Fix duplicate detection: lower default threshold and log near-misses

src/Admin/Admin.php

@@ -914,7 +914,7 @@ public function handle_save_settings(): void {
 				'ai_categorize'          => isset( $_POST['ai_categorize'] ),
 				'hub_review'             => isset( $_POST['hub_review'] ),
 				'strategic_mode'         => isset( $_POST['strategic_mode'] ),
-				'similarity_threshold'   => min( 1, max( 0, (float) ( $_POST['similarity_threshold'] ?? 0.82 ) ) ),
+				'similarity_threshold'   => min( 1, max( 0, (float) ( $_POST['similarity_threshold'] ?? 0.72 ) ) ),
 				'cluster_window_days'    => max( 1, (int) ( $_POST['cluster_window_days'] ?? 7 ) ),
 				'import_interval_minutes' => max( 1, (int) ( $_POST['import_interval_minutes'] ?? 30 ) ),
 				'processing_enabled'     => isset( $_POST['processing_enabled'] ),

src/Admin/views/settings.php

@@ -175,7 +175,8 @@
 		<table class="form-table" role="presentation">
 			<tr>
 				<th><label for="similarity_threshold"><?php esc_html_e( 'Duplicate sensitivity', 'aggregate-it' ); ?></label></th>
-				<td><input name="similarity_threshold" id="similarity_threshold" type="number" step="0.01" min="0" max="1" class="small-text" value="<?php echo esc_attr( (string) $settings->similarity_threshold() ); ?>"></td>
+				<td><input name="similarity_threshold" id="similarity_threshold" type="number" step="0.01" min="0" max="1" class="small-text" value="<?php echo esc_attr( (string) $settings->similarity_threshold() ); ?>">
+					<p class="description"><?php esc_html_e( 'How similar two articles must be to merge into one story (0-1). Lower merges more aggressively; raise it if unrelated stories are being combined. Default 0.72.', 'aggregate-it' ); ?></p></td>
 			</tr>
 			<tr>
 				<th><label for="cluster_window_days"><?php esc_html_e( 'Keep updating a story for (days)', 'aggregate-it' ); ?></label></th>

src/Cluster/Clusterer.php

@@ -41,6 +41,14 @@ public function match( array $vector, string $content ): ?int {
 
 		foreach ( $candidates as $candidate ) {
 			if ( $candidate['score'] < $threshold ) {
+				EventLog::info(
+					sprintf(
+						'Nearest existing story #%d scored %.3f, under the %.2f duplicate threshold, so this published as a new story. Lower "Duplicate sensitivity" to merge stories like this.',
+						$candidate['owner_id'],
+						$candidate['score'],
+						$threshold
+					)
+				);
 				return null; // sorted desc — nothing below the top will clear it either
 			}
 

src/Settings.php

@@ -117,7 +117,7 @@ public function strategic_mode(): bool {
 	}
 
 	public function similarity_threshold(): float {
-		return (float) $this->get( 'similarity_threshold', 0.82 );
+		return (float) $this->get( 'similarity_threshold', 0.72 );
 	}
 
 	public function cluster_window_days(): int {