feat: Add Content Generation Engine v2 with quality scoring

Major improvements to AI content generation: ## New Components (app/services/ai/) - PromptLibrary: YAML-based prompt templates with inheritance - ContextEngine: Anti-repetition and best performers tracking - ContentGeneratorV2: Enhanced generation with dynamic parameters - PlatformAdapter: Platform-specific content adaptation - ContentValidator: AI-powered quality scoring (0-100) ## Prompt Library (app/prompts/) - 3 personalities: default, educational, promotional - 5 templates: tip_tech, product_post, service_post, thread, response - 4 platform configs: x, threads, instagram, facebook - Few-shot examples by category: ia, productividad, seguridad ## Database Changes - New table: content_memory (tracks generated content) - New columns in posts: quality_score, score_breakdown, generation_attempts ## New API Endpoints (/api/v2/generate/) - POST /generate - Generation with quality check - POST /generate/batch - Batch generation - POST /quality/evaluate - Evaluate content quality - GET /templates, /personalities, /platforms - List configs ## Celery Tasks - update_engagement_scores (every 6h) - cleanup_old_memory (monthly) - refresh_best_posts_yaml (weekly) ## Tests - Comprehensive tests for all AI engine components Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
2026-01-28 20:55:28 +00:00
parent f458f809ca
commit 11b0ba46fa
36 changed files with 6266 additions and 55 deletions
--- a/tests/test_ai_engine.py
+++ b/tests/test_ai_engine.py
@@ -0,0 +1,447 @@
+"""
+Tests para el Content Generation Engine v2.
+
+Tests para:
+- PromptLibrary: Carga de YAMLs, renderizado
+- ContextEngine: Anti-repetición, best performers
+- PlatformAdapter: Adaptación por plataforma
+- ContentValidator: Validación y scoring
+"""
+
+import pytest
+from pathlib import Path
+from unittest.mock import MagicMock, patch, AsyncMock
+import json
+
+
+# ============================================================
+# PromptLibrary Tests
+# ============================================================
+
+class TestPromptLibrary:
+    """Tests para PromptLibrary."""
+
+    @pytest.fixture
+    def prompt_lib(self):
+        """Create PromptLibrary instance."""
+        from app.services.ai.prompt_library import PromptLibrary
+        return PromptLibrary()
+
+    def test_list_templates(self, prompt_lib):
+        """Test listado de templates disponibles."""
+        templates = prompt_lib.list_templates()
+
+        assert isinstance(templates, list)
+        assert "tip_tech" in templates
+        assert "product_post" in templates
+        assert "service_post" in templates
+        assert "thread" in templates
+
+    def test_list_personalities(self, prompt_lib):
+        """Test listado de personalidades disponibles."""
+        personalities = prompt_lib.list_personalities()
+
+        assert isinstance(personalities, list)
+        assert "default" in personalities
+        assert "educational" in personalities
+        assert "promotional" in personalities
+
+    def test_list_platforms(self, prompt_lib):
+        """Test listado de plataformas configuradas."""
+        platforms = prompt_lib.list_platforms()
+
+        assert isinstance(platforms, list)
+        assert "x" in platforms
+        assert "threads" in platforms
+        assert "instagram" in platforms
+        assert "facebook" in platforms
+
+    def test_get_personality_default(self, prompt_lib):
+        """Test carga de personalidad default."""
+        personality = prompt_lib.get_personality("default")
+
+        assert "name" in personality
+        assert personality["name"] == "default"
+        assert "voice" in personality
+        assert "system_prompt" in personality
+
+    def test_get_personality_with_inheritance(self, prompt_lib):
+        """Test que personalidades heredan correctamente."""
+        educational = prompt_lib.get_personality("educational")
+
+        # Debe tener propiedades de base (default)
+        assert "system_prompt" in educational
+
+        # Debe tener propiedades propias
+        assert "teaching_techniques" in educational
+
+    def test_get_template(self, prompt_lib):
+        """Test carga de template."""
+        template = prompt_lib.get_template("tip_tech")
+
+        assert "name" in template
+        assert template["name"] == "tip_tech"
+        assert "template" in template
+        assert "variables" in template
+        assert "parameters" in template
+
+    def test_render_template_basic(self, prompt_lib):
+        """Test renderizado básico de template."""
+        with patch('app.services.ai.prompt_library.settings') as mock_settings:
+            mock_settings.BUSINESS_NAME = "Test Corp"
+            mock_settings.BUSINESS_LOCATION = "Test City"
+            mock_settings.BUSINESS_WEBSITE = "test.com"
+            mock_settings.CONTENT_TONE = "Professional"
+
+            rendered = prompt_lib.render_template(
+                "tip_tech",
+                {"category": "productividad"}
+            )
+
+        assert "system_prompt" in rendered
+        assert "user_prompt" in rendered
+        assert "parameters" in rendered
+        assert "productividad" in rendered["user_prompt"]
+
+    def test_render_template_with_defaults(self, prompt_lib):
+        """Test que variables con defaults funcionan."""
+        with patch('app.services.ai.prompt_library.settings') as mock_settings:
+            mock_settings.BUSINESS_NAME = "Test Corp"
+            mock_settings.BUSINESS_LOCATION = "Test City"
+            mock_settings.BUSINESS_WEBSITE = "test.com"
+            mock_settings.CONTENT_TONE = "Professional"
+
+            # Solo category es requerida, difficulty_level tiene default
+            rendered = prompt_lib.render_template(
+                "tip_tech",
+                {"category": "ia"}
+            )
+
+        # Debe funcionar sin error y usar default
+        assert "user_prompt" in rendered
+
+    def test_get_platform_config(self, prompt_lib):
+        """Test carga de configuración de plataforma."""
+        config = prompt_lib.get_platform_config("x")
+
+        assert "platform" in config
+        assert config["platform"] == "x"
+        assert "limits" in config
+        assert "max_characters" in config["limits"]
+        assert config["limits"]["max_characters"] == 280
+
+    def test_get_platform_limits(self, prompt_lib):
+        """Test obtención de límites de plataforma."""
+        limits = prompt_lib.get_platform_limits("instagram")
+
+        assert "max_characters" in limits
+        assert limits["max_characters"] == 2200
+        assert "max_hashtags" in limits
+
+    def test_cache_works(self, prompt_lib):
+        """Test que el cache evita lecturas repetidas."""
+        # Primera carga
+        prompt_lib.get_template("tip_tech")
+
+        # Debe estar en cache
+        assert any("tip_tech" in key for key in prompt_lib._cache.keys())
+
+        # Segunda carga usa cache (no debería fallar)
+        template = prompt_lib.get_template("tip_tech")
+        assert template["name"] == "tip_tech"
+
+    def test_clear_cache(self, prompt_lib):
+        """Test limpieza de cache."""
+        prompt_lib.get_template("tip_tech")
+        assert len(prompt_lib._cache) > 0
+
+        prompt_lib.clear_cache()
+        assert len(prompt_lib._cache) == 0
+
+
+# ============================================================
+# PlatformAdapter Tests
+# ============================================================
+
+class TestPlatformAdapter:
+    """Tests para PlatformAdapter."""
+
+    @pytest.fixture
+    def adapter(self):
+        """Create PlatformAdapter instance."""
+        from app.services.ai.platform_adapter import PlatformAdapter
+        return PlatformAdapter()
+
+    def test_get_limits(self, adapter):
+        """Test obtención de límites."""
+        limits = adapter.get_limits("x")
+
+        assert limits["max_characters"] == 280
+        assert limits["max_hashtags"] == 2
+
+    def test_adapt_short_content(self, adapter):
+        """Test adaptación de contenido corto."""
+        content = "Tip corto #Tech"
+        result = adapter.adapt(content, "x")
+
+        assert result.content == content.strip()
+        assert not result.truncated
+        assert result.platform == "x"
+
+    def test_adapt_long_content_truncates(self, adapter):
+        """Test que contenido largo se trunca."""
+        # Contenido de 400 caracteres
+        content = "Este es un contenido muy largo. " * 15 + "#Tech #AI"
+
+        result = adapter.adapt(content, "x")
+
+        assert len(result.content) <= 280
+        assert result.truncated
+        assert "Contenido truncado" in str(result.changes_made)
+
+    def test_adapt_reduces_hashtags(self, adapter):
+        """Test que hashtags se reducen al límite."""
+        content = "Tip importante #Tech #AI #Python #Tips #Code #Dev"
+
+        result = adapter.adapt(content, "x")
+
+        # X solo permite 2 hashtags
+        hashtag_count = result.content.count("#")
+        assert hashtag_count <= 2
+        assert result.hashtags_adjusted
+
+    def test_validate_for_platform_valid(self, adapter):
+        """Test validación de contenido válido."""
+        content = "Tip corto #Tech"
+        validation = adapter.validate_for_platform(content, "x")
+
+        assert validation["valid"]
+        assert len(validation["issues"]) == 0
+
+    def test_validate_for_platform_too_long(self, adapter):
+        """Test validación de contenido muy largo."""
+        content = "X" * 300  # Excede límite de X
+
+        validation = adapter.validate_for_platform(content, "x")
+
+        assert not validation["valid"]
+        assert any(i["type"] == "length" for i in validation["issues"])
+
+    def test_adapt_for_all_platforms(self, adapter):
+        """Test adaptación para múltiples plataformas."""
+        content = "Contenido de prueba con mucho texto. " * 5 + "#Tech"
+
+        results = adapter.adapt_for_all_platforms(
+            content,
+            ["x", "threads", "instagram"]
+        )
+
+        assert "x" in results
+        assert "threads" in results
+        assert "instagram" in results
+
+        # X debe ser más corto
+        assert len(results["x"].content) <= len(results["instagram"].content)
+
+
+# ============================================================
+# ContextEngine Tests
+# ============================================================
+
+class TestContextEngine:
+    """Tests para ContextEngine."""
+
+    @pytest.fixture
+    def context_engine(self):
+        """Create ContextEngine instance."""
+        from app.services.ai.context_engine import ContextEngine
+        return ContextEngine()
+
+    def test_analyze_content_detects_hook(self, context_engine):
+        """Test detección de tipo de hook."""
+        # Pregunta retórica
+        content1 = "¿Sabías que el 90% de developers usan IA? #Tech"
+        analysis1 = context_engine.analyze_content(content1, "tip_tech", "x")
+        assert analysis1["hook_type"] == "pregunta_retórica"
+
+        # Dato impactante
+        content2 = "El 73% de empresas ya usan IA. #Tech"
+        analysis2 = context_engine.analyze_content(content2, "tip_tech", "x")
+        assert analysis2["hook_type"] == "dato_impactante"
+
+        # Tip directo
+        content3 = "Tip: usa ChatGPT para debugging. #Tech"
+        analysis3 = context_engine.analyze_content(content3, "tip_tech", "x")
+        assert analysis3["hook_type"] == "tip_directo"
+
+    def test_analyze_content_extracts_topics(self, context_engine):
+        """Test extracción de temas."""
+        content = "ChatGPT y Claude son herramientas de IA geniales para productividad. #Tech"
+        analysis = context_engine.analyze_content(content, "tip_tech", "x")
+
+        assert "ia" in analysis["topics"]
+        assert "productividad" in analysis["topics"]
+
+    def test_suggest_hook_type_varies(self, context_engine):
+        """Test que sugiere hooks variados."""
+        # Sin DB, debería retornar el primer preferido
+        mock_db = MagicMock()
+        mock_db.query.return_value.filter.return_value.order_by.return_value.limit.return_value.all.return_value = []
+
+        suggested = context_engine.suggest_hook_type(mock_db)
+
+        # Debería ser uno de los hooks preferidos
+        assert suggested in [
+            "pregunta_retórica", "dato_impactante", "tip_directo",
+            "afirmación_bold", "historia_corta"
+        ]
+
+
+# ============================================================
+# ContentValidator Tests
+# ============================================================
+
+class TestContentValidator:
+    """Tests para ContentValidator."""
+
+    @pytest.fixture
+    def validator(self):
+        """Create ContentValidator instance."""
+        from app.services.ai.validator import ContentValidator
+        return ContentValidator()
+
+    def test_validate_length_passes(self, validator):
+        """Test validación de longitud válida."""
+        content = "Contenido corto #Tech"
+        result = validator.validate(content, "x")
+
+        assert result.passed
+        assert not any(i["type"] == "length" for i in result.issues)
+
+    def test_validate_length_fails(self, validator):
+        """Test validación de longitud excedida."""
+        content = "X" * 300  # Excede 280 de X
+        result = validator.validate(content, "x")
+
+        assert not result.passed
+        assert any(i["type"] == "length" and i["severity"] == "error" for i in result.issues)
+
+    def test_validate_prohibited_content(self, validator):
+        """Test validación de contenido prohibido."""
+        # Modificar config temporalmente para test
+        validator.config["validations"]["prohibited_content"]["prohibited_words"] = ["test_banned"]
+
+        content = "Este contenido tiene test_banned palabra"
+        result = validator.validate(content, "x")
+
+        assert not result.passed
+        assert any(i["type"] == "prohibited_content" for i in result.issues)
+
+    def test_validate_empty_content_fails(self, validator):
+        """Test que contenido vacío falla."""
+        content = "   "
+        result = validator.validate(content, "x")
+
+        assert not result.passed
+        assert any(i["type"] == "empty_content" for i in result.issues)
+
+    def test_should_regenerate_respects_max_attempts(self, validator):
+        """Test que regeneración respeta máximo de intentos."""
+        from app.services.ai.validator import ContentQualityResult, ValidationResult, ScoringResult
+
+        mock_quality = ContentQualityResult(
+            validation=ValidationResult(passed=True, issues=[], content="test"),
+            scoring=ScoringResult(
+                total_score=55,
+                breakdown={},
+                feedback="Needs improvement",
+                is_top_performer=False,
+                action="regenerate"
+            ),
+            final_decision="regenerate",
+            content="test"
+        )
+
+        # Primer intento: debería regenerar
+        assert validator.should_regenerate(mock_quality, attempt=1)
+
+        # Segundo intento: debería regenerar
+        assert validator.should_regenerate(mock_quality, attempt=2)
+
+        # Tercer intento: NO debería regenerar (max_attempts=2)
+        assert not validator.should_regenerate(mock_quality, attempt=3)
+
+
+# ============================================================
+# Integration Tests
+# ============================================================
+
+class TestAIEngineIntegration:
+    """Tests de integración del motor completo."""
+
+    @pytest.mark.asyncio
+    async def test_full_generation_flow_mocked(self, mock_openai_client):
+        """Test flujo completo de generación con mock."""
+        with patch('app.services.ai.generator.settings') as mock_settings:
+            mock_settings.DEEPSEEK_API_KEY = "test-key"
+            mock_settings.DEEPSEEK_BASE_URL = "https://api.deepseek.com"
+            mock_settings.BUSINESS_NAME = "Test Corp"
+            mock_settings.BUSINESS_LOCATION = "Test City"
+            mock_settings.BUSINESS_WEBSITE = "test.com"
+            mock_settings.CONTENT_TONE = "Professional"
+
+            from app.services.ai.generator import ContentGeneratorV2
+
+            generator = ContentGeneratorV2()
+            generator._client = mock_openai_client
+
+            result = await generator.generate(
+                template_name="tip_tech",
+                variables={"category": "productividad"},
+                platform="x",
+                use_context=False,
+                use_few_shot=False
+            )
+
+        assert "content" in result
+        assert "adapted_content" in result
+        assert "metadata" in result
+        assert result["metadata"]["template"] == "tip_tech"
+        assert result["metadata"]["platform"] == "x"
+
+    def test_prompt_library_to_adapter_flow(self):
+        """Test flujo de PromptLibrary a PlatformAdapter."""
+        from app.services.ai.prompt_library import PromptLibrary
+        from app.services.ai.platform_adapter import PlatformAdapter
+
+        lib = PromptLibrary()
+        adapter = PlatformAdapter()
+
+        # Obtener límites de plataforma desde library
+        limits_from_lib = lib.get_platform_limits("x")
+
+        # Obtener límites desde adapter
+        limits_from_adapter = adapter.get_limits("x")
+
+        # Deberían ser consistentes
+        assert limits_from_lib["max_characters"] == limits_from_adapter["max_characters"]
+
+
+# ============================================================
+# Fixtures adicionales
+# ============================================================
+
+@pytest.fixture
+def mock_openai_client():
+    """Mock OpenAI client for DeepSeek API tests."""
+    mock_client = MagicMock()
+
+    mock_response = MagicMock()
+    mock_response.choices = [MagicMock()]
+    mock_response.choices[0].message.content = "Generated test content #TechTip #AI"
+    mock_response.usage = MagicMock()
+    mock_response.usage.total_tokens = 100
+
+    mock_client.chat.completions.create.return_value = mock_response
+
+    return mock_client