';
            }

            return $raw;
        }, $xml) ?? $xml;
        // Remove zero-width and soft hyphen
        $xml = str_replace(["\xE2\x80\x8B", "\xC2\xAD"], '', $xml);

        return $xml;
    }
}