Full-stack analysis covering Django backend, frontend JS/CSS, infrastructure, and test suite. Identifies: - 4 critical issues (XSS, debug code, mass assignment, N+1 queries) - 10 high priority improvements (fat models, missing indexes, inconsistent API responses, exception handling) - 9 medium priority items (frontend tooling, test gaps, Celery config) - Architecture recommendations and action plan
16 KiB
ThrillWiki Codebase Quality Review
Date: January 2026 Scope: Full-stack analysis (Django backend, frontend, infrastructure, tests)
Executive Summary
This codebase is a well-architected Django 5.2 application with HTMX/Alpine.js frontend, PostgreSQL/PostGIS database, Redis caching, and Celery task queue. The project demonstrates strong engineering fundamentals but has accumulated technical debt in several areas that, if addressed, would significantly improve maintainability, performance, and security.
Overall Assessment
| Area | Score | Notes |
|---|---|---|
| Architecture | ⭐⭐⭐⭐ | Well-organized modular Django apps |
| Code Quality | ⭐⭐⭐ | Good patterns but inconsistencies exist |
| Security | ⭐⭐⭐ | Solid foundation with some XSS risks |
| Performance | ⭐⭐⭐ | Good caching but N+1 queries present |
| Testing | ⭐⭐⭐ | 70% coverage with gaps |
| Frontend | ⭐⭐⭐ | Clean JS but no tooling/types |
| Infrastructure | ⭐⭐⭐⭐ | Comprehensive CI/CD and deployment |
🔴 Critical Issues (Fix Immediately)
1. XSS Vulnerabilities in Admin Panel
Location: backend/apps/moderation/admin.py
# Line 228 - changes_preview() method
return mark_safe("".join(html)) # User data not escaped!
# Line 740 - context_preview() method
return mark_safe("".join(html)) # Context data not escaped!
Risk: Attackers could inject malicious JavaScript through edit submissions.
Fix:
from django.utils.html import escape
# In changes_preview():
html.append(f"<td>{escape(str(old))}</td><td>{escape(str(new))}</td>")
2. Debug Print Statements in Production Code
Location: backend/apps/parks/models/parks.py:375-426
print(f"\nLooking up slug: {slug}") # DEBUG CODE IN PRODUCTION
print(f"Found current park with slug: {slug}")
print(f"Checking historical slugs...")
Fix: Remove or convert to logging.debug().
3. Mass Assignment Vulnerability in Serializers
Location: backend/apps/api/v1/accounts/serializers.py
class UserProfileUpdateInputSerializer(serializers.ModelSerializer):
class Meta:
model = UserProfile
fields = "__all__" # DANGEROUS - exposes all fields for update
Fix: Explicitly list allowed fields:
fields = ["display_name", "bio", "location", "website", "social_links"]
4. N+1 Query in Trip Planning Views
Location: backend/apps/parks/views.py:577-583, 635-639, 686-690
# Current (N+1 problem - one query per park):
for tid in _get_session_trip(request):
try:
parks.append(Park.objects.get(id=tid))
except Park.DoesNotExist:
continue
# Fix (single query):
park_ids = _get_session_trip(request)
parks = list(Park.objects.filter(id__in=park_ids))
🟠 High Priority Issues
5. Fat Models with Business Logic
The following models have 200+ lines of business logic that should be extracted to services:
| Model | Location | Lines | Issue |
|---|---|---|---|
Park |
parks/models/parks.py |
220-428 | FSM transitions, slug resolution, computed fields |
EditSubmission |
moderation/models.py |
76-371 | Full approval workflow |
PhotoSubmission |
moderation/models.py |
668-903 | Photo approval workflow |
Recommendation: Create service classes:
apps/parks/services/
├── park_service.py # FSM transitions, computed fields
├── slug_service.py # Historical slug resolution
└── ...
apps/moderation/services/
├── submission_service.py # EditSubmission workflow
├── photo_service.py # PhotoSubmission workflow
└── ...
6. Missing Database Indexes
Critical indexes to add:
# ParkPhoto - No index on frequently-filtered FK
class ParkPhoto(models.Model):
park = models.ForeignKey(Park, on_delete=models.CASCADE, db_index=True) # ADD db_index
# UserNotification - Missing composite index
class Meta:
indexes = [
models.Index(fields=["user", "created_at"]), # ADD for sorting
]
# RideCredit - Missing index for ordering
class Meta:
indexes = [
models.Index(fields=["user", "display_order"]), # ADD
]
# Company - Missing status filter index
class Meta:
indexes = [
models.Index(fields=["status", "founded_year"]), # ADD
]
7. Inconsistent API Response Formats
Current state (3+ different formats):
# Format 1: rides endpoint
{"rides": [...], "total_count": X, "strategy": "...", "has_more": bool}
# Format 2: parks endpoint
{"parks": [...], "total_count": X, "strategy": "..."}
# Format 3: DRF paginator
{"results": [...], "count": X, "next": "...", "previous": "..."}
# Format 4: Success responses
{"success": True, "data": {...}} # vs
{"detail": "Success message"} # vs
{"message": "Success"}
Recommendation: Create a standard response wrapper:
# apps/core/api/responses.py
class StandardResponse:
@staticmethod
def success(data=None, message=None, meta=None):
return {
"success": True,
"data": data,
"message": message,
"meta": meta # pagination, counts, etc.
}
@staticmethod
def error(message, code=None, details=None):
return {
"success": False,
"error": {
"message": message,
"code": code,
"details": details
}
}
8. Overly Broad Exception Handling
Pattern found 15+ times:
# BAD - masks actual errors
try:
queryset = self.apply_filters(queryset)
except Exception as e:
log_exception(e)
return Park.objects.none() # Silent failure!
Fix: Catch specific exceptions:
from django.core.exceptions import ValidationError
from django.db import DatabaseError
try:
queryset = self.apply_filters(queryset)
except ValidationError as e:
messages.warning(request, f"Invalid filter: {e}")
return base_queryset
except DatabaseError as e:
logger.error("Database error in filter", exc_info=True)
raise # Let it bubble up or return error response
9. Duplicated Permission Checks
Found in 6+ locations:
# Repeated pattern in views:
if not (request.user == instance.uploaded_by or request.user.is_staff):
raise PermissionDenied()
Fix: Create reusable permission class:
# apps/core/permissions.py
class IsOwnerOrStaff(permissions.BasePermission):
def has_object_permission(self, request, view, obj):
if request.method in permissions.SAFE_METHODS:
return True
owner_field = getattr(view, 'owner_field', 'user')
owner = getattr(obj, owner_field, None)
return owner == request.user or request.user.is_staff
🟡 Medium Priority Issues
10. Frontend Has No Build Tooling
Current state:
- No
package.jsonor npm dependencies - No TypeScript (vanilla JS only)
- No ESLint/Prettier configuration
- No minification/bundling pipeline
- No source maps for debugging
Impact:
- No type safety in 8,000+ lines of JavaScript
- Manual debugging without source maps
- No automated code quality checks
Recommendation: Add minimal tooling:
// package.json
{
"scripts": {
"lint": "eslint backend/static/js/",
"format": "prettier --write 'backend/static/js/**/*.js'",
"typecheck": "tsc --noEmit"
},
"devDependencies": {
"eslint": "^8.0.0",
"prettier": "^3.0.0",
"typescript": "^5.0.0"
}
}
11. Test Coverage Gaps
Disabled tests (technical debt):
tests_disabled/test_models.py- Park model teststests_disabled/test_filters.py- Filter teststests_disabled/test_search.py- Search/autocomplete tests
Missing test coverage:
- Celery async tasks (not tested)
- Cache hit/miss behavior
- Concurrent operations/race conditions
- Performance benchmarks
- Component-level accessibility
Recommendation:
- Re-enable disabled tests with updated model references
- Add Celery task tests with
CELERY_TASK_ALWAYS_EAGER = True - Implement Page Object Model for E2E tests
12. Celery Configuration Issues
Location: backend/config/celery.py
# Issue 1: No retry policy visible
# Tasks that fail don't have automatic retry with backoff
# Issue 2: Beat schedule lacks jitter
# All daily tasks run at midnight - thundering herd problem
CELERYBEAT_SCHEDULE = {
"daily-ban-expiry": {"schedule": crontab(hour=0, minute=0)},
"daily-deletion-processing": {"schedule": crontab(hour=0, minute=0)},
"daily-closing-checks": {"schedule": crontab(hour=0, minute=0)},
# All at midnight!
}
Fix:
# Add retry policy to tasks
@app.task(bind=True, max_retries=3, default_retry_delay=60)
def process_submission(self, submission_id):
try:
# ... task logic
except TransientError as e:
raise self.retry(exc=e, countdown=60 * (2 ** self.request.retries))
# Stagger beat schedule
CELERYBEAT_SCHEDULE = {
"daily-ban-expiry": {"schedule": crontab(hour=0, minute=0)},
"daily-deletion-processing": {"schedule": crontab(hour=0, minute=15)},
"daily-closing-checks": {"schedule": crontab(hour=0, minute=30)},
}
13. Rate Limiting Only on Auth Endpoints
Location: backend/apps/core/middleware/rate_limiting.py
RATE_LIMITED_PATHS = {
"/api/v1/auth/login/": {...},
"/api/v1/auth/signup/": {...},
# Missing: file uploads, form submissions, search endpoints
}
Recommendation: Extend rate limiting:
RATE_LIMITED_PATHS = {
# Auth
"/api/v1/auth/login/": {"per_minute": 5, "per_hour": 30},
"/api/v1/auth/signup/": {"per_minute": 3, "per_hour": 10},
"/api/v1/auth/password-reset/": {"per_minute": 3, "per_hour": 10},
# File uploads
"/api/v1/photos/upload/": {"per_minute": 10, "per_hour": 100},
# Search (prevent abuse)
"/api/v1/search/": {"per_minute": 30, "per_hour": 500},
# Submissions
"/api/v1/submissions/": {"per_minute": 5, "per_hour": 50},
}
14. Inconsistent Status Field Implementations
Three different patterns used:
# Pattern 1: RichFSMField (Park)
status = RichFSMField(default=ParkStatus.OPERATING, ...)
# Pattern 2: CharField with choices (Company)
status = models.CharField(max_length=20, choices=STATUS_CHOICES, ...)
# Pattern 3: RichChoiceField (User role)
role = RichChoiceField(choices=UserRole.choices, ...)
Recommendation: Standardize on one approach (RichFSMField for state machines, RichChoiceField for enums).
15. Magic Numbers Throughout Code
Examples found:
# auth/views.py
get_random_string(64) # Why 64?
timeout=300 # Why 300 seconds?
MAX_AVATAR_SIZE = 10 * 1024 * 1024 # Inline constant
# Various files
page_size = 20 # vs 24 in other places
Fix: Create constants module:
# apps/core/constants.py
class Security:
MFA_TOKEN_LENGTH = 64
MFA_TOKEN_TIMEOUT_SECONDS = 300
MAX_AVATAR_SIZE_BYTES = 10 * 1024 * 1024
class Pagination:
DEFAULT_PAGE_SIZE = 20
MAX_PAGE_SIZE = 100
🟢 Low Priority / Nice-to-Have
16. Deprecated Code Not Removed
Location: backend/static/js/moderation/history.js
- File marked as DEPRECATED but still present
- Should be removed or migrated
17. Unused Imports
Multiple files have duplicate or unused imports:
backend/apps/api/v1/rides/views.py-Qimported twice
18. Missing Docstrings on Complex Methods
Many service methods and complex views lack docstrings explaining:
- Expected inputs/outputs
- Business rules applied
- Side effects
19. Template |safe Filter Usage
Files using |safe that should use |sanitize:
templates/components/ui/icon.html:61templates/components/navigation/breadcrumbs.html:116
Architecture Recommendations
1. Adopt Service Layer Pattern Consistently
apps/
├── parks/
│ ├── models/ # Data models only
│ ├── services/ # Business logic
│ │ ├── park_service.py
│ │ ├── slug_service.py
│ │ └── media_service.py
│ ├── selectors/ # Read queries (already exists)
│ └── api/ # Serializers, viewsets
2. Create Shared Response/Error Handling
# apps/core/api/
├── responses.py # StandardResponse class
├── exceptions.py # Custom exceptions with codes
├── error_handlers.py # DRF exception handler
└── mixins.py # Reusable view mixins
3. Implement Repository Pattern for Complex Queries
# apps/parks/repositories/park_repository.py
class ParkRepository:
@staticmethod
def get_by_slug_with_history(slug: str) -> Park | None:
"""Resolve slug including historical slugs."""
# Move 60+ lines from Park.get_by_slug() here
4. Add Event-Driven Architecture for Cross-App Communication
# Instead of direct imports between apps:
from apps.parks.models import Park # Tight coupling
# Use signals/events:
# apps/core/events.py
park_status_changed = Signal()
# apps/parks/services/park_service.py
park_status_changed.send(sender=Park, park=park, old_status=old, new_status=new)
# apps/notifications/handlers.py
@receiver(park_status_changed)
def notify_followers(sender, park, **kwargs):
...
Performance Optimization Opportunities
1. Database Query Optimization
| Issue | Location | Impact |
|---|---|---|
| N+1 in trip views | parks/views.py:577 |
High - loops with .get() |
| Missing indexes | Multiple models | Medium - slow filters |
| No query count monitoring | Production | Unknown query count |
2. Caching Strategy Improvements
# Add cache key versioning
CACHE_VERSION = "v1"
def get_park_cache_key(park_id):
return f"park:{CACHE_VERSION}:{park_id}"
# Add cache tags for invalidation
from django.core.cache import cache
def invalidate_park_caches(park_id):
cache.delete_pattern(f"park:*:{park_id}:*")
3. Frontend Performance
- Add
loading="lazy"to images below the fold - Implement virtual scrolling for long lists
- Add service worker for offline capability
Security Hardening Checklist
- Fix XSS in admin
mark_safe()calls - Replace
fields = "__all__"in serializers - Add rate limiting to file upload endpoints
- Review
|safetemplate filter usage - Add Content Security Policy headers
- Implement API request signing for sensitive operations
- Add audit logging for admin actions
- Review OAuth state management consistency
Recommended Action Plan
Phase 1: Critical Fixes (This Sprint)
- Fix XSS vulnerabilities in admin
- Remove debug print statements
- Fix mass assignment in serializers
- Fix N+1 queries in trip views
- Add missing database indexes
Phase 2: High Priority (Next 2 Sprints)
- Extract business logic to services
- Standardize API response format
- Fix overly broad exception handling
- Re-enable disabled tests
- Add rate limiting to more endpoints
Phase 3: Medium Priority (Next Quarter)
- Add frontend build tooling
- Implement TypeScript for type safety
- Improve Celery configuration
- Standardize status field patterns
- Add comprehensive E2E tests
Phase 4: Ongoing
- Remove deprecated code
- Add missing docstrings
- Monitor and optimize queries
- Security audits
Conclusion
This codebase has a solid foundation with good architectural decisions (modular apps, service layer beginnings, comprehensive configuration). The main areas needing attention are:
- Security: XSS vulnerabilities and mass assignment risks
- Performance: N+1 queries and missing indexes
- Maintainability: Fat models and inconsistent patterns
- Testing: Re-enable disabled tests and expand coverage
Addressing the critical and high-priority issues would significantly improve code quality and reduce technical debt. The codebase is well-positioned for scaling with relatively minor refactoring efforts.