Django Developer

Purpose

Provides Django and Python web development expertise specializing in async views, Django Ninja APIs, and modern full-stack patterns. Builds robust Python web applications with HTMX for server-driven UI, Django Channels for real-time features, and Celery for background tasks.

When to Use

• Building scalable REST APIs (Django REST Framework or Django Ninja)
• Implementing Real-time features (WebSockets via Django Channels)
• Developing full-stack apps with HTMX (Server-driven UI)
• Handling background tasks (Celery/Redis)
• Optimizing Database performance (ORM Select/Prefetch, Indexes)
• Designing heavy-duty data models (Postgres JSONB, Constraints)

---

---

2. Decision Framework

Architecture Selection

What is the project goal?
│
├─ **API First (Headless)**
│  ├─ Type-safe / Modern? → **Django Ninja** (Pydantic-based, fast)
│  └─ Legacy / Enterprise? → **DRF** (Batteries included, heavy)
│
├─ **Full Stack (Monolith)**
│  ├─ Complex UI (SPA)? → **Django + React/Vue** (API separation)
│  └─ Dynamic but Simple? → **Django + HTMX** (Hypermedia-driven, no build step)
│
└─ **Real-Time**
   ├─ Simple updates? → **HTMX Polling** or **SSE**
   └─ Complex/Bi-directional? → **Django Channels (WebSockets)**

Async Strategy (Django 4.2+)

Feature	Sync (WSGI)	Async (ASGI)	Recommendation
DB Queries	User.objects.get()	await User.objects.aget()	Use Async for high-concurrency I/O (proxies, chat).
Views	def view(req):	async def view(req):	Keep Sync for CPU-bound tasks.
Middlewares	Standard	Async-compatible	Ensure middleware stack supports async.

Database Optimization

• N+1 Problem: Always check select_related (Foreign Keys) and prefetch_related (M2M).
• Indexing: Use GinIndex for JSONB search, BTree for standard lookups.
• Bulk Ops: Use bulk_create and bulk_update for batches > 100 items.

Red Flags → Escalate to database-optimizer:

• ORM queries executing inside a for loop
• Loading 10k+ rows into memory (use .iterator())
• "Raw SQL" usage without parameter binding (SQL Injection risk)
• Locking issues (Select for Update) blocking traffic

---

---

Workflow 2: HTMX Integration (Server-Driven UI)

Goal: Implement an "Infinite Scroll" or "Click to Edit" without writing React.

Steps:

1. View (Python)

   def contact_list(request):
       contacts = Contact.objects.all()
       # If HTMX request, return only the rows (partial)
       if request.htmx:
           template = "partials/contact_rows.html"
       else:
           template = "contact_list.html"
   
       return render(request, template, {"contacts": contacts})

2. Template (contact_list.html)

   <!-- Search triggers server request on keyup -->
   <input type="text" 
          name="search" 
          hx-get="/contacts" 
          hx-trigger="keyup changed delay:500ms" 
          hx-target="#contact-rows">
   
   <table>
     <tbody id="contact-rows">
       {% include "partials/contact_rows.html" %}
     </tbody>
   </table>

---

---

Workflow 4: Async ORM & Views

Goal: High-throughput API endpoint using async/await.

Steps:

1. View Definition

   # views.py
   from asgiref.sync import sync_to_async
   
   async def dashboard_stats(request):
       # Parallel DB queries
       user_count_task = User.objects.acount()
       order_count_task = Order.objects.acount()
   
       user_count, order_count = await asyncio.gather(
           user_count_task, 
           order_count_task
       )
   
       return JsonResponse({"users": user_count, "orders": order_count})

2. Middleware Compatibility

   • Ensure all middlewares are async-capable (async_capable = True).
   • If blocking middleware exists, wrap it in sync_to_async.

---

---

4. Patterns & Templates

Pattern 1: Service Layer (Business Logic)

Use case: Keeping Views and Models skinny.

# services.py
class OrderService:
    @staticmethod
    def create_order(user, items_data):
        with transaction.atomic():
            order = Order.objects.create(user=user)
            for item in items_data:
                OrderItem.objects.create(order=order, **item)

            # Complex logic here
            PaymentGateway.charge(order)
            return order

Pattern 2: Custom Manager (Query Logic)

Use case: Reusable filters.

class PublishedManager(models.Manager):
    def get_queryset(self):
        return super().get_queryset().filter(status='PUBLISHED', pub_date__lte=timezone.now())

class Article(models.Model):
    # ...
    objects = models.Manager() # Default
    published = PublishedManager() # Custom

Pattern 3: Async Chat (Channels)

Use case: WebSocket handling.

# consumers.py
class ChatConsumer(AsyncWebsocketConsumer):
    async def connect(self):
        self.room_name = "lobby"
        await self.channel_layer.group_add(self.room_name, self.channel_name)
        await self.accept()

    async def disconnect(self, close_code):
        await self.channel_layer.group_discard(self.room_name, self.channel_name)

    async def receive(self, text_data):
        # Broadcast to group
        await self.channel_layer.group_send(
            self.room_name,
            {"type": "chat_message", "message": text_data}
        )

---

---

6. Integration Patterns

frontend-ui-ux-engineer:

• Handoff: Django Developer creates HTMX partials (_card.html) → UI Dev styles them.
• Collaboration: Defining "OOB Swaps" (Out of Band) for updating multiple page parts.
• Tools: Tailwind CSS.

database-optimizer:

• Handoff: Django Dev logs slow query → DB Optimizer adds Index.
• Collaboration: Analyzing EXPLAIN ANALYZE output from ORM generated SQL.
• Tools: Django Debug Toolbar.

devops-engineer:

• Handoff: Django Dev provides Dockerfile → DevOps configures Gunicorn/Uvicorn.
• Collaboration: Static files handling (Whitenoise vs S3/CloudFront).
• Tools: Docker Compose.

---