Python 3.13 was released today. Datasette 1.0a16 is compatible with Python 3.13, but Datasette 0.64.8 was not. The new Datasette 0.65 release fixes compatibility with the new version of Python.

Datasette 1.0a14 includes some breaking changes to how metadata works for plugins, described in detail in the new upgrade guide. See also the annotated release notes that accompany this release.

Datasette 1.0a10 is a focused alpha that changes some internal details about how Datasette handles transactions. The datasette.execute_write_fn() internal method now wraps the function in a database transaction unless you pass transaction=False.

Datasette 1.0a9 adds basic alter table support to the JSON API, tweaks how permissions works and introduces some new plugin debugging utilities.

Datasette 1.0a8 introduces several new plugin hooks, a JavaScript plugin system and moves plugin configuration from metadata.yaml to datasette.yaml. Read more about the release in the annotated release notes for 1.0a8.

Datasette Enrichments is a new feature for Datasette that supports enriching data by running custom code against every selected row in a table. Read Datasette Enrichments: a new plugin framework for augmenting your data for more details, plus a video demo of enrichments for geocoding addresses and processing text and images using GPT-4.

datasette-comments is a new plugin by Alex Garcia which adds collaborative commenting to Datasette. Alex built the plugin for Datasette Cloud, but it's also available as an open source package for people who are hosting their own Datasette instances. See Annotate and explore your data with datasette-comments on the Datasette Cloud blog for more details.

Datasette 1.0a4 has a fix for a security vulnerability in the Datasette 1.0 alpha series: the API explorer interface exposed the names of private databases and tables in public instances that were protected by a plugin such as datasette-auth-passwords, though not the actual content of those tables. See the security advisory for more details and workarounds for if you can't upgrade immediately. The latest edition of the Datasette Newsletter also talks about this issue.

datasette-write-ui: a Datasette plugin for editing, inserting, and deleting rows introduces a new plugin adding add/edit/delete functionality to Datasette, developed by Alex Garcia. Alex built this for Datasette Cloud, and this post is the first announcement made on the new Datasette Cloud blog - see also Welcome to Datasette Cloud.

Datasette 1.0a3 is an alpha release of Datasette that previews the new default JSON API design that’s coming in version 1.0 - the single most significant change planned for that 1.0 release.

New tutorial: Data analysis with SQLite and Python. This tutorial, originally presented at PyCon 2023, includes a 2h45m video and an extensive handout that should be useful with or without the video. Topics covered include Python's sqlite3 module, sqlite-utils, Datasette, Datasette Lite, advanced SQL patterns and more.

I built a ChatGPT plugin to answer questions about data hosted in Datasette describes a new experimental Datasette plugin to enable people to query data hosted in a Datasette interface via ChatGPT, asking human language questions that are automatically converted to SQL and used to generate a readable response.

Using Datasette in GitHub Codespaces is a new tutorial showing how Datasette can be run in GitHub's free Codespaces browser-based development environments, using the new datasette-codespaces plugin.

Examples of sites built using Datasette now includes screenshots of Datasette deployments that illustrate a variety of problems that can be addressed using Datasette and its plugins.

Semantic search answers: Q&A against documentation with GPT3 + OpenAI embeddings shows how Datasette can be used to implement semantic search and build a system for answering questions against an existing corpus of text, using two new plugins: datasette-openai and datasette-faiss, and a new tool: openai-to-sqlite.

• Plugins can now reuse the sqlite-utils memory command with the new return_db=True parameter. #643

• Fixed a bug where llm chat crashes if a follow-up prompt is provided. #601

Support for attachments, allowing multi-modal models to accept images, audio, video and other formats. #578

The default OpenAI gpt-4o and gpt-4o-mini models can both now be prompted with JPEG, GIF, PNG and WEBP images.

Attachments in the CLI can be URLs:

llm -m gpt-4o "describe this image" \
  -a https://static.simonwillison.net/static/2024/pelicans.jpg

Or file paths:

llm -m gpt-4o-mini "extract text" -a image1.jpg -a image2.jpg

Or binary data, which may need to use --attachment-type to specify the MIME type:

cat image | llm -m gpt-4o-mini "extract text" --attachment-type - image/jpeg

Attachments are also available in the Python API:

model = llm.get_model("gpt-4o-mini")
response = model.prompt(
    "Describe these images",
    attachments=[
        llm.Attachment(path="pelican.jpg"),
        llm.Attachment(url="https://static.simonwillison.net/static/2024/pelicans.jpg"),
    ]
)

Plugins that provide alternative models can support attachments, see Attachments for multi-modal models for details.

The latest llm-claude-3 plugin now supports attachments for Anthropic's Claude 3 and 3.5 models. The llm-gemini plugin supports attachments for Google's Gemini 1.5 models.

Also in this release: OpenAI models now record their "usage" data in the database even when the response was streamed. These records can be viewed using llm logs --json. #591

Alpha support for attachments, allowing multi-modal models to accept images, audio, video and other formats. #578

Attachments in the CLI can be URLs:

llm "describe this image" \
  -a https://static.simonwillison.net/static/2024/pelicans.jpg

Or file paths:

llm "extract text" -a image1.jpg -a image2.jpg

Or binary data, which may need to use --attachment-type to specify the MIME type:

cat image | llm "extract text" --attachment-type - image/jpeg

Attachments are also available in the Python API:

model = llm.get_model("gpt-4o-mini")
response = model.prompt(
    "Describe these images",
    attachments=[
        llm.Attachment(path="pelican.jpg"),
        llm.Attachment(url="https://static.simonwillison.net/static/2024/pelicans.jpg"),
    ]
)

Plugins that provide alternative models can support attachments, see Attachments for multi-modal models for details.

• Upgrade for compatibility with Python 3.13 (by vendoring Pint dependency). (#2434)
• Dropped support for Python 3.8.

• Several new features for the YAML configuration used by shot-scraper multi:
• You can now add a - server: python -m http.server 8003 block to start a server running before screenshots are taken. The PID for this server will be recorded and the server automatically terminated when the command completes, unless you specify the --leave-server option in which case it will be left running, useful for debugging. #156
• The sh: shell command or python: python code blocks can specify Python or shell commands to run before a screenshot is taken. This means a YAML script can make modifications to the environment in between screenshots, useful for things like progressive tutorials. #155
• Fixed a bug that occurred if a max-width was accidentally applied to the <div> used for region screenshots. Thanks, Johann Klähn. #143
• Documented that shot-scraper will quit with an error if a --wait-for expression has not resolved in 30s.

• OpenAI models now use the internal self.get_key() mechanism, which means they can be used from Python code in a way that will pick up keys that have been configured using llm keys set or the OPENAI_API_KEY environment variable. #552. This code now works correctly: python import llm print(llm.get_model("gpt-4o-mini").prompt("hi"))
• New documented API methods: llm.get_default_model(), llm.set_default_model(alias), llm.get_default_embedding_model(alias), llm.set_default_embedding_model(). #553
• Support for OpenAI's new o1 family of preview models, llm -m o1-preview "prompt" and llm -m o1-mini "prompt". These models are currently only available to tier 5 OpenAI API users, though this may change in the future. #570

• New feature: --extra key "python format string", for adding additional output keys to the human-readable version of the diff. #38
• Don't crash in JSON format mode if some JSON keys are missing. #13
• Now includes a Dockerfile and instructions for building and running it that way. Thanks, @gourk. #11

This release focuses on performance, in particular against large tables, and introduces some minor breaking changes for CSS styling in Datasette plugins. - Removed the unit conversions feature and its dependency, Pint. This means Datasette is now compatible with the upcoming Python 3.13. (#2400, #2320) - The datasette --pdb option now uses the ipdb debugger if it is installed. You can install it using datasette install ipdb. Thanks, Tiago Ilieve. (#2342) - Fixed a confusing error that occurred if metadata.json contained nested objects. (#2403) - Fixed a bug with ?_trace=1 where it returned a blank page if the response was larger than 256KB. (#2404) - Tracing mechanism now also displays SQL queries that returned errors or ran out of time. datasette-pretty-traces 0.5 includes support for displaying this new type of trace. (#2405) - Fixed a text spacing with table descriptions on the homepage. (#2399) - Performance improvements for large tables: - Suggested facets now only consider the first 1000 rows. (#2406) - Improved performance of date facet suggestion against large tables. (#2407) - Row counts stop at 10,000 rows when listing tables. (#2398) - On table page the count stops at 10,000 rows too, with a "count all" button to execute the full count. (#2408) - New .dicts() internal method on Results that returns a list of dictionaries representing the results from a SQL query: (#2414) python rows = (await db.execute("select * from t")).dicts() - Default Datasette core CSS that styles inputs and buttons now requires a class of "core" on the element or a containing element, for example <form class="core">. (#2415) - Similarly, default table styles now only apply to <table class="rows-and-columns">. (#2420)

• More CSS compatibility with future Datasette alpha releases. #21

• Fixed CSS for compatibility with upcoming Datasette 1.0 alpha. datasette#2417

• Now skips tables ending in _fts to avoid duplicate search results in most recent Datasette alphas.
• Fixed CSS for compatibility with upcoming Datasette 1.0 alpha. #20, datasette#2417

• Database write page URLs are now /db/-/write, the old /-/write?database=db page redirects to the new location. #12
• Removed the application menu item so as not to duplicate the database action item.
• New row action menu item (on Datasette 1.0a13 and higher) linking to a SQL query to edit the current values of that row. #10
• Fixed CSS for compatibility with upcoming Datasette 1.0 alpha. datasette#2417

• Fix for compatibility with the base_url setting. Thanks, @meowcat #25
• autocomplete="off" on password field. Thanks, @brandonrobertz #24
• Fixed CSS for compatibility with upcoming Datasette 1.0 alpha. datasette#2417

• Fixed CSS for compatibility with upcoming Datasette 1.0 alpha. datasette#2417