Django Llm Feature Builder

Output format

For implementation requests, produce output in this order:

1. short architecture note
2. file-by-file change plan
3. code for each file
4. migration note if needed
5. admin registration note
6. test checklist

Coding rules

• use django and drf conventions
• use class-based apis unless the project clearly uses function-based views
• keep serializers explicit
• avoid putting llm client logic in serializers or views
• prefer dependency injection through service parameters or settings access wrappers
• if the project supports multiple providers, implement provider selection through a routing service
• include structured error handling
• avoid hardcoding secrets; assume they come from env vars or encrypted admin-managed references

Provider integration rules

When building multi-provider support:

• define a stable internal interface for providers
• map provider-specific parameters into a normalized payload
• normalize response fields into a common schema
• preserve raw provider response optionally for audit/debug logs
• include model capability checks where relevant

Admin rules

When admin-managed configuration is involved:

• optimize admin usability with search, filters, readonly fields where appropriate
• protect sensitive fields
• separate active/inactive configurations
• support priority or default selection for providers/models

References

Read these when relevant:

• references/architecture.md for suggested layering
• references/conventions.md for naming and error-handling conventions

Script usage

Use scripts/scaffold_checklist.py if you need to generate or verify a feature implementation checklist before writing code.39:["$","$L41",null,{"content":"$42","frontMatter":{"name":"django-llm-feature-builder","description":"build and extend django rest framework features for llm-based applications with multi-provider support configured from django admin. use when the user wants to add a new api endpoint, service, serializer, admin configuration, provider routing logic, conversation flow, usage logging, or a reusable backend pattern for an ai/chat application built with django and drf."}}]