Simon Willison’s Weblog

Finally, a replacement for BERT: Introducing ModernBERT (via) BERT was an early language model released by Google in October 2018. Unlike modern LLMs it wasn't designed for generating text. BERT was trained for masked token prediction and was generally applied to problems like Named Entity Recognition or Sentiment Analysis. BERT also wasn't very useful on its own - most applications required you to fine-tune a model on top of it.

In exploring BERT I decided to try out dslim/distilbert-NER, a popular Named Entity Recognition model fine-tuned on top of DistilBERT (a smaller distilled version of the original BERT model). Here are my notes on running that using uv run.

Jeremy Howard's Answer.AI research group, LightOn and friends supported the development of ModernBERT, a brand new BERT-style model that applies many enhancements from the past six years of advances in this space.

While BERT was trained on 3.3 billion tokens, producing 110 million and 340 million parameter models, ModernBERT trained on 2 trillion tokens, resulting in 140 million and 395 million parameter models. The parameter count hasn't increased much because it's designed to run on lower-end hardware. It has a 8192 token context length, a significant improvement on BERT's 512.

I was able to run one of the demos from the announcement post using uv run like this (I'm not sure why I had to use numpy<2.0 but without that I got an error about cannot import name 'ComplexWarning' from 'numpy.core.numeric'):

uv run --with 'numpy<2.0' --with torch --with 'git+https://github.com/huggingface/transformers.git' python

Then this Python:

import torch
from transformers import pipeline
from pprint import pprint
pipe = pipeline(
    "fill-mask",
    model="answerdotai/ModernBERT-base",
    torch_dtype=torch.bfloat16,
)
input_text = "He walked to the [MASK]."
results = pipe(input_text)
pprint(results)

Which downloaded 573MB to ~/.cache/huggingface/hub/models--answerdotai--ModernBERT-base and output:

[{'score': 0.11669921875,
  'sequence': 'He walked to the door.',
  'token': 3369,
  'token_str': ' door'},
 {'score': 0.037841796875,
  'sequence': 'He walked to the office.',
  'token': 3906,
  'token_str': ' office'},
 {'score': 0.0277099609375,
  'sequence': 'He walked to the library.',
  'token': 6335,
  'token_str': ' library'},
 {'score': 0.0216064453125,
  'sequence': 'He walked to the gate.',
  'token': 7394,
  'token_str': ' gate'},
 {'score': 0.020263671875,
  'sequence': 'He walked to the window.',
  'token': 3497,
  'token_str': ' window'}]

I'm looking forward to trying out models that use ModernBERT as their base. The model release is accompanied by a paper (Smarter, Better, Faster, Longer: A Modern Bidirectional Encoder for Fast, Memory Efficient, and Long Context Finetuning and Inference) and new documentation for using it with the Transformers library.

# 24th December 2024, 6:21 am / ai, jeremy-howard, transformers, hugging-face, uv, python, nlp

Introducing Limbo: A complete rewrite of SQLite in Rust (via) This looks absurdly ambitious:

Our goal is to build a reimplementation of SQLite from scratch, fully compatible at the language and file format level, with the same or higher reliability SQLite is known for, but with full memory safety and on a new, modern architecture.

The Turso team behind it have been maintaining their libSQL fork for two years now, so they're well equipped to take on a challenge of this magnitude.

SQLite is justifiably famous for its meticulous approach to testing. Limbo plans to take an entirely different approach based on "Deterministic Simulation Testing" - a modern technique pioneered by FoundationDB and now spearheaded by Antithesis, the company Turso have been working with on their previous testing projects.

Another bold claim (emphasis mine):

We have both added DST facilities to the core of the database, and partnered with Antithesis to achieve a level of reliability in the database that lives up to SQLite’s reputation.

[...] With DST, we believe we can achieve an even higher degree of robustness than SQLite, since it is easier to simulate unlikely scenarios in a simulator, test years of execution with different event orderings, and upon finding issues, reproduce them 100% reliably.

The two most interesting features that Limbo is planning to offer are first-party WASM support and fully asynchronous I/O:

SQLite itself has a synchronous interface, meaning driver authors who want asynchronous behavior need to have the extra complication of using helper threads. Because SQLite queries tend to be fast, since no network round trips are involved, a lot of those drivers just settle for a synchronous interface. [...]

Limbo is designed to be asynchronous from the ground up. It extends sqlite3_step, the main entry point API to SQLite, to be asynchronous, allowing it to return to the caller if data is not ready to consume immediately.

Datasette provides an async API for executing SQLite queries which is backed by all manner of complex thread management - I would be very interested in a native asyncio Python library for talking to SQLite database files.

I successfully tried out Limbo's Python bindings against a demo SQLite test database using uv like this:

uv run --with pylimbo python
>>> import limbo
>>> conn = limbo.connect("/tmp/demo.db")
>>> cursor = conn.cursor()
>>> print(cursor.execute("select * from foo").fetchall())

It crashed when I tried against a more complex SQLite database that included SQLite FTS tables.

The Python bindings aren't yet documented, so I piped them through LLM and had the new google-exp-1206 model write this initial documentation for me:

files-to-prompt limbo/bindings/python -c | llm -m gemini-exp-1206 -s 'write extensive usage documentation in markdown, including realistic usage examples'

# 10th December 2024, 7:25 pm / rust, sqlite, uv, open-source, python, llm, ai-assisted-programming, documentation

Transferring Python Build Standalone Stewardship to Astral. Gregory Szorc's Python Standalone Builds have been quietly running an increasing portion of the Python ecosystem for a few years now, but really accelerated in importance when uv started using them for new Python installations managed by that tool. The releases (shipped via GitHub) have now been downloaded over 70 million times, 50 million of those since uv's initial release in March of this year.

uv maintainers Astral have been helping out with PSB maintenance for a while:

When I told Charlie I could use assistance supporting PBS, Astral employees started contributing to the project. They have built out various functionality, including Python 3.13 support (including free-threaded builds), turnkey automated release publishing, and debug symbol stripped builds to further reduce the download/install size. Multiple Astral employees now have GitHub permissions to approve/merge PRs and publish releases. All releases since April have been performed by Astral employees.

As-of December 17th Gregory will be transferring the project to the Astral organization, while staying on as a maintainer and advisor. Here's Astral's post about this: A new home for python-build-standalone.

# 3rd December 2024, 11:18 pm / uv, astral, python

SmolVLM—small yet mighty Vision Language Model. I've been having fun playing with this new vision model from the Hugging Face team behind SmolLM. They describe it as:

[...] a 2B VLM, SOTA for its memory footprint. SmolVLM is small, fast, memory-efficient, and fully open-source. All model checkpoints, VLM datasets, training recipes and tools are released under the Apache 2.0 license.

I've tried it in a few flavours but my favourite so far is the mlx-vlm approach, via mlx-vlm author Prince Canuma. Here's the uv recipe I'm using to run it:

uv run \
  --with mlx-vlm \
  --with torch \
  python -m mlx_vlm.generate \
    --model mlx-community/SmolVLM-Instruct-bf16 \
    --max-tokens 500 \
    --temp 0.5 \
    --prompt "Describe this image in detail" \
    --image IMG_4414.JPG

If you run into an error using Python 3.13 (torch compatibility) try uv run --python 3.11 instead.

This one-liner installs the necessary dependencies, downloads the model (about 4.2GB, saved to ~/.cache/huggingface/hub/models--mlx-community--SmolVLM-Instruct-bf16) and executes the prompt and displays the result.

I ran that against this Pelican photo:

The model replied:

In the foreground of this photograph, a pelican is perched on a pile of rocks. The pelican’s wings are spread out, and its beak is open. There is a small bird standing on the rocks in front of the pelican. The bird has its head cocked to one side, and it seems to be looking at the pelican. To the left of the pelican is another bird, and behind the pelican are some other birds. The rocks in the background of the image are gray, and they are covered with a variety of textures. The rocks in the background appear to be wet from either rain or sea spray.

There are a few spatial mistakes in that description but the vibes are generally in the right direction.

On my 64GB M2 MacBook pro it read the prompt at 7.831 tokens/second and generated that response at an impressive 74.765 tokens/second.

# 28th November 2024, 8:29 pm / vision-llms, uv, mlx, ai, edge-llms, llms, python, generative-ai, smollm

follow_theirs.py. Hamel Husain wrote this Python script on top of the atproto Python library for interacting with Bluesky, which lets you specify another user and then follows every account that user is following.

I forked it and added two improvements: inline PEP 723 dependencies and input() and getpass.getpass() to interactively ask for the credentials needed to run the script.

This means you can run my version using uv run like this:

uv run https://gist.githubusercontent.com/simonw/848a3b91169a789bc084a459aa7ecf83/raw/397ad07c8be0601eaf272d9d5ab7675c7fd3c0cf/follow_theirs.py

I really like this pattern of being able to create standalone Python scripts with dependencies that can be run from a URL as a one-liner. Here's the comment section at the top of the script that makes it work:

# /// script
# dependencies = [
#   "atproto"
# ]
# ///

# 24th November 2024, 6:57 pm / uv, hamel-husain, bluesky, python

open-interpreter (via) This "natural language interface for computers" open source ChatGPT Code Interpreter alternative has been around for a while, but today I finally got around to trying it out.

Here's how I ran it (without first installing anything) using uv:

uvx --from open-interpreter interpreter

The default mode asks you for an OpenAI API key so it can use gpt-4o - there are a multitude of other options, including the ability to use local models with interpreter --local.

It runs in your terminal and works by generating Python code to help answer your questions, asking your permission to run it and then executing it directly on your computer.

I pasted in an API key and then prompted it with this:

find largest files on my desktop

Here's the full transcript.

Since code is run directly on your machine there are all sorts of ways things could go wrong if you don't carefully review the generated code before hitting "y". The team have an experimental safe mode in development which works by scanning generated code with semgrep. I'm not convinced by that approach, I think executing code in a sandbox would be a much more robust solution here - but sandboxing Python is still a very difficult problem.

They do at least have an experimental Docker integration.

# 24th November 2024, 6:29 pm / llms, ai, generative-ai, uv, sandboxing, code-interpreter, openai, ai-assisted-programming, python

Is async Django ready for prime time? (via) Jonathan Adly reports on his experience using Django to build ColiVara, a hosted RAG API that uses ColQwen2 visual embeddings, inspired by the ColPali paper.

In a breach of Betteridge's law of headlines the answer to the question posed by this headline is “yes”.

We believe async Django is ready for production. In theory, there should be no performance loss when using async Django instead of FastAPI for the same tasks.

The ColiVara application is itself open source, and you can see how it makes use of Django’s relatively new asynchronous ORM features in the api/views.py module.

I also picked up a useful trick from their Dockerfile: if you want uv in a container you can install it with this one-liner:

COPY --from=ghcr.io/astral-sh/uv:latest /uv /bin/uv

# 24th November 2024, 5:47 pm / embeddings, asynchronous, django, uv, rag, python

Using uv with PyTorch (via) PyTorch is a notoriously tricky piece of Python software to install, due to the need to provide separate wheels for different combinations of Python version and GPU accelerator (e.g. different CUDA versions).

uv now has dedicated documentation for PyTorch which I'm finding really useful - it clearly explains the challenge and then shows exactly how to configure a pyproject.toml such that uv knows which version of each package it should install from where.

# 19th November 2024, 11:20 pm / packaging, python, uv, pytorch, pip

uv 0.5.0. The first backwards-incompatible (in minor ways) release after 30 releases without a breaking change.

I found out about this release this morning when I filed an issue about a fiddly usability problem I had encountered with the combo of uv and conda... and learned that the exact problem had already been fixed in the brand new version!

# 8th November 2024, 11:54 pm / uv, packaging, python

ChainForge. I'm still on the hunt for good options for running evaluations against prompts. ChainForge offers an interesting approach, calling itself "an open-source visual programming environment for prompt engineering".

The interface is one of those boxes-and-lines visual programming tools, which reminds me of Yahoo Pipes.

It's open source (from a team at Harvard) and written in Python, which means you can run a local copy instantly via uvx like this:

uvx chainforge serve

You can then configure it with API keys to various providers (OpenAI worked for me, Anthropic models returned JSON parsing errors due to a 500 page from the ChainForge proxy) and start trying it out.

The "Add Node" menu shows the full list of capabilities.

The JavaScript and Python evaluation blocks are particularly interesting: the JavaScript one runs outside of a sandbox using plain eval(), while the Python one still runs in your browser but uses Pyodide in a Web Worker.

# 8th November 2024, 8:52 pm / pyodide, evals, uv, ai, llms, prompt-engineering, prompt-injection, python, javascript, generative-ai

Docling. MIT licensed document extraction Python library from the Deep Search team at IBM, who released Docling v2 on October 16th.

Here's the Docling Technical Report paper from August, which provides details of two custom models: a layout analysis model for figuring out the structure of the document (sections, figures, text, tables etc) and a TableFormer model specifically for extracting structured data from tables.

Those models are available on Hugging Face.

Here's how to try out the Docling CLI interface using uvx (avoiding the need to install it first - though since it downloads models it will take a while to run the first time):

uvx docling mydoc.pdf --to json --to md

This will output a mydoc.json file with complex layout information and a mydoc.md Markdown file which includes Markdown tables where appropriate.

The Python API is a lot more comprehensive. It can even extract tables as Pandas DataFrames:

from docling.document_converter import DocumentConverter
converter = DocumentConverter()
result = converter.convert("document.pdf")
for table in result.document.tables:
    df = table.export_to_dataframe()
    print(df)

I ran that inside uv run --with docling python. It took a little while to run, but it demonstrated that the library works.

# 3rd November 2024, 4:57 am / ocr, ai, pdf, python, ibm, hugging-face, uv

python-imgcat (via) I was investigating options for displaying images in a terminal window (for multi-modal logging output of LLM) and I found this neat Python library for displaying images using iTerm 2.

It includes a CLI tool, which means you can run it without installation using uvx like this:

uvx imgcat filename.png

# 28th October 2024, 5:13 am / llm, cli, python, uv

TIL: Using uv to develop Python command-line applications. I've been increasingly using uv to try out new software (via uvx) and experiment with new ideas, but I hadn't quite figured out the right way to use it for developing my own projects.

It turns out I was missing a few things - in particular the fact that there's no need to use uv pip at all when working with a local development environment, you can get by entirely on uv run (and maybe uv sync --extra test to install test dependencies) with no direct invocations of uv pip at all.

I bounced a few questions off Charlie Marsh and filled in the missing gaps - this TIL shows my new uv-powered process for hacking on Python CLI apps built using Click and my simonw/click-app cookecutter template.

# 24th October 2024, 5:56 am / uv, astral, charlie-marsh, python, cookiecutter, packaging, pip, til

sudoku-in-python-packaging (via) Absurdly clever hack by konsti: solve a Sudoku puzzle entirely using the Python package resolver!

First convert the puzzle into a requirements.in file representing the current state of the board:

git clone https://github.com/konstin/sudoku-in-python-packaging
cd sudoku-in-python-packaging
echo '5,3,_,_,7,_,_,_,_                                        
6,_,_,1,9,5,_,_,_
_,9,8,_,_,_,_,6,_
8,_,_,_,6,_,_,_,3
4,_,_,8,_,3,_,_,1
7,_,_,_,2,_,_,_,6
_,6,_,_,_,_,2,8,_
_,_,_,4,1,9,_,_,5
_,_,_,_,8,_,_,7,9' > sudoku.csv
python csv_to_requirements.py sudoku.csv requirements.in

That requirements.in file now contains lines like this for each of the filled-in cells:

sudoku_0_0 == 5
sudoku_1_0 == 3
sudoku_4_0 == 7

Then run uv pip compile to convert that into a fully fleshed out requirements.txt file that includes all of the resolved dependencies, based on the wheel files in the packages/ folder:

uv pip compile \
  --find-links packages/ \
  --no-annotate \
  --no-header \
  requirements.in > requirements.txt

The contents of requirements.txt is now the fully solved board:

sudoku-0-0==5
sudoku-0-1==6
sudoku-0-2==1
sudoku-0-3==8
...

The trick is the 729 wheel files in packages/ - each with a name like sudoku_3_4-8-py3-none-any.whl. I decompressed that wheel and it included a sudoku_3_4-8.dist-info/METADATA file which started like this:

Name: sudoku_3_4
Version: 8
Metadata-Version: 2.2
Requires-Dist: sudoku_3_0 != 8
Requires-Dist: sudoku_3_1 != 8
Requires-Dist: sudoku_3_2 != 8
Requires-Dist: sudoku_3_3 != 8
...

With a !=8 line for every other cell on the board that cannot contain the number 8 due to the rules of Sudoku (if 8 is in the 3, 4 spot). Visualized:

So the trick here is that the Python dependency resolver (now lightning fast thanks to uv) reads those dependencies and rules out every package version that represents a number in an invalid position. The resulting version numbers represent the cell numbers for the solution.

How much faster? I tried the same thing with the pip-tools pip-compile command:

time pip-compile \   
  --find-links packages/ \
  --no-annotate \
  --no-header \
  requirements.in > requirements.txt

That took 17.72s. On the same machine the time pip uv compile... command took 0.24s.

Update: Here's an earlier implementation of the same idea by Artjoms Iškovs in 2022.

# 21st October 2024, 6:59 pm / uv, packaging, python

[red-knot] type inference/checking test framework (via) Ruff maintainer Carl Meyer recently landed an interesting new design for a testing framework. It's based on Markdown, and could be described as a form of "literate testing" - the testing equivalent of Donald Knuth's literate programming.

A markdown test file is a suite of tests, each test can contain one or more Python files, with optionally specified path/name. The test writes all files to an in-memory file system, runs red-knot, and matches the resulting diagnostics against Type: and Error: assertions embedded in the Python source as comments.

Test suites are Markdown documents with embedded fenced blocks that look like this:

```py
reveal_type(1.0) # revealed: float
```

Tests can optionally include a path= specifier, which can provide neater messages when reporting test failures:

```py path=branches_unify_to_non_union_type.py
def could_raise_returns_str() -> str:
    return 'foo'
...
```

A larger example test suite can be browsed in the red_knot_python_semantic/resources/mdtest directory.

This document on control flow for exception handlers (from this PR) is the best example I've found of detailed prose documentation to accompany the tests.

The system is implemented in Rust, but it's easy to imagine an alternative version of this idea written in Python as a pytest plugin. This feels like an evolution of the old Python doctest idea, except that tests are embedded directly in Markdown rather than being embedded in Python code docstrings.

... and it looks like such plugins exist already. Here are two that I've found so far:

• pytest-markdown-docs by Elias Freider and Modal Labs.
• sphinx.ext.doctest is a core Sphinx extension for running test snippets in documentation.
• pytest-doctestplus from the Scientific Python community, first released in 2011.

I tried pytest-markdown-docs by creating a doc.md file like this:

# Hello test doc

```py
assert 1 + 2 == 3
```

But this fails:

```py
assert 1 + 2 == 4
```

And then running it with uvx like this:

uvx --with pytest-markdown-docs pytest --markdown-docs

I got one pass and one fail:

_______ docstring for /private/tmp/doc.md __________
Error in code block:
```
10   assert 1 + 2 == 4
11   
```
Traceback (most recent call last):
  File "/private/tmp/tt/doc.md", line 10, in <module>
    assert 1 + 2 == 4
AssertionError

============= short test summary info ==============
FAILED doc.md::/private/tmp/doc.md
=========== 1 failed, 1 passed in 0.02s ============

I also just learned that the venerable Python doctest standard library module has the ability to run tests in documentation files too, with doctest.testfile("example.txt"): "The file content is treated as if it were a single giant docstring; the file doesn’t need to contain a Python program!"

# 16th October 2024, 8:43 pm / testing, rust, python, astral, markdown, ruff, pytest, uv

jefftriplett/django-startproject (via) Django's django-admin startproject and startapp commands include a --template option which can be used to specify an alternative template for generating the initial code.

Jeff Triplett actively maintains his own template for new projects, which includes the pattern that I personally prefer of keeping settings and URLs in a config/ folder. It also configures the development environment to run using Docker Compose.

The latest update adds support for Python 3.13, Django 5.1 and uv. It's neat how you can get started without even installing Django using uv run like this:

uv run --with=django django-admin startproject \
  --extension=ini,py,toml,yaml,yml \
  --template=https://github.com/jefftriplett/django-startproject/archive/main.zip \
  example_project

# 12th October 2024, 11:19 pm / uv, jeff-triplett, django, python, docker

otterwiki (via) It's been a while since I've seen a new-ish Wiki implementation, and this one by Ralph Thesen is really nice. It's written in Python (Flask + SQLAlchemy + mistune for Markdown + GitPython) and keeps all of the actual wiki content as Markdown files in a local Git repository.

The installation instructions are a little in-depth as they assume a production installation with Docker or systemd - I figured out this recipe for trying it locally using uv:

git clone https://github.com/redimp/otterwiki.git
cd otterwiki

mkdir -p app-data/repository
git init app-data/repository

echo "REPOSITORY='${PWD}/app-data/repository'" >> settings.cfg
echo "SQLALCHEMY_DATABASE_URI='sqlite:///${PWD}/app-data/db.sqlite'" >> settings.cfg
echo "SECRET_KEY='$(echo $RANDOM | md5sum | head -c 16)'" >> settings.cfg

export OTTERWIKI_SETTINGS=$PWD/settings.cfg
uv run --with gunicorn gunicorn --bind 127.0.0.1:8080 otterwiki.server:app

# 9th October 2024, 3:22 pm / python, wikis, uv, markdown, git, flask, sqlalchemy, sqlite

UV with GitHub Actions to run an RSS to README project. Jeff Triplett demonstrates a very neat pattern for using uv to run Python scripts with their dependencies inside of GitHub Actions. First, add uv to the workflow using the setup-uv action:

- uses: astral-sh/setup-uv@v3
  with:
    enable-cache: true
    cache-dependency-glob: "*.py"

This enables the caching feature, which stores uv's own cache of downloads from PyPI between runs. The cache-dependency-glob key ensures that this cache will be invalidated if any .py file in the repository is updated.

Now you can run Python scripts using steps that look like this:

- run: uv run fetch-rss.py

If that Python script begins with some dependency definitions (PEP 723) they will be automatically installed by uv run on the first run and reused from the cache in the future. From the start of fetch-rss.py:

# /// script
# requires-python = ">=3.11"
# dependencies = [
#     "feedparser",
#     "typer",
# ]
# ///

uv will download the required Python version and cache that as well.

# 5th October 2024, 11:39 pm / uv, jeff-triplett, github-actions, python

mlx-vlm (via) The MLX ecosystem of libraries for running machine learning models on Apple Silicon continues to expand. Prince Canuma is actively developing this library for running vision models such as Qwen-2 VL and Pixtral and LLaVA using Python running on a Mac.

I used uv to run it against this image with this shell one-liner:

uv run --with mlx-vlm \
  python -m mlx_vlm.generate \
  --model Qwen/Qwen2-VL-2B-Instruct \
  --max-tokens 1000 \
  --temp 0.0 \
  --image https://static.simonwillison.net/static/2024/django-roadmap.png \
  --prompt "Describe image in detail, include all text"

The --image option works equally well with a URL or a path to a local file on disk.

This first downloaded 4.1GB to my ~/.cache/huggingface/hub/models--Qwen--Qwen2-VL-2B-Instruct folder and then output this result, which starts:

The image is a horizontal timeline chart that represents the release dates of various software versions. The timeline is divided into years from 2023 to 2029, with each year represented by a vertical line. The chart includes a legend at the bottom, which distinguishes between different types of software versions. [...]

# 29th September 2024, 9:38 pm / vision-llms, apple, python, generative-ai, uv, ai, llms, mlx, qwen

simonw/docs cookiecutter template. Over the last few years I’ve settled on the combination of Sphinx, the Furo theme and the myst-parser extension (enabling Markdown in place of reStructuredText) as my documentation toolkit of choice, maintained in GitHub and hosted using ReadTheDocs.

My LLM and shot-scraper projects are two examples of that stack in action.

Today I wanted to spin up a new documentation site so I finally took the time to construct a cookiecutter template for my preferred configuration. You can use it like this:

pipx install cookiecutter
cookiecutter gh:simonw/docs

Or with uv:

uv tool run cookiecutter gh:simonw/docs

Answer a few questions:

[1/3] project (): shot-scraper
[2/3] author (): Simon Willison
[3/3] docs_directory (docs):

And it creates a docs/ directory ready for you to start editing docs:

cd docs
pip install -r requirements.txt
make livehtml

# 23rd September 2024, 9:45 pm / uv, markdown, sphinx-docs, cookiecutter, read-the-docs, python, projects, documentation

Moshi (via) Moshi is "a speech-text foundation model and full-duplex spoken dialogue framework". It's effectively a text-to-text model - like an LLM but you input audio directly to it and it replies with its own audio.

It's fun to play around with, but it's not particularly useful in comparison to other pure text models: I tried to talk to it about California Brown Pelicans and it gave me some very basic hallucinated thoughts about California Condors instead.

It's very easy to run locally, at least on a Mac (and likely on other systems too). I used uv and got the 8 bit quantized version running as a local web server using this one-liner:

uv run --with moshi_mlx python -m moshi_mlx.local_web -q 8

That downloads ~8.17G of model to a folder in ~/.cache/huggingface/hub/ - or you can use -q 4 and get a 4.81G version instead (albeit even lower quality).

# 19th September 2024, 6:20 pm / generative-ai, uv, text-to-speech, ai, llms, mlx

Serializing package requirements in marimo notebooks. The latest release of Marimo - a reactive alternative to Jupyter notebooks - has a very neat new feature enabled by its integration with uv:

One of marimo’s goals is to make notebooks reproducible, down to the packages used in them. To that end, it’s now possible to create marimo notebooks that have their package requirements serialized into them as a top-level comment.

This takes advantage of the PEP 723 inline metadata mechanism, where a code comment at the top of a Python file can list package dependencies (and their versions).

I tried this out by installing marimo using uv:

uv tool install --python=3.12 marimo

Then grabbing one of their example notebooks:

wget 'https://raw.githubusercontent.com/marimo-team/spotlights/main/001-anywidget/tldraw_colorpicker.py'

And running it in a fresh dependency sandbox like this:

marimo run --sandbox tldraw_colorpicker.py

Also neat is that when editing a notebook using marimo edit:

marimo edit --sandbox notebook.py

Just importing a missing package is enough for Marimo to prompt to add that to the dependencies - at which point it automatically adds that package to the comment at the top of the file:

# 17th September 2024, 6:06 pm / uv, python, marimo

UV — I am (somewhat) sold (via) Oliver Andrich's detailed notes on adopting uv. Oliver has some pretty specific requirements:

I need to have various Python versions installed locally to test my work and my personal projects. Ranging from Python 3.8 to 3.13. [...] I also require decent dependency management in my projects that goes beyond manually editing a pyproject.toml file. Likewise, I am way too accustomed to poetry add .... And I run a number of Python-based tools --- djhtml, poetry, ipython, llm, mkdocs, pre-commit, tox, ...

He's braver than I am!

I started by removing all Python installations, pyenv, pipx and Homebrew from my machine. Rendering me unable to do my work.

Here's a neat trick: first install a specific Python version with uv like this:

uv python install 3.11

Then create an alias to run it like this:

alias python3.11 'uv run --python=3.11 python3'

And install standalone tools with optional extra dependencies like this (a replacement for pipx and pipx inject):

uv tool install --python=3.12 --with mkdocs-material mkdocs

Oliver also links to Anže Pečar's handy guide on using UV with Django.

# 15th September 2024, 2:54 pm / uv, astral, packaging, python, django

uv under discussion on Mastodon. Jacob Kaplan-Moss kicked off this fascinating conversation about uv on Mastodon recently. It's worth reading the whole thing, which includes input from a whole range of influential Python community members such as Jeff Triplett, Glyph Lefkowitz, Russell Keith-Magee, Seth Michael Larson, Hynek Schlawack, James Bennett and others. (Mastodon is a pretty great place for keeping up with the Python community these days.)

The key theme of the conversation is that, while uv represents a huge set of potential improvements to the Python ecosystem, it comes with additional risks due its attachment to a VC-backed company - and its reliance on Rust rather than Python.

Here are a few comments that stood out to me.

Russell:

As enthusiastic as I am about the direction uv is going, I haven't adopted them anywhere - because I want very much to understand Astral’s intended business model before I hook my wagon to their tools. It's definitely not clear to me how they're going to stay liquid once the VC money runs out. They could get me onboard in a hot second if they published a "This is what we're planning to charge for" blog post.

Hynek:

As much as I hate VC, [...] FOSS projects flame out all the time too. If Frost loses interest, there’s no PDM anymore. Same for Ofek and Hatch(ling).

I fully expect Astral to flame out and us having to fork/take over—it’s the circle of FOSS. To me uv looks like a genius sting to trick VCs into paying to fix packaging. We’ll be better off either way.

Glyph:

Even in the best case, Rust is more expensive and difficult to maintain, not to mention "non-native" to the average customer here. [...] And the difficulty with VC money here is that it can burn out all the other projects in the ecosystem simultaneously, creating a risk of monoculture, where previously, I think we can say that "monoculture" was the least of Python's packaging concerns.

Hynek on Rust:

I don’t think y’all quite grok what uv makes so special due to your seniority. The speed is really cool, but the reason Rust is elemental is that it’s one compiled blob that can be used to bootstrap and maintain a Python development. A blob that will never break because someone upgraded Homebrew, ran pip install or any other creative way people found to fuck up their installations. Python has shown to be a terrible tech to maintain Python.

Christopher Neugebauer:

Just dropping in here to say that corporate capture of the Python ecosystem is the #1 keeps-me-up-at-night subject in my community work, so I watch Astral with interest, even if I'm not yet too worried.

I'm reminded of this note from Armin Ronacher, who created Rye and later donated it to uv maintainers Astral:

However having seen the code and what uv is doing, even in the worst possible future this is a very forkable and maintainable thing. I believe that even in case Astral shuts down or were to do something incredibly dodgy licensing wise, the community would be better off than before uv existed.

I'm currently inclined to agree with Armin and Hynek: while the risk of corporate capture for a crucial aspect of the Python packaging and onboarding ecosystem is a legitimate concern, the amount of progress that has been made here in a relatively short time combined with the open license and quality of the underlying code keeps me optimistic that uv will be a net positive for Python overall.

Update: uv creator Charlie Marsh joined the conversation:

I don't want to charge people money to use our tools, and I don't want to create an incentive structure whereby our open source offerings are competing with any commercial offerings (which is what you see with a lost of hosted-open-source-SaaS business models).

What I want to do is build software that vertically integrates with our open source tools, and sell that software to companies that are already using Ruff, uv, etc. Alternatives to things that companies already pay for today.

An example of what this might look like (we may not do this, but it's helpful to have a concrete example of the strategy) would be something like an enterprise-focused private package registry. A lot of big companies use uv. We spend time talking to them. They all spend money on private package registries, and have issues with them. We could build a private registry that integrates well with uv, and sell it to those companies. [...]

But the core of what I want to do is this: build great tools, hopefully people like them, hopefully they grow, hopefully companies adopt them; then sell software to those companies that represents the natural next thing they need when building with Python. Hopefully we can build something better than the alternatives by playing well with our OSS, and hopefully we are the natural choice if they're already using our OSS.

# 8th September 2024, 4:23 pm / uv, glyph, russell-keith-magee, jacob-kaplan-moss, packaging, python, hynek-schlawack, armin-ronacher, mastodon, open-source, astral, rust, charlie-marsh

Docker images using uv’s python (via) Michael Kennedy interviewed uv/Ruff lead Charlie Marsh on his Talk Python podcast, and was inspired to try uv with Talk Python's own infrastructure, a single 8 CPU server running 17 Docker containers (status page here).

The key line they're now using is this:

RUN uv venv --python 3.12.5 /venv

Which downloads the uv selected standalone Python binary for Python 3.12.5 and creates a virtual environment for it at /venv all in one go.

# 6th September 2024, 11:54 pm / docker, uv, python, charlie-marsh

Why I Still Use Python Virtual Environments in Docker (via) Hynek Schlawack argues for using virtual environments even when running Python applications in a Docker container. This argument was most convincing to me:

I'm responsible for dozens of services, so I appreciate the consistency of knowing that everything I'm deploying is in /app, and if it's a Python application, I know it's a virtual environment, and if I run /app/bin/python, I get the virtual environment's Python with my application ready to be imported and run.

Also:

It’s good to use the same tools and primitives in development and in production.

Also worth a look: Hynek's guide to Production-ready Docker Containers with uv, an actively maintained guide that aims to reflect ongoing changes made to uv itself.

# 2nd September 2024, 11:57 pm / docker, python, hynek-schlawack, uv, virtualenv, packaging