VITEC-website/packages/vitec/Classes/Import/MarketHeroPageImporter.php

<?php

declare(strict_types=1);

namespace Evomedien\Vitec\Import;

use PhpOffice\PhpSpreadsheet\IOFactory;
use TYPO3\CMS\Core\Database\ConnectionPool;
use TYPO3\CMS\Core\DataHandling\DataHandler;
use TYPO3\CMS\Core\Resource\StorageRepository;
use TYPO3\CMS\Core\Utility\GeneralUtility;

/**
 * Imports ONE agency "Markets hero sub page" workbook (Joe's Dropbox
 * delivery, structure decision Olli 2026-10-23) into its market hero page -
 * the eleven pages directly below /markets after vitec:align-market-pages.
 *
 * DELIBERATE SIBLING of the other workbook importers - pipelines evolve
 * independently.
 *
 * Workbook format: sheets Read me / Page layout / Content / Sources / Meta;
 * the Content sheet is a label table (column A component, column B copy,
 * column C/D guidance). The market is identified by the FILE name (minus a
 * " market page" suffix); the hero page must already exist below /markets.
 *
 * Target template is the NEW layout 23 "Market Hero Page" (zones hero=1,
 * main=0). Element plan, in page order:
 *
 *   colPos 1: vitec_markethero        (CTA -> #submarkets, bg slot 01)
 *             vitec_introparagraph    (header = Hero H1, body = hero intro)
 *   colPos 0: vitec_marketintro       (Market intro H2/copy + CTA label)
 *             vitec_introparagraph    (Sub-markets H2)
 *             vitec_card xN           (sub-market cards, text_only, CTA links
 *                                      to the child page when it exists -
 *                                      re-run after the sub-market import
 *                                      resolves the rest)
 *             vitec_customerlogos     (Trusted by - HIDDEN shell, logos are
 *                                      editorial)
 *             vitec_usecaselist       (Case studies - HIDDEN shell, selection
 *                                      is editorial)
 *             vitec_ctabanner         (prefooter, two CTA labels)
 *
 * Page properties as on the solution pages (og/twitter/canonical/abstract
 * from the hero intro copy, keywords from MARKET_KEYWORDS).
 *
 * NOTE norm(): '&' maps to 'and' here (unlike the sub-market/sub-solution
 * importers) so the file "Broadcast & Telecoms" matches the page
 * "Broadcast and Telecoms". Both spellings occur in Joe's delivery.
 */
final class MarketHeroPageImporter
{
    private const LAYOUT = 23;
    private const COLPOS_HERO = 1;
    private const COLPOS_MAIN = 0;
    private const CARD_ROWS = 6;

    /** Curated page keywords per market (Claude, 2026-10-23). Keys per norm(). */
    private const MARKET_KEYWORDS = [
        'hospitality' => 'hospitality iptv, hotel tv systems, guest entertainment, digital signage hotels, cruise video systems',
        'transport' => 'transport video solutions, airport iptv, passenger information displays, transport signage, station video systems',
        'sportsandentertainment' => 'stadium iptv, venue video distribution, sports video solutions, fan engagement displays, casino video systems',
        'broadcastandtelecoms' => 'broadcast video solutions, telecom video distribution, contribution encoding, media networks, video over ip broadcast',
        'education' => 'campus iptv, education video distribution, lecture capture, student digital signage, university video systems',
        'corporation' => 'corporate iptv, enterprise video distribution, workplace digital signage, internal communications video',
        'industrial' => 'industrial video monitoring, mining control rooms, manufacturing video systems, logistics monitoring',
        'criticalnationalinfrastructure' => 'critical infrastructure video, utility control rooms, energy grid monitoring, data centre video walls',
        'government' => 'government video solutions, public sector iptv, traffic control rooms, agency video distribution',
        'publicsafety' => 'public safety video, emergency operations centre, rtcc video wall, real time crime center technology',
        'defenseandsecurity' => 'defense video solutions, military video distribution, isr video, command and control video, secure video systems',
    ];

    /** @var string[] */
    private array $report = [];
    /** @var array<string,array<string|int,array<string,mixed>>> */
    private array $datamap = [];
    /** @var array<string,array<int,array<string,int>>> */
    private array $cmdmap = [];
    /** @var array<int,string[]> */
    private array $sequence = [];
    /** @var array<string,int[]> */
    private array $slotCursor = [];
    private int $newCounter = 0;

    /**
     * @return array{file:string,market:string,pageUid:int,elements:int,applied:bool,report:string[]}
     */
    public function importFile(string $file, bool $apply): array
    {
        $this->report = [];
        $this->datamap = [];
        $this->cmdmap = [];
        $this->sequence = [];
        $this->slotCursor = [];

        $market = trim((string)preg_replace('/\s+market page$/i', '', (string)preg_replace('/\.xlsx$/i', '', basename($file))));
        if ($this->norm($market) === 'marketsheropage') {
            throw new \RuntimeException('"Markets Hero Page" is the /markets main page (template 2) - own flow, skipped here.');
        }
        $values = $this->readContentSheet($file);

        // --- resolve the hero page below /markets --------------------------------
        $connection = GeneralUtility::makeInstance(ConnectionPool::class)->getConnectionForTable('pages');
        $root = $connection->fetchAssociative("SELECT uid FROM pages WHERE deleted = 0 AND slug = '/markets'");
        if (!$root) {
            throw new \RuntimeException('Page /markets not found.');
        }
        $page = $this->childByTitle((int)$root['uid'], $market);
        if ($page === null) {
            throw new \RuntimeException('Market hero page "' . $market . '" not found below /markets - run vitec:align-market-pages first.');
        }
        $pageUid = (int)$page['uid'];
        $pageId = (string)$pageUid;
        if ((int)$page['layout'] !== self::LAYOUT) {
            $this->datamap['pages'][$pageId]['layout'] = self::LAYOUT;
            $this->datamap['pages'][$pageId]['backend_layout'] = 'vitec__' . self::LAYOUT;
            $this->report[] = 'page: switch layout -> ' . self::LAYOUT . ' (was ' . (int)$page['layout'] . ')';
        }

        // --- SEO / page properties ------------------------------------------------
        $seoTitle = trim((string)($values['seo title'] ?? ''));
        $seoDescription = trim((string)($values['seo description'] ?? ''));
        if ($seoTitle !== '') {
            $this->setFields('pages', $pageId, [
                'seo_title' => $seoTitle, 'og_title' => $seoTitle, 'twitter_title' => $seoTitle,
            ]);
        }
        if ($seoDescription !== '') {
            $this->setFields('pages', $pageId, [
                'description' => $seoDescription, 'og_description' => $seoDescription, 'twitter_description' => $seoDescription,
            ]);
        }
        $this->datamap['pages'][$pageId]['canonical_link'] = 't3://page?uid=' . $pageId;
        $abstract = $this->abstractFrom((string)($values['hero intro copy'] ?? ''), $seoDescription);
        if ($abstract !== '') {
            $this->datamap['pages'][$pageId]['abstract'] = $abstract;
        }
        $keywords = self::MARKET_KEYWORDS[$this->norm($market)] ?? '';
        if ($keywords !== '') {
            $this->datamap['pages'][$pageId]['keywords'] = $keywords;
        }
        $this->report[] = 'page meta: og/twitter + canonical + abstract (' . mb_strlen($abstract) . ' chars) + keywords' . ($keywords === '' ? ' MISSING (no curated set for "' . $market . '")' : '');

        // --- hero zone -------------------------------------------------------------
        $heroId = $this->slotNext($pageUid, self::COLPOS_HERO, 'vitec_markethero');
        $this->setFields('tt_content', $heroId, [
            'vitec_cta_label' => (string)($values['hero primary cta'] ?? ''),
            'vitec_cta_link' => '#submarkets',
        ]);
        $this->attachPlaceholder($heroId, 'vitec_bgimage', $apply, $pageUid, 1);

        $heroIntroId = $this->slotNext($pageUid, self::COLPOS_HERO, 'vitec_introparagraph');
        $this->setFields('tt_content', $heroIntroId, [
            'header' => (string)($values['hero h1'] ?? ''),
            'bodytext' => $this->p((string)($values['hero intro copy'] ?? '')),
        ]);

        // --- main: market intro -----------------------------------------------------
        $introId = $this->slotNext($pageUid, self::COLPOS_MAIN, 'vitec_marketintro');
        $this->setFields('tt_content', $introId, [
            'vitec_intro_headline' => (string)($values['market intro h2'] ?? ''),
            'vitec_intro_text' => $this->p((string)($values['market intro copy'] ?? '')),
            'vitec_learnmore_label' => (string)($values['market intro secondary cta'] ?? ''),
            'vitec_show_portfolio' => 0,
        ]);
        if (trim((string)($values['market intro secondary cta'] ?? '')) !== '') {
            $this->report[] = 'intro CTA "' . trim((string)$values['market intro secondary cta']) . '": set the link target editorially (contact page)';
        }

        // --- main: sub-market cards --------------------------------------------------
        $subHeaderId = $this->slotNext($pageUid, self::COLPOS_MAIN, 'vitec_introparagraph');
        $this->setFields('tt_content', $subHeaderId, [
            'header' => (string)($values['sub-markets h2'] ?? 'Explore ' . $market),
        ]);
        foreach (range(1, self::CARD_ROWS) as $i) {
            $title = trim((string)($values['sub-market card ' . $i . ' title'] ?? ''));
            if ($title === '') {
                continue;
            }
            $cardId = $this->slotNext($pageUid, self::COLPOS_MAIN, 'vitec_card');
            $fields = [
                'vitec_layout_variant' => 'text_only',
                'header' => $title,
                'bodytext' => $this->p((string)($values['sub-market card ' . $i . ' body'] ?? '')),
                'vitec_primary_cta_label' => (string)($values['sub-market card ' . $i . ' cta'] ?? 'Learn more'),
            ];
            $child = $this->childByTitle($pageUid, $title);
            if ($child !== null) {
                $fields['vitec_primary_cta'] = 't3://page?uid=' . (int)$child['uid'];
                $this->report[] = 'sub-market card "' . $title . '" -> page #' . (int)$child['uid'];
            } else {
                $this->report[] = 'sub-market card "' . $title . '": child page missing - re-run after the sub-market import to set the link';
            }
            $this->setFields('tt_content', $cardId, $fields);
        }

        // --- main: trusted by (hidden shell) + case studies (hidden shell) -----------
        $logosExisting = $this->slotElements($pageUid, self::COLPOS_MAIN, 'vitec_customerlogos') !== [];
        $logosId = $this->slotNext($pageUid, self::COLPOS_MAIN, 'vitec_customerlogos');
        $fields = ['header' => (string)($values['trusted by h2'] ?? 'Trusted by')];
        if (!$logosExisting) {
            $fields['hidden'] = 1;
            $this->report[] = 'trusted-by: created HIDDEN - pick the customer logos editorially, then unhide';
        }
        $this->setFields('tt_content', $logosId, $fields);

        $casesExisting = $this->slotElements($pageUid, self::COLPOS_MAIN, 'vitec_usecaselist') !== [];
        $casesId = $this->slotNext($pageUid, self::COLPOS_MAIN, 'vitec_usecaselist');
        $fields = ['header' => (string)($values['case studies h2'] ?? 'Case studies')];
        if (!$casesExisting) {
            $fields['hidden'] = 1;
            $fields['pi_flexform'] = $this->flex([
                'settings.usecases' => '',
                'settings.showtoolbar' => '0',
            ]);
            $this->report[] = 'case studies: created HIDDEN - pick the stories editorially, then unhide';
        }
        $this->setFields('tt_content', $casesId, $fields);

        // --- main: prefooter -----------------------------------------------------------
        $preId = $this->slotNext($pageUid, self::COLPOS_MAIN, 'vitec_ctabanner');
        $this->setFields('tt_content', $preId, [
            'header' => (string)($values['prefooter h2'] ?? ''),
            'bodytext' => $this->p((string)($values['prefooter supporting copy'] ?? '')),
            'vitec_primary_cta_label' => (string)($values['prefooter primary cta'] ?? ''),
            'vitec_secondary_cta_label' => (string)($values['prefooter secondary cta'] ?? ''),
            'vitec_layout_variant' => 'centered',
            'vitec_background_variant' => 'orange',
        ]);
        $this->report[] = 'prefooter CTAs: set the link targets editorially (contact / solutions)';

        $result = [
            'file' => basename($file),
            'market' => $market,
            'pageUid' => $pageUid,
            'elements' => count($this->datamap['tt_content'] ?? []),
            'applied' => $apply,
            'report' => $this->report,
        ];
        if (!$apply) {
            return $result;
        }

        $dataHandler = GeneralUtility::makeInstance(DataHandler::class);
        $dataHandler->start($this->datamap, $this->cmdmap);
        $dataHandler->process_datamap();
        if ($this->cmdmap !== []) {
            $dataHandler->process_cmdmap();
        }
        if ($dataHandler->errorLog !== []) {
            throw new \RuntimeException('DataHandler: ' . implode(' | ', $dataHandler->errorLog));
        }
        $this->applySequence($dataHandler->substNEWwithIDs);
        $result['report'] = $this->report;
        return $result;
    }

    // ---- workbook reading --------------------------------------------------------

    /** @return array<string,mixed> lower-cased column-A label => column-B value, first hit wins */
    private function readContentSheet(string $file): array
    {
        $spreadsheet = IOFactory::load($file);
        $content = $spreadsheet->getSheetByName('Content');
        if ($content === null) {
            $spreadsheet->disconnectWorksheets();
            throw new \RuntimeException('No "Content" sheet - not a market hero workbook?');
        }
        $values = [];
        $highestRow = min(200, (int)$content->getHighestRow());
        for ($rowNum = 1; $rowNum <= $highestRow; $rowNum++) {
            $label = mb_strtolower(trim((string)$content->getCell('A' . $rowNum)->getValue()));
            $label = rtrim($label, ' *');
            if ($label === '' || str_starts_with($label, 'section ')) {
                continue;
            }
            if (!array_key_exists($label, $values)) {
                $values[$label] = $content->getCell('B' . $rowNum)->getValue();
            }
        }
        $spreadsheet->disconnectWorksheets();
        return $values;
    }

    private function p(string $text): string
    {
        $text = trim($text);
        return $text === '' ? '' : '<p>' . htmlspecialchars($text, ENT_QUOTES | ENT_HTML5) . '</p>';
    }

    /** @return array{uid:int,layout:int}|null */
    private function childByTitle(int $parentUid, string $title): ?array
    {
        $connection = GeneralUtility::makeInstance(ConnectionPool::class)->getConnectionForTable('pages');
        foreach ($connection->fetchAllAssociative(
            'SELECT uid, title, layout FROM pages WHERE deleted = 0 AND pid = ? AND sys_language_uid IN (0,-1)', [$parentUid]
        ) as $row) {
            if ($this->norm((string)$row['title']) === $this->norm($title)) {
                return ['uid' => (int)$row['uid'], 'layout' => (int)$row['layout']];
            }
        }
        return null;
    }

    // ---- element plumbing (pattern of the sibling importers) -----------------------

    private function slotNext(int $pageUid, int $colPos, string $ctype): string
    {
        $key = $colPos . '/' . $ctype;
        if (!isset($this->slotCursor[$key])) {
            $this->slotCursor[$key] = $this->slotElements($pageUid, $colPos, $ctype);
        }
        $uid = array_shift($this->slotCursor[$key]);
        if ($uid !== null) {
            $this->sequence[$colPos][] = (string)$uid;
            return (string)$uid;
        }
        $id = $this->newId();
        $this->datamap['tt_content'][$id] = [
            'pid' => $pageUid,
            'colPos' => $colPos,
            'CType' => $ctype,
        ];
        $this->sequence[$colPos][] = $id;
        return $id;
    }

    /** @return int[] uids in sorting order */
    private function slotElements(int $pageUid, int $colPos, string $ctype): array
    {
        $connection = GeneralUtility::makeInstance(ConnectionPool::class)->getConnectionForTable('tt_content');
        return array_map('intval', $connection->fetchFirstColumn(
            'SELECT uid FROM tt_content WHERE deleted = 0 AND pid = ? AND colPos = ? AND CType = ? ORDER BY sorting',
            [$pageUid, $colPos, $ctype]
        ));
    }

    /**
     * Writes tt_content.sorting in page order (DataHandler puts NEW records
     * with a positive pid at the TOP of their column - lesson 2026-10-07).
     *
     * @param array<string,int> $substNEWwithIDs
     */
    private function applySequence(array $substNEWwithIDs): void
    {
        $connection = GeneralUtility::makeInstance(ConnectionPool::class)->getConnectionForTable('tt_content');
        foreach ($this->sequence as $ids) {
            $sorting = 0;
            foreach ($ids as $id) {
                $uid = str_starts_with($id, 'NEW') ? (int)($substNEWwithIDs[$id] ?? 0) : (int)$id;
                if ($uid > 0) {
                    $sorting += 256;
                    $connection->update('tt_content', ['sorting' => $sorting], ['uid' => $uid]);
                }
            }
        }
    }

    /** @param array<string,mixed> $fields */
    private function setFields(string $table, string $id, array $fields): void
    {
        foreach ($fields as $field => $value) {
            $this->datamap[$table][$id][$field] = $value;
        }
    }

    /** Grey numbered stand-in, same rules as the sibling importers. */
    private function attachPlaceholder(string $elementId, string $fieldName, bool $apply, int $pid, int $slot): void
    {
        if (!str_starts_with($elementId, 'NEW') && $this->liveReferences((int)$elementId, $fieldName) !== []) {
            return;
        }
        $storage = GeneralUtility::makeInstance(StorageRepository::class)->getDefaultStorage();
        $identifier = null;
        if ($slot > 0) {
            $candidate = 'placeholders/pages/' . $pid . '-' . str_pad((string)$slot, 2, '0', STR_PAD_LEFT) . '.png';
            if ($storage->hasFile($candidate)) {
                $identifier = $candidate;
            }
        }
        $identifier ??= 'placeholders/' . ($fieldName === 'vitec_bgimage' ? 'placeholder-grey-wide.png' : 'placeholder-grey.png');
        if (!$storage->hasFile($identifier)) {
            $this->report[] = $fieldName . ': empty in the workbook and fileadmin/' . $identifier . ' is missing - left empty';
            return;
        }
        if (!$apply) {
            $this->report[] = $fieldName . ': empty in the workbook - would attach ' . $identifier;
            return;
        }
        if ($this->attachFalFile($elementId, $fieldName, $storage->getFile($identifier), $pid) !== 'unchanged') {
            $this->report[] = $fieldName . ': placeholder ' . basename($identifier) . ' attached - swap in the real asset later';
        }
    }

    /** Content-idempotent FAL attach (lesson 2026-10-02), as in the siblings. */
    private function attachFalFile(string $elementId, string $fieldName, \TYPO3\CMS\Core\Resource\File $file, int|string $pid): string
    {
        if (!str_starts_with($elementId, 'NEW')) {
            $existingRefs = $this->liveReferences((int)$elementId, $fieldName);
            if ($existingRefs !== []) {
                $first = array_shift($existingRefs);
                $state = 'unchanged';
                if ((int)$first['uid_local'] !== $file->getUid()) {
                    $this->datamap['sys_file_reference'][(int)$first['uid']] = ['uid_local' => $file->getUid()];
                    $state = 'swapped';
                }
                foreach ($existingRefs as $surplus) {
                    $this->cmdmap['sys_file_reference'][(int)$surplus['uid']]['delete'] = 1;
                    $this->report[] = $fieldName . ': removed duplicate reference #' . $surplus['uid'];
                }
                return $state;
            }
        }
        $refId = $this->newId();
        // pid, tablenames and uid_foreign are REQUIRED - without them
        // DataHandler silently rejects the reference (lesson 2026-10-02).
        $this->datamap['sys_file_reference'][$refId] = [
            'pid' => $pid,
            'table_local' => 'sys_file',
            'uid_local' => $file->getUid(),
            'tablenames' => 'tt_content',
            'fieldname' => $fieldName,
            'uid_foreign' => str_starts_with($elementId, 'NEW') ? $elementId : (int)$elementId,
        ];
        $this->datamap['tt_content'][$elementId][$fieldName] = $refId;
        return 'attached';
    }

    /** @return array<int,array{uid:int,uid_local:int}> live refs on one field, oldest first */
    private function liveReferences(int $elementUid, string $fieldName): array
    {
        $connection = GeneralUtility::makeInstance(ConnectionPool::class)->getConnectionForTable('sys_file_reference');
        return array_map(
            static fn(array $r): array => ['uid' => (int)$r['uid'], 'uid_local' => (int)$r['uid_local']],
            $connection->fetchAllAssociative(
                "SELECT uid, uid_local FROM sys_file_reference
                 WHERE deleted = 0 AND hidden = 0 AND tablenames = 'tt_content'
                   AND fieldname = ? AND uid_foreign = ? ORDER BY uid",
                [$fieldName, $elementUid]
            )
        );
    }

    /** Abstract from the hero intro copy, word-boundary cut near 300 chars. */
    private function abstractFrom(string $introHtml, string $fallback): string
    {
        $plain = trim((string)preg_replace('/\s+/u', ' ', html_entity_decode(strip_tags($introHtml), ENT_QUOTES | ENT_HTML5)));
        if ($plain === '') {
            return $fallback;
        }
        if (mb_strlen($plain) <= 300) {
            return $plain;
        }
        $cut = mb_substr($plain, 0, 300);
        $pos = mb_strrpos($cut, ' ');
        return rtrim(mb_substr($cut, 0, $pos ?: 300), ' ,;:.') . "\u{2026}";
    }

    /**
     * @param array<string,string> $settings
     * @return array<string,mixed>
     */
    private function flex(array $settings): array
    {
        $fields = [];
        foreach ($settings as $key => $value) {
            $fields[$key] = ['vDEF' => $value];
        }
        return ['data' => ['sDEF' => ['lDEF' => $fields]]];
    }

    private function newId(): string
    {
        return 'NEW' . str_pad((string)(++$this->newCounter), 6, '0', STR_PAD_LEFT) . substr(md5(uniqid('', true)), 0, 6);
    }

    /** '&' -> 'and' so "Broadcast & Telecoms" matches "Broadcast and Telecoms". */
    private function norm(string $value): string
    {
        return (string)preg_replace('/[^a-z0-9]+/', '', str_replace(['+', '&'], ['plus', 'and'], mb_strtolower($value)));
    }
}