Merge pull request #61 from ubc/feat/fixbugs

add unit tests and new feature

Author: fanxiaotuGod

routes/create/services/promptTemplateInitializer.js

@@ -0,0 +1,152 @@
+import SystemPromptTemplate from '../models/SystemPromptTemplate.js';
+
+const OUTER_PROMPT = `Create a question distribution plan for a quiz.
+
+Learning Objectives:
+{{learningObjectives}}
+
+{{materialsContext}}
+
+TASK: Distribute EXACTLY {{totalQuestions}} questions across the learning objectives.
+
+{{innerPrompt}}
+
+CRITICAL REQUIREMENTS:
+1. The sum of ALL counts MUST equal EXACTLY {{totalQuestions}}
+2. Distribute across ALL learning objectives proportionally
+3. Each item must have count >= 1
+4. Return ONLY the JSON object below (no text, no markdown, no explanations)
+
+{
+  "planItems": [
+    { "type": "multiple-choice", "learningObjectiveIndex": 0, "count": 5 },
+    { "type": "flashcard", "learningObjectiveIndex": 1, "count": 3 }
+  ]
+}`;
+
+const TEMPLATES = [
+  {
+    approach: 'support',
+    outerPrompt: OUTER_PROMPT,
+    innerPrompt: `Pedagogical Approach: Support Learning
+This approach emphasizes helping students memorize and understand key concepts through flashcards and summaries.
+
+Question Type Rules:
+- ONLY use these types: flashcard, summary
+- Flashcards should be the MAJORITY (approximately 80-90%)
+- Summaries should be the MINORITY (approximately 10-20%)
+- Each learning objective should have AT MOST 1 summary question
+- Distribute flashcards evenly across all learning objectives
+
+Design Guidelines:
+- Flashcards for key terms, definitions, and concepts
+- Summaries for broader understanding and synthesis`,
+    questionTypeRules: {
+      allowedTypes: ['flashcard', 'summary'],
+      distribution: new Map([['flashcard', 0.85], ['summary', 0.15]]),
+      maxPerLO: new Map([['summary', 1]])
+    },
+    description: 'Support Learning: Focus on memorization and understanding through flashcards and summaries',
+    exampleOutput: 'Mostly flashcards (85%) with occasional summaries (15%), max 1 summary per LO',
+    isActive: true
+  },
+  {
+    approach: 'assess',
+    outerPrompt: OUTER_PROMPT,
+    innerPrompt: `Pedagogical Approach: Assess Understanding
+This approach emphasizes evaluating student comprehension through multiple-choice and true/false questions.
+
+Question Type Rules:
+- ONLY use these types: multiple-choice, true-false
+- Multiple-choice should be the MAJORITY (approximately 60-70%)
+- True/false should complement (approximately 30-40%)
+- Distribute both types across all learning objectives
+
+Design Guidelines:
+- Multiple-choice for complex understanding and application
+- True/false for fundamental concepts and quick checks
+- Mix question types within each learning objective for variety`,
+    questionTypeRules: {
+      allowedTypes: ['multiple-choice', 'true-false'],
+      distribution: new Map([['multiple-choice', 0.65], ['true-false', 0.35]]),
+      maxPerLO: new Map()
+    },
+    description: 'Assess Understanding: Evaluate comprehension with multiple-choice and true/false questions',
+    exampleOutput: 'Mix of multiple-choice (65%) and true/false (35%), distributed across all LOs',
+    isActive: true
+  },
+  {
+    approach: 'gamify',
+    outerPrompt: OUTER_PROMPT,
+    innerPrompt: `Pedagogical Approach: Gamify Learning
+This approach emphasizes creating engaging, game-like experiences through interactive question types.
+
+Question Type Rules:
+- Use DIVERSE types: matching, ordering, cloze, discussion
+- DO NOT use: multiple-choice, true-false, flashcard, summary
+- Create a VARIED distribution - avoid having all questions of the same type
+- Each learning objective should have AT LEAST 2 different question types
+
+Design Guidelines:
+- Matching for connecting concepts and relationships
+- Ordering for sequences, processes, and hierarchies
+- Cloze for fill-in-the-blank and context understanding
+- Discussion for critical thinking and analysis
+- Aim for visual variety and engagement
+- Distribute question types so no single type dominates`,
+    questionTypeRules: {
+      allowedTypes: ['matching', 'ordering', 'cloze', 'discussion'],
+      distribution: new Map([
+        ['matching', 0.3],
+        ['ordering', 0.25],
+        ['cloze', 0.25],
+        ['discussion', 0.2]
+      ]),
+      maxPerLO: new Map()
+    },
+    description: 'Gamify Learning: Create engaging experiences with interactive question types',
+    exampleOutput: 'Diverse mix of matching (30%), ordering (25%), cloze (25%), and discussion (20%)',
+    isActive: true
+  }
+];
+
+/**
+ * Initialize system prompt templates if they don't exist
+ * This runs automatically on server startup
+ */
+export async function initializePromptTemplates() {
+  try {
+    // Check if templates already exist
+    const existingCount = await SystemPromptTemplate.countDocuments();
+    
+    if (existingCount >= 3) {
+      console.log(`✅ System prompt templates already initialized (${existingCount} templates found)`);
+      return;
+    }
+
+    console.log('🔧 Initializing system prompt templates...');
+
+    // Check which templates are missing
+    const existingApproaches = await SystemPromptTemplate.find({}, 'approach').lean();
+    const existingApproachNames = new Set(existingApproaches.map(t => t.approach));
+
+    let createdCount = 0;
+    for (const template of TEMPLATES) {
+      if (!existingApproachNames.has(template.approach)) {
+        await SystemPromptTemplate.create(template);
+        console.log(`✅ Created template for approach: ${template.approach}`);
+        createdCount++;
+      }
+    }
+
+    if (createdCount > 0) {
+      console.log(`🎉 Successfully initialized ${createdCount} system prompt template(s)`);
+    }
+  } catch (error) {
+    console.error('❌ Error initializing system prompt templates:', error.message);
+    console.error('💡 You may need to run the initialization script manually:');
+    console.error('   node routes/create/scripts/initializeSystemPrompts.js');
+  }
+}
+
+export default { initializePromptTemplates };

server.js

@@ -46,6 +46,17 @@ import('./routes/create/services/ragService.js')
     console.error('💡 RAG features may not work properly');
   });
 
+// Initialize system prompt templates on server startup
+console.log('🔧 Initializing system prompt templates...');
+import('./routes/create/services/promptTemplateInitializer.js')
+  .then(module => {
+    module.initializePromptTemplates();
+  })
+  .catch(err => {
+    console.error('❌ Failed to initialize prompt templates:', err.message);
+    console.error('💡 Prompt template features may not work properly');
+  });
+
 const app = express();
 const PORT = process.env.PORT || 7736;