django-cfg 1.1.82__py3-none-any.whl → 1.2.1__py3-none-any.whl

django_cfg/apps/knowbase/management/commands/knowbase_stats.py

@@ -0,0 +1,158 @@
+"""
+Knowledge Base statistics command.
+"""
+
+from django.core.management.base import BaseCommand
+from django.db.models import Count, Sum, Avg, Q, F, ExpressionWrapper
+from django.db.models.functions import Extract
+from django.contrib.auth import get_user_model
+from django.db import models
+
+User = get_user_model()
+
+
+class Command(BaseCommand):
+    """Display Knowledge Base statistics."""
+    
+    help = 'Display Knowledge Base usage statistics'
+    
+    def add_arguments(self, parser):
+        parser.add_argument(
+            '--user',
+            type=str,
+            help='Show statistics for specific user (username)',
+        )
+        parser.add_argument(
+            '--detailed',
+            action='store_true',
+            help='Show detailed statistics',
+        )
+    
+    def handle(self, *args, **options):
+        """Display statistics."""
+        
+        from ...models import Document, DocumentChunk, ChatSession, ChatMessage
+        
+        self.stdout.write(
+            self.style.SUCCESS('📊 Knowledge Base Statistics')
+        )
+        self.stdout.write('=' * 50)
+        
+        # Filter by user if specified
+        user_filter = {}
+        if options['user']:
+            try:
+                user = User.objects.get(username=options['user'])
+                user_filter['user'] = user
+                self.stdout.write(f"👤 User: {user.username}")
+                self.stdout.write('-' * 30)
+            except User.DoesNotExist:
+                self.stdout.write(
+                    self.style.ERROR(f"User '{options['user']}' not found")
+                )
+                return
+        
+        # Document statistics
+        # Calculate processing duration in database
+        processing_time_expr = ExpressionWrapper(
+            Extract(F('processing_completed_at') - F('processing_started_at'), 'epoch'),
+            output_field=models.FloatField()
+        )
+        
+        doc_stats = Document.objects.filter(**user_filter).aggregate(
+            total_docs=Count('id'),
+            completed_docs=Count('id', filter=Q(processing_status='completed')),
+            total_chunks=Sum('chunks_count'),
+            total_tokens=Sum('total_tokens'),
+            total_cost=Sum('total_cost_usd'),
+            avg_processing_time=Avg(processing_time_expr, filter=Q(
+                processing_started_at__isnull=False,
+                processing_completed_at__isnull=False
+            ))
+        )
+        
+        self.stdout.write("📄 Documents:")
+        self.stdout.write(f"  Total: {doc_stats['total_docs'] or 0}")
+        self.stdout.write(f"  Completed: {doc_stats['completed_docs'] or 0}")
+        self.stdout.write(f"  Success Rate: {((doc_stats['completed_docs'] or 0) / (doc_stats['total_docs'] or 1) * 100):.1f}%")
+        
+        self.stdout.write("\n📝 Content:")
+        self.stdout.write(f"  Total Chunks: {doc_stats['total_chunks'] or 0}")
+        self.stdout.write(f"  Total Tokens: {doc_stats['total_tokens'] or 0}")
+        
+        self.stdout.write("\n💰 Costs:")
+        self.stdout.write(f"  Total Cost: ${(doc_stats['total_cost'] or 0):.6f}")
+        
+        # Chat statistics
+        chat_stats = ChatSession.objects.filter(**user_filter).aggregate(
+            total_sessions=Count('id'),
+            active_sessions=Count('id', filter=Q(is_active=True)),
+            total_messages=Sum('messages_count'),
+            total_chat_tokens=Sum('total_tokens_used'),
+            total_chat_cost=Sum('total_cost_usd')
+        )
+        
+        self.stdout.write("\n💬 Chat:")
+        self.stdout.write(f"  Total Sessions: {chat_stats['total_sessions'] or 0}")
+        self.stdout.write(f"  Active Sessions: {chat_stats['active_sessions'] or 0}")
+        self.stdout.write(f"  Total Messages: {chat_stats['total_messages'] or 0}")
+        self.stdout.write(f"  Chat Tokens: {chat_stats['total_chat_tokens'] or 0}")
+        self.stdout.write(f"  Chat Cost: ${(chat_stats['total_chat_cost'] or 0):.6f}")
+        
+        # Detailed statistics
+        if options['detailed']:
+            self.show_detailed_stats(user_filter)
+    
+    def show_detailed_stats(self, user_filter):
+        """Show detailed statistics."""
+        from ...models import Document, ChatSession
+        
+        self.stdout.write("\n" + "=" * 50)
+        self.stdout.write("📈 Detailed Statistics")
+        self.stdout.write("=" * 50)
+        
+        # Processing status breakdown
+        status_counts = Document.objects.filter(**user_filter).values(
+            'processing_status'
+        ).annotate(count=Count('id'))
+        
+        self.stdout.write("\n📊 Document Status Breakdown:")
+        for status in status_counts:
+            self.stdout.write(f"  {status['processing_status']}: {status['count']}")
+        
+        # Model usage breakdown
+        model_counts = ChatSession.objects.filter(**user_filter).values(
+            'model_name'
+        ).annotate(count=Count('id'))
+        
+        self.stdout.write("\n🤖 Model Usage:")
+        for model in model_counts:
+            self.stdout.write(f"  {model['model_name']}: {model['count']} sessions")
+        
+        # Top documents by cost
+        top_docs = Document.objects.filter(
+            **user_filter
+        ).order_by('-total_cost_usd')[:5]
+        
+        self.stdout.write("\n💸 Most Expensive Documents:")
+        for doc in top_docs:
+            self.stdout.write(f"  {doc.title[:40]}...: ${doc.total_cost_usd:.6f}")
+        
+        # Recent activity
+        from django.utils import timezone
+        from datetime import timedelta
+        
+        week_ago = timezone.now() - timedelta(days=7)
+        recent_docs = Document.objects.filter(
+            **user_filter,
+            created_at__gte=week_ago
+        ).count()
+        
+        recent_sessions = ChatSession.objects.filter(
+            **user_filter,
+            created_at__gte=week_ago
+        ).count()
+        
+        self.stdout.write("\n📅 Recent Activity (Last 7 Days):")
+        self.stdout.write(f"  New Documents: {recent_docs}")
+        self.stdout.write(f"  New Chat Sessions: {recent_sessions}")

django_cfg/apps/knowbase/management/commands/setup_knowbase.py

@@ -0,0 +1,59 @@
+"""
+Setup command for Knowledge Base application.
+"""
+
+from django.core.management.base import BaseCommand
+from django.db import connection
+from django.core.management import call_command
+
+
+class Command(BaseCommand):
+    """Setup Knowledge Base with pgvector extension and initial data."""
+    
+    help = 'Setup Knowledge Base with pgvector extension and run migrations'
+    
+    def add_arguments(self, parser):
+        parser.add_argument(
+            '--skip-extensions',
+            action='store_true',
+            help='Skip creating PostgreSQL extensions',
+        )
+    
+    def handle(self, *args, **options):
+        """Execute the setup process."""
+        
+        self.stdout.write(
+            self.style.SUCCESS('🚀 Setting up Knowledge Base...')
+        )
+        
+        # Step 1: Create PostgreSQL extensions
+        if not options['skip_extensions']:
+            self.create_extensions()
+        
+        
+        self.stdout.write(
+            self.style.SUCCESS('✅ Knowledge Base setup completed!')
+        )
+    
+    def create_extensions(self):
+        """Create required PostgreSQL extensions."""
+        self.stdout.write('📦 Creating PostgreSQL extensions...')
+        
+        try:
+            with connection.cursor() as cursor:
+                # Create pgvector extension
+                cursor.execute("CREATE EXTENSION IF NOT EXISTS vector;")
+                self.stdout.write('  ✓ pgvector extension created')
+                
+                # Create pg_trgm for text search
+                cursor.execute("CREATE EXTENSION IF NOT EXISTS pg_trgm;")
+                self.stdout.write('  ✓ pg_trgm extension created')
+                
+        except Exception as e:
+            self.stdout.write(
+                self.style.ERROR(f'  ❌ Failed to create extensions: {e}')
+            )
+            self.stdout.write(
+                self.style.WARNING('  ⚠️  You may need to create extensions manually as superuser')
+            )
+    

django_cfg/apps/knowbase/managers/__init__.py

@@ -0,0 +1,22 @@
+"""
+Custom managers for knowledge base models.
+"""
+
+from .base import *
+from .document import *
+from .chat import *
+from .archive import *
+from .external_data import *
+
+__all__ = [
+    'BaseKnowbaseManager',
+    'DocumentManager',
+    'DocumentChunkManager',
+    'ChatSessionManager',
+    'ChatMessageManager',
+    'DocumentArchiveManager',
+    'ArchiveItemManager',
+    'ArchiveItemChunkManager',
+    'ExternalDataManager',
+    'ExternalDataChunkManager',
+]

django_cfg/apps/knowbase/managers/archive.py

@@ -0,0 +1,426 @@
+"""
+Archive managers for document archive models.
+"""
+
+import logging
+from django.db import models, transaction
+from django.db.models import Count, Sum, Avg, Q
+from typing import Optional, List, Dict, Any
+from django.contrib.auth import get_user_model
+
+User = get_user_model()
+
+
+class DocumentArchiveManager(models.Manager):
+    """Custom manager for DocumentArchive model."""
+    
+    def for_user(self, user):
+        """Explicitly filter by specific user."""
+        return self.get_queryset().filter(user=user)
+    
+    def all_users(self):
+        """Get unfiltered queryset (admin use)."""
+        return self.get_queryset()
+    
+    def processed(self):
+        """Get only processed archives."""
+        from ..models.base import ProcessingStatus
+        return self.get_queryset().filter(
+            processing_status=ProcessingStatus.COMPLETED
+        )
+    
+    def pending_processing(self):
+        """Get archives pending processing."""
+        from ..models.base import ProcessingStatus
+        return self.get_queryset().filter(
+            processing_status=ProcessingStatus.PENDING
+        )
+    
+    def failed_processing(self):
+        """Get archives that failed processing."""
+        from ..models.base import ProcessingStatus
+        return self.get_queryset().filter(
+            processing_status=ProcessingStatus.FAILED
+        )
+    
+    def by_content_hash(self, content_hash: str):
+        """Find archives by content hash."""
+        return self.get_queryset().filter(content_hash=content_hash)
+    
+    def by_archive_type(self, archive_type: str):
+        """Get archives by type (zip, tar, etc.)."""
+        return self.get_queryset().filter(archive_type=archive_type)
+    
+    def with_stats(self):
+        """Get archives with item and chunk statistics."""
+        return self.get_queryset().select_related('user').prefetch_related(
+            'items', 'chunks', 'categories'
+        )
+    
+    def get_processing_statistics(self, user=None) -> Dict[str, Any]:
+        """Get archive processing statistics."""
+        queryset = self.for_user(user) if user else self.get_queryset()
+        
+        # Get basic statistics
+        stats = queryset.aggregate(
+            total_archives=Count('id'),
+            processed_archives=Count('id', filter=Q(processing_status='completed')),
+            failed_archives=Count('id', filter=Q(processing_status='failed')),
+            total_items=Sum('total_items'),
+            total_chunks=Sum('total_chunks'),
+            total_tokens=Sum('total_tokens'),
+            total_cost=Sum('total_cost_usd'),
+            avg_processing_time=Avg('processing_duration_ms')
+        )
+        
+        # Calculate averages manually to avoid aggregate on aggregate error
+        if stats['total_archives'] and stats['total_archives'] > 0:
+            stats['avg_items_per_archive'] = (stats['total_items'] or 0) / stats['total_archives']
+            stats['avg_chunks_per_archive'] = (stats['total_chunks'] or 0) / stats['total_archives']
+        else:
+            stats['avg_items_per_archive'] = 0
+            stats['avg_chunks_per_archive'] = 0
+            
+        return stats
+    
+    def check_duplicate_before_save(self, user, title, file_size, exclude_id=None):
+        """Check for duplicate archive before saving. Returns (is_duplicate, existing_archive)."""
+        if not title or not file_size:
+            return False, None
+        
+        # Use all_users() to bypass user filtering
+        query = self.all_users().filter(
+            user=user, 
+            title=title,
+            file_size=file_size
+        )
+        
+        if exclude_id:
+            query = query.exclude(pk=exclude_id)
+        
+        existing_archive = query.first()
+        return existing_archive is not None, existing_archive
+    
+    def reprocess(self, archive_id: str) -> bool:
+        """
+        Reset and reprocess an archive.
+        
+        Args:
+            archive_id: ID of the archive to reprocess
+            
+        Returns:
+            bool: True if reprocessing was initiated successfully
+            
+        Raises:
+            ValueError: If archive not found or has no file
+        """
+        from ..models.base import ProcessingStatus
+        from ..tasks.archive_tasks import process_archive_task
+        
+        logger = logging.getLogger(__name__)
+        
+        try:
+            # Import the model directly to avoid queryset issues
+            from ..models.archive import DocumentArchive
+            # Use Django's default manager to avoid custom queryset issues
+            archive = DocumentArchive.objects.get(pk=archive_id)
+        except DocumentArchive.DoesNotExist:
+            raise ValueError(f"Archive with ID {archive_id} not found")
+        except Exception as e:
+            raise
+        
+        # Check if archive has a file
+        if not archive.archive_file:
+            raise ValueError("Archive has no file to process")
+        
+        # Check if archive is already being processed
+        if archive.processing_status == ProcessingStatus.PROCESSING:
+            raise ValueError(f"Archive {archive.id} is already being processed")
+        
+        # Set processing status immediately to prevent concurrent reprocessing
+        # Use select_for_update to prevent race conditions
+        with transaction.atomic():
+            archive = DocumentArchive.objects.select_for_update().get(pk=archive_id)
+            if archive.processing_status == ProcessingStatus.PROCESSING:
+                raise ValueError(f"Archive {archive.id} is already being processed by another process")
+            
+            archive.processing_status = ProcessingStatus.PROCESSING
+            archive.save(update_fields=['processing_status'])
+            logger.info(f"🔒 Locked archive {archive.id} for reprocessing")
+        
+        logger.info(f"🔄 Starting reprocessing for archive {archive.id} ({archive.title})")
+        
+        # Reset processing status and clear error
+        archive.processing_status = ProcessingStatus.PENDING
+        archive.processing_error = ""
+        archive.processing_duration_ms = 0
+        archive.processed_at = None
+        
+        # Clear existing items and chunks using Django ORM with proper transaction handling
+        from ..models.archive import ArchiveItem, ArchiveItemChunk
+        
+        # Count existing records first
+        items_count = ArchiveItem.objects.filter(archive=archive).count()
+        chunks_count = ArchiveItemChunk.objects.filter(item__archive=archive).count()
+        
+        logger.info(f"🗑️ Found {items_count} items and {chunks_count} chunks to delete")
+        
+        if items_count > 0 or chunks_count > 0:
+            # Delete in separate transaction to ensure complete removal before new processing
+            try:
+                with transaction.atomic():
+                    # Delete chunks first (foreign key dependency)
+                    chunks_deleted, _ = ArchiveItemChunk.objects.filter(item__archive=archive).delete()
+                    
+                    # Delete items
+                    items_deleted, _ = ArchiveItem.objects.filter(archive=archive).delete()
+                
+                # Verify deletion outside transaction with retry logic
+                import time
+                max_retries = 3
+                for retry in range(max_retries):
+                    remaining_items = ArchiveItem.objects.filter(archive=archive).count()
+                    remaining_chunks = ArchiveItemChunk.objects.filter(item__archive=archive).count()
+                    
+                    if remaining_items == 0 and remaining_chunks == 0:
+                        break
+                        
+                    if retry < max_retries - 1:
+                        logger.warning(f"⚠️ Retry {retry + 1}: Still {remaining_items} items and {remaining_chunks} chunks remaining, waiting...")
+                        time.sleep(0.2)
+                    else:
+                        logger.error(f"❌ Failed to delete all records after {max_retries} retries! {remaining_items} items and {remaining_chunks} chunks still remain")
+                        raise ValueError(f"Failed to clear existing archive data after {max_retries} retries. {remaining_items} items and {remaining_chunks} chunks still exist.")
+                
+                logger.info(f"🗑️ Successfully deleted {items_deleted} items and {chunks_deleted} chunks")
+                
+            except Exception as e:
+                logger.error(f"❌ Error during deletion: {e}")
+                # Reset processing status on error
+                archive.processing_status = ProcessingStatus.FAILED
+                archive.processing_error = f"Failed to clear existing data: {str(e)}"
+                archive.save(update_fields=['processing_status', 'processing_error'])
+                raise
+        else:
+            logger.info("🗑️ No existing records to delete")
+        
+        # Reset statistics
+        archive.total_items = 0
+        archive.processed_items = 0
+        archive.total_chunks = 0
+        archive.vectorized_chunks = 0
+        archive.total_cost_usd = 0.0
+        
+        archive.save()
+        logger.info(f"💾 Archive {archive.id} reset to PENDING status")
+        
+        # Trigger reprocessing directly via task
+        process_archive_task.send(str(archive.id), str(archive.user.id))
+        logger.info(f"🚀 Queued reprocessing task for archive {archive.id}")
+        
+        return True
+    
+    def get_vectorization_progress(self, archive_id):
+        """Get vectorization progress for an archive."""
+        try:
+            from ..models.archive import ArchiveItemChunk
+            from django.db.models import Count, Q
+            
+            # Count chunks using Django ORM
+            
+            # Get all chunks for this archive
+            chunks_qs = ArchiveItemChunk.objects.filter(archive_id=archive_id)
+            
+            total = chunks_qs.count()
+            
+            # Count vectorized chunks by checking if embedding has non-zero values
+            # We'll check if the first element is not 0.0 as a proxy for non-zero vector
+            vectorized = 0
+            if total > 0:
+                for chunk in chunks_qs.only('embedding'):
+                    if chunk.embedding is not None and len(chunk.embedding) > 0 and any(x != 0.0 for x in chunk.embedding):
+                        vectorized += 1
+            
+            return {
+                'total': total,
+                'vectorized': vectorized,
+                'percentage': round((vectorized / total * 100) if total > 0 else 0, 1)
+            }
+        except Exception as e:
+            import logging
+            logger = logging.getLogger(__name__)
+            logger.error(f"Error getting vectorization progress for archive {archive_id}: {e}")
+            return {
+                'total': 0,
+                'vectorized': 0,
+                'percentage': 0
+            }
+
+
+class ArchiveItemManager(models.Manager):
+    """Custom manager for ArchiveItem model."""
+    
+    def for_user(self, user):
+        """Explicitly filter by specific user."""
+        return self.get_queryset().filter(user=user)
+    
+    def all_users(self):
+        """Get unfiltered queryset (admin use)."""
+        return self.get_queryset()
+    
+    def for_archive(self, archive_id: str):
+        """Get items for specific archive."""
+        return self.get_queryset().filter(archive_id=archive_id)
+    
+    def by_content_type(self, content_type: str):
+        """Get items by content type."""
+        return self.get_queryset().filter(content_type=content_type)
+    
+    def by_language(self, language: str):
+        """Get items by programming language."""
+        return self.get_queryset().filter(language=language)
+    
+    def processable(self):
+        """Get only processable items."""
+        return self.get_queryset().filter(is_processable=True)
+    
+    def code_files(self):
+        """Get only code files."""
+        return self.get_queryset().filter(content_type='code')
+    
+    def document_files(self):
+        """Get only document files."""
+        return self.get_queryset().filter(content_type='document')
+    
+    def data_files(self):
+        """Get only data files."""
+        return self.get_queryset().filter(content_type='data')
+    
+    def with_chunks(self):
+        """Get items with their chunks."""
+        return self.get_queryset().prefetch_related('chunks')
+    
+    def get_content_type_distribution(self, archive_id: Optional[str] = None) -> Dict[str, int]:
+        """Get distribution of content types."""
+        queryset = self.get_queryset()
+        if archive_id:
+            queryset = queryset.filter(archive_id=archive_id)
+        
+        return dict(
+            queryset.values('content_type').annotate(
+                count=Count('id')
+            ).values_list('content_type', 'count')
+        )
+    
+    def get_language_distribution(self, archive_id: Optional[str] = None) -> Dict[str, int]:
+        """Get distribution of programming languages."""
+        queryset = self.get_queryset().filter(language__isnull=False).exclude(language='')
+        if archive_id:
+            queryset = queryset.filter(archive_id=archive_id)
+        
+        return dict(
+            queryset.values('language').annotate(
+                count=Count('id')
+            ).values_list('language', 'count')
+        )
+
+
+class ArchiveItemChunkManager(models.Manager):
+    """Custom manager for ArchiveItemChunk model."""
+    
+    def for_user(self, user):
+        """Explicitly filter by specific user."""
+        return self.get_queryset().filter(user=user)
+    
+    def all_users(self):
+        """Get unfiltered queryset (admin use)."""
+        return self.get_queryset()
+    
+    def for_archive(self, archive_id: str):
+        """Get chunks for specific archive."""
+        return self.get_queryset().filter(archive_id=archive_id)
+    
+    def for_item(self, item_id: str):
+        """Get chunks for specific item."""
+        return self.get_queryset().filter(item_id=item_id)
+    
+    def by_chunk_type(self, chunk_type: str):
+        """Get chunks by type."""
+        return self.get_queryset().filter(chunk_type=chunk_type)
+    
+    def vectorized(self):
+        """Get only vectorized chunks."""
+        return self.get_queryset().filter(embedding__isnull=False)
+    
+    def pending_vectorization(self):
+        """Get chunks pending vectorization."""
+        return self.get_queryset().filter(embedding__isnull=True)
+    
+    def by_content_type(self, content_type: str):
+        """Get chunks by parent item content type."""
+        return self.get_queryset().filter(item__content_type=content_type)
+    
+    def by_language(self, language: str):
+        """Get chunks by parent item language."""
+        return self.get_queryset().filter(item__language=language)
+    
+    def semantic_search(
+        self,
+        query_embedding: List[float],
+        limit: int = 5,
+        similarity_threshold: float = 0.7,
+        content_types: Optional[List[str]] = None,
+        languages: Optional[List[str]] = None,
+        chunk_types: Optional[List[str]] = None
+    ):
+        """Perform semantic search with advanced filtering."""
+        from pgvector.django import CosineDistance
+        
+        queryset = self.get_queryset().filter(embedding__isnull=False)
+        
+        # Apply filters
+        if content_types:
+            queryset = queryset.filter(item__content_type__in=content_types)
+        
+        if languages:
+            queryset = queryset.filter(item__language__in=languages)
+        
+        if chunk_types:
+            queryset = queryset.filter(chunk_type__in=chunk_types)
+        
+        return queryset.annotate(
+            similarity=1 - CosineDistance('embedding', query_embedding)
+        ).filter(
+            similarity__gte=similarity_threshold
+        ).order_by('-similarity')[:limit]
+    
+    def with_context(self):
+        """Get chunks with archive and item context."""
+        return self.get_queryset().select_related('archive', 'item')
+    
+    def get_vectorization_statistics(self, archive_id: Optional[str] = None) -> Dict[str, Any]:
+        """Get vectorization statistics."""
+        queryset = self.get_queryset()
+        if archive_id:
+            queryset = queryset.filter(archive_id=archive_id)
+        
+        return queryset.aggregate(
+            total_chunks=Count('id'),
+            vectorized_chunks=Count('id', filter=Q(embedding__isnull=False)),
+            total_tokens=Sum('token_count'),
+            total_cost=Sum('embedding_cost'),
+            avg_tokens_per_chunk=Avg('token_count'),
+            avg_cost_per_chunk=Avg('embedding_cost')
+        )
+    
+    def get_chunk_type_distribution(self, archive_id: Optional[str] = None) -> Dict[str, int]:
+        """Get distribution of chunk types."""
+        queryset = self.get_queryset()
+        if archive_id:
+            queryset = queryset.filter(archive_id=archive_id)
+        
+        return dict(
+            queryset.values('chunk_type').annotate(
+                count=Count('id')
+            ).values_list('chunk_type', 'count')
+        )