Upload folder using huggingface_hub

README.md

@@ -3,7 +3,7 @@ title: AnkiGen
 emoji: 📚
 app_file: app.py
 requirements: requirements.txt
-python: 3.10
+python: 3.12
 sdk: gradio
 sdk_version: 5.38.1
 ---

ankigen_core/agents/generators.py

@@ -115,6 +115,11 @@ class SubjectExpertAgent(BaseAgentWrapper):
                 user_input = (
                     f"Generate {cards_in_this_batch} flashcards for the topic: {topic}"
                 )
+
+                # Add cloze generation instruction if enabled
+                if context and context.get("generate_cloze"):
+                    user_input += "\n\nIMPORTANT: Generate a mix of card types including cloze cards. For code examples, syntax, and fill-in-the-blank concepts, use cloze cards (card_type='cloze'). Aim for roughly 50% cloze cards when dealing with technical/programming content."
+
                 if context:
                     user_input += f"\n\nAdditional context: {context}"
 

ankigen_core/agents/integration.py

@@ -9,8 +9,7 @@ from ankigen_core.models import Card
 from ankigen_core.llm_interface import OpenAIClientManager
 from ankigen_core.context7 import Context7Client
 
-from .generators import SubjectExpertAgent, QualityReviewAgent
-from ankigen_core.agents.config import get_config_manager
+from .generators import SubjectExpertAgent
 
 
 class AgentOrchestrator:
@@ -21,7 +20,6 @@ class AgentOrchestrator:
         self.openai_client = None
 
         self.subject_expert = None
-        self.quality_reviewer = None
 
     async def initialize(self, api_key: str, model_overrides: Dict[str, str] = None):
         """Initialize the agent system"""
@@ -50,10 +48,10 @@ class AgentOrchestrator:
         subject: str = "general",
         num_cards: int = 5,
         difficulty: str = "intermediate",
-        enable_quality_pipeline: bool = True,
         context: Dict[str, Any] = None,
         library_name: Optional[str] = None,
         library_topic: Optional[str] = None,
+        generate_cloze: bool = False,
     ) -> Tuple[List[Card], Dict[str, Any]]:
         """Generate cards using the agent system"""
         start_time = datetime.now()
@@ -117,18 +115,14 @@ class AgentOrchestrator:
                 num_cards=num_cards,
                 difficulty=difficulty,
                 context=enhanced_context,
+                generate_cloze=generate_cloze,
             )
 
-            review_results = {}
-            if enable_quality_pipeline:
-                cards, review_results = await self._quality_review_phase(cards)
-
             # Collect metadata
             metadata = {
                 "generation_method": "agent_system",
                 "generation_time": (datetime.now() - start_time).total_seconds(),
                 "cards_generated": len(cards),
-                "review_results": review_results,
                 "topic": topic,
                 "subject": subject,
                 "difficulty": difficulty,
@@ -152,16 +146,18 @@ class AgentOrchestrator:
         num_cards: int,
         difficulty: str,
         context: Dict[str, Any] = None,
+        generate_cloze: bool = False,
     ) -> List[Card]:
         """Execute the card generation phase"""
 
         if not self.subject_expert or self.subject_expert.subject != subject:
             self.subject_expert = SubjectExpertAgent(self.openai_client, subject)
 
-        # Add difficulty to context if needed
+        # Add difficulty and cloze preference to context
         if context is None:
             context = {}
         context["difficulty"] = difficulty
+        context["generate_cloze"] = generate_cloze
 
         cards = await self.subject_expert.generate_cards(
             topic=topic, num_cards=num_cards, context=context
@@ -170,58 +166,6 @@ class AgentOrchestrator:
         logger.info(f"Generation phase complete: {len(cards)} cards generated")
         return cards
 
-    async def _quality_review_phase(
-        self, cards: List[Card]
-    ) -> Tuple[List[Card], Dict[str, Any]]:
-        """Perform a single quality-review pass with optional fixes."""
-
-        if not cards:
-            return cards, {"message": "No cards to review"}
-
-        logger.info(f"Performing quality review for {len(cards)} cards")
-
-        if not self.quality_reviewer:
-            # Use the same model as the subject expert by default.
-            subject_config = get_config_manager().get_agent_config("subject_expert")
-            reviewer_model = subject_config.model if subject_config else "gpt-4.1"
-            self.quality_reviewer = QualityReviewAgent(
-                self.openai_client, reviewer_model
-            )
-
-        reviewed_cards: List[Card] = []
-        approvals: List[Dict[str, Any]] = []
-
-        for card in cards:
-            reviewed_card, approved, reason = await self.quality_reviewer.review_card(
-                card
-            )
-            if approved:
-                reviewed_cards.append(reviewed_card)
-            else:
-                approvals.append(
-                    {
-                        "question": card.front.question if card.front else "",
-                        "reason": reason,
-                    }
-                )
-
-        review_results = {
-            "total_cards_reviewed": len(cards),
-            "approved_cards": len(reviewed_cards),
-            "rejected_cards": approvals,
-        }
-
-        if approvals:
-            logger.warning(
-                "Quality review rejected cards: %s",
-                "; ".join(
-                    f"{entry['question'][:50]}… ({entry['reason']})"
-                    for entry in approvals
-                ),
-            )
-
-        return reviewed_cards, review_results
-
     def get_performance_metrics(self) -> Dict[str, Any]:
         """Get performance metrics for the agent system"""
 

ankigen_core/agents/templates/generators.j2

@@ -2,7 +2,7 @@
 {
     "subject_expert": {
         "name": "subject_expert",
-        "instructions": "You are a world-class expert in {{ subject | default('the subject area') }} with deep pedagogical knowledge. \nYour role is to generate high-quality flashcards that demonstrate mastery of {{ subject | default('the subject') }} concepts.\n\nKey responsibilities:\n- Create ATOMIC cards: extremely short (1-9 words on back), break complex info into multiple simple cards\n- Use standardized, bland prompts without fancy formatting or unusual words\n- Design prompts that match real-life recall situations\n- Put ALL to-be-learned information on the BACK of cards, never in prompts\n- Ensure technical accuracy and depth appropriate for the target level\n- Use domain-specific terminology correctly\n- Connect concepts to prerequisite knowledge\n\nPrioritize atomic simplicity over comprehensive single cards. Generate cards that test understanding through simple, direct recall.",
+        "instructions": "You are a world-class expert in {{ subject | default('the subject area') }} with deep pedagogical knowledge. \nYour role is to generate high-quality flashcards that demonstrate mastery of {{ subject | default('the subject') }} concepts.\n\nKey responsibilities:\n- Create ATOMIC cards: extremely short (1-9 words on back), break complex info into multiple simple cards\n- Use standardized, bland prompts without fancy formatting or unusual words\n- Design prompts that match real-life recall situations\n- Put ALL to-be-learned information on the BACK of cards, never in prompts\n- Ensure technical accuracy and depth appropriate for the target level\n- Use domain-specific terminology correctly\n- Connect concepts to prerequisite knowledge\n\nCard Types:\n- Basic cards (card_type='basic'): Standard Q&A format for concepts and facts\n- Cloze cards (card_type='cloze'): Fill-in-the-blank format using {{c1::answer}} syntax for code, syntax, formulas\n\nFor cloze cards, wrap the answer in {{c1::text}} format. Example: 'The vLLM class for inference is {{c1::LLM}}'\n\nPrioritize atomic simplicity over comprehensive single cards. Generate cards that test understanding through simple, direct recall.",
         "model": "{{ subject_expert_model }}",
         "temperature": 0.7,
         "timeout": 120.0,

ankigen_core/card_generator.py

@@ -128,10 +128,10 @@ async def orchestrate_card_generation(  # MODIFIED: Added async
                 subject=agent_subject,
                 num_cards=total_cards_needed,
                 difficulty="intermediate",
-                enable_quality_pipeline=True,
                 context=context,
                 library_name=library_name,
                 library_topic=library_topic,
+                generate_cloze=generate_cloze,
             )
 
             # Get token usage from session

pyproject.toml

@@ -10,7 +10,7 @@ authors = [
     { name = "Justin", email = "[email protected]" },
 ]
 readme = "README.md"
-requires-python = ">=3.10"
+requires-python = ">=3.12"
 dependencies = [
     "openai>=1.109.1",
     "openai-agents>=0.3.2",