This commit is contained in:
Javier Casares 2026-06-03 06:26:23 +00:00
commit 672d0f295f
11 changed files with 438 additions and 73 deletions

View file

@ -1,5 +1,33 @@
== Changelog == == Changelog ==
= 1.3.0 =
_Release date: 2026-05-05_
**Added**
* Per-provider scan tracking table (`mra_provider_status`) with composite primary key `(attachment_id, provider)`. Each API provider now has an independent row tracking its scan status, so adding a new provider only queues the missing scans rather than resetting everything.
* "Sync providers" operation in the Tools page: inserts rows for any provider that has no existing entry for a given attachment, then schedules a background scan batch for the new work only.
**Changed**
* `mra_external_results.provider` column changed from `ENUM('google_vision','tineye')` to `varchar(100)` to allow third-party providers registered via the `mra/external/providers` filter to store results.
* External scan batch processing now reads from `mra_provider_status` and processes each provider independently. Aggregate `external_status` on `mra_media_index` is recomputed after each scan and used only for display.
* `ExternalScanner::get_pending_count()` counts queued rows from `mra_provider_status`.
* DB schema version bumped to `1.1.0`. Migration runs automatically on admin_init when updating from an older version; existing data is backfilled into `mra_provider_status`.
**Compatibility**
* WordPress: 6.8 - 7.0
* PHP: 8.2 - 8.4
* WP-CLI: 2.x
**Tests**
* PHP Coding Standards: WPCS 3.x / PHPCS 3.x
* PHPStan: level 9
* PHPCompatibility: PHP 8.2 - 8.4
= 1.2.0 = = 1.2.0 =
_Release date: 2026-05-02_ _Release date: 2026-05-02_

View file

@ -203,6 +203,10 @@ class ToolsPage {
$count = ExternalScanner::requeue_errors(); $count = ExternalScanner::requeue_errors();
break; break;
case 'sync_providers':
$count = ExternalScanner::sync_providers();
break;
case 'rescan_usage': case 'rescan_usage':
$count = UsageScanner::reset_all(); $count = UsageScanner::reset_all();
UsageScanner::schedule(); UsageScanner::schedule();
@ -358,6 +362,7 @@ class ToolsPage {
'schedule_internal' => __( 'Internal scan scheduled.', 'robotstxt-mediaaudit' ), 'schedule_internal' => __( 'Internal scan scheduled.', 'robotstxt-mediaaudit' ),
'schedule_external' => __( 'External scan scheduled.', 'robotstxt-mediaaudit' ), 'schedule_external' => __( 'External scan scheduled.', 'robotstxt-mediaaudit' ),
'requeue_errors' => __( 'Failed scans re-queued. Background scan scheduled.', 'robotstxt-mediaaudit' ), 'requeue_errors' => __( 'Failed scans re-queued. Background scan scheduled.', 'robotstxt-mediaaudit' ),
'sync_providers' => __( 'Provider registry synced. New scans scheduled.', 'robotstxt-mediaaudit' ),
'rescan_usage' => __( 'Usage data cleared. Background scan scheduled.', 'robotstxt-mediaaudit' ), 'rescan_usage' => __( 'Usage data cleared. Background scan scheduled.', 'robotstxt-mediaaudit' ),
'reindex_all' => __( 'Index cleared. Fresh indexing scheduled.', 'robotstxt-mediaaudit' ), 'reindex_all' => __( 'Index cleared. Fresh indexing scheduled.', 'robotstxt-mediaaudit' ),
'purge_external' => __( 'External scan data purged.', 'robotstxt-mediaaudit' ), 'purge_external' => __( 'External scan data purged.', 'robotstxt-mediaaudit' ),
@ -632,6 +637,13 @@ class ToolsPage {
'class' => 'button', 'class' => 'button',
'confirm' => '', 'confirm' => '',
), ),
array(
'op' => 'sync_providers',
'label' => __( 'Sync provider registry', 'robotstxt-mediaaudit' ),
'desc' => __( 'Creates scan entries for any newly registered external providers. Run this after adding a new provider via the mra/external/providers filter. Already-scanned attachments are not affected.', 'robotstxt-mediaaudit' ),
'class' => 'button',
'confirm' => '',
),
); );
$destructive = array( $destructive = array(

View file

@ -23,6 +23,7 @@ class Database {
public static function create_tables(): void { public static function create_tables(): void {
self::migration_100(); self::migration_100();
self::migration_101(); self::migration_101();
self::migration_110();
} }
/** /**
@ -43,6 +44,10 @@ class Database {
if ( version_compare( $current, '1.0.1', '<' ) ) { if ( version_compare( $current, '1.0.1', '<' ) ) {
self::migration_101(); self::migration_101();
} }
if ( version_compare( $current, '1.1.0', '<' ) ) {
self::migration_110();
}
} }
/** /**
@ -52,6 +57,7 @@ class Database {
*/ */
public static function drop_tables(): void { public static function drop_tables(): void {
global $wpdb; global $wpdb;
$wpdb->query( "DROP TABLE IF EXISTS `{$wpdb->prefix}mra_provider_status`" ); // phpcs:ignore WordPress.DB.DirectDatabaseQuery
$wpdb->query( "DROP TABLE IF EXISTS `{$wpdb->prefix}mra_external_results`" ); // phpcs:ignore WordPress.DB.DirectDatabaseQuery $wpdb->query( "DROP TABLE IF EXISTS `{$wpdb->prefix}mra_external_results`" ); // phpcs:ignore WordPress.DB.DirectDatabaseQuery
$wpdb->query( "DROP TABLE IF EXISTS `{$wpdb->prefix}mra_media_usage`" ); // phpcs:ignore WordPress.DB.DirectDatabaseQuery $wpdb->query( "DROP TABLE IF EXISTS `{$wpdb->prefix}mra_media_usage`" ); // phpcs:ignore WordPress.DB.DirectDatabaseQuery
$wpdb->query( "DROP TABLE IF EXISTS `{$wpdb->prefix}mra_media_index`" ); // phpcs:ignore WordPress.DB.DirectDatabaseQuery $wpdb->query( "DROP TABLE IF EXISTS `{$wpdb->prefix}mra_media_index`" ); // phpcs:ignore WordPress.DB.DirectDatabaseQuery
@ -68,6 +74,7 @@ class Database {
global $wpdb; global $wpdb;
// phpcs:disable WordPress.DB.DirectDatabaseQuery // phpcs:disable WordPress.DB.DirectDatabaseQuery
$wpdb->query( "TRUNCATE TABLE `{$wpdb->prefix}mra_provider_status`" );
$wpdb->query( "TRUNCATE TABLE `{$wpdb->prefix}mra_external_results`" ); $wpdb->query( "TRUNCATE TABLE `{$wpdb->prefix}mra_external_results`" );
$wpdb->query( "TRUNCATE TABLE `{$wpdb->prefix}mra_media_usage`" ); $wpdb->query( "TRUNCATE TABLE `{$wpdb->prefix}mra_media_usage`" );
$wpdb->query( "TRUNCATE TABLE `{$wpdb->prefix}mra_media_index`" ); $wpdb->query( "TRUNCATE TABLE `{$wpdb->prefix}mra_media_index`" );
@ -125,7 +132,7 @@ class Database {
"CREATE TABLE {$wpdb->prefix}mra_external_results ( "CREATE TABLE {$wpdb->prefix}mra_external_results (
result_id bigint(20) unsigned NOT NULL auto_increment, result_id bigint(20) unsigned NOT NULL auto_increment,
attachment_id bigint(20) unsigned NOT NULL, attachment_id bigint(20) unsigned NOT NULL,
provider enum('google_vision','tineye') NOT NULL, provider varchar(100) NOT NULL,
raw_response json DEFAULT NULL, raw_response json DEFAULT NULL,
match_count int(10) unsigned NOT NULL DEFAULT 0, match_count int(10) unsigned NOT NULL DEFAULT 0,
top_domains json DEFAULT NULL, top_domains json DEFAULT NULL,
@ -155,4 +162,57 @@ class Database {
); );
// phpcs:enable WordPress.DB.DirectDatabaseQuery,WordPress.DB.DirectDatabaseQuery.SchemaChange // phpcs:enable WordPress.DB.DirectDatabaseQuery,WordPress.DB.DirectDatabaseQuery.SchemaChange
} }
/**
* Migration 1.1.0 widens provider column and adds mra_provider_status table.
*
* Changes:
* - ALTER mra_external_results.provider from ENUM to varchar(100) for extensibility.
* - CREATE mra_provider_status for per-provider scan tracking.
* - Backfill mra_provider_status from existing mra_external_results data.
*
* Safe to re-run: ALTER TABLE is a no-op if already varchar(100);
* dbDelta skips existing tables; INSERT IGNORE skips duplicate rows.
*
* @return void
*/
private static function migration_110(): void {
global $wpdb;
$charset_collate = $wpdb->get_charset_collate();
// phpcs:disable WordPress.DB.DirectDatabaseQuery,WordPress.DB.DirectDatabaseQuery.SchemaChange
$wpdb->query(
"ALTER TABLE `{$wpdb->prefix}mra_external_results`
MODIFY COLUMN provider varchar(100) NOT NULL"
);
// phpcs:enable WordPress.DB.DirectDatabaseQuery,WordPress.DB.DirectDatabaseQuery.SchemaChange
require_once ABSPATH . 'wp-admin/includes/upgrade.php';
dbDelta(
"CREATE TABLE {$wpdb->prefix}mra_provider_status (
attachment_id bigint(20) unsigned NOT NULL,
provider varchar(100) NOT NULL,
status enum('pending','queued','scanned','error') NOT NULL DEFAULT 'pending',
scanned_at datetime DEFAULT NULL,
PRIMARY KEY (attachment_id,provider),
KEY status (status)
) {$charset_collate};"
);
// Backfill from existing data. Idempotent: INSERT IGNORE skips duplicates.
// phpcs:disable WordPress.DB.DirectDatabaseQuery,WordPress.DB.PreparedSQL.InterpolatedNotPrepared
$wpdb->query(
"INSERT IGNORE INTO {$wpdb->prefix}mra_provider_status (attachment_id, provider, status, scanned_at)
SELECT er.attachment_id, er.provider,
CASE WHEN mi.external_status = 'error' THEN 'error'
WHEN mi.external_status = 'queued' THEN 'queued'
ELSE 'scanned' END,
mi.external_scanned_at
FROM {$wpdb->prefix}mra_external_results er
JOIN {$wpdb->prefix}mra_media_index mi ON mi.attachment_id = er.attachment_id"
);
// phpcs:enable WordPress.DB.DirectDatabaseQuery,WordPress.DB.PreparedSQL.InterpolatedNotPrepared
}
} }

View file

@ -13,6 +13,9 @@ use MediaRightsAudit\Core\Queue\Scheduler;
* Processes the external scan queue: runs each pending attachment through * Processes the external scan queue: runs each pending attachment through
* all active providers and stores results in mra_external_results. * all active providers and stores results in mra_external_results.
* *
* Per-provider scan state is tracked in mra_provider_status. The aggregate
* external_status in mra_media_index is updated after each scan for display.
*
* Active providers are resolved at runtime via the mra/external/providers filter. * Active providers are resolved at runtime via the mra/external/providers filter.
* Scanning is triggered manually by an administrator (Run External Scan bulk action * Scanning is triggered manually by an administrator (Run External Scan bulk action
* or wp mra scan-external CLI) never automatically, to control API costs. * or wp mra scan-external CLI) never automatically, to control API costs.
@ -27,11 +30,14 @@ class ExternalScanner {
const AS_HOOK = 'mra_external_scan_batch'; const AS_HOOK = 'mra_external_scan_batch';
/** /**
* Scans one batch of pending attachments with all active providers. * Scans one batch of queued attachments with their pending providers.
* *
* @param int $batch_size Maximum number of attachments to process. * Queries mra_provider_status for queued entries and processes each
* attachment with the intersection of active providers and queued slugs.
* *
* @return int Number of attachments attempted. * @param int $batch_size Maximum number of attachment rows to process.
*
* @return int Number of attachment rows attempted.
*/ */
public static function scan_batch( int $batch_size = 10 ): int { public static function scan_batch( int $batch_size = 10 ): int {
global $wpdb; global $wpdb;
@ -41,13 +47,20 @@ class ExternalScanner {
return 0; return 0;
} }
// Index providers by slug for fast intersection lookup.
$provider_map = array();
foreach ( $providers as $provider ) {
$provider_map[ $provider->provider_slug() ] = $provider;
}
// phpcs:disable WordPress.DB.DirectDatabaseQuery,WordPress.DB.PreparedSQL.InterpolatedNotPrepared // phpcs:disable WordPress.DB.DirectDatabaseQuery,WordPress.DB.PreparedSQL.InterpolatedNotPrepared
$rows = $wpdb->get_results( $rows = $wpdb->get_results(
$wpdb->prepare( $wpdb->prepare(
"SELECT attachment_id, file_url "SELECT DISTINCT ps.attachment_id, mi.file_url
FROM {$wpdb->prefix}mra_media_index FROM {$wpdb->prefix}mra_provider_status ps
WHERE external_status = 'queued' JOIN {$wpdb->prefix}mra_media_index mi ON mi.attachment_id = ps.attachment_id
ORDER BY attachment_id ASC WHERE ps.status = 'queued'
ORDER BY ps.attachment_id ASC
LIMIT %d", LIMIT %d",
$batch_size $batch_size
), ),
@ -69,14 +82,41 @@ class ExternalScanner {
continue; continue;
} }
self::process_attachment( $aid, $url, $providers ); // Get queued provider slugs for this attachment.
// phpcs:disable WordPress.DB.DirectDatabaseQuery,WordPress.DB.PreparedSQL.InterpolatedNotPrepared
$queued_slugs = $wpdb->get_col(
$wpdb->prepare(
"SELECT provider FROM {$wpdb->prefix}mra_provider_status
WHERE attachment_id = %d AND status = 'queued'",
$aid
)
);
// phpcs:enable WordPress.DB.DirectDatabaseQuery,WordPress.DB.PreparedSQL.InterpolatedNotPrepared
if ( empty( $queued_slugs ) ) {
continue;
}
// Intersect queued slugs with currently active providers.
$batch_providers = array();
foreach ( $queued_slugs as $slug ) {
if ( is_string( $slug ) && isset( $provider_map[ $slug ] ) ) {
$batch_providers[] = $provider_map[ $slug ];
}
}
if ( empty( $batch_providers ) ) {
continue;
}
self::process_attachment( $aid, $url, $batch_providers );
} }
return count( $rows ); return count( $rows );
} }
/** /**
* Returns the count of attachments still waiting for external scanning. * Returns the count of provider-attachment rows still waiting for external scanning.
* *
* @return int * @return int
*/ */
@ -84,7 +124,7 @@ class ExternalScanner {
global $wpdb; global $wpdb;
$result = $wpdb->get_var( // phpcs:ignore WordPress.DB.DirectDatabaseQuery $result = $wpdb->get_var( // phpcs:ignore WordPress.DB.DirectDatabaseQuery
"SELECT COUNT(*) FROM {$wpdb->prefix}mra_media_index WHERE external_status = 'queued'" "SELECT COUNT(*) FROM {$wpdb->prefix}mra_provider_status WHERE status = 'queued'"
); );
return is_numeric( $result ) ? (int) $result : 0; return is_numeric( $result ) ? (int) $result : 0;
@ -106,40 +146,22 @@ class ExternalScanner {
} }
/** /**
* Promotes all attachments with external_status = 'pending' to 'queued'. * Marks all pending provider rows as queued and schedules a background scan batch.
* *
* Does not schedule an Action Scheduler action. Useful in CLI context where * Calls queue_provider_attachments() for each active provider to populate
* processing runs synchronously in a loop rather than via background jobs. * mra_provider_status, then syncs the aggregate status on mra_media_index.
* * Only schedules the Action Scheduler action if there are queued items and
* @return void * no batch is already running.
*/
public static function queue_all_pending(): void {
global $wpdb;
$wpdb->query( // phpcs:ignore WordPress.DB.DirectDatabaseQuery
"UPDATE {$wpdb->prefix}mra_media_index
SET external_status = 'queued'
WHERE external_status = 'pending'"
);
}
/**
* Marks all pending attachments as queued and schedules a background scan batch.
*
* Moves every attachment with external_status = 'pending' to 'queued' so the
* status change is visible immediately in the admin list. Only schedules the
* Action Scheduler action if there are now queued items and no batch is already running.
* *
* @return void * @return void
*/ */
public static function schedule(): void { public static function schedule(): void {
global $wpdb; $providers = self::get_active_providers();
foreach ( $providers as $provider ) {
self::queue_provider_attachments( $provider->provider_slug() );
}
$wpdb->query( // phpcs:ignore WordPress.DB.DirectDatabaseQuery self::sync_aggregate_queued_status();
"UPDATE {$wpdb->prefix}mra_media_index
SET external_status = 'queued'
WHERE external_status = 'pending'"
);
if ( self::get_pending_count() > 0 && ! Scheduler::has_pending( self::AS_HOOK ) ) { if ( self::get_pending_count() > 0 && ! Scheduler::has_pending( self::AS_HOOK ) ) {
Scheduler::schedule_single( self::AS_HOOK ); Scheduler::schedule_single( self::AS_HOOK );
@ -147,10 +169,75 @@ class ExternalScanner {
} }
/** /**
* Marks specific attachments as pending and schedules a scan batch. * Promotes all pending provider rows to queued for all active providers.
* *
* Idempotent: already-pending attachments are left unchanged; * Does not schedule an Action Scheduler action. Useful in CLI context where
* previously-scanned or error attachments are reset to pending. * processing runs synchronously in a loop rather than via background jobs.
*
* @return void
*/
public static function queue_all_pending(): void {
$providers = self::get_active_providers();
foreach ( $providers as $provider ) {
self::queue_provider_attachments( $provider->provider_slug() );
}
self::sync_aggregate_queued_status();
}
/**
* Creates scan entries for newly registered providers.
*
* For each active provider, inserts mra_provider_status rows for attachments
* that have no existing row for that provider. Already-scanned or error rows
* are not touched. If new rows are created, syncs aggregate status and
* schedules a background scan batch.
*
* @return int Total number of new rows inserted.
*/
public static function sync_providers(): int {
global $wpdb;
$providers = self::get_active_providers();
$total = 0;
foreach ( $providers as $provider ) {
$slug = $provider->provider_slug();
// phpcs:disable WordPress.DB.DirectDatabaseQuery,WordPress.DB.PreparedSQL.InterpolatedNotPrepared
$wpdb->query(
$wpdb->prepare(
"INSERT IGNORE INTO {$wpdb->prefix}mra_provider_status (attachment_id, provider, status)
SELECT mi.attachment_id, %s, 'queued'
FROM {$wpdb->prefix}mra_media_index mi
LEFT JOIN {$wpdb->prefix}mra_provider_status ps
ON ps.attachment_id = mi.attachment_id AND ps.provider = %s
WHERE ps.attachment_id IS NULL",
$slug,
$slug
)
);
// phpcs:enable WordPress.DB.DirectDatabaseQuery,WordPress.DB.PreparedSQL.InterpolatedNotPrepared
$total += (int) $wpdb->rows_affected;
}
if ( $total > 0 ) {
self::sync_aggregate_queued_status();
if ( ! Scheduler::has_pending( self::AS_HOOK ) ) {
Scheduler::schedule_single( self::AS_HOOK );
}
}
return $total;
}
/**
* Marks specific attachments as queued for all active providers and schedules a scan batch.
*
* Idempotent: already-queued or scanned rows are left unchanged;
* error rows are reset to queued.
* *
* @param array<int, int> $attachment_ids Attachment IDs to queue. Must be non-empty. * @param array<int, int> $attachment_ids Attachment IDs to queue. Must be non-empty.
* *
@ -165,49 +252,58 @@ class ExternalScanner {
$ids = array_map( 'intval', $attachment_ids ); $ids = array_map( 'intval', $attachment_ids );
$placeholders = implode( ',', array_fill( 0, count( $ids ), '%d' ) ); $placeholders = implode( ',', array_fill( 0, count( $ids ), '%d' ) );
$providers = self::get_active_providers();
foreach ( $providers as $provider ) {
$slug = $provider->provider_slug();
// phpcs:disable WordPress.DB.DirectDatabaseQuery,WordPress.DB.PreparedSQL.InterpolatedNotPrepared,WordPress.DB.PreparedSQLPlaceholders.UnfinishedPrepare // phpcs:disable WordPress.DB.DirectDatabaseQuery,WordPress.DB.PreparedSQL.InterpolatedNotPrepared,WordPress.DB.PreparedSQLPlaceholders.UnfinishedPrepare
$wpdb->query( $wpdb->query(
$wpdb->prepare( $wpdb->prepare(
"UPDATE {$wpdb->prefix}mra_media_index "INSERT INTO {$wpdb->prefix}mra_provider_status (attachment_id, provider, status)
SET external_status = 'queued', external_scanned_at = NULL SELECT attachment_id, %s, 'queued'
FROM {$wpdb->prefix}mra_media_index
WHERE attachment_id IN ({$placeholders}) WHERE attachment_id IN ({$placeholders})
AND external_status != 'queued'", ON DUPLICATE KEY UPDATE status = IF(status NOT IN ('scanned','queued'),'queued',status)",
...$ids array_merge( array( $slug ), $ids )
) )
); );
// phpcs:enable WordPress.DB.DirectDatabaseQuery,WordPress.DB.PreparedSQL.InterpolatedNotPrepared,WordPress.DB.PreparedSQLPlaceholders.UnfinishedPrepare // phpcs:enable WordPress.DB.DirectDatabaseQuery,WordPress.DB.PreparedSQL.InterpolatedNotPrepared,WordPress.DB.PreparedSQLPlaceholders.UnfinishedPrepare
}
self::sync_aggregate_queued_status();
// Schedule directly — do not call schedule() which would also mark all 'pending' items.
if ( ! Scheduler::has_pending( self::AS_HOOK ) ) { if ( ! Scheduler::has_pending( self::AS_HOOK ) ) {
Scheduler::schedule_single( self::AS_HOOK ); Scheduler::schedule_single( self::AS_HOOK );
} }
} }
/** /**
* Re-queues all attachments that previously failed with external_status = 'error'. * Re-queues all provider rows that previously failed with status = 'error'.
* *
* Does not touch pending, queued, scanned, or matched attachments. * Does not touch pending, queued, or scanned rows.
* Schedules a background scan batch if any items were re-queued and * Schedules a background scan batch if any items were re-queued and
* no batch is already pending. * no batch is already pending.
* *
* @return int Number of attachments re-queued. * @return int Number of provider rows re-queued.
*/ */
public static function requeue_errors(): int { public static function requeue_errors(): int {
global $wpdb; global $wpdb;
// phpcs:disable WordPress.DB.DirectDatabaseQuery // phpcs:disable WordPress.DB.DirectDatabaseQuery
$count = (int) $wpdb->get_var( $count = (int) $wpdb->get_var(
"SELECT COUNT(*) FROM {$wpdb->prefix}mra_media_index WHERE external_status = 'error'" "SELECT COUNT(*) FROM {$wpdb->prefix}mra_provider_status WHERE status = 'error'"
); );
if ( $count > 0 ) { if ( $count > 0 ) {
$wpdb->query( $wpdb->query(
"UPDATE {$wpdb->prefix}mra_media_index "UPDATE {$wpdb->prefix}mra_provider_status
SET external_status = 'queued', external_scanned_at = NULL SET status = 'queued', scanned_at = NULL
WHERE external_status = 'error'" WHERE status = 'error'"
); );
self::sync_aggregate_queued_status();
if ( ! Scheduler::has_pending( self::AS_HOOK ) ) { if ( ! Scheduler::has_pending( self::AS_HOOK ) ) {
Scheduler::schedule_single( self::AS_HOOK ); Scheduler::schedule_single( self::AS_HOOK );
} }
@ -243,6 +339,14 @@ class ExternalScanner {
) )
); );
$wpdb->query(
$wpdb->prepare(
"DELETE FROM {$wpdb->prefix}mra_provider_status
WHERE attachment_id IN ({$placeholders})",
...$ids
)
);
$wpdb->query( $wpdb->query(
$wpdb->prepare( $wpdb->prepare(
"UPDATE {$wpdb->prefix}mra_media_index "UPDATE {$wpdb->prefix}mra_media_index
@ -257,7 +361,7 @@ class ExternalScanner {
/** /**
* Deletes all external scan results and resets every attachment's external_status to pending. * Deletes all external scan results and resets every attachment's external_status to pending.
* *
* @return int Number of result rows deleted. * @return int Number of result rows deleted from mra_external_results.
*/ */
public static function purge_all(): int { public static function purge_all(): int {
global $wpdb; global $wpdb;
@ -265,6 +369,7 @@ class ExternalScanner {
// phpcs:disable WordPress.DB.DirectDatabaseQuery // phpcs:disable WordPress.DB.DirectDatabaseQuery
$count = (int) $wpdb->get_var( "SELECT COUNT(*) FROM {$wpdb->prefix}mra_external_results" ); $count = (int) $wpdb->get_var( "SELECT COUNT(*) FROM {$wpdb->prefix}mra_external_results" );
$wpdb->query( "DELETE FROM {$wpdb->prefix}mra_external_results" ); $wpdb->query( "DELETE FROM {$wpdb->prefix}mra_external_results" );
$wpdb->query( "DELETE FROM {$wpdb->prefix}mra_provider_status" );
$wpdb->query( $wpdb->query(
"UPDATE {$wpdb->prefix}mra_media_index "UPDATE {$wpdb->prefix}mra_media_index
SET external_status = 'pending', external_scanned_at = NULL" SET external_status = 'pending', external_scanned_at = NULL"
@ -277,6 +382,8 @@ class ExternalScanner {
/** /**
* Returns a count of attachments grouped by external_status. * Returns a count of attachments grouped by external_status.
* *
* Reads from the mra_media_index aggregate column for display purposes.
*
* @return array<string, int> * @return array<string, int>
*/ */
public static function get_status_counts(): array { public static function get_status_counts(): array {
@ -315,35 +422,55 @@ class ExternalScanner {
// ------------------------------------------------------------------------- // -------------------------------------------------------------------------
/** /**
* Scans a single attachment with all providers and updates the index. * Scans a single attachment with specified providers and updates per-provider status.
*
* After scanning all providers, calls update_aggregate_status() to sync
* the mra_media_index aggregate column.
* *
* @param int $attachment_id WordPress attachment ID. * @param int $attachment_id WordPress attachment ID.
* @param string $file_url Public URL of the image. * @param string $file_url Public URL of the image.
* @param array<AbstractProvider> $providers Resolved provider list. * @param array<AbstractProvider> $providers Resolved provider list for this attachment.
* *
* @return void * @return void
*/ */
private static function process_attachment( int $attachment_id, string $file_url, array $providers ): void { private static function process_attachment( int $attachment_id, string $file_url, array $providers ): void {
global $wpdb; global $wpdb;
$any_success = false;
$any_match = false;
foreach ( $providers as $provider ) { foreach ( $providers as $provider ) {
$slug = $provider->provider_slug();
try { try {
$result = $provider->scan( $attachment_id, $file_url ); $result = $provider->scan( $attachment_id, $file_url );
self::save_result( $attachment_id, $provider->provider_slug(), $result ); self::save_result( $attachment_id, $slug, $result );
$any_success = true;
if ( $result->match_count() > 0 ) { $wpdb->query( // phpcs:ignore WordPress.DB.DirectDatabaseQuery
$any_match = true; $wpdb->prepare(
} // phpcs:ignore WordPress.DB.PreparedSQL.InterpolatedNotPrepared
"UPDATE {$wpdb->prefix}mra_provider_status
SET status = 'scanned', scanned_at = NOW()
WHERE attachment_id = %d AND provider = %s",
$attachment_id,
$slug
)
);
} catch ( \RuntimeException $e ) { } catch ( \RuntimeException $e ) {
$wpdb->query( // phpcs:ignore WordPress.DB.DirectDatabaseQuery
$wpdb->prepare(
// phpcs:ignore WordPress.DB.PreparedSQL.InterpolatedNotPrepared
"UPDATE {$wpdb->prefix}mra_provider_status
SET status = 'error', scanned_at = NOW()
WHERE attachment_id = %d AND provider = %s",
$attachment_id,
$slug
)
);
if ( defined( 'WP_DEBUG_LOG' ) && WP_DEBUG_LOG ) { if ( defined( 'WP_DEBUG_LOG' ) && WP_DEBUG_LOG ) {
// phpcs:ignore WordPress.PHP.DevelopmentFunctions.error_log_error_log // phpcs:ignore WordPress.PHP.DevelopmentFunctions.error_log_error_log
error_log( error_log(
sprintf( sprintf(
'[MRA] %s scan failed for attachment %d: %s', '[MRA] %s scan failed for attachment %d: %s',
$provider->provider_slug(), $slug,
$attachment_id, $attachment_id,
$e->getMessage() $e->getMessage()
) )
@ -352,14 +479,69 @@ class ExternalScanner {
} }
} }
if ( ! $any_success ) { self::update_aggregate_status( $attachment_id );
$new_status = 'error';
} elseif ( $any_match ) {
$new_status = 'matches';
} else {
$new_status = 'scanned';
} }
/**
* Recomputes and persists the aggregate external_status for a single attachment.
*
* Logic:
* - Any provider still queued 'queued'
* - No provider succeeded 'error'
* - At least one match found 'matches'
* - Otherwise 'scanned'
*
* @param int $attachment_id WordPress attachment ID.
*
* @return void
*/
private static function update_aggregate_status( int $attachment_id ): void {
global $wpdb;
// phpcs:disable WordPress.DB.DirectDatabaseQuery,WordPress.DB.PreparedSQL.InterpolatedNotPrepared
$row = $wpdb->get_row(
$wpdb->prepare(
"SELECT
SUM(status = 'queued') AS queued_count,
SUM(status = 'error') AS error_count,
SUM(status = 'scanned') AS scanned_count
FROM {$wpdb->prefix}mra_provider_status
WHERE attachment_id = %d",
$attachment_id
),
ARRAY_A
);
// phpcs:enable WordPress.DB.DirectDatabaseQuery,WordPress.DB.PreparedSQL.InterpolatedNotPrepared
if ( ! is_array( $row ) ) {
return;
}
$queued_count = isset( $row['queued_count'] ) && is_numeric( $row['queued_count'] ) ? (int) $row['queued_count'] : 0;
$error_count = isset( $row['error_count'] ) && is_numeric( $row['error_count'] ) ? (int) $row['error_count'] : 0;
$scanned_count = isset( $row['scanned_count'] ) && is_numeric( $row['scanned_count'] ) ? (int) $row['scanned_count'] : 0;
if ( $queued_count > 0 ) {
$new_status = 'queued';
} elseif ( 0 === $scanned_count ) {
$new_status = 'error';
} else {
// Check for matches across scanned providers.
// phpcs:disable WordPress.DB.DirectDatabaseQuery,WordPress.DB.PreparedSQL.InterpolatedNotPrepared
$match_sum = $wpdb->get_var(
$wpdb->prepare(
"SELECT SUM(match_count) FROM {$wpdb->prefix}mra_external_results WHERE attachment_id = %d",
$attachment_id
)
);
// phpcs:enable WordPress.DB.DirectDatabaseQuery,WordPress.DB.PreparedSQL.InterpolatedNotPrepared
$new_status = ( is_numeric( $match_sum ) && (int) $match_sum > 0 ) ? 'matches' : 'scanned';
}
// Suppress unused variable inspection: $error_count is available for future use.
unset( $error_count );
$wpdb->update( // phpcs:ignore WordPress.DB.DirectDatabaseQuery $wpdb->update( // phpcs:ignore WordPress.DB.DirectDatabaseQuery
$wpdb->prefix . 'mra_media_index', $wpdb->prefix . 'mra_media_index',
array( array(
@ -372,6 +554,57 @@ class ExternalScanner {
); );
} }
/**
* Updates mra_media_index.external_status to 'queued' for all attachments
* that have at least one queued row in mra_provider_status.
*
* Only updates rows not already marked 'queued', to avoid unnecessary writes.
*
* @return void
*/
private static function sync_aggregate_queued_status(): void {
global $wpdb;
$wpdb->query( // phpcs:ignore WordPress.DB.DirectDatabaseQuery
// phpcs:ignore WordPress.DB.PreparedSQL.InterpolatedNotPrepared
"UPDATE {$wpdb->prefix}mra_media_index mi
INNER JOIN (
SELECT DISTINCT attachment_id
FROM {$wpdb->prefix}mra_provider_status
WHERE status = 'queued'
) AS q ON q.attachment_id = mi.attachment_id
SET mi.external_status = 'queued'
WHERE mi.external_status != 'queued'"
);
}
/**
* Inserts or updates mra_provider_status rows for all indexed attachments.
*
* For attachments without an existing row for this provider, inserts 'queued'.
* For attachments with an existing non-scanned, non-queued row, resets to 'queued'.
* Scanned and already-queued rows are left unchanged.
*
* @param string $provider_slug Provider machine-readable slug.
*
* @return void
*/
private static function queue_provider_attachments( string $provider_slug ): void {
global $wpdb;
// phpcs:disable WordPress.DB.DirectDatabaseQuery,WordPress.DB.PreparedSQL.InterpolatedNotPrepared
$wpdb->query(
$wpdb->prepare(
"INSERT INTO {$wpdb->prefix}mra_provider_status (attachment_id, provider, status)
SELECT attachment_id, %s, 'queued'
FROM {$wpdb->prefix}mra_media_index
ON DUPLICATE KEY UPDATE status = IF(status NOT IN ('scanned','queued'),'queued',status)",
$provider_slug
)
);
// phpcs:enable WordPress.DB.DirectDatabaseQuery,WordPress.DB.PreparedSQL.InterpolatedNotPrepared
}
/** /**
* Persists a provider's scan result (upsert via REPLACE INTO). * Persists a provider's scan result (upsert via REPLACE INTO).
* *

Binary file not shown.

View file

@ -973,6 +973,18 @@ msgstr "Última exploració (externa)"
msgid "Scanned At" msgid "Scanned At"
msgstr "Explorat el" msgstr "Explorat el"
#: includes/Admin/ToolsPage.php:365
msgid "Provider registry synced. New scans scheduled."
msgstr "Registre de proveïdors sincronitzat. Nous escanejos programats."
#: includes/Admin/ToolsPage.php:642
msgid "Sync provider registry"
msgstr "Sincronitza el registre de proveïdors"
#: includes/Admin/ToolsPage.php:643
msgid "Creates scan entries for any newly registered external providers. Run this after adding a new provider via the mra/external/providers filter. Already-scanned attachments are not affected."
msgstr "Crea entrades d'escaneig per als proveïdors externs registrats recentment. Executa-ho després d'afegir un nou proveïdor mitjançant el filtre mra/external/providers. Els fitxers ja escanejats no es veuen afectats."
#~ msgid "https://robotstxt.es/plugins/media-audit/" #~ msgid "https://robotstxt.es/plugins/media-audit/"
#~ msgstr "https://robotstxt.es/plugins/media-audit/" #~ msgstr "https://robotstxt.es/plugins/media-audit/"

View file

@ -969,6 +969,18 @@ msgstr "Último escaneo (externo)"
msgid "Scanned At" msgid "Scanned At"
msgstr "Escaneado el" msgstr "Escaneado el"
#: includes/Admin/ToolsPage.php:365
msgid "Provider registry synced. New scans scheduled."
msgstr "Registro de proveedores sincronizado. Nuevos escaneos programados."
#: includes/Admin/ToolsPage.php:642
msgid "Sync provider registry"
msgstr "Sincronizar registro de proveedores"
#: includes/Admin/ToolsPage.php:643
msgid "Creates scan entries for any newly registered external providers. Run this after adding a new provider via the mra/external/providers filter. Already-scanned attachments are not affected."
msgstr "Crea entradas de escaneo para los proveedores externos recién registrados. Ejecútalo después de añadir un nuevo proveedor mediante el filtro mra/external/providers. Los archivos ya escaneados no se ven afectados."
#~ msgid "https://robotstxt.es/plugins/media-audit/" #~ msgid "https://robotstxt.es/plugins/media-audit/"
#~ msgstr "https://robotstxt.es/plugins/media-audit/" #~ msgstr "https://robotstxt.es/plugins/media-audit/"

View file

@ -5,7 +5,7 @@ Requires at least: 6.8
Tested up to: 7.0 Tested up to: 7.0
Requires PHP: 8.2 Requires PHP: 8.2
Requires Plugins: action-scheduler Requires Plugins: action-scheduler
Stable tag: 1.2.0 Stable tag: 1.3.0
License: GPL-3.0-or-later License: GPL-3.0-or-later
License URI: https://www.gnu.org/licenses/gpl-3.0.txt License URI: https://www.gnu.org/licenses/gpl-3.0.txt
@ -98,6 +98,13 @@ By default, no. Enable **Settings → Delete data on uninstall** if you want all
Only the 3 latest versions. The full changelog is in [changelog.txt](changelog.txt). Only the 3 latest versions. The full changelog is in [changelog.txt](changelog.txt).
= 1.3.0 =
* Added per-provider scan tracking table (`mra_provider_status`) so each API is tracked independently.
* Added "Sync providers" operation: registers newly configured providers for all existing attachments without rescanning already-completed work.
* Changed `mra_external_results.provider` from a fixed ENUM to `varchar(100)` for extensibility.
* DB schema bumped to 1.1.0 with an automatic migration on update.
= 1.2.0 = = 1.2.0 =
* Added browser-based AJAX scan runner in the Tools page (runs without WP-Cron or Action Scheduler). * Added browser-based AJAX scan runner in the Tools page (runs without WP-Cron or Action Scheduler).

View file

@ -3,7 +3,7 @@
* Plugin Name: Media Audit (by ROBOTSTXT) * Plugin Name: Media Audit (by ROBOTSTXT)
* Plugin URI: https://git.robotstxt.es/ROBOTSTXT/robotstxt-mediaaudit * Plugin URI: https://git.robotstxt.es/ROBOTSTXT/robotstxt-mediaaudit
* Description: Internal media library usage auditing and external reverse image search to detect potential copyright issues. * Description: Internal media library usage auditing and external reverse image search to detect potential copyright issues.
* Version: 1.2.0 * Version: 1.3.0
* Requires at least: 6.8 * Requires at least: 6.8
* Tested up to: 7.0 * Tested up to: 7.0
* Requires PHP: 8.2 * Requires PHP: 8.2
@ -23,8 +23,8 @@ if ( ! defined( 'ABSPATH' ) ) {
exit; exit;
} }
define( 'ROBOTSTXT_MEDIAAUDIT_VERSION', '1.2.0' ); define( 'ROBOTSTXT_MEDIAAUDIT_VERSION', '1.3.0' );
define( 'ROBOTSTXT_MEDIAAUDIT_DB_VERSION', '1.0.1' ); define( 'ROBOTSTXT_MEDIAAUDIT_DB_VERSION', '1.1.0' );
define( 'ROBOTSTXT_MEDIAAUDIT_PLUGIN_FILE', __FILE__ ); define( 'ROBOTSTXT_MEDIAAUDIT_PLUGIN_FILE', __FILE__ );
define( 'ROBOTSTXT_MEDIAAUDIT_PLUGIN_DIR', plugin_dir_path( __FILE__ ) ); define( 'ROBOTSTXT_MEDIAAUDIT_PLUGIN_DIR', plugin_dir_path( __FILE__ ) );
define( 'ROBOTSTXT_MEDIAAUDIT_PLUGIN_URL', plugin_dir_url( __FILE__ ) ); define( 'ROBOTSTXT_MEDIAAUDIT_PLUGIN_URL', plugin_dir_url( __FILE__ ) );

View file

@ -23,6 +23,7 @@ if ( ! $delete ) {
global $wpdb; global $wpdb;
// Drop tables in reverse dependency order. // Drop tables in reverse dependency order.
$wpdb->query( "DROP TABLE IF EXISTS `{$wpdb->prefix}mra_provider_status`" ); // phpcs:ignore WordPress.DB.DirectDatabaseQuery
$wpdb->query( "DROP TABLE IF EXISTS `{$wpdb->prefix}mra_external_results`" ); // phpcs:ignore WordPress.DB.DirectDatabaseQuery $wpdb->query( "DROP TABLE IF EXISTS `{$wpdb->prefix}mra_external_results`" ); // phpcs:ignore WordPress.DB.DirectDatabaseQuery
$wpdb->query( "DROP TABLE IF EXISTS `{$wpdb->prefix}mra_media_usage`" ); // phpcs:ignore WordPress.DB.DirectDatabaseQuery $wpdb->query( "DROP TABLE IF EXISTS `{$wpdb->prefix}mra_media_usage`" ); // phpcs:ignore WordPress.DB.DirectDatabaseQuery
$wpdb->query( "DROP TABLE IF EXISTS `{$wpdb->prefix}mra_media_index`" ); // phpcs:ignore WordPress.DB.DirectDatabaseQuery $wpdb->query( "DROP TABLE IF EXISTS `{$wpdb->prefix}mra_media_index`" ); // phpcs:ignore WordPress.DB.DirectDatabaseQuery