Added Downloads to search index.

This commit is contained in:
2026-08-21 15:44:03 +02:00
parent 710f95794c
commit cf7a04ea15
6 changed files with 295 additions and 22 deletions

View File

@@ -36,6 +36,7 @@ class SolrIndexCommand extends Command
$this->addOption('limit', 'l', InputOption::VALUE_REQUIRED, 'Max queue items to index in this run', '20');
$this->addOption('root', 'r', InputOption::VALUE_REQUIRED, 'Root page uid of the site', '1');
$this->addOption('debug', 'd', InputOption::VALUE_NONE, 'Single-step the first queue item verbosely');
$this->addOption('initialize', 'i', InputOption::VALUE_REQUIRED, 'Initialize the index queue for these configurations (comma list or *) before indexing');
}
protected function execute(InputInterface $input, OutputInterface $output): int
@@ -68,7 +69,21 @@ class SolrIndexCommand extends Command
return Command::FAILURE;
}
$initialize = (string)($input->getOption('initialize') ?? '');
if ($initialize !== '') {
$names = $initialize === '*' ? ['*'] : array_map('trim', explode(',', $initialize));
$initService = GeneralUtility::makeInstance(\ApacheSolrForTypo3\Solr\Domain\Index\Queue\QueueInitializationService::class);
$result = $initService->initializeBySiteAndIndexConfigurations($site, $names);
foreach ($result as $name => $ok) {
$output->writeln(' initialized ' . $name . ': ' . var_export($ok, true));
}
}
if ($input->getOption('debug')) {
$queueConfig = $site->getSolrConfiguration()->getObjectByPathOrDefault('plugin.tx_solr.index.queue.');
$output->writeln('index.queue keys: ' . implode(', ', array_keys($queueConfig)));
$output->writeln('products.table: ' . var_export($queueConfig['products.']['table'] ?? null, true));
$queue = GeneralUtility::makeInstance(Queue::class);
$items = $queue->getItemsToIndex($site, 5);
$output->writeln('getItemsToIndex(5): ' . count($items) . ' items');

View File

@@ -13,17 +13,21 @@ use TYPO3\CMS\Core\Http\Response;
use TYPO3\CMS\Core\Http\Stream;
/**
* Forced file download for download records: /download/file/<uid>
* Forced file download for download records, two routes:
*
* This is the frontend target of the "Download" record links from the link
* browser (config.recordLinks.download builds exactly this path). A direct
* fileadmin URL would open PDFs inline; this endpoint streams the file with
* Content-Disposition: attachment, so the browser saves it.
* /download/<slug> - canonical, human-readable (search results, sharing)
* /download/file/<uid> - legacy, kept because the "Download" record links
* from the link browser build exactly this path
* (config.recordLinks.download)
*
* A direct fileadmin URL would open PDFs inline; this endpoint streams the
* file with Content-Disposition: attachment, so the browser saves it.
*
* The file lookup lives in Service\DownloadFileResolver (FAL -> Collateral
* naming convention -> filepath column). Unknown uid, hidden record or
* missing file fall through to the regular pipeline - the reply is then the
* normal 404 page, never a broken download.
* naming convention -> filepath column), which also refuses records flagged
* private_download. Unknown uid/slug, hidden or private record or missing
* file fall through to the regular pipeline - the reply is then the normal
* 404 page, never a broken download.
*/
final class DownloadFileMiddleware implements MiddlewareInterface
{
@@ -31,21 +35,26 @@ final class DownloadFileMiddleware implements MiddlewareInterface
{
try {
$path = $request->getUri()->getPath();
$file = null;
if (preg_match('#^/download/file/(\\d+)/?$#', $path, $matches) === 1) {
$file = DownloadFileResolver::resolve((int)$matches[1]);
if ($file !== null) {
$filename = str_replace(['"', "\r", "\n"], '', $file['name']);
} elseif (preg_match('#^/download/([a-z0-9\\-]+)/?$#', $path, $matches) === 1) {
$file = DownloadFileResolver::resolveBySlug($matches[1]);
}
return new Response(
new Stream($file['path'], 'rb'),
200,
[
'Content-Type' => $file['mimeType'] !== '' ? $file['mimeType'] : 'application/octet-stream',
'Content-Length' => (string)$file['size'],
'Content-Disposition' => 'attachment; filename="' . $filename . '"',
]
);
}
if ($file !== null) {
$filename = str_replace(['"', "\r", "\n"], '', $file['name']);
return new Response(
new Stream($file['path'], 'rb'),
200,
[
'Content-Type' => $file['mimeType'] !== '' ? $file['mimeType'] : 'application/octet-stream',
'Content-Length' => (string)$file['size'],
'Content-Disposition' => 'attachment; filename="' . $filename . '"',
]
);
}
} catch (\Throwable $e) {
// fall through to the regular pipeline

View File

@@ -51,6 +51,39 @@ final class DownloadFileResolver
}
}
/**
* Same contract as resolve(), addressed by the record slug - the pretty
* /download/<slug> route. Slugs are unique (eval: uniqueInPid).
*
* @return array{path: string, name: string, size: int, mimeType: string}|null
*/
public static function resolveBySlug(string $slug): ?array
{
if ($slug === '') {
return null;
}
try {
$qb = GeneralUtility::makeInstance(ConnectionPool::class)
->getQueryBuilderForTable('tx_vitec_domain_model_download');
$uid = $qb
->select('uid')
->from('tx_vitec_domain_model_download')
->where(
$qb->expr()->eq('slug', $qb->createNamedParameter($slug, ParameterType::STRING)),
$qb->expr()->eq('deleted', 0),
$qb->expr()->eq('hidden', 0)
)
->setMaxResults(1)
->executeQuery()
->fetchOne();
return $uid ? self::resolve((int)$uid) : null;
} catch (\Throwable $e) {
return null;
}
}
/** @return array<string,mixed>|null */
private static function loadDownload(int $uid): ?array
{
@@ -62,7 +95,10 @@ final class DownloadFileResolver
->where(
$qb->expr()->eq('uid', $qb->createNamedParameter($uid, ParameterType::INTEGER)),
$qb->expr()->eq('deleted', 0),
$qb->expr()->eq('hidden', 0)
$qb->expr()->eq('hidden', 0),
// private downloads are never served publicly - the caller
// cannot distinguish this from a missing file, by design
$qb->expr()->eq('private_download', 0)
)
->setMaxResults(1)
->executeQuery()

View File

@@ -59,6 +59,19 @@ class SearchJsonRenderer
}
private const CTYPE = 'solr_pi_results';
/**
* Solr document types (index.queue config tables) to frontend-friendly
* type labels. Unknown types pass through verbatim.
*/
private const TYPE_LABELS = [
'pages' => 'page',
'tx_vitec_domain_model_product' => 'product',
'tx_vitec_domain_model_market' => 'market',
'tx_vitec_domain_model_usecase' => 'story',
'tx_news_domain_model_news' => 'news',
'tx_vitec_domain_model_download' => 'download',
];
private const TEASER_FALLBACK_LENGTH = 250;
#[AsAllowedCallable]
@@ -173,7 +186,7 @@ class SearchJsonRenderer
$results[] = [
'title' => (string)$document->getTitle(),
'url' => (string)$document->getUrl(),
'type' => (string)$document->getType(),
'type' => self::TYPE_LABELS[$document->getType()] ?? (string)$document->getType(),
'teaser' => $teaser,
];
}

View File

@@ -8,3 +8,6 @@ dependencies:
- typo3/fluid-styled-content
- friendsoftypo3/headless
- nb-headless-content-blocks/headless-content-blocks
# solr must load before us so our plugin.tx_solr overrides (statistics,
# highlighting, content extraction) survive - set order follows dependencies
- apache-solr-for-typo3/solr

View File

@@ -336,3 +336,200 @@ plugin.tx_solr.index.queue.pages.fields {
}
}
}
# Log every executed search (keywords, hit count) into tx_solr_statistics -
# feeds the Search Phrase Statistics backend module, most useful for the
# zero-hit list. anonymizeIP masks the last two octets before storing (GDPR).
plugin.tx_solr.statistics = 1
plugin.tx_solr.statistics.anonymizeIP = 2
# =============================================================================
# Record indexing: custom domain models as their own search documents.
# The configuration name becomes the document `type` in the search JSON, so the
# frontend can facet/label results. Every type needs a resolvable `url` -
# a hit without a link is useless, hence the where-clauses below.
# NOT indexed (yet): solutions (no slug, no detail_page - nothing to link to),
# downloads/events/locations (no public detail pages), Datapath news (pid 442,
# import awaiting an editorial decision).
# =============================================================================
plugin.tx_solr.index.queue {
products = 1
products {
type = tx_vitec_domain_model_product
fields {
title = title
content = SOLR_CONTENT
content {
cObject = COA
cObject {
10 = TEXT
10.field = subtitle
10.noTrimWrap = || |
20 = TEXT
20.field = teaser
20.noTrimWrap = || |
30 = TEXT
30.field = description
30.noTrimWrap = || |
40 = TEXT
40.field = description2
40.noTrimWrap = || |
50 = TEXT
50.field = capabilities
50.noTrimWrap = || |
60 = TEXT
60.field = highlights
60.noTrimWrap = || |
70 = TEXT
70.field = applications
70.noTrimWrap = || |
}
}
url = TEXT
url {
typolink {
parameter = 10
additionalParams = &tx_vitec_productshow[product]={field:uid}
additionalParams.insertData = 1
returnLast = url
}
}
}
}
markets = 1
markets {
type = tx_vitec_domain_model_market
# only markets with an editorially assigned detail page are linkable;
# grows automatically as editors fill the field (4 of 34 right now)
additionalWhereClause = detail_page > 0
fields {
title = title
content = SOLR_CONTENT
content {
cObject = COA
cObject {
10 = TEXT
10.field = subtitle
10.noTrimWrap = || |
20 = TEXT
20.field = teaser
20.noTrimWrap = || |
30 = TEXT
30.field = description
30.noTrimWrap = || |
}
}
url = TEXT
url {
typolink {
parameter.field = detail_page
returnLast = url
}
}
}
}
usecases = 1
usecases {
type = tx_vitec_domain_model_usecase
fields {
title = title
content = SOLR_CONTENT
content {
cObject = COA
cObject {
10 = TEXT
10.field = subtitle
10.noTrimWrap = || |
20 = TEXT
20.field = teaser
20.noTrimWrap = || |
}
}
url = TEXT
url {
typolink {
# page 26 carries the vitec_usecaseshow plugin; the route
# enhancer turns this into /success-stories/story/<slug>
parameter = 26
additionalParams = &tx_vitec_usecaseshow[usecase]={field:uid}
additionalParams.insertData = 1
returnLast = url
}
}
}
}
news = 1
news {
type = tx_news_domain_model_news
# type 1 ("page as news") points at pages that are already indexed as
# pages - indexing them again would duplicate every one of them
additionalWhereClause = pid = 11 AND type = '0'
fields {
title = title
content = SOLR_CONTENT
content {
cObject = COA
cObject {
10 = TEXT
10.field = teaser
10.noTrimWrap = || |
20 = TEXT
20.field = bodytext
20.noTrimWrap = || |
}
}
url = TEXT
url {
typolink {
parameter = 21
additionalParams = &tx_news_pi1[controller]=News&tx_news_pi1[action]=detail&tx_news_pi1[news]={field:uid}
additionalParams.insertData = 1
returnLast = url
}
}
}
}
}
# Products are what visitors of a product site are usually looking for -
# without this boost, decades of news articles dominate the first result page
# for product-ish terms like "encoder". Stories get a smaller nudge.
plugin.tx_solr.search.query.boostQuery (
(type:tx_vitec_domain_model_product)^10.0 (type:tx_vitec_domain_model_usecase)^2.0
)
# Downloads have no detail page but a public file endpoint served by
# DownloadFileMiddleware. The where-clause keeps private / website-hidden
# downloads out of the public index (the middleware itself does not enforce
# private_download - the index must not advertise those files).
plugin.tx_solr.index.queue {
downloads = 1
downloads {
type = tx_vitec_domain_model_download
additionalWhereClause = private_download = 0 AND hideonwebsite = 0
fields {
title = title
content = SOLR_CONTENT
content {
cObject = COA
cObject {
10 = TEXT
10.field = teaser
10.noTrimWrap = || |
20 = TEXT
20.field = keywords
20.noTrimWrap = || |
30 = TEXT
30.field = description
30.noTrimWrap = || |
}
}
url = TEXT
url.dataWrap = /download/{field:slug}
}
}
}