""" Analytics Console UI views """ from datetime import datetime from django.contrib.auth.decorators import login_required from django.core.cache import cache from django.core.paginator import Paginator from django.db.models import Avg, Count, F, Q, Value from django.db.models.functions import Concat from django.http import JsonResponse from django.shortcuts import render from apps.complaints.models import Complaint from apps.organizations.models import Department, Hospital from apps.px_action_center.models import PXAction from apps.surveys.models import SurveyInstance from apps.physicians.models import PhysicianMonthlyRating from .models import KPI, KPIValue from .services import UnifiedAnalyticsService, ExportService from .services.ai_analytics import ( ExecutiveSummaryGenerator, EarlyWarningSystem, ComplaintVolumeForecaster, SLABreachPredictor, ActionRecommendationEngine, ) from apps.core.decorators import block_source_user import json def serialize_queryset_values(queryset): """Properly serialize QuerySet values to JSON string.""" if queryset is None: return "[]" data = list(queryset) result = [] for item in data: row = {} for key, value in item.items(): # Convert UUID to string if hasattr(value, "hex"): # UUID object row[key] = str(value) # Convert Python None to JavaScript null elif value is None: row[key] = None else: row[key] = value result.append(row) return json.dumps(result, default=str) @block_source_user @login_required def analytics_dashboard(request): """ Analytics dashboard with KPIs and charts. Comprehensive dashboard showing: - KPI cards with current values for Complaints, Actions, Surveys, Feedback - Trend charts - Department rankings - Source distribution - Status breakdown """ from apps.feedback.models import Feedback from django.utils import timezone from datetime import timedelta from django.db.models.functions import TruncDate, TruncMonth user = request.user # Build cache key based on user and hospital cache_key = f"analytics_dashboard_{user.id}_{request.GET.get('hospital', 'all')}" cached = cache.get(cache_key) if cached: return render(request, "analytics/dashboard.html", cached) # Get hospital filter hospital_filter = request.GET.get("hospital") if hospital_filter: hospital = Hospital.objects.filter(id=hospital_filter).first() elif user.is_px_admin() and hasattr(request, "tenant_hospital") and request.tenant_hospital: hospital = request.tenant_hospital elif user.hospital: hospital = user.hospital else: hospital = None # Base querysets complaints_queryset = Complaint.objects.all() actions_queryset = PXAction.objects.all() surveys_queryset = SurveyInstance.objects.filter(status="completed") feedback_queryset = Feedback.objects.all() if hospital: complaints_queryset = complaints_queryset.filter(hospital=hospital) actions_queryset = actions_queryset.filter(hospital=hospital) surveys_queryset = surveys_queryset.filter(survey_template__hospital=hospital) feedback_queryset = feedback_queryset.filter(hospital=hospital) # ============ COMPLAINTS KPIs ============ # Single query for all status counts status_counts_qs = complaints_queryset.values("status").annotate(count=Count("id")) status_map = {item["status"]: item["count"] for item in status_counts_qs} total_complaints = sum(status_map.values()) open_complaints = status_map.get("open", 0) in_progress_complaints = status_map.get("in_progress", 0) resolved_complaints = status_map.get("resolved", 0) closed_complaints = status_map.get("closed", 0) overdue_complaints = complaints_queryset.filter(is_overdue=True).count() # Complaint source types (internal vs external) — single query source_type_counts = complaints_queryset.values("complaint_source_type").annotate(count=Count("id")) source_type_map = {item["complaint_source_type"]: item["count"] for item in source_type_counts} internal_complaints = source_type_map.get("internal", 0) external_complaints = source_type_map.get("external", 0) # Complaint sources (by PXSource name) complaint_sources = ( complaints_queryset.filter(source__isnull=False) .values("source__name_en") .annotate(count=Count("id")) .order_by("-count")[:6] ) # Complaint domains (Level 1) top_domains = ( complaints_queryset.filter(domain__isnull=False) .values("domain__name_en") .annotate(count=Count("id")) .order_by("-count")[:5] ) # Complaint categories (Level 2) top_categories = ( complaints_queryset.filter(category__isnull=False) .values("category__name_en") .annotate(count=Count("id")) .order_by("-count")[:5] ) # Complaint severity — single query severity_counts_qs = complaints_queryset.values("severity").annotate(count=Count("id")) severity_map = {item["severity"]: item["count"] for item in severity_counts_qs} critical_complaints = severity_map.get("critical", 0) high_complaints = severity_map.get("high", 0) medium_complaints = severity_map.get("medium", 0) low_complaints = severity_map.get("low", 0) # Severity breakdown for JSON severity_breakdown = severity_counts_qs.order_by("-count") # Status breakdown status_breakdown = status_counts_qs.order_by("-count") # ============ ACTIONS KPIs ============ action_status_counts = actions_queryset.values("status").annotate(count=Count("id")) action_status_map = {item["status"]: item["count"] for item in action_status_counts} total_actions = sum(action_status_map.values()) open_actions = action_status_map.get("open", 0) in_progress_actions = action_status_map.get("in_progress", 0) approved_actions = action_status_map.get("approved", 0) closed_actions = action_status_map.get("closed", 0) pending_actions = action_status_map.get("pending_approval", 0) overdue_actions = actions_queryset.filter(is_overdue=True).count() # Action sources action_sources = ( actions_queryset.filter(source_type__isnull=False) .values("source_type") .annotate(count=Count("id")) .order_by("-count")[:6] ) # Action categories - build explicit counts action_categories = ( actions_queryset.exclude(category="").values("category").annotate(count=Count("id")).order_by("-count")[:5] ) action_category_map = {item["category"]: item["count"] for item in action_categories} training_actions = action_category_map.get("training", 0) process_actions = action_category_map.get("process_improvement", 0) policy_actions = action_category_map.get("policy", 0) facility_actions = action_category_map.get("facility", 0) other_actions = action_category_map.get("other", 0) # ============ SURVEYS KPIs ============ total_surveys = surveys_queryset.count() avg_survey_score = surveys_queryset.aggregate(avg=Avg("total_score"))["avg"] or 0 negative_surveys = surveys_queryset.filter(is_negative=True).count() # Survey completion rate — single query all_surveys = SurveyInstance.objects.all() if hospital: all_surveys = all_surveys.filter(survey_template__hospital=hospital) survey_status_counts = all_surveys.values("status").annotate(count=Count("id")) survey_status_map = {item["status"]: item["count"] for item in survey_status_counts} total_sent = sum(survey_status_map.values()) completed_surveys = survey_status_map.get("completed", 0) completion_rate = (completed_surveys / total_sent * 100) if total_sent > 0 else 0 # Survey types survey_types = all_surveys.values("survey_template__survey_type").annotate(count=Count("id")).order_by("-count")[:5] # ============ FEEDBACK KPIs ============ feedback_type_counts = feedback_queryset.values("feedback_type").annotate(count=Count("id")) feedback_type_map = {item["feedback_type"]: item["count"] for item in feedback_type_counts} total_feedback = sum(feedback_type_map.values()) compliments = feedback_type_map.get("compliment", 0) suggestions = feedback_type_map.get("suggestion", 0) # Sentiment analysis sentiment_breakdown = feedback_queryset.values("sentiment").annotate(count=Count("id")).order_by("-count") # Feedback categories feedback_categories = feedback_queryset.values("category").annotate(count=Count("id")).order_by("-count")[:5] # Average rating avg_rating = feedback_queryset.filter(rating__isnull=False).aggregate(avg=Avg("rating"))["avg"] or 0 # ============ TRENDS (Last 30 days) ============ thirty_days_ago = timezone.now() - timedelta(days=30) # Complaint trends complaint_trend = ( complaints_queryset.filter(created_at__gte=thirty_days_ago) .annotate(day=TruncDate("created_at")) .values("day") .annotate(count=Count("id")) .order_by("day") ) # Survey score trend - last 6 months for chart six_months_ago = timezone.now() - timedelta(days=180) survey_score_trend_6m = ( surveys_queryset.filter(completed_at__gte=six_months_ago) .annotate(month=TruncMonth("completed_at")) .values("month") .annotate(avg_score=Avg("total_score")) .order_by("month") ) # Build survey trend array for last 6 months (pad with zeros if missing) from calendar import month_name now = timezone.now() survey_trend_values = [] survey_trend_labels = [] for i in range(5, -1, -1): target_month = now.month - i target_year = now.year while target_month <= 0: target_month += 12 target_year -= 1 survey_trend_labels.append(month_name[target_month][:3]) # Find matching data point found = None for item in survey_score_trend_6m: if item["month"].month == target_month and item["month"].year == target_year: found = round(item["avg_score"], 2) if item["avg_score"] else 0 break survey_trend_values.append(found if found is not None else 0) # ============ DEPARTMENT RANKINGS ============ dept_base_qs = Department.objects.filter(status="active") if hospital: dept_base_qs = dept_base_qs.filter(hospital=hospital) department_rankings = ( dept_base_qs.annotate( avg_score=Avg( "journey_instances__surveys__total_score", filter=Q(journey_instances__surveys__status="completed") ), survey_count=Count("journey_instances__surveys", filter=Q(journey_instances__surveys__status="completed")), complaint_count=Count("complaints"), resolved_count=Count("complaints", filter=Q(complaints__status__in=["resolved", "closed"])), action_count=Count("px_actions"), ) .filter(survey_count__gt=0) .order_by("-avg_score")[:7] ) # Build department_stats list — all data now comes from annotations, zero extra queries department_stats = [] for dept in department_rankings: resolution_rate = ( round((dept.resolved_count / dept.complaint_count * 100), 1) if dept.complaint_count > 0 else 0 ) department_stats.append( { "name_en": dept.name_en if hasattr(dept, "name_en") else str(dept), "name_ar": dept.name_ar if hasattr(dept, "name_ar") else (dept.name_en if hasattr(dept, "name_en") else str(dept)), "complaints": dept.complaint_count, "actions": dept.action_count, "survey_avg": round(dept.avg_score, 2) if dept.avg_score else 0, "resolution_rate": resolution_rate, } ) # ============ TIME-BASED CALCULATIONS ============ # Average resolution time (complaints) resolved_with_time = complaints_queryset.filter( status__in=["resolved", "closed"], resolved_at__isnull=False, created_at__isnull=False ) if resolved_with_time.exists(): avg_resolution_hours = resolved_with_time.annotate( resolution_time=F("resolved_at") - F("created_at") ).aggregate(avg=Avg("resolution_time"))["avg"] if avg_resolution_hours: avg_resolution_hours = avg_resolution_hours.total_seconds() / 3600 else: avg_resolution_hours = 0 else: avg_resolution_hours = 0 # Average action completion time closed_actions_with_time = actions_queryset.filter( status="closed", closed_at__isnull=False, created_at__isnull=False ) if closed_actions_with_time.exists(): avg_action_days = closed_actions_with_time.annotate(completion_time=F("closed_at") - F("created_at")).aggregate( avg=Avg("completion_time") )["avg"] if avg_action_days: avg_action_days = avg_action_days.days else: avg_action_days = 0 else: avg_action_days = 0 # ============ SLA COMPLIANCE ============ total_with_sla = complaints_queryset.filter(due_at__isnull=False).count() resolved_within_sla = complaints_queryset.filter( status__in=["resolved", "closed"], resolved_at__lte=F("due_at") ).count() sla_compliance = (resolved_within_sla / total_with_sla * 100) if total_with_sla > 0 else 0 # ============ NPS CALCULATION ============ # NPS = % Promoters (9-10) - % Detractors (0-6) nps_surveys = surveys_queryset.filter(survey_template__survey_type="nps", total_score__isnull=False) if nps_surveys.exists(): promoters = nps_surveys.filter(total_score__gte=9).count() detractors = nps_surveys.filter(total_score__lte=6).count() total_nps = nps_surveys.count() nps_score = ((promoters - detractors) / total_nps * 100) if total_nps > 0 else 0 else: nps_score = 0 kpis = { "total_complaints": total_complaints, "open_complaints": open_complaints, "in_progress_complaints": in_progress_complaints, "resolved_complaints": resolved_complaints, "closed_complaints": closed_complaints, "overdue_complaints": overdue_complaints, "internal_complaints": internal_complaints, "external_complaints": external_complaints, "critical_complaints": critical_complaints, "high_complaints": high_complaints, "medium_complaints": medium_complaints, "low_complaints": low_complaints, "avg_resolution_hours": round(avg_resolution_hours, 1), "sla_compliance": round(sla_compliance, 1), "total_actions": total_actions, "open_actions": open_actions, "in_progress_actions": in_progress_actions, "approved_actions": approved_actions, "closed_actions": closed_actions, "pending_actions": pending_actions, "overdue_actions": overdue_actions, "training_actions": training_actions, "process_actions": process_actions, "policy_actions": policy_actions, "facility_actions": facility_actions, "other_actions": other_actions, "avg_action_days": round(avg_action_days, 1), "total_surveys": total_surveys, "avg_survey_score": round(avg_survey_score, 2), "nps_score": round(nps_score, 1), "negative_surveys": negative_surveys, "completion_rate": round(completion_rate, 1), "total_feedback": total_feedback, "compliments": compliments, "suggestions": suggestions, "avg_rating": round(avg_rating, 2), "survey_trend_1": survey_trend_values[0] if len(survey_trend_values) > 0 else 0, "survey_trend_2": survey_trend_values[1] if len(survey_trend_values) > 1 else 0, "survey_trend_3": survey_trend_values[2] if len(survey_trend_values) > 2 else 0, "survey_trend_4": survey_trend_values[3] if len(survey_trend_values) > 3 else 0, "survey_trend_5": survey_trend_values[4] if len(survey_trend_values) > 4 else 0, "survey_trend_6": survey_trend_values[5] if len(survey_trend_values) > 5 else 0, } # ============ AI-POWERED ANALYTICS ============ hospital_id = str(hospital.id) if hospital else None # Trigger async Celery tasks to refresh cache in background from .tasks import ( generate_executive_summary_task, generate_action_recommendations_task, ) generate_executive_summary_task.delay(user_id=str(user.id), hospital_id=hospital_id, period="30d") generate_action_recommendations_task.delay(user_id=str(user.id), hospital_id=hospital_id) # Read AI analytics from cache ONLY (populated hourly by Celery beat). # If cache miss, return lightweight placeholders so page loads instantly. exec_summary = cache.get(ExecutiveSummaryGenerator._cache_key(hospital_id, None, "30d")) if not exec_summary: exec_summary = { "summary_en": "Executive summary is being computed in the background...", "summary_ar": "جاري حساب الملخص التنفيذي في الخلفية...", "key_findings_en": [], "key_findings_ar": [], "recommendations_en": [], "recommendations_ar": [], "risk_level": "medium", "_data": {}, } early_warnings = cache.get(EarlyWarningSystem._cache_key(hospital_id, 5)) if early_warnings is None: early_warnings = [] complaint_forecast = cache.get(ComplaintVolumeForecaster._cache_key(hospital_id, 30)) if not complaint_forecast: complaint_forecast = ComplaintVolumeForecaster._insufficient_data_response(30) sla_breach_predictions = cache.get(SLABreachPredictor._cache_key(hospital_id, 10)) if sla_breach_predictions is None: sla_breach_predictions = [] action_recommendations = cache.get(ActionRecommendationEngine._cache_key(hospital_id, None, 5)) if not action_recommendations: action_recommendations = ActionRecommendationEngine._no_data_response() context = { "kpis": kpis, "selected_hospital": hospital, "complaint_sources": serialize_queryset_values(complaint_sources), "top_domains": serialize_queryset_values(top_domains), "top_categories": serialize_queryset_values(top_categories), "severity_breakdown": serialize_queryset_values(severity_breakdown), "status_breakdown": serialize_queryset_values(status_breakdown), "complaint_trend": serialize_queryset_values(complaint_trend), "action_sources": serialize_queryset_values(action_sources), "action_categories": serialize_queryset_values(action_categories), "survey_types": serialize_queryset_values(survey_types), "survey_score_trend": serialize_queryset_values(survey_score_trend_6m), "sentiment_breakdown": serialize_queryset_values(sentiment_breakdown), "feedback_categories": serialize_queryset_values(feedback_categories), "department_rankings": department_rankings, "department_stats": department_stats, "survey_trend_labels": json.dumps(survey_trend_labels), # AI-powered features "exec_summary": exec_summary, "early_warnings": early_warnings, "complaint_forecast": complaint_forecast, "sla_breach_predictions": sla_breach_predictions, "action_recommendations": action_recommendations, } # Cache the full dashboard context for 5 minutes so next load is instant cache.set(cache_key, context, 300) return render(request, "analytics/dashboard.html", context) @block_source_user @login_required def refresh_ai_analytics(request): """ API endpoint: Trigger async AI analytics refresh and return status. POST to trigger, GET to check if cache is fresh. """ if request.method == "POST": from .tasks import ( generate_executive_summary_task, generate_action_recommendations_task, precompute_dashboard_cache_task, ) hospital_id = request.POST.get("hospital") or request.GET.get("hospital") user = request.user # Trigger async tasks generate_executive_summary_task.delay( user_id=str(user.id), hospital_id=hospital_id, period="30d", force_refresh=True ) generate_action_recommendations_task.delay(user_id=str(user.id), hospital_id=hospital_id) # Also clear caches so next page load triggers fresh computation from apps.analytics.services.ai_analytics import ( ExecutiveSummaryGenerator, ActionRecommendationEngine, ) cache.delete(ExecutiveSummaryGenerator._cache_key(hospital_id, None, "30d")) cache.delete(ActionRecommendationEngine._cache_key(hospital_id, None, 5)) return JsonResponse( {"status": "triggered", "message": "AI analytics refresh queued. Results will be available in ~30 seconds."} ) # GET — check cache freshness hospital_id = request.GET.get("hospital") or ( str(request.tenant_hospital.id) if hasattr(request, "tenant_hospital") and request.tenant_hospital else None ) user = request.user from apps.analytics.services.ai_analytics import ( ExecutiveSummaryGenerator, ActionRecommendationEngine, ) summary_cached = cache.get(ExecutiveSummaryGenerator._cache_key(hospital_id, None, "30d")) recommendations_cached = cache.get(ActionRecommendationEngine._cache_key(hospital_id, None, 5)) return JsonResponse( { "cached": { "executive_summary": summary_cached is not None, "action_recommendations": recommendations_cached is not None, }, "risk_level": summary_cached.get("risk_level", "unknown") if summary_cached else None, } ) @block_source_user @login_required def refresh_dashboard_cache(request): """ API endpoint: Trigger dashboard cache refresh on demand. POST to trigger refresh, returns immediately with task status. """ if request.method != "POST": return JsonResponse({"error": "POST method required"}, status=405) from .tasks import precompute_dashboard_cache_task user = request.user # Trigger async cache refresh task = precompute_dashboard_cache_task.delay() # Clear user's dashboard cache so next load gets fresh data cache.delete(f"analytics_dashboard_{user.id}_all") if hasattr(request, "tenant_hospital") and request.tenant_hospital: cache.delete(f"analytics_dashboard_{user.id}_{request.tenant_hospital.id}") return JsonResponse( { "status": "triggered", "message": "Dashboard cache refresh queued. Please reload the page in a few seconds.", "task_id": str(task.id), } ) @block_source_user @login_required def kpi_list(request): """KPI definitions list view""" queryset = KPI.objects.all() # Apply filters category_filter = request.GET.get("category") if category_filter: queryset = queryset.filter(category=category_filter) is_active = request.GET.get("is_active") if is_active == "true": queryset = queryset.filter(is_active=True) elif is_active == "false": queryset = queryset.filter(is_active=False) # Ordering queryset = queryset.order_by("category", "name") # Pagination page_size = int(request.GET.get("page_size", 25)) paginator = Paginator(queryset, page_size) page_number = request.GET.get("page", 1) page_obj = paginator.get_page(page_number) context = { "page_obj": page_obj, "kpis": page_obj.object_list, "filters": request.GET, } return render(request, "analytics/kpi_list.html", context) @block_source_user @login_required def command_center(request): """ PX Command Center - Unified Dashboard Comprehensive dashboard showing all PX360 metrics: - Complaints, Surveys, Actions KPIs - Interactive charts with ApexCharts - Department and Physician rankings - Export to Excel/PDF """ user = request.user # Get filter parameters filters = { "date_range": request.GET.get("date_range", "30d"), "hospital": request.GET.get("hospital", ""), "department": request.GET.get("department", ""), "kpi_category": request.GET.get("kpi_category", ""), "custom_start": request.GET.get("custom_start", ""), "custom_end": request.GET.get("custom_end", ""), } # Get hospitals for filter hospitals = Hospital.objects.filter(status="active") if not user.is_px_admin() and user.hospital: hospitals = hospitals.filter(id=user.hospital.id) # Get departments for filter departments = Department.objects.filter(status="active") if filters.get("hospital"): departments = departments.filter(hospital_id=filters["hospital"]) elif not user.is_px_admin() and user.hospital: departments = departments.filter(hospital=user.hospital) # Get initial KPIs custom_start = None custom_end = None if filters["custom_start"] and filters["custom_end"]: custom_start = datetime.strptime(filters["custom_start"], "%Y-%m-%d") custom_end = datetime.strptime(filters["custom_end"], "%Y-%m-%d") kpis = UnifiedAnalyticsService.get_all_kpis( user=user, date_range=filters["date_range"], hospital_id=filters["hospital"] if filters["hospital"] else None, department_id=filters["department"] if filters["department"] else None, custom_start=custom_start, custom_end=custom_end, ) # Initial AI data for server-side render from .services.ai_analytics import ( ExecutiveSummaryGenerator, EarlyWarningSystem, ComplaintVolumeForecaster, SLABreachPredictor, ActionRecommendationEngine, ) hospital_id = filters["hospital"] if filters["hospital"] else None department_id = filters["department"] if filters["department"] else None if not hospital_id and user.is_px_admin(): tenant = getattr(request, "tenant_hospital", None) if tenant: hospital_id = str(tenant.id) # Trigger async refresh from .tasks import generate_executive_summary_task, generate_action_recommendations_task generate_executive_summary_task.delay( user_id=str(user.id), hospital_id=hospital_id, department_id=department_id, period=filters["date_range"] ) generate_action_recommendations_task.delay( user_id=str(user.id), hospital_id=hospital_id, department_id=department_id ) # Read AI analytics from cache ONLY (never block request with LLM calls) exec_summary = cache.get(ExecutiveSummaryGenerator._cache_key(hospital_id, department_id, filters["date_range"])) if not exec_summary: exec_summary = { "summary_en": "Executive summary is being computed in the background...", "summary_ar": "جاري حساب الملخص التنفيذي في الخلفية...", "key_findings_en": [], "key_findings_ar": [], "recommendations_en": [], "recommendations_ar": [], "risk_level": "medium", "_data": {}, } early_warnings = cache.get(EarlyWarningSystem._cache_key(hospital_id, 5)) if early_warnings is None: early_warnings = [] complaint_forecast = cache.get(ComplaintVolumeForecaster._cache_key(hospital_id, 30)) if not complaint_forecast: complaint_forecast = ComplaintVolumeForecaster._insufficient_data_response(30) sla_breach_predictions = cache.get(SLABreachPredictor._cache_key(hospital_id, 10)) if sla_breach_predictions is None: sla_breach_predictions = [] action_recommendations = cache.get(ActionRecommendationEngine._cache_key(hospital_id, department_id, 5)) if not action_recommendations: action_recommendations = ActionRecommendationEngine._no_data_response() context = { "filters": filters, "departments": departments, "kpis": kpis, "exec_summary": exec_summary, "early_warnings": early_warnings, "complaint_forecast": complaint_forecast, "sla_breach_predictions": sla_breach_predictions, "action_recommendations": action_recommendations, } return render(request, "analytics/command_center.html", context) @block_source_user @login_required def command_center_api(request): """ API endpoint for Command Center data Returns JSON data for KPIs, charts, and tables based on filters. Used by JavaScript to dynamically update dashboard. """ if request.method != "GET": return JsonResponse({"error": "Only GET requests allowed"}, status=405) user = request.user # Get filter parameters date_range = request.GET.get("date_range", "30d") hospital_id = request.GET.get("hospital") department_id = request.GET.get("department") kpi_category = request.GET.get("kpi_category") custom_start_str = request.GET.get("custom_start") custom_end_str = request.GET.get("custom_end") # Parse custom dates custom_start = None custom_end = None if custom_start_str and custom_end_str: try: custom_start = datetime.strptime(custom_start_str, "%Y-%m-%d") custom_end = datetime.strptime(custom_end_str, "%Y-%m-%d") except ValueError: pass # Handle hospital_id (can be integer or UUID string) hospital_id = hospital_id if hospital_id else None # Handle department_id (UUID string) department_id = department_id if department_id else None if not hospital_id and user.is_px_admin(): tenant = getattr(request, "tenant_hospital", None) if tenant: hospital_id = str(tenant.id) # Get KPIs kpis = UnifiedAnalyticsService.get_all_kpis( user=user, date_range=date_range, hospital_id=hospital_id, department_id=department_id, kpi_category=kpi_category, custom_start=custom_start, custom_end=custom_end, ) # Ensure numeric KPIs are proper Python types for JSON serialization numeric_kpis = [ "total_complaints", "open_complaints", "overdue_complaints", "high_severity_complaints", "resolved_complaints", "total_actions", "open_actions", "overdue_actions", "escalated_actions", "resolved_actions", "total_surveys", "negative_surveys", "avg_survey_score", "negative_social_mentions", "low_call_ratings", "total_sentiment_analyses", ] for key in numeric_kpis: if key in kpis: value = kpis[key] if value is None: kpis[key] = 0.0 if key == "avg_survey_score" else 0 elif isinstance(value, (int, float)): # Already a number - ensure floats for specific fields if key == "avg_survey_score": kpis[key] = float(value) else: # Try to convert to number try: kpis[key] = float(value) except (ValueError, TypeError): kpis[key] = 0.0 if key == "avg_survey_score" else 0 # Handle nested trend data if "complaints_trend" in kpis and isinstance(kpis["complaints_trend"], dict): trend = kpis["complaints_trend"] trend["current"] = int(trend.get("current", 0)) trend["previous"] = int(trend.get("previous", 0)) trend["percentage_change"] = float(trend.get("percentage_change", 0)) # Get chart data chart_types = [ "complaints_trend", "complaints_by_category", "survey_satisfaction_trend", "survey_distribution", "department_performance", "physician_leaderboard", ] charts = {} for chart_type in chart_types: charts[chart_type] = UnifiedAnalyticsService.get_chart_data( user=user, chart_type=chart_type, date_range=date_range, hospital_id=hospital_id, department_id=department_id, custom_start=custom_start, custom_end=custom_end, ) # Get table data tables = {} # Overdue complaints table complaints_qs = Complaint.objects.filter(is_overdue=True) if hospital_id: complaints_qs = complaints_qs.filter(hospital_id=hospital_id) if department_id: complaints_qs = complaints_qs.filter(department_id=department_id) # Apply role-based filtering if not user.is_px_admin() and user.hospital: complaints_qs = complaints_qs.filter(hospital=user.hospital) if user.is_department_manager() and user.department: complaints_qs = complaints_qs.filter(department=user.department) tables["overdue_complaints"] = list( complaints_qs.select_related("hospital", "department", "patient", "source") .order_by("due_at")[:20] .values( "id", "title", "severity", "due_at", "complaint_source_type", hospital_name=F("hospital__name"), department_name=F("department__name"), patient_full_name=Concat("patient__first_name", Value(" "), "patient__last_name"), source_name=F("source__name_en"), assigned_to_full_name=Concat("assigned_to__first_name", Value(" "), "assigned_to__last_name"), ) ) # Physician leaderboard table physician_data = charts.get("physician_leaderboard", {}).get("metadata", []) tables["physician_leaderboard"] = [ { "physician_id": p["physician_id"], "name": p["name"], "specialization": p["specialization"], "department": p["department"], "rating": float(p["rating"]) if p["rating"] is not None else 0.0, "surveys": int(p["surveys"]) if p["surveys"] is not None else 0, "positive": int(p["positive"]) if p["positive"] is not None else 0, "neutral": int(p["neutral"]) if p["neutral"] is not None else 0, "negative": int(p["negative"]) if p["negative"] is not None else 0, } for p in physician_data ] # ============ AI-POWERED ANALYTICS ============ from .services.ai_analytics import ( ExecutiveSummaryGenerator, EarlyWarningSystem, ComplaintVolumeForecaster, SLABreachPredictor, ActionRecommendationEngine, ) # Trigger async Celery tasks for background refresh from .tasks import ( generate_executive_summary_task, generate_action_recommendations_task, ) generate_executive_summary_task.delay( user_id=str(user.id), hospital_id=hospital_id, department_id=department_id, period=date_range.replace("d", "") if date_range.endswith("d") else "30d", ) generate_action_recommendations_task.delay( user_id=str(user.id), hospital_id=hospital_id, department_id=department_id ) # AI features — read from cache ONLY (never block request with LLM calls) exec_summary = cache.get(ExecutiveSummaryGenerator._cache_key(hospital_id, department_id, date_range)) if not exec_summary: exec_summary = { "summary_en": "Executive summary is being computed in the background...", "summary_ar": "جاري حساب الملخص التنفيذي في الخلفية...", "key_findings_en": [], "key_findings_ar": [], "recommendations_en": [], "recommendations_ar": [], "risk_level": "medium", "_data": {}, } early_warnings = cache.get(EarlyWarningSystem._cache_key(hospital_id, 5)) if early_warnings is None: early_warnings = [] complaint_forecast = cache.get(ComplaintVolumeForecaster._cache_key(hospital_id, 30)) if not complaint_forecast: complaint_forecast = ComplaintVolumeForecaster._insufficient_data_response(30) sla_breach_predictions = cache.get(SLABreachPredictor._cache_key(hospital_id, 10)) if sla_breach_predictions is None: sla_breach_predictions = [] action_recommendations = cache.get(ActionRecommendationEngine._cache_key(hospital_id, department_id, 5)) if not action_recommendations: action_recommendations = ActionRecommendationEngine._no_data_response() ai_data = { "executive_summary": exec_summary, "early_warnings": early_warnings, "complaint_forecast": complaint_forecast, "sla_breach_predictions": sla_breach_predictions, "action_recommendations": action_recommendations, } return JsonResponse({"kpis": kpis, "charts": charts, "tables": tables, "ai": ai_data}) @block_source_user @login_required def export_command_center(request, export_format): """ Export Command Center data to Excel or PDF Args: export_format: 'excel' or 'pdf' Returns: HttpResponse with file download """ if export_format not in ["excel", "pdf"]: return JsonResponse({"error": "Invalid export format"}, status=400) user = request.user # Get filter parameters date_range = request.GET.get("date_range", "30d") hospital_id = request.GET.get("hospital") department_id = request.GET.get("department") kpi_category = request.GET.get("kpi_category") custom_start_str = request.GET.get("custom_start") custom_end_str = request.GET.get("custom_end") # Parse custom dates custom_start = None custom_end = None if custom_start_str and custom_end_str: try: custom_start = datetime.strptime(custom_start_str, "%Y-%m-%d") custom_end = datetime.strptime(custom_end_str, "%Y-%m-%d") except ValueError: pass # Handle hospital_id and department_id (can be integer or UUID string) hospital_id = hospital_id if hospital_id else None department_id = department_id if department_id else None # Get all data kpis = UnifiedAnalyticsService.get_all_kpis( user=user, date_range=date_range, hospital_id=hospital_id, department_id=department_id, kpi_category=kpi_category, custom_start=custom_start, custom_end=custom_end, ) chart_types = [ "complaints_trend", "complaints_by_category", "survey_satisfaction_trend", "survey_distribution", "department_performance", "physician_leaderboard", ] charts = {} for chart_type in chart_types: charts[chart_type] = UnifiedAnalyticsService.get_chart_data( user=user, chart_type=chart_type, date_range=date_range, hospital_id=hospital_id, department_id=department_id, custom_start=custom_start, custom_end=custom_end, ) # Get table data tables = {} # Overdue complaints complaints_qs = Complaint.objects.filter(is_overdue=True) if hospital_id: complaints_qs = complaints_qs.filter(hospital_id=hospital_id) if department_id: complaints_qs = complaints_qs.filter(department_id=department_id) if not user.is_px_admin() and user.hospital: complaints_qs = complaints_qs.filter(hospital=user.hospital) if user.is_department_manager() and user.department: complaints_qs = complaints_qs.filter(department=user.department) tables["overdue_complaints"] = { "headers": ["ID", "Title", "Patient", "Severity", "Hospital", "Department", "Due Date"], "rows": list( complaints_qs.select_related("hospital", "department", "patient") .order_by("due_at")[:100] .annotate( patient_full_name=Concat("patient__first_name", Value(" "), "patient__last_name"), hospital_name=F("hospital__name"), department_name=F("department__name"), ) .values_list("id", "title", "patient_full_name", "severity", "hospital_name", "department_name", "due_at") ), } # Physician leaderboard physician_data = charts.get("physician_leaderboard", {}).get("metadata", []) tables["physician_leaderboard"] = { "headers": ["Name", "Specialization", "Department", "Rating", "Surveys", "Positive", "Neutral", "Negative"], "rows": [ [ p["name"], p["specialization"], p["department"], str(p["rating"]), str(p["surveys"]), str(p["positive"]), str(p["neutral"]), str(p["negative"]), ] for p in physician_data ], } # Prepare export data export_data = ExportService.prepare_dashboard_data(user=user, kpis=kpis, charts=charts, tables=tables) # Export based on format if export_format == "excel": return ExportService.export_to_excel(export_data) elif export_format == "pdf": return ExportService.export_to_pdf(export_data) return JsonResponse({"error": "Export failed"}, status=500)