Ephes Blog

Don't use a big word when a singularly unloquacious and diminutive linguistic expression will satisfactorily accomplish the contemporary necessity. --Dgar

This week brought some improvements and a new release for podcast-transcripts. Overall, though, it was more work and less play than usual—a pattern I suspect will continue through the end of the year.

Articles

• Playground Wisdom: Threads Beat Async/Await | I agree that it's all very confusing 😵‍💫
• Don't Look Down on Print Debugging | Yep, using print debugging all the time - ftw!

Weeknotes

• Weeknotes: asynchronous LLMs, synchronous embeddings, and I kind of started a podcast | Simon Willison
• Weeklog for Week 46: November 11 to November 17 | Johannes

Software

• django-removals - a maintainer's best friend for finding removed features in your Django project | Fediverse 🧵
• Proposal for a Django project template

Videos

• jamellebouie tiktok

Out of Context Images

Hey folks! I recently built a little command-line tool called podcast-transcript that turns audio into text. While it started as a podcast transcription project during the PyDDF autumn sprint, it works great for any speech audio. The coolest part? It can transcribe a 2-hour podcast in about 90 seconds!

Quick Start 🚀

Why Groq?

After trying different approaches, I landed on using the Groq API for transcription. Here's why:

• It's blazing fast
• Getting an API key is free and API usage is free (with reasonable limits: 8 hours of audio per day, 2 hours per hour)
• The Whisper large-v3 model handles multiple languages well (especially noticeable for German content)

Technical Bits

The tool handles some interesting challenges under the hood:

• Automatically resamples audio to 16kHz mono before upload (if you don't do it before, Groq will after upload)
• Splits files larger than 25MB into chunks and stitches the transcripts back together
• Uses httpx for direct API calls to get detailed JSON responses inspired by Simon Willison’s approach
• Outputs in multiple formats: DOTe JSON, Podlove JSON, WebVTT, and plain text

Future Plans

I'm planning to add support for local transcription using the OpenAI Whisper model. While Whisper v2 works well enough for English content, v3 shows notable improvements for other languages (especially German). I initially skipped local processing because of the PyTorch dependency, but it's on the roadmap! I also plan to add multitrack support for handling audio files with separate speaker tracks.

The code is open source and contributions are welcome. Let me know if you try it out!

I learned all my HTML<br>
From a code camp held yearly in hell<br>
We started on `<s>`<br>
Then striked through the rest<br>
Now I have a career in DevRel --Heydon Pickering

Published a new podcast episode discussing Python 3.13. I’ve been using Jupyter notebooks to generate transcripts for the episodes and have started consolidating that workflow into a dedicated Python command-line application.

Articles

• The br element
• return in finally considered harmful
• Three Django wishes
• Qwen2.5-Coder-32B is an LLM that can code well that runs on my Mac

Books

• Boost Your Django DX updated again | Cool! Great book!

Software

• Recraft V3. Recraft are a generative AI design tool startup based out of London who released their v3 model a few weeks ago | It does SVG, too
• Mesa: Agent-based modeling in Python

Out of Context Images

AAAAAAAAAAAAAAAAAAAAAAAHHHHHHHH --Endless Screaming

This is fine 🔥.

In other news, I attended the PyDDF autumn sprint! After putting off the transcripts feature for django-cast for quite some time, I finally started working on it. You can already see some transcript functionality in the web player on my podcast website. There’s still plenty to do, but at least it’s underway now.

Articles

• The button element
• HTML link, or button, that is the question
• Don't return named tuples in new APIs
• Der Prototype Fund geht weiter – und wie! | Great news!
• My IndieWeb Journey
• We've moved to Hetzner
• Ten simple rules for writing a technical book

Videos

• When and how to start coding with kids [PyCon DE & PyData Berlin 2024]
• I bought a freeze dryer so you don't have to

Software

• Optuna: A hyperparameter optimization framework
• Beacon v0.1 and the Future | CMS for elixir
• dsd-flyio - A plugin for deploying Django projects to Fly.io, using django-simple-deploy
• Rails 8.0: No PaaS Required | Rails seems do do quite well

Fediverse

• The VERDAD prompts are pretty complex - Rajiv shared this example of a conversation he had with Claude 3.5 Sonnet to further iterate on the existing prompt

Out of Context Images

By age 35, your 500-foot burial pyramid should already be 80% complete. --batkaren

I’ve implemented some performance improvements on my podcast page and published an article about them: “TIL: Podlove Web Player Performance Improvements.” Additionally, I released a new version of django-cast that incorporates the new podlove-player web component and offers official support for Wagtail 6.3.

Articles

• How Google is Killing Bloggers and Small Publishers – And Why | Revival of the web ring or blogroll?
• Text fragments allow linking directly to a specific portion of text in a web document | I cannot believe I didn't know about that 😳
• The body element | Organ donor

Software

• llm-whisper-api | Might come in handy for processing podcast transcripts..
• Form validation styling | Nice to know how to do this
• Level up your image accessibility and dashboards with Wagtail 6.3 | 6.3 released 🥳
• xml2invoice | Might gonna need this in 2025 - receipts is also great!

Books

• Flexible Typesetting | Now free!

Fediverse

• Whenever I think I might be coming around to Python typing, I end up stumbling across something like this. | 🧵
• In a world where everyone does most internetting on their phones, how are PDFs still a thing? | I find myself asking this question repeatedly. And why do people use office software—which is intended for printing—when they never plan to print anything and just send me their junk via email? Why?

Weeknotes

• W̶e̶e̶k̶n̶o̶t̶e̶s̶ Monthnotes for October | Simon Willison
• Weeklog for Week 43: October 21 to October 27 | Johannes

Videos

• Fixing Latent Unsound Abstract Operators in the eBPF Verifier of the Linux Kernel Santosh Nagarakatte @ SAS 2024, Rutgers University | eBPF is soo cool!
• Midjourney Editor | Nice

Out of Context Images