mirror of
https://github.com/thecyberlearn/quantum-ai-v3.git
synced 2026-08-18 10:12:58 +00:00
MAJOR ARCHITECTURAL CHANGES: - Remove Agent/AgentCategory database models entirely - Update AgentExecution/ChatSession to use agent_slug instead of foreign keys - Eliminate hybrid complexity and AgentCompat workaround classes - Enhanced AgentFileService with production-ready caching CORE IMPROVEMENTS: - Pure file-based architecture eliminates database sync complexity - Enhanced caching: 5min in dev, 1hr in production, graceful fallback - Fixed validation logic for 0.0 price fields - Added database indexes for optimal query performance - Updated admin interface to work with new slug-based fields TECHNICAL BENEFITS: - Simplified agent execution without database record creation - Eliminated get_or_create_agent_db_record() complexity - Streamlined imports across core and agents apps - Better error handling and cache availability detection FILE CHANGES: - agents/models.py: Removed Agent/AgentCategory models, updated execution models - agents/services.py: Enhanced caching, improved validation, removed DB sync - agents/views.py: Updated execute_agent to use direct agent_data - agents/admin.py: Updated for slug-based fields - core/views.py: Updated to use AgentFileService instead of Agent model All 8 agents remain fully operational with significantly reduced codebase complexity. Migration applied successfully with proper defaults for existing data. 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <noreply@anthropic.com>
13 lines
459 B
Python
13 lines
459 B
Python
from rest_framework import serializers
|
|
from .models import AgentExecution
|
|
|
|
class AgentExecutionSerializer(serializers.ModelSerializer):
|
|
# Agent data comes from files via AgentFileService
|
|
|
|
class Meta:
|
|
model = AgentExecution
|
|
fields = [
|
|
'id', 'agent_slug', 'agent_name', 'input_data', 'output_data', 'status',
|
|
'fee_charged', 'error_message', 'execution_time',
|
|
'created_at', 'completed_at'
|
|
] |