Merge pull request #51 from wahyd4/feat/vector-search-fast-track

feat: rapid vector search UI

links/search_urls.py

@@ -1,9 +1,10 @@
 from django.urls import path
-from .search_views import SearchReactView, search, search_aliases, search_api_v2
+from .search_views import SearchReactView, search, search_aliases, search_api_v2, search_vector_api
 
 urlpatterns = [
     path('', SearchReactView.as_view(), name='search_main'),
     path('api/', search, name='search_api'),
     path('api/v2/', search_api_v2, name='search_api_v2'),
+    path('api/vector/', search_vector_api, name='search_vector_api'),
     path('aliases/', search_aliases, name='search_aliases'),
 ]

links/search_views.py

@@ -11,9 +11,63 @@ from .search_backend import search_backend
 from django.views.decorators.http import require_http_methods
 from django.views.decorators.csrf import csrf_exempt
 import json
+import requests
+import os
+from qdrant_client import QdrantClient
 
 logger = logging.getLogger(__name__)
 
+def get_ollama_embedding(text):
+    ollama_url = os.getenv("OLLAMA_URL", "http://ollama-service.ollama.svc.cluster.local:11434")
+    model = os.getenv("OLLAMA_MODEL", "llama3")
+    try:
+        r = requests.post(f"{ollama_url}/api/embeddings", json={"model": model, "prompt": text}, timeout=5)
+        r.raise_for_status()
+        return r.json()["embedding"]
+    except Exception as e:
+        logger.error(f"Ollama embedding error: {e}")
+        return None
+
+@require_http_methods(["GET"])
+def search_vector_api(request):
+    query = request.GET.get('q', '').strip()
+    if not query:
+        return JsonResponse({'results': [], 'total': 0})
+
+    qdrant_host = os.getenv("QDRANT_HOST", "192.168.1.2")
+    qdrant_port = int(os.getenv("QDRANT_PORT", "6333"))
+    collection = os.getenv("COLLECTION_NAME", "links")
+
+    try:
+        vector = get_ollama_embedding(query)
+        if not vector:
+            return JsonResponse({'error': 'Failed to generate embedding'}, status=500)
+
+        client = QdrantClient(host=qdrant_host, port=qdrant_port)
+        search_result = client.search(
+            collection_name=collection,
+            query_vector=vector,
+            limit=20
+        )
+
+        results = []
+        for hit in search_result:
+            p = hit.payload
+            results.append({
+                'id': p.get('id'),
+                'type': 'post' if 'content' in p else 'link',
+                'title': p.get('title'),
+                'url': p.get('url'),
+                'summary': p.get('description') or p.get('summary', ''),
+                'score': hit.score,
+                'tags': p.get('tags', [])
+            })
+        
+        return JsonResponse({'results': results, 'total': len(results)})
+    except Exception as e:
+        logger.error(f"Vector search error: {e}")
+        return JsonResponse({'error': str(e)}, status=500)
+
 class SearchView(ListView):
     template_name = 'links/search.html'
     context_object_name = 'results'
@@ -32,7 +86,6 @@ class SearchView(ListView):
 
         if query:
             try:
-                # Use Whoosh search backend for full-text search
                 search_results = search_backend.search(
                     query_string=query,
                     model_type=type_filter if type_filter else None,
@@ -41,14 +94,12 @@ class SearchView(ListView):
                     sort_by=sort_by
                 )
 
-                # Enrich results with actual model data
                 enriched_results = []
                 for result in search_results['results']:
                     enriched_result = self._enrich_result(result)
                     if enriched_result:
                         enriched_results.append(enriched_result)
 
-                # Store search metadata for pagination
                 self.search_total = search_results['total']
                 self.search_has_next = search_results['has_next']
                 self.search_has_prev = search_results['has_prev']
@@ -56,205 +107,80 @@ class SearchView(ListView):
                 return enriched_results
             except Exception as e:
                 logger.error(f"Search error: {e}", exc_info=True)
-                # Store error for display
                 self.search_error = str(e)
                 return []
         return []
 
     def _enrich_result(self, result):
-        """Fetch full model data for search result"""
         model_type = result['model_type']
         model_id = result['model_id']
-
         try:
             if model_type == 'link':
                 link = Link.objects.get(id=model_id)
-                return {
-                    'id': link.id,
-                    'model_type': 'link',
-                    'alias': link.alias,
-                    'original_url': link.original_url,
-                    'description': link.description,
-                    'click_count': link.click_count,
-                    'created_at': link.created_at,
-                    'tags': list(link.tags.all())
-                }
+                return {'id': link.id, 'model_type': 'link', 'alias': link.alias, 'original_url': link.original_url, 'description': link.description, 'click_count': link.click_count, 'created_at': link.created_at, 'tags': list(link.tags.all())}
             elif model_type == 'page':
                 page = Page.objects.get(id=model_id)
-                return {
-                    'id': page.id,
-                    'model_type': 'page',
-                    'title': page.title,
-                    'url': page.url,
-                    'summary': page.summary,
-                    'created_at': page.created_at,
-                    'tags': list(page.tags.all())
-                }
+                return {'id': page.id, 'model_type': 'page', 'title': page.title, 'url': page.url, 'summary': page.summary, 'created_at': page.created_at, 'tags': list(page.tags.all())}
             elif model_type == 'post':
                 post = Post.objects.get(id=model_id)
-                return {
-                    'id': post.id,
-                    'model_type': 'post',
-                    'title': post.title,
-                    'summary': post.summary,
-                    'created_at': post.created_at,
-                    'tags': list(post.tags.all())
-                }
+                return {'id': post.id, 'model_type': 'post', 'title': post.title, 'summary': post.summary, 'created_at': post.created_at, 'tags': list(post.tags.all())}
         except Exception as e:
             logger.error(f"Error enriching result {model_type} {model_id}: {e}")
             return None
 
     def get_context_data(self, **kwargs):
         context = super().get_context_data(**kwargs)
-        context.update({
-            'query': self.request.GET.get('q', ''),
-            'selected_type': self.request.GET.get('type', ''),
-            'sort_by': self.request.GET.get('sort', 'relevance'),
-            'total_results': getattr(self, 'search_total', 0),
-            'search_error': getattr(self, 'search_error', None),
-        })
+        context.update({'query': self.request.GET.get('q', ''), 'selected_type': self.request.GET.get('type', ''), 'sort_by': self.request.GET.get('sort', 'relevance'), 'total_results': getattr(self, 'search_total', 0), 'search_error': getattr(self, 'search_error', None)})
         return context
 
 def search(request):
     query = request.GET.get('q', '').strip()
     if query:
         results = []
-
-        # Search links
-        links = Link.objects.filter(
-            Q(alias__icontains=query) |
-            Q(original_url__icontains=query)
-        )[:5]
-        results.extend([{
-            'type': 'link',
-            'alias': link.alias,
-            'url': link.original_url
-        } for link in links])
-
-        # Search posts
-        posts = Post.objects.filter(
-            Q(title__icontains=query) |
-            Q(summary__icontains=query) |
-            Q(content__icontains=query)
-        )[:5]
-        results.extend([{
-            'type': 'post',
-            'title': post.title,
-            'url': f'/ui/posts/{post.id}'
-        } for post in posts])
-
+        links = Link.objects.filter(Q(alias__icontains=query) | Q(original_url__icontains=query))[:5]
+        results.extend([{'type': 'link', 'alias': link.alias, 'url': link.original_url} for link in links])
+        posts = Post.objects.filter(Q(title__icontains=query) | Q(summary__icontains=query) | Q(content__icontains=query))[:5]
+        results.extend([{'type': 'post', 'title': post.title, 'url': f'/ui/posts/{post.id}'} for post in posts])
     else:
         results = []
     return JsonResponse(results, safe=False)
 
 @require_http_methods(["GET"])
 def search_api_v2(request):
-    """
-    Modern API endpoint for React search interface
-    """
     query = request.GET.get('q', '').strip()
     type_filter = request.GET.get('type', '')
     sort_by = request.GET.get('sort', 'relevance')
     page = int(request.GET.get('page', 1))
     per_page = int(request.GET.get('per_page', 20))
-    
     if not query:
-        return JsonResponse({
-            'results': [],
-            'total': 0,
-            'page': page,
-            'per_page': per_page,
-            'has_next': False,
-            'has_prev': False
-        })
-    
+        return JsonResponse({'results': [], 'total': 0, 'page': page, 'per_page': per_page, 'has_next': False, 'has_prev': False})
     try:
-        # Use Whoosh search backend
-        search_results = search_backend.search(
-            query_string=query,
-            model_type=type_filter if type_filter else None,
-            page=page,
-            per_page=per_page,
-            sort_by=sort_by
-        )
-        
-        # Enrich results with actual model data
+        search_results = search_backend.search(query_string=query, model_type=type_filter if type_filter else None, page=page, per_page=per_page, sort_by=sort_by)
         enriched_results = []
         for result in search_results['results']:
             model_type = result['model_type']
             model_id = result['model_id']
-            
             try:
                 if model_type == 'link':
                     link = Link.objects.get(id=model_id)
-                    enriched_results.append({
-                        'id': link.id,
-                        'type': 'link',
-                        'title': link.alias,
-                        'url': f'/{link.alias}',
-                        'original_url': link.original_url,
-                        'description': link.description or '',
-                        'click_count': link.click_count,
-                        'created_at': link.created_at.isoformat(),
-                        'tags': [{'id': tag.id, 'name': tag.name, 'slug': tag.slug} for tag in link.tags.all()]
-                    })
+                    enriched_results.append({'id': link.id, 'type': 'link', 'title': link.alias, 'url': f'/{link.alias}', 'original_url': link.original_url, 'description': link.description or '', 'click_count': link.click_count, 'created_at': link.created_at.isoformat(), 'tags': [{'id': tag.id, 'name': tag.name, 'slug': tag.slug} for tag in link.tags.all()]})
                 elif model_type == 'page':
                     page_obj = Page.objects.get(id=model_id)
-                    enriched_results.append({
-                        'id': page_obj.id,
-                        'type': 'page',
-                        'title': page_obj.title or 'Untitled',
-                        'url': page_obj.url,
-                        'detail_url': f'/ui/pages/{page_obj.id}/',
-                        'summary': page_obj.summary or '',
-                        'created_at': page_obj.created_at.isoformat(),
-                        'tags': [{'id': tag.id, 'name': tag.name, 'slug': tag.slug} for tag in page_obj.tags.all()]
-                    })
+                    enriched_results.append({'id': page_obj.id, 'type': 'page', 'title': page_obj.title or 'Untitled', 'url': page_obj.url, 'detail_url': f'/ui/pages/{page_obj.id}/', 'summary': page_obj.summary or '', 'created_at': page_obj.created_at.isoformat(), 'tags': [{'id': tag.id, 'name': tag.name, 'slug': tag.slug} for tag in page_obj.tags.all()]})
                 elif model_type == 'post':
                     post = Post.objects.get(id=model_id)
-                    enriched_results.append({
-                        'id': post.id,
-                        'type': 'post',
-                        'title': post.title,
-                        'url': f'/ui/posts/{post.id}/',
-                        'summary': post.summary or '',
-                        'created_at': post.created_at.isoformat(),
-                        'tags': [{'id': tag.id, 'name': tag.name, 'slug': tag.slug} for tag in post.tags.all()]
-                    })
+                    enriched_results.append({'id': post.id, 'type': 'post', 'title': post.title, 'url': f'/ui/posts/{post.id}/', 'summary': post.summary or '', 'created_at': post.created_at.isoformat(), 'tags': [{'id': tag.id, 'name': tag.name, 'slug': tag.slug} for tag in post.tags.all()]})
             except Exception as e:
                 logger.error(f"Error enriching {model_type} {model_id}: {e}")
                 continue
-        
-        return JsonResponse({
-            'results': enriched_results,
-            'total': search_results['total'],
-            'page': search_results['page'],
-            'per_page': search_results['per_page'],
-            'has_next': search_results['has_next'],
-            'has_prev': search_results['has_prev']
-        })
-    
+        return JsonResponse({'results': enriched_results, 'total': search_results['total'], 'page': search_results['page'], 'per_page': search_results['per_page'], 'has_next': search_results['has_next'], 'has_prev': search_results['has_prev']})
     except Exception as e:
         logger.error(f"Search API error: {e}", exc_info=True)
-        return JsonResponse({
-            'error': str(e),
-            'results': [],
-            'total': 0
-        }, status=500)
+        return JsonResponse({'error': str(e), 'results': [], 'total': 0}, status=500)
 
 def search_aliases(request):
     query = request.GET.get('q', '').strip()
-    
-    # Prioritize results: exact match first, then starts with, then contains
-    links = Link.objects.filter(alias__icontains=query).annotate(
-        priority=Case(
-            When(alias__iexact=query, then=Value(1)),
-            When(alias__istartswith=query, then=Value(2)),
-            default=Value(3),
-            output_field=IntegerField()
-        )
-    ).order_by('priority', 'alias')[:10]
-
+    links = Link.objects.filter(alias__icontains=query).annotate(priority=Case(When(alias__iexact=query, then=Value(1)), When(alias__istartswith=query, then=Value(2)), default=Value(3), output_field=IntegerField())).order_by('priority', 'alias')[:10]
     results = []
     for link in links:
         try:
@@ -263,9 +189,7 @@ def search_aliases(request):
         except Exception as e:
             logger.error(f"Error creating URL for alias {link.alias}: {str(e)}")
             continue
-
     return JsonResponse(results, safe=False)
 
 class SearchReactView(TemplateView):
-    """Modern React-based search interface"""
     template_name = 'links/search_react.html'

links/templates/links/search_react.html

@@ -200,6 +200,7 @@
             const [query, setQuery] = useState('');
             const [type, setType] = useState('');
             const [sort, setSort] = useState('relevance');
+            const [isVector, setIsVector] = useState(false);
             const [results, setResults] = useState([]);
             const [loading, setLoading] = useState(false);
             const [total, setTotal] = useState(0);
@@ -226,7 +227,8 @@
                         per_page: perPage
                     });
 
-                    const response = await fetch(`/search/api/v2/?${params}`);
+                    const endpoint = isVector ? "/search/api/vector/" : "/search/api/v2/";
+                    const response = await fetch(`${endpoint}?${params}`);
                     const data = await response.json();
 
                     if (response.ok) {
@@ -336,7 +338,8 @@
                                         <label className="block text-sm font-medium text-gray-700 mb-2">
                                             Sort By
                                         </label>
-                                        <Select value={sort} onChange={(e) => setSort(e.target.value)}>
+                                        <div className="flex items-center gap-2 mb-2"><label className="text-sm font-medium text-gray-700">AI Vector Search</label><input type="checkbox" checked={isVector} onChange={(e) => setIsVector(e.target.checked)} className="h-4 w-4 text-blue-600 focus:ring-blue-500 border-gray-300 rounded" /></div>
+                                        <Select value={sort} onChange={(e) => setSort(e.target.value)} disabled={isVector}>
                                             <option value="relevance">Most Relevant</option>
                                             <option value="newest">Newest First</option>
                                             <option value="oldest">Oldest First</option>