feat: Enhance Support Reply Service with tone instructions and article details

- Added tone instruction retrieval to SupportReplyService.
- Improved user feedback when no relevant article is found.
- Included article URL and tone instruction in LLM prompt.
- Updated response format to include source information.
- Enhanced article management UI with search functionality and editing capabilities.
- Introduced a new API endpoint for nearest articles based on vector search.
- Added confidence badge component to display article confidence levels.
- Implemented tests for article searching, editing, and nearest article API.
- Removed obsolete .htaccess file.

app/Http/Controllers/Api/NearestArticleController.php

@@ -0,0 +1,81 @@
+<?php
+
+namespace App\Http\Controllers\Api;
+
+use App\Exceptions\OllamaUnavailableException;
+use App\Http\Controllers\Controller;
+use App\Repositories\Contracts\ArticleRepositoryInterface;
+use App\Services\EmbeddingService;
+use Illuminate\Http\JsonResponse;
+use Illuminate\Http\Request;
+
+class NearestArticleController extends Controller
+{
+    public function __invoke(
+        Request $request,
+        EmbeddingService $embeddingService,
+        ArticleRepositoryInterface $articleRepository
+    ): JsonResponse {
+        $validated = $request->validate([
+            'query' => ['required', 'string', 'min:2', 'max:1000'],
+            'limit' => ['sometimes', 'integer', 'min:1', 'max:20'],
+            'min_similarity' => ['sometimes', 'numeric', 'min:0', 'max:1'],
+            'include_content' => ['sometimes', 'boolean'],
+        ]);
+
+        $query = trim($validated['query']);
+        $limit = (int) ($validated['limit'] ?? 5);
+        $minSimilarity = (float) ($validated['min_similarity'] ?? 0);
+        $includeContent = $request->boolean('include_content', false);
+
+        try {
+            $embedding = $embeddingService->embed($query);
+        } catch (OllamaUnavailableException $exception) {
+            return response()->json([
+                'message' => 'Embedding provider is unavailable.',
+                'error' => $exception->getMessage(),
+            ], 503);
+        }
+
+        $embeddingContext = $embeddingService->context();
+        $candidates = $articleRepository->findSimilarByEmbedding(
+            embedding: $embedding,
+            limit: $limit,
+            embeddingContext: $embeddingContext,
+            filters: ['published_only' => true]
+        );
+
+        $results = collect($candidates)
+            ->map(function ($candidate) use ($includeContent) {
+                $similarity = max(0, min(1, 1 - $candidate->distance));
+                $content = trim($candidate->content);
+
+                return [
+                    'article_id' => $candidate->articleId,
+                    'title' => $candidate->title,
+                    'similarity' => round($similarity, 4),
+                    'distance' => round($candidate->distance, 4),
+                    'snippet' => str($content)->limit(220)->toString(),
+                    'content' => $includeContent ? $content : null,
+                    'source_url' => $candidate->sourceUrl,
+                    'source_article_id' => $candidate->sourceArticleId,
+                    'note' => $candidate->note,
+                    'allowed_actions' => $candidate->allowedActions,
+                ];
+            })
+            ->filter(fn (array $result) => $result['similarity'] >= $minSimilarity)
+            ->values();
+
+        return response()->json([
+            'data' => $results,
+            'meta' => [
+                'query' => $query,
+                'limit' => $limit,
+                'min_similarity' => $minSimilarity,
+                'published_only' => true,
+                'embedding_provider_instance_id' => $embeddingContext['provider_instance_id'] ?? null,
+                'embedding_model' => $embeddingContext['embedding_model'] ?? null,
+            ],
+        ]);
+    }
+}