feat(admin): Sekcja Baza Wiedzy AI w panelu ZOPK

- Dodano sekcję "🧠 Baza Wiedzy AI" do /admin/zopk
- Statystyki scrapingu: zatwierdzonych, zescrapowanych, oczekujących
- Statystyki wiedzy: chunks, fakty, encje
- Statystyki embeddingów: z/bez embeddingów
- Przyciski: Scrapuj treść, Ekstrakcja AI, Generuj embeddingi
- Top encje z kolorowymi pillami (firmy, osoby, miejsca, projekty)
- Rozszerzono get_extraction_statistics() o pending_scrape i embeddings

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

templates/admin/zopk_dashboard.html

@@ -1263,6 +1263,92 @@
     {% endif %}
 </div>
 
+<!-- AI Knowledge Base -->
+<div class="panel-section">
+    <h2>🧠 Baza Wiedzy AI</h2>
+
+    <div id="knowledge-stats-loading" class="text-center py-4">
+        <span class="spinner"></span> Ładowanie statystyk...
+    </div>
+
+    <div id="knowledge-stats-content" style="display: none;">
+        <!-- Content Scraping Stats -->
+        <div class="stats-section">
+            <h3 class="stats-section-title">Scraping treści</h3>
+            <div class="stats-grid stats-grid-small">
+                <div class="stat-card info-only">
+                    <div class="stat-value" id="kb-total-approved">-</div>
+                    <div class="stat-label">Zatwierdzonych newsów</div>
+                </div>
+                <div class="stat-card info-only success">
+                    <div class="stat-value" id="kb-scraped">-</div>
+                    <div class="stat-label">Zescrapowanych</div>
+                </div>
+                <div class="stat-card info-only warning">
+                    <div class="stat-value" id="kb-pending-scrape">-</div>
+                    <div class="stat-label">Oczekuje na scraping</div>
+                </div>
+                <button class="stat-card filter-card ai-action" onclick="scrapeContent()" id="scrapeBtn" title="Scrapuj treść artykułów">
+                    <div class="stat-value" style="font-size: var(--font-size-xl);">📄</div>
+                    <div class="stat-label">Scrapuj treść</div>
+                </button>
+            </div>
+        </div>
+
+        <!-- Knowledge Extraction Stats -->
+        <div class="stats-section">
+            <h3 class="stats-section-title">Ekstrakcja wiedzy (Gemini AI)</h3>
+            <div class="stats-grid stats-grid-small">
+                <div class="stat-card info-only">
+                    <div class="stat-value" id="kb-total-chunks">-</div>
+                    <div class="stat-label">Chunks</div>
+                </div>
+                <div class="stat-card info-only">
+                    <div class="stat-value" id="kb-total-facts">-</div>
+                    <div class="stat-label">Faktów</div>
+                </div>
+                <div class="stat-card info-only">
+                    <div class="stat-value" id="kb-total-entities">-</div>
+                    <div class="stat-label">Encji</div>
+                </div>
+                <button class="stat-card filter-card ai-action" onclick="extractKnowledge()" id="extractBtn" title="Wyekstrahuj wiedzę z zescrapowanych artykułów">
+                    <div class="stat-value" style="font-size: var(--font-size-xl);">🤖</div>
+                    <div class="stat-label">Ekstrakcja AI</div>
+                </button>
+            </div>
+        </div>
+
+        <!-- Embeddings Stats -->
+        <div class="stats-section">
+            <h3 class="stats-section-title">Semantic Search (Embeddings)</h3>
+            <div class="stats-grid stats-grid-small">
+                <div class="stat-card info-only success">
+                    <div class="stat-value" id="kb-with-embeddings">-</div>
+                    <div class="stat-label">Z embeddingami</div>
+                </div>
+                <div class="stat-card info-only warning">
+                    <div class="stat-value" id="kb-pending-embeddings">-</div>
+                    <div class="stat-label">Bez embeddingów</div>
+                </div>
+                <button class="stat-card filter-card ai-action" onclick="generateEmbeddings()" id="embeddingsBtn" title="Wygeneruj embeddingi dla semantic search">
+                    <div class="stat-value" style="font-size: var(--font-size-xl);">🔍</div>
+                    <div class="stat-label">Generuj embeddingi</div>
+                </button>
+            </div>
+        </div>
+
+        <!-- Top Entities -->
+        <div class="stats-section" id="top-entities-section" style="display: none;">
+            <h3 class="stats-section-title">Najczęściej wymieniane encje</h3>
+            <div id="top-entities-list" class="entity-pills"></div>
+        </div>
+    </div>
+
+    <div id="knowledge-stats-error" style="display: none;" class="alert alert-danger">
+        Błąd ładowania statystyk
+    </div>
+</div>
+
 <!-- Projects -->
 <div class="panel-section">
     <h2>Projekty strategiczne</h2>
@@ -1447,6 +1533,60 @@
     from { transform: translateX(0); opacity: 1; }
     to { transform: translateX(100%); opacity: 0; }
 }
+
+/* Spinner */
+.spinner, .spinner-small {
+    display: inline-block;
+    border: 3px solid var(--border);
+    border-top-color: var(--primary);
+    border-radius: 50%;
+    animation: spin 0.8s linear infinite;
+}
+.spinner {
+    width: 24px;
+    height: 24px;
+}
+.spinner-small {
+    width: 16px;
+    height: 16px;
+    border-width: 2px;
+}
+@keyframes spin {
+    to { transform: rotate(360deg); }
+}
+
+/* Entity Pills */
+.entity-pills {
+    display: flex;
+    flex-wrap: wrap;
+    gap: var(--spacing-sm);
+}
+.entity-pill {
+    display: inline-flex;
+    align-items: center;
+    gap: var(--spacing-xs);
+    padding: var(--spacing-xs) var(--spacing-sm);
+    border-radius: var(--radius-full, 9999px);
+    font-size: var(--font-size-sm);
+    background: var(--surface-secondary, #f5f5f5);
+    border: 1px solid var(--border);
+}
+.entity-pill small {
+    color: var(--text-muted);
+}
+.entity-pill.entity-company { background: #dbeafe; border-color: #93c5fd; }
+.entity-pill.entity-person { background: #fef3c7; border-color: #fcd34d; }
+.entity-pill.entity-place { background: #d1fae5; border-color: #6ee7b7; }
+.entity-pill.entity-organization { background: #e0e7ff; border-color: #a5b4fc; }
+.entity-pill.entity-project { background: #fce7f3; border-color: #f9a8d4; }
+
+/* Text utilities */
+.text-center { text-align: center; }
+.py-4 { padding-top: 1rem; padding-bottom: 1rem; }
+
+/* Alert */
+.alert { padding: var(--spacing-md); border-radius: var(--radius); }
+.alert-danger { background: #fef2f2; border: 1px solid #fecaca; color: #dc2626; }
 </style>
 {% endblock %}
 
@@ -2357,4 +2497,168 @@ document.getElementById('addNewsModal').addEventListener('click', function(e) {
         closeAddNewsModal();
     }
 });
+
+// ===========================================
+// AI Knowledge Base Functions
+// ===========================================
+
+async function loadKnowledgeStats() {
+    try {
+        const response = await fetch('/admin/zopk/knowledge/stats');
+        const data = await response.json();
+
+        if (data.success) {
+            document.getElementById('knowledge-stats-loading').style.display = 'none';
+            document.getElementById('knowledge-stats-content').style.display = 'block';
+
+            // Scraping stats
+            document.getElementById('kb-total-approved').textContent = data.articles?.total_approved || 0;
+            document.getElementById('kb-scraped').textContent = data.articles?.scraped || 0;
+            document.getElementById('kb-pending-scrape').textContent = data.articles?.pending_scrape || 0;
+
+            // Knowledge base stats
+            document.getElementById('kb-total-chunks').textContent = data.knowledge_base?.total_chunks || 0;
+            document.getElementById('kb-total-facts').textContent = data.knowledge_base?.total_facts || 0;
+            document.getElementById('kb-total-entities').textContent = data.knowledge_base?.total_entities || 0;
+
+            // Embeddings stats
+            document.getElementById('kb-with-embeddings').textContent = data.knowledge_base?.chunks_with_embeddings || 0;
+            document.getElementById('kb-pending-embeddings').textContent = data.knowledge_base?.chunks_without_embeddings || 0;
+
+            // Top entities
+            if (data.top_entities && data.top_entities.length > 0) {
+                document.getElementById('top-entities-section').style.display = 'block';
+                const list = document.getElementById('top-entities-list');
+                list.innerHTML = data.top_entities.map(e =>
+                    `<span class="entity-pill entity-${e.type}">${e.name} <small>(${e.mentions})</small></span>`
+                ).join('');
+            }
+        } else {
+            throw new Error(data.error || 'Unknown error');
+        }
+    } catch (error) {
+        console.error('Error loading knowledge stats:', error);
+        document.getElementById('knowledge-stats-loading').style.display = 'none';
+        document.getElementById('knowledge-stats-error').style.display = 'block';
+        document.getElementById('knowledge-stats-error').textContent = `Błąd: ${error.message}`;
+    }
+}
+
+async function scrapeContent() {
+    const btn = document.getElementById('scrapeBtn');
+    const originalContent = btn.innerHTML;
+
+    const confirmed = await showConfirm(
+        'Czy chcesz rozpocząć scrapowanie treści artykułów?<br><br>' +
+        '<small>Proces pobierze pełną treść z zatwierdzonych newsów które jeszcze nie mają treści.</small>',
+        { icon: '📄', title: 'Scraping treści', okText: 'Rozpocznij', okClass: 'btn-primary' }
+    );
+
+    if (!confirmed) return;
+
+    btn.disabled = true;
+    btn.innerHTML = '<span class="spinner-small"></span>';
+
+    try {
+        const response = await fetch('/admin/zopk/news/scrape-content', {
+            method: 'POST',
+            headers: { 'Content-Type': 'application/json', 'X-CSRFToken': csrfToken },
+            body: JSON.stringify({ limit: 30 })
+        });
+        const data = await response.json();
+
+        if (data.success) {
+            showToast(data.message || `Zescrapowano ${data.scraped || 0} artykułów`, 'success');
+            loadKnowledgeStats();
+        } else {
+            showToast(data.error || 'Błąd scrapowania', 'error');
+        }
+    } catch (error) {
+        showToast(`Błąd: ${error.message}`, 'error');
+    } finally {
+        btn.disabled = false;
+        btn.innerHTML = originalContent;
+    }
+}
+
+async function extractKnowledge() {
+    const btn = document.getElementById('extractBtn');
+    const originalContent = btn.innerHTML;
+
+    const confirmed = await showConfirm(
+        'Czy chcesz uruchomić ekstrakcję wiedzy przez AI?<br><br>' +
+        '<small>Gemini AI przeanalizuje zescrapowane artykuły i wyekstrahuje:<br>' +
+        '• Chunks (fragmenty tekstu)<br>' +
+        '• Fakty (daty, liczby, decyzje)<br>' +
+        '• Encje (firmy, osoby, projekty)</small>',
+        { icon: '🤖', title: 'Ekstrakcja wiedzy', okText: 'Uruchom AI', okClass: 'btn-primary' }
+    );
+
+    if (!confirmed) return;
+
+    btn.disabled = true;
+    btn.innerHTML = '<span class="spinner-small"></span>';
+
+    try {
+        const response = await fetch('/admin/zopk/knowledge/extract', {
+            method: 'POST',
+            headers: { 'Content-Type': 'application/json', 'X-CSRFToken': csrfToken },
+            body: JSON.stringify({ limit: 10 })
+        });
+        const data = await response.json();
+
+        if (data.success) {
+            showToast(data.message || 'Ekstrakcja zakończona', 'success');
+            loadKnowledgeStats();
+        } else {
+            showToast(data.error || 'Błąd ekstrakcji', 'error');
+        }
+    } catch (error) {
+        showToast(`Błąd: ${error.message}`, 'error');
+    } finally {
+        btn.disabled = false;
+        btn.innerHTML = originalContent;
+    }
+}
+
+async function generateEmbeddings() {
+    const btn = document.getElementById('embeddingsBtn');
+    const originalContent = btn.innerHTML;
+
+    const confirmed = await showConfirm(
+        'Czy chcesz wygenerować embeddingi dla semantic search?<br><br>' +
+        '<small>Google Text Embedding API przekształci tekst w wektory 768-wymiarowe.<br>' +
+        'Embeddingi umożliwiają inteligentne wyszukiwanie w bazie wiedzy.</small>',
+        { icon: '🔍', title: 'Generowanie embeddingów', okText: 'Generuj', okClass: 'btn-primary' }
+    );
+
+    if (!confirmed) return;
+
+    btn.disabled = true;
+    btn.innerHTML = '<span class="spinner-small"></span>';
+
+    try {
+        const response = await fetch('/admin/zopk/knowledge/embeddings', {
+            method: 'POST',
+            headers: { 'Content-Type': 'application/json', 'X-CSRFToken': csrfToken },
+            body: JSON.stringify({ limit: 50 })
+        });
+        const data = await response.json();
+
+        if (data.success) {
+            showToast(data.message || 'Embeddingi wygenerowane', 'success');
+            loadKnowledgeStats();
+        } else {
+            showToast(data.error || 'Błąd generowania', 'error');
+        }
+    } catch (error) {
+        showToast(`Błąd: ${error.message}`, 'error');
+    } finally {
+        btn.disabled = false;
+        btn.innerHTML = originalContent;
+    }
+}
+
+// Load knowledge stats on page load
+document.addEventListener('DOMContentLoaded', loadKnowledgeStats);
 {% endblock %}

zopk_knowledge_service.py

@@ -740,20 +740,33 @@ class ZOPKKnowledgeService:
             ZOPKNews.scrape_status == 'scraped'
         ).scalar()
 
+        # Pending scrape: approved but not yet scraped
+        pending_scrape = self.db.query(func.count(ZOPKNews.id)).filter(
+            ZOPKNews.status.in_(['approved', 'auto_approved']),
+            ZOPKNews.scrape_status.in_(['pending', None])
+        ).scalar()
+
         extracted = self.db.query(func.count(ZOPKNews.id)).filter(
             ZOPKNews.knowledge_extracted == True
         ).scalar()
 
-        pending = self.db.query(func.count(ZOPKNews.id)).filter(
+        pending_extract = self.db.query(func.count(ZOPKNews.id)).filter(
             ZOPKNews.scrape_status == 'scraped',
             ZOPKNews.knowledge_extracted == False
         ).scalar()
 
         # Knowledge base stats
-        chunks_count = self.db.query(func.count(ZOPKKnowledgeChunk.id)).scalar()
-        facts_count = self.db.query(func.count(ZOPKKnowledgeFact.id)).scalar()
-        entities_count = self.db.query(func.count(ZOPKKnowledgeEntity.id)).scalar()
-        relations_count = self.db.query(func.count(ZOPKKnowledgeRelation.id)).scalar()
+        total_chunks = self.db.query(func.count(ZOPKKnowledgeChunk.id)).scalar()
+        total_facts = self.db.query(func.count(ZOPKKnowledgeFact.id)).scalar()
+        total_entities = self.db.query(func.count(ZOPKKnowledgeEntity.id)).scalar()
+        total_relations = self.db.query(func.count(ZOPKKnowledgeRelation.id)).scalar()
+
+        # Embeddings stats
+        chunks_with_embeddings = self.db.query(func.count(ZOPKKnowledgeChunk.id)).filter(
+            ZOPKKnowledgeChunk.embedding.isnot(None)
+        ).scalar()
+
+        chunks_without_embeddings = (total_chunks or 0) - (chunks_with_embeddings or 0)
 
         # Top entities by mentions
         top_entities = self.db.query(
@@ -768,14 +781,17 @@ class ZOPKKnowledgeService:
             'articles': {
                 'total_approved': total_approved or 0,
                 'scraped': scraped or 0,
+                'pending_scrape': pending_scrape or 0,
                 'extracted': extracted or 0,
-                'pending': pending or 0
+                'pending_extract': pending_extract or 0
             },
             'knowledge_base': {
-                'chunks': chunks_count or 0,
-                'facts': facts_count or 0,
-                'entities': entities_count or 0,
-                'relations': relations_count or 0
+                'total_chunks': total_chunks or 0,
+                'total_facts': total_facts or 0,
+                'total_entities': total_entities or 0,
+                'total_relations': total_relations or 0,
+                'chunks_with_embeddings': chunks_with_embeddings or 0,
+                'chunks_without_embeddings': chunks_without_embeddings or 0
             },
             'top_entities': [
                 {'name': e[0], 'type': e[1], 'mentions': e[2]}