The Complete Guide to Python/Django Web Service Development with React (Based on Django 4.2)

Hello. This is Jinseok Lee.

Last weekend, I gave a keynote presentation at PyCon Korea 2025 held at Dongguk University Seoul Campus on the topic "Django! The Most Practical Choice for the AI Era" for about an hour. The presentation hall was packed, with many people standing to listen. Over 300 attendees really focused and listened intently. So even though I went 20 minutes over the scheduled 40-minute time slot, the staff didn't cut me off and helped me finish the presentation to the end.

I've introduced the core topics covered in this course and informed you that they can be useful for AI application development as well. I'm sharing the relevant slides with you, so please take a look. :-)

https://docs.google.com/presentation/d/1hQlaXP4qKudjkaIr-RyP03q-ofrKjlxsNUT9-iqLLlQ/edit?usp=sharing

I will be your Python/Django pacemaker.

Python Love Room, Lee Jin-seok

hello.

In the era of AI/LLM/agents, I think there are more things you can do with Django. So, to spread Django, I am developing and sharing various AI-related tutorials.

For the third tutorial, I'm sharing a Django-based "RAG from Scratch to Web Chat" tutorial.
https://ai.pyhub.kr/rag-01/

These days, RAG and RAG-based web chat are firmly established as use cases in AI applications. RAG is a core technology for obtaining knowledge data when creating services such as ChatGPT these days. Not only general companies, but also many public institutions are competing to develop RAG chatbots. Web chat and various functions required for RAG are functions that Django already supports well. You can implement RAG web chat with high productivity based on Django.

The YouTube live broadcast that was planned to be a supplementary Q&A session to the lecture that we previously informed you about requires more preparation, so we will proceed with it after the Lunar New Year holiday.

It's more fun to do it together than to follow the tutorial alone. On Saturday, January 25th at 9 PM , I'm going to hold < Follow along. "RAG from the bottom to web chat" Tutorial Live #1 >. We'll build our understanding by implementing RAG directly with Python, and I'll explain the tutorial on creating a web chat UI using Django/htmx live, and we'll have a Q&A. We'll build our understanding by practicing together. If you're interested in RAG and Django-based RAG development, come see me on YouTube this Saturday. :-)
https://www.youtube.com/live/aI-Dt4wSgLY

thank you

Python Love Room Lee Jin-seok Dream

We are sharing with you the analysis of the 2024 Django Developer Survey conducted by JetBrains.
https://blog.jetbrains.com/ko/pycharm/2024/09/the-state-of-django/

• Async: Used by 61% of Django developers

• Frontend: Django developers increasingly prefer HTMX, Alpine.js, and Tailwind CSS.

• Database: 75% of Django developers prefer PostgreSQL, 50% use Redis for caching

• Orchestration: Over 50% of Django developers use container orchestration

• CI System: GitHub Actions Leads the Industry

• IaC: 39% of Django developers use Infrastructure as Code (IaC)

I'll leave a few comments here.

1. I've been actively using asynchrony lately. Almost all codes are async def. In the era of AI, there are many external API calls, so I/O optimization is necessary. Accordingly, if you apply async in Django, you can operate the server more efficiently. Since Django models and some decorators do not support asynchrony, you have to implement an asynchronous version of the decorator separately. Please refer to the login_required, csrf_exept, and require_http_methods decorator codes of the asynchronous versions that I implemented . In Python, there is a convention of adding a in front of the asynchronous function name.

2. htmx/alpine.js/tailwindcss are also actively utilized in the django-pyhub-ai library that I developed. htmx is also actively covered in this lecture.

3. PostgreSQL is rapidly gaining popularity worldwide. The sqlite3 database, which is starting in the ruby on rails camp, is said to be able to serve well with a single server for a service targeting about 100,000 users. It is easy to think of the sqlite3 database as a test DB, but in these days when disk I/O is rapidly increasing and CPU/memory is more than 100 times better than 20 years ago, it is said to perform well enough.

Thank you for being with us in 2024. I look forward to working with you in 2025 as well.

I will be your Python/Django pacemaker.

thank you

Lee Jin Seok Dream

hello.

< Create a data analysis agent chatbot in 30 minutes with Python/Django > We share the tutorial video and documentation.

• Tutorial video: https://www.youtube.com/watch?v=10Fp78n3jSw

• Tutorial documentation:https://django-pyhub-ai.readthedocs.io

This article contains the process of quickly creating an LLM situational chatbot and a data analysis agent chatbot using the django-pyhub-ai library that I created. This library is based on the DRY (Don't Repeat Yourself) philosophy, one of Django's core philosophies, to eliminate repetitive and cumbersome tasks and to help easily build an efficient agent-based chat service. You can implement an agent chatbot with code close to the settings without worrying about the web front end. It is based on Django Channels and HTMX, and uses Langchain, famous for its LLM library, internally.

Chatbot operation screen

I believe that by utilizing Django's models/cache/templates/storage/API/authentication, we can create more productive and valuable AI agents. We will continue to strive to deliver various Django news in the future.

Please share this widely.
I will be your Python/Django pacemaker.

thank you

Python Love Room, Lee Jin-seok Dream

Infraon's <Introductory Python Django Course Starting with Django Design Philosophy> is now available for free.
https://inf.run/RDjT

If there are people around you who are interested in Django, please share it widely. However, it may be difficult for those who are new to web development. For those who are new to web development, please refer to the daily coding I recommend that you study the YouTube lectures first.
https://www.youtube.com/playlist?list=PLuHgQVnccGMDLp4GH-rgQhVKqqZawlNwG

thank you. :-)

A little while ago, a video testing the OpenAI Realtime API was uploaded to Teddy Note YouTube .

He showed me the code from the https://github.com/langchain-ai/react-voice-agent repository as is. The code in this repository is written based on Starlette and langchain.

I ported this to django/channels + langchain version. ;-)
https://github.com/pyhub-kr/pyhub-django-react-voice-agent

It is functionally identical to the react-voice-agent project. Just set the OPENAI_API_KEY environment variable and access the http://localhost:8000 page to make it work.

If you haven't seen the Teddy Note video above, watch it first, then try turning the code.

I used AsyncWebsocketConsumer from django channels.

#openai #realtime

A special lecture on the topic of "Web Front-end Development Using Django Effectively" will be held offline at the Gangnam branch of Everyone's Research Institute for two hours starting at 3 p.m. I will share the presentation slides and code repository.

1. How to create various dynamic components using the django-cotton library in Django

- Icon components, map components, image file preview widgets, etc.)

- Comparison with React components

- Utilizing HTMX

2. Create a ChatGPT service using django view (No JS)

- Sending image/text input values (multimodal)

- How to implement streaming response and apply it to the screen

- Generate responses using LLM

- How to refactor to class-based views

- Implementing SSE as a Django view without web socket, implementing multi-party chat

3. Freely configure the form layout

- Use custom templates

- Utilizing django-widget-tweaks and django-crispy-forms

4. How to use Django's default session authentication instead of JWT in web API

- Comparison with JWT authentication

- Handling login/logout requests from other hosts in Django LoginView/LogoutView

- How to share session cookies with other hosts

Slides: https://drive.google.com/file/d/1tcdHEiadFuFb-HOSEVIQkCvgJ8ssY4wv/view?usp=drive_link

Code repository: https://github.com/pyhub-kr/pyhub-htmx-examples

I will film it separately and share it with you. :-)

Hello. This is Jinseok Lee.

I uploaded all the videos last night.
Thank you for waiting for a long time. Thanks to you, I was able to finish it successfully.

The last video of the distribution covers issuing HTTPS certificates. To obtain an HTTPS certificate, you need a domain. For those who do not have a domain, as a benefit for students taking this course, a pyhub-django.com subdomain will be issued for practice purposes only once. You can check the details and apply through the related notice .

The production of this lecture does not end here, and we will do our best to continuously update the content to make it a better lecture.

We ask for your continued interest and support.

I will be your Python/Django pacemaker.

Python Love Room, Lee Jin-seok Dream