Datasette¶
A tool for exploring and publishing data
Datasette is a tool for exploring and publishing data. It helps people take data of any shape or size and publish that as an interactive, explorable website and accompanying API.
Datasette is aimed at data journalists, museum curators, archivists, local governments and anyone else who has data that they wish to share with the world. It is part of a wider ecosystem of tools dedicated to make working with structured data as productive as possible.
Explore a demo, watch a presentation about the project or Try Datasette without installing anything using Glitch.
More examples: https://github.com/simonw/datasette/wiki/Datasettes
Contents¶
Getting started¶
Play with a live demo¶
The best way to experience Datasette for the first time is with a demo:
- fivethirtyeight.datasettes.com shows Datasette running against over 400 datasets imported from the FiveThirtyEight GitHub repository.
- sf-trees.datasettes.com demonstrates the datasette-cluster-map plugin running against 190,000 trees imported from data.sfgov.org.
Try Datasette without installing anything using Glitch¶
Glitch is a free online tool for building web apps directly from your web browser. You can use Glitch to try out Datasette without needing to install any software on your own computer.
Here’s a demo project on Glitch which you can use as the basis for your own experiments:
Glitch allows you to “remix” any project to create your own copy and start editing it in your browser. You can remix the datasette-csvs
project by clicking this button:
Find a CSV file and drag it onto the Glitch file explorer panel - datasette-csvs
will automatically convert it to a SQLite database (using sqlite-utils) and allow you to start exploring it using Datasette.
If your CSV file has a latitude
and longitude
column you can visualize it on a map by uncommenting the datasette-cluster-map
line in the requirements.txt
file using the Glitch file editor.
Need some data? Try this Public Art Data for the city of Seattle - hit “Export” and select “CSV” to download it as a CSV file.
For more on how this works, see Running Datasette on Glitch.
Using Datasette on your own computer¶
First, follow the Installation instructions. Now you can run Datasette against a SQLite file on your computer using the following command:
datasette serve path/to/database.db
This will start a web server on port 8001 - visit http://localhost:8001/ to access the web interface.
serve
is the default subcommand, you can omit it if you like.
Use Chrome on OS X? You can run datasette against your browser history like so:
datasette ~/Library/Application\ Support/Google/Chrome/Default/History
Now visiting http://localhost:8001/History/downloads will show you a web interface to browse your downloads data:
http://localhost:8001/History/downloads.json will return that data as JSON:
{
"database": "History",
"columns": [
"id",
"current_path",
"target_path",
"start_time",
"received_bytes",
"total_bytes",
...
],
"rows": [
[
1,
"/Users/simonw/Downloads/DropboxInstaller.dmg",
"/Users/simonw/Downloads/DropboxInstaller.dmg",
13097290269022132,
626688,
0,
...
]
]
}
http://localhost:8001/History/downloads.json?_shape=objects will return that data as JSON in a more convenient but less efficient format:
{
...
"rows": [
{
"start_time": 13097290269022132,
"interrupt_reason": 0,
"hash": "",
"id": 1,
"site_url": "",
"referrer": "https://www.dropbox.com/downloading?src=index",
...
}
]
}
datasette serve options¶
$ datasette serve --help
Usage: datasette serve [OPTIONS] [FILES]...
Serve up specified SQLite database files with a web UI
Options:
-i, --immutable PATH Database files to open in immutable mode
-h, --host TEXT Host for server. Defaults to 127.0.0.1 which means only
connections from the local machine will be allowed. Use
0.0.0.0 to listen to all IPs and allow access from other
machines.
-p, --port INTEGER Port for server, defaults to 8001
--debug Enable debug mode - useful for development
--reload Automatically reload if database or code change detected -
useful for development
--cors Enable CORS by serving Access-Control-Allow-Origin: *
--load-extension PATH Path to a SQLite extension to load
--inspect-file TEXT Path to JSON file created using "datasette inspect"
-m, --metadata FILENAME Path to JSON file containing license/source metadata
--template-dir DIRECTORY Path to directory containing custom templates
--plugins-dir DIRECTORY Path to directory containing custom plugins
--static STATIC MOUNT mountpoint:path-to-directory for serving static files
--memory Make :memory: database available
--config CONFIG Set config option using configname:value
datasette.readthedocs.io/en/latest/config.html
--version-note TEXT Additional note to show on /-/versions
--help-config Show available config options
--help Show this message and exit.
Installation¶
Note
If you just want to try Datasette out you don’t need to install anything: see Try Datasette without installing anything using Glitch
There are two main options for installing Datasette. You can install it directly on to your machine, or you can install it using Docker.
Using Docker¶
A Docker image containing the latest release of Datasette is published to Docker Hub here: https://hub.docker.com/r/datasetteproject/datasette/
If you have Docker installed (for example with Docker for Mac on OS X) you can download and run this image like so:
docker run -p 8001:8001 -v `pwd`:/mnt \
datasetteproject/datasette \
datasette -p 8001 -h 0.0.0.0 /mnt/fixtures.db
This will start an instance of Datasette running on your machine’s port 8001,
serving the fixtures.db
file in your current directory.
Now visit http://127.0.0.1:8001/ to access Datasette.
(You can download a copy of fixtures.db
from
https://latest.datasette.io/fixtures.db )
Loading Spatialite¶
The datasetteproject/datasette
image includes a recent version of the
SpatiaLite extension for SQLite. To load and enable that
module, use the following command:
docker run -p 8001:8001 -v `pwd`:/mnt \
datasetteproject/datasette \
datasette -p 8001 -h 0.0.0.0 /mnt/fixtures.db \
--load-extension=/usr/local/lib/mod_spatialite.so
You can confirm that SpatiaLite is successfully loaded by visiting http://127.0.0.1:8001/-/versions
Installing plugins¶
If you want to install plugins into your local Datasette Docker image you can do
so using the following recipe. This will install the plugins and then save a
brand new local image called datasette-with-plugins
:
docker run datasetteproject/datasette \
pip install datasette-vega
docker commit $(docker ps -lq) datasette-with-plugins
You can now run the new custom image like so:
docker run -p 8001:8001 -v `pwd`:/mnt \
datasette-with-plugins \
datasette -p 8001 -h 0.0.0.0 /mnt/fixtures.db
You can confirm that the plugins are installed by visiting http://127.0.0.1:8001/-/plugins
Install using pip¶
To run Datasette without Docker you will need Python 3.6 or higher.
You can install Datasette and its dependencies using pip
:
pip install datasette
The last version to support Python 3.5 was 0.30.2 - you can install that version like so:
pip install datasette==0.30.2
If you want to install Datasette in its own virtual environment, use this:
python -mvenv datasette-venv
source datasette-venv/bin/activate
pip install datasette
You can now run Datasette like so:
datasette fixtures.db
If you want to start making contributions to the Datasette project by installing a copy that lets you directly modify the code, take a look at our guide to Setting up a development environment.
The Datasette Ecosystem¶
Datasette sits at the center of a growing ecosystem of open source tools aimed at making it as easy as possible to gather, analyze and publish interesting data.
These tools are divided into two main groups: tools for building SQLite databases (for use with Datasette) and plugins that extend Datasette’s functionality.
Tools for creating SQLite databases¶
csvs-to-sqlite¶
csvs-to-sqlite lets you take one or more CSV files and load them into a SQLite database. It can also extract repeated columns out into a separate table and configure SQLite full-text search against the contents of specific columns.
sqlite-utils¶
sqlite-utils is a Python library and CLI tool that provides shortcuts for loading data into SQLite. It can be used programmatically (e.g. in a Jupyter notebook) to load data, and will automatically create SQLite tables with the necessary schema.
The CLI tool can consume JSON streams directly and use them to create tables. It can also be used to query SQLite databases and output the results as CSV or JSON.
See sqlite-utils: a Python library and CLI tool for building SQLite databases for more.
db-to-sqlite¶
db-to-sqlite is a CLI tool that builds on top of SQLAlchemy and allows you to connect to any database supported by that library (including MySQL, oracle and PostgreSQL), run a SQL query and save the results to a new table in a SQLite database.
You can mirror an entire database (including copying foreign key relationships) with the --all
option:
$ db-to-sqlite --connection="postgresql://simonw@localhost/myblog" --all blog.db
dbf-to-sqlite¶
dbf-to-sqlite works with dBase files such as those produced by Visual FoxPro. It is a command-line tool that can convert one or more .dbf
file to tables in a SQLite database.
markdown-to-sqlite¶
markdown-to-sqlite reads Markdown files with embedded YAML metadata (e.g. for Jekyll Front Matter) and creates a SQLite table with a schema matching the metadata. This is useful if you want to keep structured data in text form in a GitHub repository and use that to build a SQLite database.
socrata2sql¶
socrata2sql is a tool by Andrew Chavez at the Dallas Morning News. It works with Socrata, a widely used platform for local and national government open data portals. It uses the Socrata API to pull down government datasets and store them in a local SQLite database (it can also export data to PostgreSQL, MySQL and other SQLAlchemy-supported databases).
For example, to create a SQLite database of the City of Dallas Payment Register you would run the following command:
$ socrata2sql insert www.dallasopendata.com 64pp-jeba
Datasette Plugins¶
Datasette’s plugin system makes it easy to enhance Datasette with additional functionality.
datasette-cluster-map¶
datasette-cluster-map is the original Datasette plugin, described in Datasette plugins, and building a clustered map visualization.
The plugin works against any table with latitude and longitude columns. It can load over 100,000 points onto a map to visualize the geographical distribution of the underlying data.
datasette-vega¶
datasette-vega exposes the powerful Vega charting library, allowing you to construct line, bar and scatter charts against your data and share links to your visualizations.
datasette-auth-github¶
datasette-auth-github adds an authentication layer to Datasette. Users will have to sign in using their GitHub account before they can view data or interact with Datasette. You can also use it to restrict access to specific GitHub users, or to members of specified GitHub organizations or teams.
datasette-json-html¶
datasette-json-html renders HTML in Datasette’s table view driven by JSON returned from your SQL queries. This provides a way to embed images, links and lists of links directly in Datasette’s main interface, defined using custom SQL statements.
datasette-jellyfish¶
datasette-jellyfish exposes custom SQL functions for a range of common fuzzy string matching functions, including soundex, porter stemming and levenshtein distance. It builds on top of the Jellyfish Python library.
datasette-doublemetaphone¶
datasette-doublemetaphone by Matthew Somerville adds custom SQL functions for applying the Double Metaphone fuzzy “sounds like” algorithm.
datasette-jq¶
datasette-jq adds a custom SQL function for filtering and transforming values from JSON columns using the jq expression language.
datasette-render-images¶
datasette-render-images works with SQLite tables that contain binary image data in BLOB columns. It converts any images it finds into data-uri
image elements, allowing you to view them directly in the Datasette interface.
datasette-render-binary¶
datasette-render-binary renders binary data in a slightly more readable fashion: it shows ASCII characters as they are, and shows all other data as monospace octets. Useful as a tool for exploring new unfamiliar databases as it makes it easier to spot if a binary column may contain a decipherable binary format.
datasette-pretty-json¶
datasette-pretty-json seeks out JSON values in Datasette’s table browsing interface and pretty-prints them, making them easier to read.
datasette-sqlite-fts4¶
datasette-sqlite-fts4 provides search relevance ranking algorithms that can be used with SQLite’s FTS4 search module. You can read more about it in Exploring search relevance algorithms with SQLite.
datasette-bplist¶
datasette-bplist provides tools for working with Apple’s binary plist format embedded in SQLite database tables. If you use OS X you already have dozens of SQLite databases hidden away in your ~/Library
folder that include data in this format - this plugin allows you to view the decoded data and run SQL queries against embedded values using a bplist_to_json(value)
custom SQL function.
datasette-cors¶
datasette-cors allows you to configure CORS headers for your Datasette instance. You can use this to enable JavaScript running on a whitelisted set of domains to make fetch()
calls to the JSON API provided by your Datasette instance.
Pages and API endpoints¶
The Datasette web application offers a number of different pages that can be accessed to explore the data in question, each of which is accompanied by an equivalent JSON API.
Top-level index¶
The root page of any Datasette installation is an index page that lists all of the currently attached databases. Some examples:
- fivethirtyeight.datasettes.com
- global-power-plants.datasettes.com
- register-of-members-interests.datasettes.com
Add /.json
to the end of the URL for the JSON version of the underlying data:
Database¶
Each database has a page listing the tables, views and canned queries available for that database. If the allow_sql config option is enabled (it’s turned on by default) there will also be an interface for executing arbitrary SQL select queries against the data.
Examples:
- fivethirtyeight.datasettes.com/fivethirtyeight-ac35616
- global-power-plants.datasettes.com/global-power-plants-9e55be2
The JSON version of this page provides programmatic access to the underlying data:
- fivethirtyeight.datasettes.com/fivethirtyeight-ac35616.json
- global-power-plants.datasettes.com/global-power-plants-9e55be2.json
Note that these URLs end in a 7 character hash. This hash is derived from the contents of the database, and ensures that each URL is immutable: the data returned from a URL containing the hash will always be the same, since if the contents of the database file changes by even a single byte a new hash will be generated.
If you access one of these URLs with an incorrect hash (say because a new version of the underlying database has been published) Datasette will 302 redirect you to the correct URL. This happens for all URLs below the database page as well.
Thanks to this hashing scheme, Datasette URLs can all be returned with far-future cache expiry headers. This means browsers will cache the data (including data from the JSON APIs) for a long time, and CDNs such as Cloudflare or Fastly can be used to dramatically improve the performance of a Datasette hosted API.
Table¶
The table page is the heart of Datasette: it allows users to interactively explore the contents of a database table, including sorting, filtering, Full-text search and applying Facets.
The HTML interface is worth spending some time exploring. As with other pages, you can return the JSON data by appending .json
to the URL path, before any ? querystring arguments.
The querystring arguments are described in more detail here: Table arguments
You can also use the table page to interactively construct a SQL query - by applying different filters and a sort order for example - and then click the “View and edit SQL” link to see the SQL query that was used for the page and edit and re-submit it.
Some examples:
- ../items lists all of the line-items registered by UK MPs as potential conflicts of interest. It demonstrates Datasette’s support for Full-text search.
- ../antiquities-act%2Factions_under_antiquities_act is an interface for exploring the “actions under the antiquities act” data table published by FiveThirtyEight.
- ../global-power-plants?country_long=United+Kingdom&primary_fuel=Gas is a filtered table page showing every Gas power plant in the United Kingdom. It includes some default facets (configured using its metadata.json) and uses the datasette-cluster-map plugin to show a map of the results.
Row¶
Every row in every Datasette table has its own URL. This means individual records can be linked to directly.
Table cells with extremely long text contents are truncated on the table view according to the truncate_cells_html setting. If a cell has been truncated the full length version of that cell will be available on the row page.
Rows which are the targets of foreign key references from other tables will show a link to a filtered search for all records that reference that row. Here’s an example from the Registers of Members Interests database:
../people/uk.org.publicwhip%2Fperson%2F10001
Note that this URL includes the encoded primary key of the record.
Here’s that same page as JSON:
Publishing data¶
Datasette includes tools for publishing and deploying your data to the internet. The datasette publish
command will deploy a new Datasette instance containing your databases directly to a Heroku, Google Cloud or Zeit Now hosting account. You can also use datasette package
to create a Docker image that bundles your databases together with the datasette application that is used to serve them.
datasette publish¶
Once you have created a SQLite database (e.g. using csvs-to-sqlite) you can deploy it to a hosting account using a single command.
You will need a hosting account with Heroku or Google Cloud. Once you have created your account you will need to install and configure the heroku
or gcloud
command-line tools.
Publishing to Heroku¶
To publish your data using Heroku, first create an account there and install and configure the Heroku CLI tool.
You can publish a database to Heroku using the following command:
datasette publish heroku mydatabase.db
This will output some details about the new deployment, including a URL like this one:
https://limitless-reef-88278.herokuapp.com/ deployed to Heroku
You can specify a custom app name by passing -n my-app-name
to the publish command. This will also allow you to overwrite an existing app.
$ datasette publish heroku --help
Usage: datasette publish heroku [OPTIONS] [FILES]...
Options:
-m, --metadata FILENAME Path to JSON file containing metadata to publish
--extra-options TEXT Extra options to pass to datasette serve
--branch TEXT Install datasette from a GitHub branch e.g. master
--template-dir DIRECTORY Path to directory containing custom templates
--plugins-dir DIRECTORY Path to directory containing custom plugins
--static STATIC MOUNT mountpoint:path-to-directory for serving static files
--install TEXT Additional packages (e.g. plugins) to install
--plugin-secret <TEXT TEXT TEXT>...
Secrets to pass to plugins, e.g. --plugin-secret
datasette-auth-github client_id xxx
--version-note TEXT Additional note to show on /-/versions
--title TEXT Title for metadata
--license TEXT License label for metadata
--license_url TEXT License URL for metadata
--source TEXT Source label for metadata
--source_url TEXT Source URL for metadata
--about TEXT About label for metadata
--about_url TEXT About URL for metadata
-n, --name TEXT Application name to use when deploying
--help Show this message and exit.
Publishing to Google Cloud Run¶
Google Cloud Run launched as a beta in in April 2019. It allows you to publish data in a scale-to-zero environment, so your application will start running when the first request is received and will shut down again when traffic ceases. This means you only pay for time spent serving traffic.
You will first need to install and configure the Google Cloud CLI tools by following these instructions.
You can then publish a database to Google Cloud Run using the following command:
datasette publish cloudrun mydatabase.db --service=my-database
A Cloud Run service is a single hosted application. The service name you specify will be used as part of the Cloud Run URL. If you deploy to a service name that you have used in the past your new deployment will replace the previous one.
If you omit the --service
option you will be asked to pick a service name interactively during the deploy.
You may need to interact with prompts from the tool. Once it has finished it will output a URL like this one:
Service [my-service] revision [my-service-00001] has been deployed
and is serving traffic at https://my-service-j7hipcg4aq-uc.a.run.app
$ datasette publish cloudrun --help
Usage: datasette publish cloudrun [OPTIONS] [FILES]...
Options:
-m, --metadata FILENAME Path to JSON file containing metadata to publish
--extra-options TEXT Extra options to pass to datasette serve
--branch TEXT Install datasette from a GitHub branch e.g. master
--template-dir DIRECTORY Path to directory containing custom templates
--plugins-dir DIRECTORY Path to directory containing custom plugins
--static STATIC MOUNT mountpoint:path-to-directory for serving static files
--install TEXT Additional packages (e.g. plugins) to install
--plugin-secret <TEXT TEXT TEXT>...
Secrets to pass to plugins, e.g. --plugin-secret
datasette-auth-github client_id xxx
--version-note TEXT Additional note to show on /-/versions
--title TEXT Title for metadata
--license TEXT License label for metadata
--license_url TEXT License URL for metadata
--source TEXT Source label for metadata
--source_url TEXT Source URL for metadata
--about TEXT About label for metadata
--about_url TEXT About URL for metadata
-n, --name TEXT Application name to use when building
--service TEXT Cloud Run service to deploy (or over-write)
--spatialite Enable SpatialLite extension
--show-files Output the generated Dockerfile and metadata.json
--help Show this message and exit.
Publishing to Zeit Now v1¶
Datasette can be deployed to Zeit Now’s older v1 hosting platform. They no longer accept new signups for this service, so this option is currently only available if you created an account before January 2019.
To publish your database(s) to a new instance hosted by Zeit Now v1, install the now cli tool and then run the following command:
datasette publish nowv1 mydatabase.db
This will upload your database to Zeit Now, assign you a new URL and install and start a new instance of Datasette to serve your database.
The command will output a URL that looks something like this:
https://datasette-elkksjmyfj.now.sh
You can navigate to this URL to see live logs of the deployment process. Your new Datasette instance will be available at that URL.
Once the deployment has completed, you can assign a custom URL to your instance using the now alias
command:
now alias https://datasette-elkksjmyfj.now.sh datasette-publish-demo.now.sh
You can use anything-you-like.now.sh
, provided no one else has already registered that alias.
You can also use custom domains, if you first register them with Zeit Now.
$ datasette publish nowv1 --help
Usage: datasette publish nowv1 [OPTIONS] [FILES]...
Options:
-m, --metadata FILENAME Path to JSON file containing metadata to publish
--extra-options TEXT Extra options to pass to datasette serve
--branch TEXT Install datasette from a GitHub branch e.g. master
--template-dir DIRECTORY Path to directory containing custom templates
--plugins-dir DIRECTORY Path to directory containing custom plugins
--static STATIC MOUNT mountpoint:path-to-directory for serving static files
--install TEXT Additional packages (e.g. plugins) to install
--plugin-secret <TEXT TEXT TEXT>...
Secrets to pass to plugins, e.g. --plugin-secret
datasette-auth-github client_id xxx
--version-note TEXT Additional note to show on /-/versions
--title TEXT Title for metadata
--license TEXT License label for metadata
--license_url TEXT License URL for metadata
--source TEXT Source label for metadata
--source_url TEXT Source URL for metadata
--about TEXT About label for metadata
--about_url TEXT About URL for metadata
-n, --name TEXT Application name to use when deploying
--force Pass --force option to now
--token TEXT Auth token to use for deploy
--alias TEXT Desired alias e.g. yoursite.now.sh
--spatialite Enable SpatialLite extension
--show-files Output the generated Dockerfile and metadata.json
--help Show this message and exit.
Custom metadata and plugins¶
datasette publish
accepts a number of additional options which can be used to further customize your Datasette instance.
You can define your own Metadata and deploy that with your instance like so:
datasette publish cloudrun --service=my-service mydatabase.db -m metadata.json
If you just want to set the title, license or source information you can do that directly using extra options to datasette publish
:
datasette publish cloudrun mydatabase.db --service=my-service \
--title="Title of my database" \
--source="Where the data originated" \
--source_url="http://www.example.com/"
You can also specify plugins you would like to install. For example, if you want to include the datasette-vega visualization plugin you can use the following:
datasette publish cloudrun mydatabase.db --service=my-service --install=datasette-vega
If a plugin has any Secret configuration values you can use the --plugin-secret
option to set those secrets at publish time. For example, using Heroku with datasette-auth-github you might run the following command:
$ datasette publish heroku my_database.db \
--name my-heroku-app-demo \
--install=datasette-auth-github \
--plugin-secret datasette-auth-github client_id your_client_id \
--plugin-secret datasette-auth-github client_secret your_client_secret
datasette package¶
If you have docker installed (e.g. using Docker for Mac) you can use the datasette package
command to create a new Docker image in your local repository containing the datasette app bundled together with your selected SQLite databases:
datasette package mydatabase.db
Here’s example output for the package command:
$ datasette package parlgov.db --extra-options="--config sql_time_limit_ms:2500"
Sending build context to Docker daemon 4.459MB
Step 1/7 : FROM python:3
---> 79e1dc9af1c1
Step 2/7 : COPY . /app
---> Using cache
---> cd4ec67de656
Step 3/7 : WORKDIR /app
---> Using cache
---> 139699e91621
Step 4/7 : RUN pip install datasette
---> Using cache
---> 340efa82bfd7
Step 5/7 : RUN datasette inspect parlgov.db --inspect-file inspect-data.json
---> Using cache
---> 5fddbe990314
Step 6/7 : EXPOSE 8001
---> Using cache
---> 8e83844b0fed
Step 7/7 : CMD datasette serve parlgov.db --port 8001 --inspect-file inspect-data.json --config sql_time_limit_ms:2500
---> Using cache
---> 1bd380ea8af3
Successfully built 1bd380ea8af3
You can now run the resulting container like so:
docker run -p 8081:8001 1bd380ea8af3
This exposes port 8001 inside the container as port 8081 on your host machine, so you can access the application at http://localhost:8081/
A full list of options can be seen by running datasette package --help
:
$ datasette package --help
Usage: datasette package [OPTIONS] FILES...
Package specified SQLite files into a new datasette Docker container
Options:
-t, --tag TEXT Name for the resulting Docker container, can optionally use
name:tag format
-m, --metadata FILENAME Path to JSON file containing metadata to publish
--extra-options TEXT Extra options to pass to datasette serve
--branch TEXT Install datasette from a GitHub branch e.g. master
--template-dir DIRECTORY Path to directory containing custom templates
--plugins-dir DIRECTORY Path to directory containing custom plugins
--static STATIC MOUNT mountpoint:path-to-directory for serving static files
--install TEXT Additional packages (e.g. plugins) to install
--spatialite Enable SpatialLite extension
--version-note TEXT Additional note to show on /-/versions
--title TEXT Title for metadata
--license TEXT License label for metadata
--license_url TEXT License URL for metadata
--source TEXT Source label for metadata
--source_url TEXT Source URL for metadata
--about TEXT About label for metadata
--about_url TEXT About URL for metadata
--help Show this message and exit.
JSON API¶
Datasette provides a JSON API for your SQLite databases. Anything you can do through the Datasette user interface can also be accessed as JSON via the API.
To access the API for a page, either click on the .json
link on that page or
edit the URL and add a .json
extension to it.
If you started Datasette with the --cors
option, each JSON endpoint will be
served with the following additional HTTP header:
Access-Control-Allow-Origin: *
This means JavaScript running on any domain will be able to make cross-origin requests to fetch the data.
If you start Datasette without the --cors
option only JavaScript running on
the same domain as Datasette will be able to access the API.
Different shapes¶
The default JSON representation of data from a SQLite table or custom query looks like this:
{
"database": "sf-trees",
"table": "qSpecies",
"columns": [
"id",
"value"
],
"rows": [
[
1,
"Myoporum laetum :: Myoporum"
],
[
2,
"Metrosideros excelsa :: New Zealand Xmas Tree"
],
[
3,
"Pinus radiata :: Monterey Pine"
]
],
"truncated": false,
"next": "100",
"next_url": "http://127.0.0.1:8001/sf-trees-02c8ef1/qSpecies.json?_next=100",
"query_ms": 1.9571781158447266
}
The columns
key lists the columns that are being returned, and the rows
key then returns a list of lists, each one representing a row. The order of the
values in each row corresponds to the columns.
The _shape
parameter can be used to access alternative formats for the
rows
key which may be more convenient for your application. There are three
options:
?_shape=arrays
-"rows"
is the default option, shown above?_shape=objects
-"rows"
is a list of JSON key/value objects?_shape=array
- an JSON array of objects?_shape=array&_nl=on
- a newline-separated list of JSON objects?_shape=arrayfirst
- a flat JSON array containing just the first value from each row?_shape=object
- a JSON object keyed using the primary keys of the rows
_shape=objects
looks like this:
{
"database": "sf-trees",
...
"rows": [
{
"id": 1,
"value": "Myoporum laetum :: Myoporum"
},
{
"id": 2,
"value": "Metrosideros excelsa :: New Zealand Xmas Tree"
},
{
"id": 3,
"value": "Pinus radiata :: Monterey Pine"
}
]
}
_shape=array
looks like this:
[
{
"id": 1,
"value": "Myoporum laetum :: Myoporum"
},
{
"id": 2,
"value": "Metrosideros excelsa :: New Zealand Xmas Tree"
},
{
"id": 3,
"value": "Pinus radiata :: Monterey Pine"
}
]
_shape=array&_nl=on
looks like this:
{"id": 1, "value": "Myoporum laetum :: Myoporum"}
{"id": 2, "value": "Metrosideros excelsa :: New Zealand Xmas Tree"}
{"id": 3, "value": "Pinus radiata :: Monterey Pine"}
_shape=arrayfirst
looks like this:
[1, 2, 3]
_shape=object
looks like this:
{
"1": {
"id": 1,
"value": "Myoporum laetum :: Myoporum"
},
"2": {
"id": 2,
"value": "Metrosideros excelsa :: New Zealand Xmas Tree"
},
"3": {
"id": 3,
"value": "Pinus radiata :: Monterey Pine"
}
]
The object
shape is only available for queries against tables - custom SQL
queries and views do not have an obvious primary key so cannot be returned using
this format.
The object
keys are always strings. If your table has a compound primary
key, the object
keys will be a comma-separated string.
Special JSON arguments¶
Every Datasette endpoint that can return JSON also accepts the following querystring arguments:
?_shape=SHAPE
- The shape of the JSON to return, documented above.
?_nl=on
- When used with
?_shape=array
produces newline-delimited JSON objects. ?_json=COLUMN1&_json=COLUMN2
If any of your SQLite columns contain JSON values, you can use one or more
_json=
parameters to request that those columns be returned as regular JSON. Without this argument those columns will be returned as JSON objects that have been double-encoded into a JSON string value.Compare this query without the argument to this query using the argument
?_json_infinity=on
- If your data contains infinity or -infinity values, Datasette will replace
them with None when returning them as JSON. If you pass
_json_infinity=1
Datasette will instead return them asInfinity
or-Infinity
which is invalid JSON but can be processed by some custom JSON parsers. ?_timelimit=MS
- Sets a custom time limit for the query in ms. You can use this for optimistic queries where you would like Datasette to give up if the query takes too long, for example if you want to implement autocomplete search but only if it can be executed in less than 10ms.
?_ttl=SECONDS
- For how many seconds should this response be cached by HTTP proxies? Use
?_ttl=0
to disable HTTP caching entirely for this request.
Table arguments¶
The Datasette table view takes a number of special querystring arguments.
Column filter arguments¶
You can filter the data returned by the table based on column values using a querystring argument.
?column__exact=value
or?_column=value
- Returns rows where the specified column exactly matches the value.
?column__not=value
- Returns rows where the column does not match the value.
?column__contains=value
- Rows where the string column contains the specified value (
column like "%value%"
in SQL). ?column__endswith=value
- Rows where the string column ends with the specified value (
column like "%value"
in SQL). ?column__startswith=value
- Rows where the string column starts with the specified value (
column like "value%"
in SQL). ?column__gt=value
- Rows which are greater than the specified value.
?column__gte=value
- Rows which are greater than or equal to the specified value.
?column__lt=value
- Rows which are less than the specified value.
?column__lte=value
- Rows which are less than or equal to the specified value.
?column__like=value
- Match rows with a LIKE clause, case insensitive and with
%
as the wildcard character. ?column__glob=value
- Similar to LIKE but uses Unix wildcard syntax and is case sensitive.
?column__in=value1,value2,value3
Rows where column matches any of the provided values.
You can use a comma separated string, or you can use a JSON array.
The JSON array option is useful if one of your matching values itself contains a comma:
?column__in=["value","value,with,commas"]
?column__notin=value1,value2,value3
- Rows where column does not match any of the provided values. The inverse of
__in=
. Also supports JSON arrays. ?column__arraycontains=value
Works against columns that contain JSON arrays - matches if any of the values in that array match.
This is only available if the
json1
SQLite extension is enabled.?column__date=value
- Column is a datestamp occurring on the specified YYYY-MM-DD date, e.g.
2018-01-02
. ?column__isnull=1
- Matches rows where the column is null.
?column__notnull=1
- Matches rows where the column is not null.
?column__isblank=1
- Matches rows where the column is blank, meaning null or the empty string.
?column__notblank=1
- Matches rows where the column is not blank.
Special table arguments¶
?_labels=on/off
- Expand foreign key references for every possible column. See below.
?_label=COLUMN1&_label=COLUMN2
- Expand foreign key references for one or more specified columns.
?_size=1000
or?_size=max
- Sets a custom page size. This cannot exceed the
max_returned_rows
limit passed todatasette serve
. Usemax
to getmax_returned_rows
. ?_sort=COLUMN
- Sorts the results by the specified column.
?_sort_desc=COLUMN
- Sorts the results by the specified column in descending order.
?_search=keywords
- For SQLite tables that have been configured for full-text search executes a search with the provided keywords.
?_search_COLUMN=keywords
- Like
_search=
but allows you to specify the column to be searched, as opposed to searching all columns that have been indexed by FTS. ?_where=SQL-fragment
If the allow_sql config option is enabled, this parameter can be used to pass one or more additional SQL fragments to be used in the WHERE clause of the SQL used to query the table.
This is particularly useful if you are building a JavaScript application that needs to do something creative but still wants the other conveniences provided by the table view (such as faceting) and hence would like not to have to construct a completely custom SQL query.
Some examples:
?_through={json}
This can be used to filter rows via a join against another table.
The JSON parameter must include three keys:
table
,column
andvalue
.table
must be a table that the current table is related to via a foreign key relationship.column
must be a column in that other table.value
is the value that you want to match against.For example, to filter
roadside_attractions
to just show the attractions that have a characteristic of “museum”, you would construct this JSON:{ "table": "roadside_attraction_characteristics", "column": "characteristic_id", "value": "1" }
As a URL, that looks like this:
?_through={%22table%22:%22roadside_attraction_characteristics%22,%22column%22:%22characteristic_id%22,%22value%22:%221%22}
Here’s an example.
?_next=TOKEN
- Pagination by continuation token - pass the token that was returned in the
"next"
property by the previous page. ?_trace=1
Turns on tracing for this page: SQL queries executed during the request will be gathered and included in the response, either in a new
"_traces"
key for JSON responses or at the bottom of the page if the response is in HTML.The structure of the data returned here should be considered highly unstable and very likely to change.
Expanding foreign key references¶
Datasette can detect foreign key relationships and resolve those references into
labels. The HTML interface does this by default for every detected foreign key
column - you can turn that off using ?_labels=off
.
You can request foreign keys be expanded in JSON using the _labels=on
or
_label=COLUMN
special querystring parameters. Here’s what an expanded row
looks like:
[
{
"rowid": 1,
"TreeID": 141565,
"qLegalStatus": {
"value": 1,
"label": "Permitted Site"
},
"qSpecies": {
"value": 1,
"label": "Myoporum laetum :: Myoporum"
},
"qAddress": "501X Baker St",
"SiteOrder": 1
}
]
The column in the foreign key table that is used for the label can be specified
in metadata.json
- see Specifying the label column for a table.
Running SQL queries¶
Datasette treats SQLite database files as read-only and immutable. This means it is not possible to execute INSERT or UPDATE statements using Datasette, which allows us to expose SELECT statements to the outside world without needing to worry about SQL injection attacks.
The easiest way to execute custom SQL against Datasette is through the web UI. The database index page includes a SQL editor that lets you run any SELECT query you like. You can also construct queries using the filter interface on the tables page, then click “View and edit SQL” to open that query in the custom SQL editor.
Note that this interface is only available if the allow_sql option has not been disabled.
Any Datasette SQL query is reflected in the URL of the page, allowing you to bookmark them, share them with others and navigate through previous queries using your browser back button.
You can also retrieve the results of any query as JSON by adding .json
to
the base URL.
Named parameters¶
Datasette has special support for SQLite named parameters. Consider a SQL query like this:
select * from Street_Tree_List
where "PermitNotes" like :notes
and "qSpecies" = :species
If you execute this query using the custom query editor, Datasette will extract the two named parameters and use them to construct form fields for you to provide values.
You can also provide values for these fields by constructing a URL:
/mydatabase?sql=select...&species=44
SQLite string escaping rules will be applied to values passed using named parameters - they will be wrapped in quotes and their content will be correctly escaped.
Datasette disallows custom SQL containing the string PRAGMA, as SQLite pragma statements can be used to change database settings at runtime. If you need to include the string “pragma” in a query you can do so safely using a named parameter.
Views¶
If you want to bundle some pre-written SQL queries with your Datasette-hosted database you can do so in two ways. The first is to include SQL views in your database - Datasette will then list those views on your database index page.
The easiest way to create views is with the SQLite command-line interface:
$ sqlite3 sf-trees.db
SQLite version 3.19.3 2017-06-27 16:48:08
Enter ".help" for usage hints.
sqlite> CREATE VIEW demo_view AS select qSpecies from Street_Tree_List;
<CTRL+D>
Canned queries¶
As an alternative to adding views to your database, you can define canned
queries inside your metadata.json
file. Here’s an example:
{
"databases": {
"sf-trees": {
"queries": {
"just_species": {
"sql": "select qSpecies from Street_Tree_List"
}
}
}
}
}
Then run datasette like this:
datasette sf-trees.db -m metadata.json
Each canned query will be listed on the database index page, and will also get its own URL at:
/database-name/canned-query-name
For the above example, that URL would be:
/sf-trees/just_species
You can optionally include "title"
and "description"
keys to show a
title and description on the canned query page. As with regular table metadata
you can alternatively specify "description_html"
to have your description
rendered as HTML (rather than having HTML special characters escaped).
Canned queries support named parameters, so if you include those in the SQL you will then be able to enter them using the form fields on the canned query page or by adding them to the URL. This means canned queries can be used to create custom JSON APIs based on a carefully designed SQL statement.
Here’s an example of a canned query with a named parameter:
select neighborhood, facet_cities.name, state
from facetable join facet_cities on facetable.city_id = facet_cities.id
where neighborhood like '%' || :text || '%' order by neighborhood;
In the canned query JSON it looks like this:
{
"databases": {
"fixtures": {
"queries": {
"neighborhood_search": {
"sql": "select neighborhood, facet_cities.name, state\nfrom facetable join facet_cities on facetable.city_id = facet_cities.id\nwhere neighborhood like '%' || :text || '%' order by neighborhood;",
"title": "Search neighborhoods",
"description_html": "<b>Demonstrating</b> simple like search"
}
}
}
}
}
You can try this canned query out here: https://latest.datasette.io/fixtures/neighborhood_search?text=town
Note that we are using SQLite string concatenation here - the ||
operator -
to add wildcard %
characters to the string provided by the user.
Pagination¶
Datasette’s default table pagination is designed to be extremely efficient. SQL OFFSET/LIMIT pagination can have a significant performance penalty once you get into multiple thousands of rows, as each page still requires the database to scan through every preceding row to find the correct offset.
When paginating through tables, Datasette instead orders the rows in the table by their primary key and performs a WHERE clause against the last seen primary key for the previous page. For example:
select rowid, * from Tree_List where rowid > 200 order by rowid limit 101
This represents page three for this particular table, with a page size of 100.
Note that we request 101 items in the limit clause rather than 100. This allows us to detect if we are on the last page of the results: if the query returns less than 101 rows we know we have reached the end of the pagination set. Datasette will only return the first 100 rows - the 101st is used purely to detect if there should be another page.
Since the where clause acts against the index on the primary key, the query is extremely fast even for records that are a long way into the overall pagination set.
Performance and caching¶
Datasette runs on top of SQLite, and SQLite has excellent performance. For small databases almost any query should return in just a few milliseconds, and larger databases (100s of MBs or even GBs of data) should perform extremely well provided your queries make sensible use of database indexes.
That said, there are a number of tricks you can use to improve Datasette’s performance.
Immutable mode¶
If you can be certain that a SQLite database file will not be changed by another process you can tell Datasette to open that file in immutable mode.
Doing so will disable all locking and change detection, which can result in improved query performance.
This also enables further optimizations relating to HTTP caching, described below.
To open a file in immutable mode pass it to the datasette command using the -i
option:
datasette -i data.db
When you open a file in immutable mode like this Datasette will also calculate and cache the row counts for each table in that database when it first starts up, further improving performance.
Using “datasette inspect”¶
Counting the rows in a table can be a very expensive operation on larger databases. In immutable mode Datasette performs this count only once and caches the results, but this can still cause server startup time to increase by several seconds or more.
If you know that a database is never going to change you can precalculate the table row counts once and store then in a JSON file, then use that file when you later start the server.
To create a JSON file containing the calculated row counts for a database, use the following:
datasette inspect data.db --inspect-file=counts.json
Then later you can start Datasette against the counts.json
file and use it to skip the row counting step and speed up server startup:
datasette -i data.db --inspect-file=counts.json
You need to use the -i
immutable mode against the databse file here or the counts from the JSON file will be ignored.
You will rarely need to use this optimization in every-day use, but several of the datasette publish
commands described in Publishing data use this optimization for better performance when deploying a database file to a hosting provider.
HTTP caching¶
If your database is immutable and guaranteed not to change, you can gain major performance improvements from Datasette by enabling HTTP caching.
This can work at two different levels. First, it can tell browsers to cache the results of queries and serve future requests from the browser cache.
More significantly, it allows you to run Datasette behind a caching proxy such as Varnish or use a cache provided by a hosted service such as Fastly or Cloudflare. This can provide incredible speed-ups since a query only needs to be executed by Datasette the first time it is accessed - all subsequent hits can then be served by the cache.
Using a caching proxy in this way could enable a Datasette-backed visualization to serve thousands of hits a second while running Datasette itself on extremely inexpensive hosting.
Datasette’s integration with HTTP caches can be enabled using a combination of configuration options and querystring arguments.
The default_cache_ttl setting sets the default HTTP cache TTL for all Datasette pages. This is 5 seconds unless you change it - you can set it to 0 if you wish to disable HTTP caching entirely.
You can also change the cache timeout on a per-request basis using the ?_ttl=10
querystring parameter. This can be useful when you are working with the Datasette JSON API - you may decide that a specific query can be cached for a longer time, or maybe you need to set ?_ttl=0
for some requests for example if you are running a SQL order by random()
query.
Hashed URL mode¶
When you open a database file in immutable mode using the -i
option, Datasette calculates a SHA-256 hash of the contents of that file on startup. This content hash can then optionally be used to create URLs that are guaranteed to change if the contents of the file changes in the future. This results in URLs that can then be cached indefinitely by both browsers and caching proxies - an enormous potential performance optimization.
You can enable these hashed URLs in two ways: using the hash_urls configuration setting (which affects all requests to Datasette) or via the ?_hash=1
querystring parameter (which only applies to the current request).
With hashed URLs enabled, any request to e.g. /mydatabase/mytable
will 302 redirect to mydatabase-455fe3a/mytable
. The URL containing the hash will be served with a very long cache expire header - configured using default_cache_ttl_hashed which defaults to 365 days.
Since these responses are cached for a long time, you may wish to build API clients against the non-hashed version of these URLs. These 302 redirects are served extremely quickly, so this should still be a performant way to work against the Datasette API.
If you run Datasette behind an HTTP/2 server push aware proxy such as Cloudflare Datasette will serve the 302 redirects in such a way that the redirected page will be efficiently “pushed” to the browser as part of the response, without the browser needing to make a second HTTP request to fetch the redirected resource.
Note
Prior to Datasette 0.28 hashed URL mode was the default behaviour for Datasette, since all database files were assumed to be immutable and unchanging. From 0.28 onwards the default has been to treat database files as mutable unless explicitly configured otherwise.
CSV Export¶
Any Datasette table, view or custom SQL query can be exported as CSV.
To obtain the CSV representation of the table you are looking, click the “this data as CSV” link.
You can also use the advanced export form for more control over the resulting file, which looks like this and has the following options:
- download file - instead of displaying CSV in your browser, this forces your browser to download the CSV to your downloads directory.
- expand labels - if your table has any foreign key references this option
will cause the CSV to gain additional
COLUMN_NAME_label
columns with a label for each foreign key derived from the linked table. In this example thecity_id
column is accompanied by acity_id_label
column. - stream all rows - by default CSV files only contain the first max_returned_rows records. This option will cause Datasette to loop through every matching record and return them as a single CSV file.
You can try that out on https://latest.datasette.io/fixtures/facetable?_size=4
Streaming all records¶
The stream all rows option is designed to be as efficient as possible - under the hood it takes advantage of Python 3 asyncio capabilities and Datasette’s efficient pagination to stream back the full CSV file.
Since databases can get pretty large, by default this option is capped at 100MB - if a table returns more than 100MB of data the last line of the CSV will be a truncation error message.
You can increase or remove this limit using the max_csv_mb config setting. You can also disable the CSV export feature entirely using allow_csv_stream.
A note on URLs¶
The default URL for the CSV representation of a table is that table with
.csv
appended to it:
- https://latest.datasette.io/fixtures/facetable - HTML interface
- https://latest.datasette.io/fixtures/facetable.csv - CSV export
- https://latest.datasette.io/fixtures/facetable.json - JSON API
This pattern doesn’t work for tables with names that already end in .csv
or
.json
. For those tables, you can instead use the _format=
querystring
parameter:
Facets¶
Datasette facets can be used to add a faceted browse interface to any database table. With facets, tables are displayed along with a summary showing the most common values in specified columns. These values can be selected to further filter the table.
Facets can be specified in two ways: using querystring parameters, or in metadata.json
configuration for the table.
Facets in querystrings¶
To turn on faceting for specific columns on a Datasette table view, add one or more _facet=COLUMN
parameters to the URL. For example, if you want to turn on facets for the city_id
and state
columns, construct a URL that looks like this:
/dbname/tablename?_facet=state&_facet=city_id
This works for both the HTML interface and the .json
view. When enabled, facets will cause a facet_results
block to be added to the JSON output, looking something like this:
{
"state": {
"name": "state",
"results": [
{
"value": "CA",
"label": "CA",
"count": 10,
"toggle_url": "http://...?_facet=city_id&_facet=state&state=CA",
"selected": false
},
{
"value": "MI",
"label": "MI",
"count": 4,
"toggle_url": "http://...?_facet=city_id&_facet=state&state=MI",
"selected": false
},
{
"value": "MC",
"label": "MC",
"count": 1,
"toggle_url": "http://...?_facet=city_id&_facet=state&state=MC",
"selected": false
}
],
"truncated": false
}
"city_id": {
"name": "city_id",
"results": [
{
"value": 1,
"label": "San Francisco",
"count": 6,
"toggle_url": "http://...?_facet=city_id&_facet=state&city_id=1",
"selected": false
},
{
"value": 2,
"label": "Los Angeles",
"count": 4,
"toggle_url": "http://...?_facet=city_id&_facet=state&city_id=2",
"selected": false
},
{
"value": 3,
"label": "Detroit",
"count": 4,
"toggle_url": "http://...?_facet=city_id&_facet=state&city_id=3",
"selected": false
},
{
"value": 4,
"label": "Memnonia",
"count": 1,
"toggle_url": "http://...?_facet=city_id&_facet=state&city_id=4",
"selected": false
}
],
"truncated": false
}
}
If Datasette detects that a column is a foreign key, the "label"
property will be automatically derived from the detected label column on the referenced table.
Facets in metadata.json¶
You can turn facets on by default for specific tables by adding them to a "facets"
key in a Datasette Metadata file.
Here’s an example that turns on faceting by default for the qLegalStatus
column in the Street_Tree_List
table in the sf-trees
database:
{
"databases": {
"sf-trees": {
"tables": {
"Street_Tree_List": {
"facets": ["qLegalStatus"]
}
}
}
}
}
Facets defined in this way will always be shown in the interface and returned in the API, regardless of the _facet
arguments passed to the view.
Suggested facets¶
Datasette’s table UI will suggest facets for the user to apply, based on the following criteria:
For the currently filtered data are there any columns which, if applied as a facet…
- Will return 30 or less unique options
- Will return more than one unique option
- Will return less unique options than the total number of filtered rows
- And the query used to evaluate this criteria can be completed in under 50ms
That last point is particularly important: Datasette runs a query for every column that is displayed on a page, which could get expensive - so to avoid slow load times it sets a time limit of just 50ms for each of those queries. This means suggested facets are unlikely to appear for tables with millions of records in them.
Speeding up facets with indexes¶
The performance of facets can be greatly improved by adding indexes on the columns you wish to facet by. Adding indexes can be performed using the sqlite3
command-line utility. Here’s how to add an index on the state
column in a table called Food_Trucks
:
$ sqlite3 mydatabase.db
SQLite version 3.19.3 2017-06-27 16:48:08
Enter ".help" for usage hints.
sqlite> CREATE INDEX Food_Trucks_state ON Food_Trucks("state");
Facet by JSON array¶
If your SQLite installation provides the json1
extension (you can check using /-/versions) Datasette will automatically detect columns that contain JSON arrays of values and offer a faceting interface against those columns.
This is useful for modelling things like tags without needing to break them out into a new table.
Example here: latest.datasette.io/fixtures/facetable?_facet_array=tags
Facet by date¶
If Datasette finds any columns that contain dates in the first 100 values, it will offer a faceting interface against the dates of those values. This works especially well against timestamp values such as 2019-03-01 12:44:00
.
Example here: latest.datasette.io/fixtures/facetable?_facet_date=created
Full-text search¶
SQLite includes a powerful mechanism for enabling full-text search against SQLite records. Datasette can detect if a table has had full-text search configured for it in the underlying database and display a search interface for filtering that table.
Datasette automatically detects which tables have been configured for full-text search.
FTS versions¶
There are three different versions of the SQLite FTS module: FTS3, FTS4 and FTS5. You can tell which versions are supported by your instance of Datasette by checking the /-/versions
page.
FTS5 is the most advanced module but may not be available in the SQLite version that is bundled with your Python installation. Most importantly, FTS5 is the only version that has the ability to order by search relevance without needing extra code.
If you can’t be sure that FTS5 will be available, you should use FTS4.
Adding full-text search to a SQLite table¶
Datasette takes advantage of the external content mechanism in SQLite, which allows a full-text search virtual table to be associated with the contents of another SQLite table.
To set up full-text search for a table, you need to do two things:
- Create a new FTS virtual table associated with your table
- Populate that FTS table with the data that you would like to be able to run searches against
Configuring FTS using sqlite-utils¶
sqlite-utils is a CLI utility and Python library for manipulating SQLite databases. You can use it from Python code to configure FTS search, or you can achieve the same goal using the accompanying command-line tool.
Here’s how to use sqlite-utils
to enable full-text search for an items
table across the name
and description
columns:
$ sqlite-utils enable-fts mydatabase.db items name description
Configuring FTS using csvs-to-sqlite¶
If your data starts out in CSV files, you can use Datasette’s companion tool csvs-to-sqlite to convert that file into a SQLite database and enable full-text search on specific columns. For a file called items.csv
where you want full-text search to operate against the name
and description
columns you would run the following:
$ csvs-to-sqlite items.csv items.db -f name -f description
Configuring FTS by hand¶
We recommend using sqlite-utils, but if you want to hand-roll a SQLite full-text search table you can do so using the following SQL.
To enable full-text search for a table called items
that works against the name
and description
columns, you would run this SQL to create a new items_fts
FTS virtual table:
CREATE VIRTUAL TABLE "items_fts" USING FTS4 (
name,
description,
content="items"
);
This creates a set of tables to power full-text search against items
. The new items_fts
table will be detected by Datasette as the fts_table
for the items
table.
Creating the table is not enough: you also need to populate it with a copy of the data that you wish to make searchable. You can do that using the following SQL:
INSERT INTO "items_fts" (rowid, name, description)
SELECT rowid, name, description FROM items;
If your table has columns that are foreign key references to other tables you can include that data in your full-text search index using a join. Imagine the items
table has a foreign key column called category_id
which refers to a categories
table - you could create a full-text search table like this:
CREATE VIRTUAL TABLE "items_fts" USING FTS4 (
name,
description,
category_name,
content="items"
);
And then populate it like this:
INSERT INTO "items_fts" (rowid, name, description, category_name)
SELECT items.rowid,
items.name,
items.description,
categories.name
FROM items JOIN categories ON items.category_id=categories.id;
You can use this technique to populate the full-text search index from any combination of tables and joins that makes sense for your project.
Configuring full-text search for a table or view¶
If a table has a corresponding FTS table set up using the content=
argument to CREATE VIRTUAL TABLE
shown above, Datasette will detect it automatically and add a search interface to the table page for that table.
You can also manually configure which table should be used for full-text search using querystring parameters or Metadata. You can set the associated FTS table for a specific table and you can also set one for a view - if you do that, the page for that SQL view will offer a search option.
Use ?_fts_table=x
to over-ride the FTS table for a specific page. If the primary key was something other than rowid
you can use ?_fts_pk=col
to set that as well. This is particularly useful for views, for example:
https://latest.datasette.io/fixtures/searchable_view?_fts_table=searchable_fts&_fts_pk=pk
The fts_table
metadata property can be used to specify an associated FTS table. If the primary key column in your table which was used to populate the FTS table is something other than rowid
, you can specify the column to use with the fts_pk
property.
Here is an example which enables full-text search for a display_ads
view which is defined against the ads
table and hence needs to run FTS against the ads_fts
table, using the id
as the primary key:
{
"databases": {
"russian-ads": {
"tables": {
"display_ads": {
"fts_table": "ads_fts",
"fts_pk": "id"
}
}
}
}
}
The table view API¶
Table views that support full-text search can be queried using the ?_search=TERMS
querystring parameter. This will run the search against content from all of the columns that have been included in the index.
SQLite full-text search supports wildcards. This means you can easily implement prefix auto-complete by including an asterisk at the end of the search term - for example:
/dbname/tablename/?_search=rob*
This will return all records containing at least one word that starts with the letters rob
.
You can also run searches against just the content of a specific named column by using _search_COLNAME=TERMS
- for example, this would search for just rows where the name
column in the FTS index mentions Sarah
:
/dbname/tablename/?_search_name=Sarah
Searches using custom SQL¶
You can include full-text search results in custom SQL queries. The general pattern with SQLite search is to run the search as a sub-select that returns rowid values, then include those rowids in another part of the query.
You can see the syntax for a basic search by running that search on a table page and then clicking “View and edit SQL” to see the underlying SQL. For example, consider this search for cherry trees in San Francisco:
/sf-trees/Street_Tree_List?_search=cherry
If you click View and edit SQL you’ll see that the underlying SQL looks like this:
select rowid, * from Street_Tree_List
where rowid in (
select rowid from [Street_Tree_List_fts]
where [Street_Tree_List_fts] match "cherry"
) order by rowid limit 101
SpatiaLite¶
The SpatiaLite module for SQLite adds features for handling geographic and spatial data. For an example of what you can do with it, see the tutorial Building a location to time zone API with SpatiaLite, OpenStreetMap and Datasette.
To use it with Datasette, you need to install the mod_spatialite
dynamic library. This can then be loaded into Datasette using the --load-extension
command-line option.
Installation¶
Installing SpatiaLite on OS X¶
The easiest way to install SpatiaLite on OS X is to use Homebrew.
brew update
brew install spatialite-tools
This will install the spatialite
command-line tool and the mod_spatialite
dynamic library.
You can now run Datasette like so:
datasette --load-extension=/usr/local/lib/mod_spatialite.dylib
Installing SpatiaLite on Linux¶
SpatiaLite is packaged for most Linux distributions.
apt install spatialite-bin libsqlite3-mod-spatialite
Depending on your distribution, you should be able to run Datasette something like this:
datasette --load-extension=/usr/lib/x86_64-linux-gnu/mod_spatialite.so
If you are unsure of the location of the module, try running locate mod_spatialite
and see what comes back.
Building SpatiaLite from source¶
The packaged versions of SpatiaLite usually provide SpatiaLite 4.3.0a. For an example of how to build the most recent unstable version, 4.4.0-RC0 (which includes the powerful VirtualKNN module), take a look at the Datasette Dockerfile.
Spatial indexing latitude/longitude columns¶
Here’s a recipe for taking a table with existing latitude and longitude columns, adding a SpatiaLite POINT geometry column to that table, populating the new column and then populating a spatial index:
import sqlite3
conn = sqlite3.connect('museums.db')
# Lead the spatialite extension:
conn.enable_load_extension(True)
conn.load_extension('/usr/local/lib/mod_spatialite.dylib')
# Initialize spatial metadata for this database:
conn.execute('select InitSpatialMetadata(1)')
# Add a geometry column called point_geom to our museums table:
conn.execute("SELECT AddGeometryColumn('museums', 'point_geom', 4326, 'POINT', 2);")
# Now update that geometry column with the lat/lon points
conn.execute('''
UPDATE events SET
point_geom = GeomFromText('POINT('||"longitude"||' '||"latitude"||')',4326);
''')
# Now add a spatial index to that column
conn.execute('select CreateSpatialIndex("museums", "point_geom");')
# If you don't commit your changes will not be persisted:
conn.commit()
conn.close()
Making use of a spatial index¶
SpatiaLite spatial indexes are R*Trees. They allow you to run efficient bounding box queries using a sub-select, with a similar pattern to that used for Searches using custom SQL.
In the above example, the resulting index will be called idx_museums_point_geom
. This takes the form of a SQLite virtual table. You can inspect its contents using the following query:
select * from idx_museums_point_geom limit 10;
Here’s a live example: timezones-api.now.sh/timezones/idx_timezones_Geometry
pkid | xmin | xmax | ymin | ymax |
---|---|---|---|---|
1 | -8.601725578308105 | -2.4930307865142822 | 4.162120819091797 | 10.74019718170166 |
2 | -3.2607860565185547 | 1.27329421043396 | 4.539252281188965 | 11.174856185913086 |
3 | 32.997581481933594 | 47.98238754272461 | 3.3974475860595703 | 14.894054412841797 |
4 | -8.66890811920166 | 11.997337341308594 | 18.9681453704834 | 37.296207427978516 |
5 | 36.43336486816406 | 43.300174713134766 | 12.354820251464844 | 18.070993423461914 |
You can now construct efficient bounding box queries that will make use of the index like this:
select * from museums where museums.rowid in (
SELECT pkid FROM idx_museums_point_geom
-- left-hand-edge of point > left-hand-edge of bbox (minx)
where xmin > :bbox_minx
-- right-hand-edge of point < right-hand-edge of bbox (maxx)
and xmax < :bbox_maxx
-- bottom-edge of point > bottom-edge of bbox (miny)
and ymin > :bbox_miny
-- top-edge of point < top-edge of bbox (maxy)
and ymax < :bbox_maxy
);
Spatial indexes can be created against polygon columns as well as point columns, in which case they will represent the minimum bounding rectangle of that polygon. This is useful for accelerating within
queries, as seen in the Timezones API example.
Importing shapefiles into SpatiaLite¶
The shapefile format is a common format for distributing geospatial data. You can use the spatialite
command-line tool to create a new database table from a shapefile.
Try it now with the North America shapefile available from the University of North Carolina Global River Database project. Download the file and unzip it (this will create files called narivs.dbf
, narivs.prj
, narivs.shp
and narivs.shx
in the current directory), then run the following:
$ spatialite rivers-database.db
SpatiaLite version ..: 4.3.0a Supported Extensions:
...
spatialite> .loadshp narivs rivers CP1252 23032
========
Loading shapefile at 'narivs' into SQLite table 'rivers'
...
Inserted 467973 rows into 'rivers' from SHAPEFILE
This will load the data from the narivs
shapefile into a new database table called rivers
.
Exit out of spatialite
(using Ctrl+D
) and run Datasette against your new database like this:
datasette rivers-database.db \
--load-extension=/usr/local/lib/mod_spatialite.dylib
If you browse to http://localhost:8001/rivers-database/rivers
you will see the new table… but the Geometry
column will contain unreadable binary data (SpatiaLite uses a custom format based on WKB).
The easiest way to turn this into semi-readable data is to use the SpatiaLite AsGeoJSON
function. Try the following using the SQL query interface at http://localhost:8001/rivers-database
:
select *, AsGeoJSON(Geometry) from rivers limit 10;
This will give you back an additional column of GeoJSON. You can copy and paste GeoJSON from this column into the debugging tool at geojson.io to visualize it on a map.
To see a more interesting example, try ordering the records with the longest geometry first. Since there are 467,000 rows in the table you will first need to increase the SQL time limit imposed by Datasette:
datasette rivers-database.db \
--load-extension=/usr/local/lib/mod_spatialite.dylib \
--config sql_time_limit_ms:10000
Now try the following query:
select *, AsGeoJSON(Geometry) from rivers
order by length(Geometry) desc limit 10;
Importing GeoJSON polygons using Shapely¶
Another common form of polygon data is the GeoJSON format. This can be imported into SpatiaLite directly, or by using the Shapely Python library.
Who’s On First is an excellent source of openly licensed GeoJSON polygons. Let’s import the geographical polygon for Wales. First, we can use the Who’s On First Spelunker tool to find the record for Wales:
spelunker.whosonfirst.org/id/404227475
That page includes a link to the GeoJSON record, which can be accessed here:
data.whosonfirst.org/404/227/475/404227475.geojson
Here’s Python code to create a SQLite database, enable SpatiaLite, create a places table and then add a record for Wales:
import sqlite3
conn = sqlite3.connect('places.db')
# Enable SpatialLite extension
conn.enable_load_extension(True)
conn.load_extension('/usr/local/lib/mod_spatialite.dylib')
# Create the masic countries table
conn.execute('select InitSpatialMetadata(1)')
conn.execute('create table places (id integer primary key, name text);')
# Add a MULTIPOLYGON Geometry column
conn.execute("SELECT AddGeometryColumn('places', 'geom', 4326, 'MULTIPOLYGON', 2);")
# Add a spatial index against the new column
conn.execute("SELECT CreateSpatialIndex('places', 'geom');")
# Now populate the table
from shapely.geometry.multipolygon import MultiPolygon
from shapely.geometry import shape
import requests
geojson = requests.get('https://data.whosonfirst.org/404/227/475/404227475.geojson').json()
# Convert to "Well Known Text" format
wkt = shape(geojson['geometry']).wkt
# Insert and commit the record
conn.execute("INSERT INTO places (id, name, geom) VALUES(null, ?, GeomFromText(?, 4326))", (
"Wales", wkt
))
conn.commit()
Querying polygons using within()¶
The within()
SQL function can be used to check if a point is within a geometry:
select
name
from
places
where
within(GeomFromText('POINT(-3.1724366 51.4704448)'), places.geom);
The GeomFromText()
function takes a string of well-known text. Note that the order used here is longitude
then latitude
.
To run that same within()
query in a way that benefits from the spatial index, use the following:
select
name
from
places
where
within(GeomFromText('POINT(-3.1724366 51.4704448)'), places.geom)
and rowid in (
SELECT pkid FROM idx_places_geom
where xmin < -3.1724366
and xmax > -3.1724366
and ymin < 51.4704448
and ymax > 51.4704448
);
Metadata¶
Data loves metadata. Any time you run Datasette you can optionally include a JSON file with metadata about your databases and tables. Datasette will then display that information in the web UI.
Run Datasette like this:
datasette database1.db database2.db --metadata metadata.json
Your metadata.json
file can look something like this:
{
"title": "Custom title for your index page",
"description": "Some description text can go here",
"license": "ODbL",
"license_url": "https://opendatacommons.org/licenses/odbl/",
"source": "Original Data Source",
"source_url": "http://example.com/"
}
The above metadata will be displayed on the index page of your Datasette-powered site. The source and license information will also be included in the footer of every page served by Datasette.
Any special HTML characters in description
will be escaped. If you want to
include HTML in your description, you can use a description_html
property
instead.
Per-database and per-table metadata¶
Metadata at the top level of the JSON will be shown on the index page and in the footer on every page of the site. The license and source is expected to apply to all of your data.
You can also provide metadata at the per-database or per-table level, like this:
{
"databases": {
"database1": {
"source": "Alternative source",
"source_url": "http://example.com/",
"tables": {
"example_table": {
"description_html": "Custom <em>table</em> description",
"license": "CC BY 3.0 US",
"license_url": "https://creativecommons.org/licenses/by/3.0/us/"
}
}
}
}
}
Each of the top-level metadata fields can be used at the database and table level.
Source, license and about¶
The three visible metadata fields you can apply to everything, specific databases or specific tables are source, license and about. All three are optional.
source and source_url should be used to indicate where the underlying data came from.
license and license_url should be used to indicate the license under which the data can be used.
about and about_url can be used to link to further information about the project - an accompanying blog entry for example.
For each of these you can provide just the *_url
field and Datasette will treat that as the default link label text and display the URL directly on the page.
Specifying units for a column¶
Datasette supports attaching units to a column, which will be used when displaying values from that column. SI prefixes will be used where appropriate.
Column units are configured in the metadata like so:
{
"databases": {
"database1": {
"tables": {
"example_table": {
"units": {
"column1": "metres",
"column2": "Hz"
}
}
}
}
}
}
Units are interpreted using Pint, and you can see the full list of available units in Pint’s unit registry. You can also add custom units to the metadata, which will be registered with Pint:
{
"custom_units": [
"decibel = [] = dB"
]
}
Setting which columns can be used for sorting¶
Datasette allows any column to be used for sorting by default. If you need to
control which columns are available for sorting you can do so using the optional
sortable_columns
key:
{
"databases": {
"database1": {
"tables": {
"example_table": {
"sortable_columns": [
"height",
"weight"
]
}
}
}
}
}
This will restrict sorting of example_table
to just the height
and
weight
columns.
You can also disable sorting entirely by setting "sortable_columns": []
By default, database views in Datasette do not support sorting. You can use sortable_columns
to enable specific sort orders for a view called name_of_view
in the database my_database
like so:
{
"databases": {
"my_database": {
"tables": {
"name_of_view": {
"sortable_columns": [
"clicks",
"impressions"
]
}
}
}
}
}
Specifying the label column for a table¶
Datasette’s HTML interface attempts to display foreign key references as labelled hyperlinks. By default, it looks for referenced tables that only have two columns: a primary key column and one other. It assumes that the second column should be used as the link label.
If your table has more than two columns you can specify which column should be
used for the link label with the label_column
property:
{
"databases": {
"database1": {
"tables": {
"example_table": {
"label_column": "title"
}
}
}
}
}
Hiding tables¶
You can hide tables from the database listing view (in the same way that FTS and
Spatialite tables are automatically hidden) using "hidden": true
:
{
"databases": {
"database1": {
"tables": {
"example_table": {
"hidden": true
}
}
}
}
}
Configuration¶
Datasette provides a number of configuration options. These can be set using the --config name:value
option to datasette serve
.
You can set multiple configuration options at once like this:
datasette mydatabase.db --config default_page_size:50 \
--config sql_time_limit_ms:3500 \
--config max_returned_rows:2000
To prevent rogue, long-running queries from making a Datasette instance inaccessible to other users, Datasette imposes some limits on the SQL that you can execute. These are exposed as config options which you can over-ride.
default_page_size¶
The default number of rows returned by the table page. You can over-ride this on a per-page basis using the ?_size=80
querystring parameter, provided you do not specify a value higher than the max_returned_rows
setting. You can set this default using --config
like so:
datasette mydatabase.db --config default_page_size:50
sql_time_limit_ms¶
By default, queries have a time limit of one second. If a query takes longer than this to run Datasette will terminate the query and return an error.
If this time limit is too short for you, you can customize it using the sql_time_limit_ms
limit - for example, to increase it to 3.5 seconds:
datasette mydatabase.db --config sql_time_limit_ms:3500
You can optionally set a lower time limit for an individual query using the ?_timelimit=100
querystring argument:
/my-database/my-table?qSpecies=44&_timelimit=100
This would set the time limit to 100ms for that specific query. This feature is useful if you are working with databases of unknown size and complexity - a query that might make perfect sense for a smaller table could take too long to execute on a table with millions of rows. By setting custom time limits you can execute queries “optimistically” - e.g. give me an exact count of rows matching this query but only if it takes less than 100ms to calculate.
max_returned_rows¶
Datasette returns a maximum of 1,000 rows of data at a time. If you execute a query that returns more than 1,000 rows, Datasette will return the first 1,000 and include a warning that the result set has been truncated. You can use OFFSET/LIMIT or other methods in your SQL to implement pagination if you need to return more than 1,000 rows.
You can increase or decrease this limit like so:
datasette mydatabase.db --config max_returned_rows:2000
num_sql_threads¶
Maximum number of threads in the thread pool Datasette uses to execute SQLite queries. Defaults to 3.
datasette mydatabase.db --config num_sql_threads:10
allow_facet¶
Allow users to specify columns they would like to facet on using the ?_facet=COLNAME
URL parameter to the table view.
This is enabled by default. If disabled, facets will still be displayed if they have been specifically enabled in metadata.json
configuration for the table.
Here’s how to disable this feature:
datasette mydatabase.db --config allow_facet:off
default_facet_size¶
The default number of unique rows returned by Facets is 30. You can customize it like this:
datasette mydatabase.db --config default_facet_size:50
facet_time_limit_ms¶
This is the time limit Datasette allows for calculating a facet, which defaults to 200ms:
datasette mydatabase.db --config facet_time_limit_ms:1000
facet_suggest_time_limit_ms¶
When Datasette calculates suggested facets it needs to run a SQL query for every column in your table. The default for this time limit is 50ms to account for the fact that it needs to run once for every column. If the time limit is exceeded the column will not be suggested as a facet.
You can increase this time limit like so:
datasette mydatabase.db --config facet_suggest_time_limit_ms:500
suggest_facets¶
Should Datasette calculate suggested facets? On by default, turn this off like so:
datasette mydatabase.db --config suggest_facets:off
allow_download¶
Should users be able to download the original SQLite database using a link on the database index page? This is turned on by default - to disable database downloads, use the following:
datasette mydatabase.db --config allow_download:off
allow_sql¶
Enable/disable the ability for users to run custom SQL directly against a database. To disable this feature, run:
datasette mydatabase.db --config allow_sql:off
default_cache_ttl¶
Default HTTP caching max-age header in seconds, used for Cache-Control: max-age=X
. Can be over-ridden on a per-request basis using the ?_ttl=
querystring parameter. Set this to 0
to disable HTTP caching entirely. Defaults to 5 seconds.
datasette mydatabase.db --config default_cache_ttl:60
default_cache_ttl_hashed¶
Default HTTP caching max-age for responses served using using the hashed-urls mechanism. Defaults to 365 days (31536000 seconds).
datasette mydatabase.db --config default_cache_ttl_hashed:10000
cache_size_kb¶
Sets the amount of memory SQLite uses for its per-connection cache, in KB.
datasette mydatabase.db --config cache_size_kb:5000
allow_csv_stream¶
Enables the CSV export feature where an entire table (potentially hundreds of thousands of rows) can be exported as a single CSV file. This is turned on by default - you can turn it off like this:
datasette mydatabase.db --config allow_csv_stream:off
max_csv_mb¶
The maximum size of CSV that can be exported, in megabytes. Defaults to 100MB. You can disable the limit entirely by settings this to 0:
datasette mydatabase.db --config max_csv_mb:0
truncate_cells_html¶
In the HTML table view, truncate any strings that are longer than this value. The full value will still be available in CSV, JSON and on the individual row HTML page. Set this to 0 to disable truncation.
datasette mydatabase.db --config truncate_cells_html:0
force_https_urls¶
Forces self-referential URLs in the JSON output to always use the https://
protocol. This is useful for cases where the application itself is hosted using
HTTP but is served to the outside world via a proxy that enables HTTPS.
datasette mydatabase.db --config force_https_urls:1
hash_urls¶
When enabled, this setting causes Datasette to append a content hash of the database file to the URL path for every table and query within that database.
When combined with far-future expire headers this ensures that queries can be cached forever, safe in the knowledge that any modifications to the database itself will result in new, uncachcacheed URL paths.
datasette mydatabase.db --config hash_urls:1
Introspection¶
Datasette includes some pages and JSON API endpoints for introspecting the current instance. These can be used to understand some of the internals of Datasette and to see how a particular instance has been configured.
Each of these pages can be viewed in your browser. Add .json
to the URL to get back the contents as JSON.
/-/metadata¶
Shows the contents of the metadata.json
file that was passed to datasette serve
, if any. Metadata example:
{
"license": "CC Attribution 4.0 License",
"license_url": "http://creativecommons.org/licenses/by/4.0/",
"source": "fivethirtyeight/data on GitHub",
"source_url": "https://github.com/fivethirtyeight/data",
"title": "Five Thirty Eight",
"databases": {...}
}
/-/versions¶
Shows the version of Datasette, Python and SQLite. Versions example:
{
"datasette": {
"version": "0.21"
},
"python": {
"full": "3.6.5 (default, May 5 2018, 03:07:21) \n[GCC 6.3.0 20170516]",
"version": "3.6.5"
},
"sqlite": {
"extensions": {
"json1": null
},
"fts_versions": [
"FTS5",
"FTS4",
"FTS3"
],
"compile_options": [
"COMPILER=gcc-6.3.0 20170516",
"ENABLE_FTS3",
"ENABLE_FTS4",
"ENABLE_FTS5",
"ENABLE_JSON1",
"ENABLE_RTREE",
"THREADSAFE=1"
],
"version": "3.16.2"
}
}
/-/plugins¶
Shows a list of currently installed plugins and their versions. Plugins example:
[
{
"name": "datasette_cluster_map",
"static": true,
"templates": false,
"version": "0.4"
}
]
/-/config¶
Shows the Configuration options for this instance of Datasette. Config example:
{
"default_facet_size": 30,
"default_page_size": 100,
"facet_suggest_time_limit_ms": 50,
"facet_time_limit_ms": 1000,
"max_returned_rows": 1000,
"sql_time_limit_ms": 1000
}
/-/databases¶
Shows currently attached databases. Databases example:
[
{
"hash": null,
"is_memory": false,
"is_mutable": true,
"name": "fixtures",
"path": "fixtures.db",
"size": 225280
}
]
/-/threads¶
Shows details of threads. Threads example:
{
"num_threads": 2,
"threads": [
{
"daemon": false,
"ident": 4759197120,
"name": "MainThread"
},
{
"daemon": true,
"ident": 123145319682048,
"name": "Thread-1"
},
]
}
Customization¶
Datasette provides a number of ways of customizing the way data is displayed.
Custom CSS and JavaScript¶
When you launch Datasette, you can specify a custom metadata file like this:
datasette mydb.db --metadata metadata.json
Your metadata.json
file can include links that look like this:
{
"extra_css_urls": [
"https://simonwillison.net/static/css/all.bf8cd891642c.css"
],
"extra_js_urls": [
"https://code.jquery.com/jquery-3.2.1.slim.min.js"
]
}
The extra CSS and JavaScript files will be linked in the <head>
of every page.
You can also specify a SRI (subresource integrity hash) for these assets:
{
"extra_css_urls": [
{
"url": "https://simonwillison.net/static/css/all.bf8cd891642c.css",
"sri": "sha384-9qIZekWUyjCyDIf2YK1FRoKiPJq4PHt6tp/ulnuuyRBvazd0hG7pWbE99zvwSznI"
}
],
"extra_js_urls": [
{
"url": "https://code.jquery.com/jquery-3.2.1.slim.min.js",
"sri": "sha256-k2WSCIexGzOj3Euiig+TlR8gA0EmPjuc79OEeY5L45g="
}
]
}
Modern browsers will only execute the stylesheet or JavaScript if the SRI hash matches the content served. You can generate hashes using www.srihash.org
Every default template includes CSS classes in the body designed to support custom styling.
The index template (the top level page at /
) gets this:
<body class="index">
The database template (/dbname
) gets this:
<body class="db db-dbname">
The custom SQL template (/dbname?sql=...
) gets this:
<body class="query db-dbname">
The table template (/dbname/tablename
) gets:
<body class="table db-dbname table-tablename">
The row template (/dbname/tablename/rowid
) gets:
<body class="row db-dbname table-tablename">
The db-x
and table-x
classes use the database or table names themselves if
they are valid CSS identifiers. If they aren’t, we strip any invalid
characters out and append a 6 character md5 digest of the original name, in
order to ensure that multiple tables which resolve to the same stripped
character version still have different CSS classes.
Some examples:
"simple" => "simple"
"MixedCase" => "MixedCase"
"-no-leading-hyphens" => "no-leading-hyphens-65bea6"
"_no-leading-underscores" => "no-leading-underscores-b921bc"
"no spaces" => "no-spaces-7088d7"
"-" => "336d5e"
"no $ characters" => "no--characters-59e024"
<td>
and <th>
elements also get custom CSS classes reflecting the
database column they are representing, for example:
<table>
<thead>
<tr>
<th class="col-id" scope="col">id</th>
<th class="col-name" scope="col">name</th>
</tr>
</thead>
<tbody>
<tr>
<td class="col-id"><a href="...">1</a></td>
<td class="col-name">SMITH</td>
</tr>
</tbody>
</table>
Custom templates¶
By default, Datasette uses default templates that ship with the package.
You can over-ride these templates by specifying a custom --template-dir
like
this:
datasette mydb.db --template-dir=mytemplates/
Datasette will now first look for templates in that directory, and fall back on the defaults if no matches are found.
It is also possible to over-ride templates on a per-database, per-row or per- table basis.
The lookup rules Datasette uses are as follows:
Index page (/):
index.html
Database page (/mydatabase):
database-mydatabase.html
database.html
Custom query page (/mydatabase?sql=...):
query-mydatabase.html
query.html
Canned query page (/mydatabase/canned-query):
query-mydatabase-canned-query.html
query-mydatabase.html
query.html
Table page (/mydatabase/mytable):
table-mydatabase-mytable.html
table.html
Row page (/mydatabase/mytable/id):
row-mydatabase-mytable.html
row.html
Table of rows and columns include on table page:
_table-table-mydatabase-mytable.html
_table-mydatabase-mytable.html
_table.html
Table of rows and columns include on row page:
_table-row-mydatabase-mytable.html
_table-mydatabase-mytable.html
_table.html
If a table name has spaces or other unexpected characters in it, the template
filename will follow the same rules as our custom <body>
CSS classes - for
example, a table called “Food Trucks” will attempt to load the following
templates:
table-mydatabase-Food-Trucks-399138.html
table.html
You can find out which templates were considered for a specific page by viewing source on that page and looking for an HTML comment at the bottom. The comment will look something like this:
<!-- Templates considered: *query-mydb-tz.html, query-mydb.html, query.html -->
This example is from the canned query page for a query called “tz” in the
database called “mydb”. The asterisk shows which template was selected - so in
this case, Datasette found a template file called query-mydb-tz.html
and
used that - but if that template had not been found, it would have tried for
query-mydb.html
or the default query.html
.
It is possible to extend the default templates using Jinja template
inheritance. If you want to customize EVERY row template with some additional
content you can do so by creating a row.html
template like this:
{% extends "default:row.html" %}
{% block content %}
<h1>EXTRA HTML AT THE TOP OF THE CONTENT BLOCK</h1>
<p>This line renders the original block:</p>
{{ super() }}
{% endblock %}
Note the default:row.html
template name, which ensures Jinja will inherit
from the default template.
The _table.html
template is included by both the row and the table pages,
and a list of rows. The default _table.html
template renders them as an
HTML template and can be seen here.
You can provide a custom template that applies to all of your databases and tables, or you can provide custom templates for specific tables using the template naming scheme described above.
If you want to present your data in a format other than an HTML table, you
can do so by looping through display_rows
in your own _table.html
template. You can use {{ row["column_name"] }}
to output the raw value
of a specific column.
If you want to output the rendered HTML version of a column, including any
links to foreign keys, you can use {{ row.display("column_name") }}
.
Here is an example of a custom _table.html
template:
{% for row in display_rows %}
<div>
<h2>{{ row["title"] }}</h2>
<p>{{ row["description"] }}<lp>
<p>Category: {{ row.display("category_id") }}</p>
</div>
{% endfor %}
Plugins¶
Datasette’s plugin system allows additional features to be implemented as Python code (or front-end JavaScript) which can be wrapped up in a separate Python package. The underlying mechanism uses pluggy.
Using plugins¶
If a plugin has been packaged for distribution using setuptools you can use the plugin by installing it alongside Datasette in the same virtual environment or Docker container.
You can also define one-off per-project plugins by saving them as
plugin_name.py
functions in a plugins/
folder and then passing that
folder to datasette serve
.
The datasette publish
and datasette package
commands both take an
optional --install
argument. You can use this one or more times to tell
Datasette to pip install
specific plugins as part of the process. You can
use the name of a package on PyPI or any of the other valid arguments to pip
install
such as a URL to a .zip
file:
datasette publish cloudrun mydb.db \
--install=datasette-plugin-demos \
--install=https://url-to-my-package.zip
Writing plugins¶
The easiest way to write a plugin is to create a my_plugin.py
file and
drop it into your plugins/
directory. Here is an example plugin, which
adds a new custom SQL function called hello_world()
which takes no
arguments and returns the string Hello world!
.
from datasette import hookimpl
@hookimpl
def prepare_connection(conn):
conn.create_function('hello_world', 0, lambda: 'Hello world!')
If you save this in plugins/my_plugin.py
you can then start Datasette like
this:
datasette serve mydb.db --plugins-dir=plugins/
Now you can navigate to http://localhost:8001/mydb and run this SQL:
select hello_world();
To see the output of your plugin.
Seeing what plugins are installed¶
You can see a list of installed plugins by navigating to the /-/plugins
page of your Datasette instance - for example: https://fivethirtyeight.datasettes.com/-/plugins
You can also use the datasette plugins
command:
$ datasette plugins
[
{
"name": "datasette_json_html",
"static": false,
"templates": false,
"version": "0.4.0"
}
]
If you run datasette plugins --all
it will include default plugins that ship as part of Datasette:
$ datasette plugins --all
[
{
"name": "datasette_json_html",
"static": false,
"templates": false,
"version": "0.4.0"
},
{
"name": "datasette.publish.heroku",
"static": false,
"templates": false,
"version": null
},
{
"name": "datasette.publish.now",
"static": false,
"templates": false,
"version": null
}
]
You can add the --plugins-dir=
option to include any plugins found in that directory.
Packaging a plugin¶
Plugins can be packaged using Python setuptools. You can see an example of a packaged plugin at https://github.com/simonw/datasette-plugin-demos
The example consists of two files: a setup.py
file that defines the plugin:
from setuptools import setup
VERSION = '0.1'
setup(
name='datasette-plugin-demos',
description='Examples of plugins for Datasette',
author='Simon Willison',
url='https://github.com/simonw/datasette-plugin-demos',
license='Apache License, Version 2.0',
version=VERSION,
py_modules=['datasette_plugin_demos'],
entry_points={
'datasette': [
'plugin_demos = datasette_plugin_demos'
]
},
install_requires=['datasette']
)
And a Python module file, datasette_plugin_demos.py
, that implements the
plugin:
from datasette import hookimpl
import random
@hookimpl
def prepare_jinja2_environment(env):
env.filters['uppercase'] = lambda u: u.upper()
@hookimpl
def prepare_connection(conn):
conn.create_function('random_integer', 2, random.randint)
Having built a plugin in this way you can turn it into an installable package using the following command:
python3 setup.py sdist
This will create a .tar.gz
file in the dist/
directory.
You can then install your new plugin into a Datasette virtual environment or
Docker container using pip
:
pip install datasette-plugin-demos-0.1.tar.gz
To learn how to upload your plugin to PyPI for use by other people, read the PyPA guide to Packaging and distributing projects.
Static assets¶
If your plugin has a static/
directory, Datasette will automatically
configure itself to serve those static assets from the following path:
/-/static-plugins/NAME_OF_PLUGIN_PACKAGE/yourfile.js
See the datasette-plugin-demos repository for an example of how to create a package that includes a static folder.
Custom templates¶
If your plugin has a templates/
directory, Datasette will attempt to load
templates from that directory before it uses its own default templates.
The priority order for template loading is:
- templates from the
--template-dir
argument, if specified - templates from the
templates/
directory in any installed plugins - default templates that ship with Datasette
See Customization for more details on how to write custom templates, including which filenames to use to customize which parts of the Datasette UI.
Plugin configuration¶
Plugins can have their own configuration, embedded in a Metadata file. Configuration options for plugins live within a "plugins"
key in that file, which can be included at the root, database or table level.
Here is an example of some plugin configuration for a specific table:
{
"databases: {
"sf-trees": {
"tables": {
"Street_Tree_List": {
"plugins": {
"datasette-cluster-map": {
"latitude_column": "lat",
"longitude_column": "lng"
}
}
}
}
}
}
}
This tells the datasette-cluster-map
column which latitude and longitude columns should be used for a table called Street_Tree_List
inside a database file called sf-trees.db
.
Secret configuration values¶
Any values embedded in metadata.json
will be visible to anyone who views the /-/metadata
page of your Datasette instance. Some plugins may need configuration that should stay secret - API keys for example. There are two ways in which you can store secret configuration values.
As environment variables. If your secret lives in an environment variable that is available to the Datasette process, you can indicate that the configuration value should be read from that environment variable like so:
{
"plugins": {
"datasette-auth-github": {
"client_secret": {
"$env": "GITHUB_CLIENT_SECRET"
}
}
}
}
As values in separate files. Your secrets can also live in files on disk. To specify a secret should be read from a file, provide the full file path like this:
{
"plugins": {
"datasette-auth-github": {
"client_secret": {
"$file": "/secrets/client-secret"
}
}
}
}
If you are publishing your data using the datasette publish family of commands, you can use the --plugin-secret
option to set these secrets at publish time. For example, using Heroku you might run the following command:
$ datasette publish heroku my_database.db \
--name my-heroku-app-demo \
--install=datasette-auth-github \
--plugin-secret datasette-auth-github client_id your_client_id \
--plugin-secret datasette-auth-github client_secret your_client_secret
Writing plugins that accept configuration¶
When you are writing plugins, you can access plugin configuration like this using the datasette.plugin_config()
method. If you know you need plugin configuration for a specific table, you can access it like this:
plugin_config = datasette.plugin_config(
"datasette-cluster-map", database="sf-trees", table="Street_Tree_List"
)
This will return the {"latitude_column": "lat", "longitude_column": "lng"}
in the above example.
If it cannot find the requested configuration at the table layer, it will fall back to the database layer and then the root layer. For example, a user may have set the plugin configuration option like so:
{
"databases: {
"sf-trees": {
"plugins": {
"datasette-cluster-map": {
"latitude_column": "xlat",
"longitude_column": "xlng"
}
}
}
}
}
In this case, the above code would return that configuration for ANY table within the sf-trees
database.
The plugin configuration could also be set at the top level of metadata.json
:
{
"title": "This is the top-level title in metadata.json",
"plugins": {
"datasette-cluster-map": {
"latitude_column": "xlat",
"longitude_column": "xlng"
}
}
}
Now that datasette-cluster-map
plugin configuration will apply to every table in every database.
Plugin hooks¶
When you implement a plugin hook you can accept any or all of the parameters that are documented as being passed to that hook. For example, you can implement a render_cell
plugin hook like this even though the hook definition defines more parameters than just value
and column
:
@hookimpl
def render_cell(value, column):
if column == "stars":
return "*" * int(value)
The full list of available plugin hooks is as follows.
prepare_connection(conn)¶
conn
- sqlite3 connection object- The connection that is being opened
This hook is called when a new SQLite database connection is created. You can use it to register custom SQL functions, aggregates and collations. For example:
from datasette import hookimpl
import random
@hookimpl
def prepare_connection(conn):
conn.create_function('random_integer', 2, random.randint)
This registers a SQL function called random_integer
which takes two
arguments and can be called like this:
select random_integer(1, 10);
prepare_jinja2_environment(env)¶
env
- jinja2 Environment- The template environment that is being prepared
This hook is called with the Jinja2 environment that is used to evaluate Datasette HTML templates. You can use it to do things like register custom template filters, for example:
from datasette import hookimpl
@hookimpl
def prepare_jinja2_environment(env):
env.filters['uppercase'] = lambda u: u.upper()
You can now use this filter in your custom templates like so:
Table name: {{ table|uppercase }}
extra_css_urls(template, database, table, datasette)¶
template
- string- The template that is being rendered, e.g.
database.html
database
- string or None- The name of the database
table
- string or None- The name of the table
datasette
- Datasette instance- You can use this to access plugin configuration options via
datasette.plugin_config(your_plugin_name)
Return a list of extra CSS URLs that should be included on the page. These can take advantage of the CSS class hooks described in Customization.
This can be a list of URLs:
from datasette import hookimpl
@hookimpl
def extra_css_urls():
return [
'https://stackpath.bootstrapcdn.com/bootstrap/4.1.0/css/bootstrap.min.css'
]
Or a list of dictionaries defining both a URL and an SRI hash:
from datasette import hookimpl
@hookimpl
def extra_css_urls():
return [{
'url': 'https://stackpath.bootstrapcdn.com/bootstrap/4.1.0/css/bootstrap.min.css',
'sri': 'sha384-9gVQ4dYFwwWSjIDZnLEWnxCjeSWFphJiwGPXr1jddIhOegiu1FwO5qRGvFXOdJZ4',
}]
extra_js_urls(template, database, table, datasette)¶
Same arguments as extra_css_urls
.
This works in the same way as extra_css_urls()
but for JavaScript. You can
return either a list of URLs or a list of dictionaries:
from datasette import hookimpl
@hookimpl
def extra_js_urls():
return [{
'url': 'https://code.jquery.com/jquery-3.3.1.slim.min.js',
'sri': 'sha384-q8i/X+965DzO0rT7abK41JStQIAqVgRVzpbzo5smXKp4YfRvH+8abtTE1Pi6jizo',
}]
You can also return URLs to files from your plugin’s static/
directory, if
you have one:
from datasette import hookimpl
@hookimpl
def extra_js_urls():
return [
'/-/static-plugins/your-plugin/app.js'
]
publish_subcommand(publish)¶
publish
- Click publish command group- The Click command group for the
datasette publish
subcommand
This hook allows you to create new providers for the datasette publish
command. Datasette uses this hook internally to implement the default now
and heroku
subcommands, so you can read
their source
to see examples of this hook in action.
Let’s say you want to build a plugin that adds a datasette publish my_hosting_provider --api_key=xxx mydatabase.db
publish command. Your implementation would start like this:
from datasette import hookimpl
from datasette.publish.common import add_common_publish_arguments_and_options
import click
@hookimpl
def publish_subcommand(publish):
@publish.command()
@add_common_publish_arguments_and_options
@click.option(
"-k",
"--api_key",
help="API key for talking to my hosting provider",
)
def my_hosting_provider(
files,
metadata,
extra_options,
branch,
template_dir,
plugins_dir,
static,
install,
version_note,
title,
license,
license_url,
source,
source_url,
api_key,
):
# Your implementation goes here
render_cell(value, column, table, database, datasette)¶
Lets you customize the display of values within table cells in the HTML table view.
value
- string, integer or None- The value that was loaded from the database
column
- string- The name of the column being rendered
table
- string or None- The name of the table - or
None
if this is a custom SQL query database
- string- The name of the database
datasette
- Datasette instance- You can use this to access plugin configuration options via
datasette.plugin_config(your_plugin_name)
If your hook returns None
, it will be ignored. Use this to indicate that your hook is not able to custom render this particular value.
If the hook returns a string, that string will be rendered in the table cell.
If you want to return HTML markup you can do so by returning a jinja2.Markup
object.
Datasette will loop through all available render_cell
hooks and display the value returned by the first one that does not return None
.
Here is an example of a custom render_cell()
plugin which looks for values that are a JSON string matching the following format:
{"href": "https://www.example.com/", "label": "Name"}
If the value matches that pattern, the plugin returns an HTML link element:
from datasette import hookimpl
import jinja2
import json
@hookimpl
def render_cell(value):
# Render {"href": "...", "label": "..."} as link
if not isinstance(value, str):
return None
stripped = value.strip()
if not stripped.startswith("{") and stripped.endswith("}"):
return None
try:
data = json.loads(value)
except ValueError:
return None
if not isinstance(data, dict):
return None
if set(data.keys()) != {"href", "label"}:
return None
href = data["href"]
if not (
href.startswith("/") or href.startswith("http://")
or href.startswith("https://")
):
return None
return jinja2.Markup('<a href="{href}">{label}</a>'.format(
href=jinja2.escape(data["href"]),
label=jinja2.escape(data["label"] or "") or " "
))
extra_body_script(template, database, table, view_name, datasette)¶
Extra JavaScript to be added to a <script>
block at the end of the <body>
element on the page.
template
- string- The template that is being rendered, e.g.
database.html
database
- string or None- The name of the database, or
None
if the page does not correspond to a database (e.g. the root page) table
- string or None- The name of the table, or
None
if the page does not correct to a table view_name
- string- The name of the view being displayed. (database, table, and row are the most important ones.)
datasette
- Datasette instance- You can use this to access plugin configuration options via
datasette.plugin_config(your_plugin_name)
The template
, database
and table
options can be used to return different code depending on which template is being rendered and which database or table are being processed.
The datasette
instance is provided primarily so that you can consult any plugin configuration options that may have been set, using the datasette.plugin_config(plugin_name)
method documented above.
The string that you return from this function will be treated as “safe” for inclusion in a <script>
block directly in the page, so it is up to you to apply any necessary escaping.
extra_template_vars(template, database, table, view_name, request, datasette)¶
Extra template variables that should be made available in the rendered template context.
template
- string- The template that is being rendered, e.g.
database.html
database
- string or None- The name of the database, or
None
if the page does not correspond to a database (e.g. the root page) table
- string or None- The name of the table, or
None
if the page does not correct to a table view_name
- string- The name of the view being displayed. (database, table, and row are the most important ones.)
request
- object- The current HTTP request object.
request.scope
provides access to the ASGI scope. datasette
- Datasette instance- You can use this to access plugin configuration options via
datasette.plugin_config(your_plugin_name)
This hook can return one of three different types:
- Dictionary
- If you return a dictionary its keys and values will be merged into the template context.
- Function that returns a dictionary
- If you return a function it will be executed. If it returns a dictionary those values will will be merged into the template context.
- Function that returns an awaitable function that returns a dictionary
- You can also return a function which returns an awaitable function which returns a dictionary.
Datasette runs Jinja2 in async mode, which means you can add awaitable functions to the template scope and they will be automatically awaited when they are rendered by the template.
Here’s an example plugin that returns an authentication object from the ASGI scope:
@hookimpl
def extra_template_vars(request):
return {
"auth": request.scope.get("auth")
}
And here’s an example which adds a sql_first(sql_query)
function which executes a SQL statement and returns the first column of the first row of results:
@hookimpl
def extra_template_vars(datasette, database):
async def sql_first(sql, dbname=None):
dbname = dbname or database or next(iter(datasette.databases.keys()))
return (await datasette.execute(dbname, sql)).rows[0][0]
You can then use the new function in a template like so:
SQLite version: {{ sql_first("select sqlite_version()") }}
register_output_renderer(datasette)¶
datasette
- Datasette instance- You can use this to access plugin configuration options via
datasette.plugin_config(your_plugin_name)
Allows the plugin to register a new output renderer, to output data in a custom format. The hook function should return a dictionary, or a list of dictionaries, which contain the file extension you want to handle and a callback function:
@hookimpl
def register_output_renderer(datasette):
return {
'extension': 'test',
'callback': render_test
}
This will register render_test to be called when paths with the extension .test (for example /database.test, /database/table.test, or /database/table/row.test) are requested. When a request is received, the callback function is called with three positional arguments:
args
- dictionary- The GET parameters of the request
data
- dictionary- The data to be rendered
view_name
- string- The name of the view where the renderer is being called. (database, table, and row are the most important ones.)
The callback function can return None, if it is unable to render the data, or a dictionary with the following keys:
body
- string or bytes, optional- The response body, default empty
content_type
- string, optional- The Content-Type header, default text/plain
status_code
- integer, optional- The HTTP status code, default 200
A simple example of an output renderer callback function:
def render_test(args, data, view_name):
return {
'body': 'Hello World'
}
register_facet_classes()¶
Return a list of additional Facet subclasses to be registered.
Each Facet subclass implements a new type of facet operation. The class should look like this:
class SpecialFacet(Facet):
# This key must be unique across all facet classes:
type = "special"
async def suggest(self):
# Use self.sql and self.params to suggest some facets
suggested_facets = []
suggested_facets.append({
"name": column, # Or other unique name
# Construct the URL that will enable this facet:
"toggle_url": self.ds.absolute_url(
self.request, path_with_added_args(
self.request, {"_facet": column}
)
),
})
return suggested_facets
async def facet_results(self):
# This should execute the facet operation and return results, again
# using self.sql and self.params as the starting point
facet_results = {}
facets_timed_out = []
# Do some calculations here...
for column in columns_selected_for_facet:
try:
facet_results_values = []
# More calculations...
facet_results_values.append({
"value": value,
"label": label,
"count": count,
"toggle_url": self.ds.absolute_url(self.request, toggle_path),
"selected": selected,
})
facet_results[column] = {
"name": column,
"results": facet_results_values,
"truncated": len(facet_rows_results) > facet_size,
}
except QueryInterrupted:
facets_timed_out.append(column)
return facet_results, facets_timed_out
See datasette/facets.py for examples of how these classes can work.
The plugin hook can then be used to register the new facet class like this:
@hookimpl
def register_facet_classes():
return [SpecialFacet]
asgi_wrapper(datasette)¶
Return an ASGI middleware wrapper function that will be applied to the Datasette ASGI application.
This is a very powerful hook. You can use it to manipulate the entire Datasette response, or even to configure new URL routes that will be handled by your own custom code.
You can write your ASGI code directly against the low-level specification, or you can use the middleware utilites provided by an ASGI framework such as Starlette.
This example plugin adds a x-databases
HTTP header listing the currently attached databases:
from datasette import hookimpl
from functools import wraps
@hookimpl
def asgi_wrapper(datasette):
def wrap_with_databases_header(app):
@wraps(app)
async def add_x_databases_header(scope, recieve, send):
async def wrapped_send(event):
if event["type"] == "http.response.start":
original_headers = event.get("headers") or []
event = {
"type": event["type"],
"status": event["status"],
"headers": original_headers + [
[b"x-databases",
", ".join(datasette.databases.keys()).encode("utf-8")]
],
}
await send(event)
await app(scope, recieve, wrapped_send)
return add_x_databases_header
return wrap_with_databases_header
Contributing¶
Datasette is an open source project. We welcome contributions!
This document describes how to contribute to Datasette core. You can also contribute to the wider Datasette ecosystem by creating new Plugins.
General guidelines¶
- master should always be releasable. Incomplete features should live in branches. This ensures that any small bug fixes can be quickly released.
- The ideal commit should bundle together the implementation, unit tests and associated documentation updates. The commit message should link to an associated issue.
Setting up a development environment¶
If you have Python 3.6 or higher installed on your computer (on OS X the easiest way to do this is using homebrew) you can install an editable copy of Datasette using the following steps.
If you want to use GitHub to publish your changes, first create a fork of datasette under your own GitHub account.
Now clone that repository somewhere on your computer:
git clone git@github.com:YOURNAME/datasette
If you just want to get started without creating your own fork, you can do this instead:
git clone git@github.com:simonw/datasette
The next step is to create a virtual environment for your project and use it to install Datasette’s dependencies:
cd datasette
# Create a virtual environment in ./venv
python3 -m venv ./venv
# Now activate the virtual environment, so pip can install into it
source venv/bin/activate
# Install Datasette and its testing dependencies
python3 -m pip install -e .[test]
That last line does most of the work: pip install -e
means “install this package in a way that allows me to edit the source code in place”. The .[test]
option means “use the setup.py in this directory and install the optional testing dependencies as well”.
Once you have done this, you can run the Datasette unit tests from inside your datasette/
directory using pytest like so:
pytest
To run Datasette itself, just type datasette
.
You’re going to need at least one SQLite database. An easy way to get started is to use the fixtures database that Datasette uses for its own tests.
You can create a copy of that database by running this command:
python tests/fixtures.py fixtures.db
Now you can run Datasette against the new fixtures database like so:
datasette fixtures.db
This will start a server at http://127.0.0.1:8001/
.
Any changes you make in the datasette/templates
or datasette/static
folder will be picked up immediately (though you may need to do a force-refresh in your browser to see changes to CSS or JavaScript).
If you want to change Datasette’s Python code you can use the --reload
option to cause Datasette to automatically reload any time the underlying code changes:
datasette --reload fixtures.db
You can also use the fixtures.py
script to recreate the testing version of metadata.json
used by the unit tests. To do that:
python tests/fixtures.py fixtures.db fixtures-metadata.json
(You may need to delete fixtures.db
before running this command.)
Then run Datasette like this:
datasette fixtures.db -m fixtures-metadata.json
Editing and building the documentation¶
Datasette’s documentation lives in the docs/
directory and is deployed automatically using Read The Docs.
The documentation is written using reStructuredText. You may find this article on The subset of reStructuredText worth committing to memory useful.
You can build it locally by installing sphinx
and sphinx_rtd_theme
in your Datasette development environment and then running make html
directly in the docs/
directory:
# You may first need to activate your virtual environment:
source venv/bin/activate
# Install the dependencies needed to build the docs
pip install -e .[docs]
# Now build the docs
cd docs/
make html
This will create the HTML version of the documentation in docs/_build/html
. You can open it in your browser like so:
open _build/html/index.html
Any time you make changes to a .rst
file you can re-run make html
to update the built documents, then refresh them in your browser.
For added productivity, you can use use sphinx-autobuild to run Sphinx in auto-build mode. This will run a local webserver serving the docs that automatically rebuilds them and refreshes the page any time you hit save in your editor.
sphinx-autobuild
will have been installed when you ran pip install -e .[docs]
. In your docs/
directory you can start the server by running the following:
make livehtml
Now browse to http://localhost:8000/
to view the documentation. Any edits you make should be instantly reflected in your browser.
Release process¶
Datasette releases are performed using tags. When a new version tag is pushed to GitHub, a Travis CI task will perform the following:
- Run the unit tests against all supported Python versions. If the tests pass…
- Set up https://v0-25-1.datasette.io/ (but with the new tag) to point to a live demo of this release
- Build a Docker image of the release and push a tag to https://hub.docker.com/r/datasetteproject/datasette
- Re-point the “latest” tag on Docker Hub to the new image
- Build a wheel bundle of the underlying Python source code
- Push that new wheel up to PyPI: https://pypi.org/project/datasette/
To deploy new releases you will need to have push access to the main Datasette GitHub repository.
Datasette follows Semantic Versioning:
major.minor.patch
We increment major
for backwards-incompatible releases. Datasette is currently pre-1.0 so the major version is always 0
.
We increment minor
for new features.
We increment patch
for bugfix releass.
To release a new version, first create a commit that updates the changelog with highlights of the new version. An example commit can be seen here:
# Update changelog
git commit -m "Release 0.25.2" -a
git push
For non-bugfix releases you may want to update the news section of README.md
as part of the same commit.
Wait long enough for Travis to build and deploy the demo version of that commit (otherwise the tag deployment may fail to alias to it properly). Then run the following:
git tag 0.25.2
git push --tags
Final steps once the release has deployed to https://pypi.org/project/datasette/
- Manually post the new release to GitHub releases: https://github.com/simonw/datasette/releases
- Manually kick off a build of the stable branch on Read The Docs: https://readthedocs.org/projects/datasette/builds/
Changelog¶
0.32 (2019-11-14)¶
Datasette now renders templates using Jinja async mode. This makes it easy for plugins to provide custom template functions that perform asynchronous actions, for example the new datasette-template-sql plugin which allows custom templates to directly execute SQL queries and render their results. (#628)
0.31.2 (2019-11-13)¶
0.31.1 (2019-11-12)¶
- Deployments created using
datasette publish
now usepython:3.8
base Docker image (#629)
0.31 (2019-11-11)¶
This version adds compatibility with Python 3.8 and breaks compatibility with Python 3.5.
If you are still running Python 3.5 you should stick with 0.30.2
, which you can install like this:
pip install datasette==0.30.2
- Format SQL button now works with read-only SQL queries - thanks, Tobias Kunze (#602)
- New
?column__notin=x,y,z
filter for table views (#614) - Table view now uses
select col1, col2, col3
instead ofselect *
- Database filenames can now contain spaces - thanks, Tobias Kunze (#590)
- Removed obsolete
?_group_count=col
feature (#504) - Improved user interface and documentation for
datasette publish cloudrun
(#608) - Tables with indexes now show the
CREATE INDEX
statements on the table page (#618) - Current version of uvicorn is now shown on
/-/versions
- Python 3.8 is now supported! (#622)
- Python 3.5 is no longer supported.
0.30.2 (2019-11-02)¶
/-/plugins
page now uses distribution name e.g.datasette-cluster-map
instead of the name of the underlying Python package (datasette_cluster_map
) (#606)- Array faceting is now only suggested for columns that contain arrays of strings (#562)
- Better documentation for the
--host
argument (#574) - Don’t show
None
with a broken link for the label on a nullable foreign key (#406)
0.30.1 (2019-10-30)¶
0.30 (2019-10-18)¶
- Added
/-/threads
debugging page - Allow
EXPLAIN WITH...
(#583) - Button to format SQL - thanks, Tobias Kunze (#136)
- Sort databases on homepage by argument order - thanks, Tobias Kunze (#585)
- Display metadata footer on custom SQL queries - thanks, Tobias Kunze (#589)
- Use
--platform=managed
forpublish cloudrun
(#587) - Fixed bug returning non-ASCII characters in CSV (#584)
- Fix for
/foo
v.s./foo-bar
bug (#601)
0.29.3 (2019-09-02)¶
- Fixed implementation of CodeMirror on database page (#560)
- Documentation typo fixes - thanks, Min ho Kim (#561)
- Mechanism for detecting if a table has FTS enabled now works if the table name used alternative escaping mechanisms (#570) - for compatibility with a recent change to sqlite-utils.
0.29.2 (2019-07-13)¶
0.29.1 (2019-07-11)¶
0.29 (2019-07-07)¶
ASGI, new plugin hooks, facet by date and much, much more…
ASGI¶
ASGI is the Asynchronous Server Gateway Interface standard. I’ve been wanting to convert Datasette into an ASGI application for over a year - Port Datasette to ASGI #272 tracks thirteen months of intermittent development - but with Datasette 0.29 the change is finally released. This also means Datasette now runs on top of Uvicorn and no longer depends on Sanic.
I wrote about the significance of this change in Porting Datasette to ASGI, and Turtles all the way down.
The most exciting consequence of this change is that Datasette plugins can now take advantage of the ASGI standard.
New plugin hook: asgi_wrapper¶
The asgi_wrapper(datasette) plugin hook allows plugins to entirely wrap the Datasette ASGI application in their own ASGI middleware. (#520)
Two new plugins take advantage of this hook:
- datasette-auth-github adds a authentication layer: users will have to sign in using their GitHub account before they can view data or interact with Datasette. You can also use it to restrict access to specific GitHub users, or to members of specified GitHub organizations or teams.
- datasette-cors allows you to configure CORS headers for your Datasette instance. You can use this to enable JavaScript running on a whitelisted set of domains to make
fetch()
calls to the JSON API provided by your Datasette instance.
New plugin hook: extra_template_vars¶
The extra_template_vars(template, database, table, view_name, request, datasette) plugin hook allows plugins to inject their own additional variables into the Datasette template context. This can be used in conjunction with custom templates to customize the Datasette interface. datasette-auth-github uses this hook to add custom HTML to the new top navigation bar (which is designed to be modified by plugins, see #540).
Secret plugin configuration options¶
Plugins like datasette-auth-github need a safe way to set secret configuration options. Since the default mechanism for configuring plugins exposes those settings in /-/metadata
a new mechanism was needed. Secret configuration values describes how plugins can now specify that their settings should be read from a file or an environment variable:
{
"plugins": {
"datasette-auth-github": {
"client_secret": {
"$env": "GITHUB_CLIENT_SECRET"
}
}
}
}
These plugin secrets can be set directly using datasette publish
. See Custom metadata and plugins for details. (#538 and #543)
Facet by date¶
If a column contains datetime values, Datasette can now facet that column by date. (#481)
Easier custom templates for table rows¶
If you want to customize the display of individual table rows, you can do so using a _table.html
template include that looks something like this:
{% for row in display_rows %}
<div>
<h2>{{ row["title"] }}</h2>
<p>{{ row["description"] }}<lp>
<p>Category: {{ row.display("category_id") }}</p>
</div>
{% endfor %}
This is a backwards incompatible change. If you previously had a custom template called _rows_and_columns.html
you need to rename it to _table.html
.
See Custom templates for full details.
?_through= for joins through many-to-many tables¶
The new ?_through={json}
argument to the Table view allows records to be filtered based on a many-to-many relationship. See Special table arguments for full documentation - here’s an example. (#355)
This feature was added to help support facet by many-to-many, which isn’t quite ready yet but will be coming in the next Datasette release.
Small changes¶
- Databases published using
datasette publish
now open in Immutable mode. (#469) ?col__date=
now works for columns containing spaces- Automatic label detection (for deciding which column to show when linking to a foreign key) has been improved. (#485)
- Fixed bug where pagination broke when combined with an expanded foreign key. (#489)
- Contributors can now run
pip install -e .[docs]
to get all of the dependencies needed to build the documentation, includingcd docs && make livehtml
support. - Datasette’s dependencies are now all specified using the
~=
match operator. (#532) white-space: pre-wrap
now used for table creation SQL. (#505)
Full list of commits between 0.28 and 0.29.
0.28 (2019-05-19)¶
A salmagundi of new features!
Supporting databases that change¶
From the beginning of the project, Datasette has been designed with read-only databases in mind. If a database is guaranteed not to change it opens up all kinds of interesting opportunities - from taking advantage of SQLite immutable mode and HTTP caching to bundling static copies of the database directly in a Docker container. The interesting ideas in Datasette explores this idea in detail.
As my goals for the project have developed, I realized that read-only databases are no longer the right default. SQLite actually supports concurrent access very well provided only one thread attempts to write to a database at a time, and I keep encountering sensible use-cases for running Datasette on top of a database that is processing inserts and updates.
So, as-of version 0.28 Datasette no longer assumes that a database file will not change. It is now safe to point Datasette at a SQLite database which is being updated by another process.
Making this change was a lot of work - see tracking tickets #418, #419 and #420. It required new thinking around how Datasette should calculate table counts (an expensive operation against a large, changing database) and also meant reconsidering the “content hash” URLs Datasette has used in the past to optimize the performance of HTTP caches.
Datasette can still run against immutable files and gains numerous performance benefits from doing so, but this is no longer the default behaviour. Take a look at the new Performance and caching documentation section for details on how to make the most of Datasette against data that you know will be staying read-only and immutable.
Faceting improvements, and faceting plugins¶
Datasette Facets provide an intuitive way to quickly summarize and interact with data. Previously the only supported faceting technique was column faceting, but 0.28 introduces two powerful new capabilities: facet-by-JSON-array and the ability to define further facet types using plugins.
Facet by array (#359) is only available if your SQLite installation provides the json1
extension. Datasette will automatically detect columns that contain JSON arrays of values and offer a faceting interface against those columns - useful for modelling things like tags without needing to break them out into a new table. See Facet by JSON array for more.
The new register_facet_classes() plugin hook (#445) can be used to register additional custom facet classes. Each facet class should provide two methods: suggest()
which suggests facet selections that might be appropriate for a provided SQL query, and facet_results()
which executes a facet operation and returns results. Datasette’s own faceting implementations have been refactored to use the same API as these plugins.
datasette publish cloudrun¶
Google Cloud Run is a brand new serverless hosting platform from Google, which allows you to build a Docker container which will run only when HTTP traffic is received and will shut down (and hence cost you nothing) the rest of the time. It’s similar to Zeit’s Now v1 Docker hosting platform which sadly is no longer accepting signups from new users.
The new datasette publish cloudrun
command was contributed by Romain Primet (#434) and publishes selected databases to a new Datasette instance running on Google Cloud Run.
See Publishing to Google Cloud Run for full documentation.
register_output_renderer plugins¶
Russ Garrett implemented a new Datasette plugin hook called register_output_renderer (#441) which allows plugins to create additional output renderers in addition to Datasette’s default .json
and .csv
.
Russ’s in-development datasette-geo plugin includes an example of this hook being used to output .geojson
automatically converted from SpatiaLite.
Medium changes¶
- Datasette now conforms to the Black coding style (#449) - and has a unit test to enforce this in the future
- New Special table arguments:
?columnname__in=value1,value2,value3
filter for executing SQL IN queries against a table, see Table arguments (#433)?columnname__date=yyyy-mm-dd
filter which returns rows where the spoecified datetime column falls on the specified date (583b22a)?tags__arraycontains=tag
filter which acts against a JSON array contained in a column (78e45ea)?_where=sql-fragment
filter for the table view (#429)?_fts_table=mytable
and?_fts_pk=mycolumn
querystring options can be used to specify which FTS table to use for a search query - see Configuring full-text search for a table or view (#428)
- You can now pass the same table filter multiple times - for example,
?content__not=world&content__not=hello
will return all rows where the content column is neitherhello
orworld
(#288) - You can now specify
about
andabout_url
metadata (in addition tosource
andlicense
) linking to further information about a project - see Source, license and about - New
?_trace=1
parameter now adds debug information showing every SQL query that was executed while constructing the page (#435) datasette inspect
now just calculates table counts, and does not introspect other database metadata (#462)- Removed
/-/inspect
page entirely - this will be replaced by something similar in the future, see #465 - Datasette can now run against an in-memory SQLite database. You can do this by starting it without passing any files or by using the new
--memory
option todatasette serve
. This can be useful for experimenting with SQLite queries that do not access any data, such asSELECT 1+1
orSELECT sqlite_version()
.
Small changes¶
- We now show the size of the database file next to the download link (#172)
- New
/-/databases
introspection page shows currently connected databases (#470) - Binary data is no longer displayed on the table and row pages (#442 - thanks, Russ Garrett)
- New show/hide SQL links on custom query pages (#415)
- The extra_body_script plugin hook now accepts an optional
view_name
argument (#443 - thanks, Russ Garrett) - Bumped Jinja2 dependency to 2.10.1 (#426)
- All table filters are now documented, and documentation is enforced via unit tests (2c19a27)
- New project guideline: master should stay shippable at all times! (31f36e1)
- Fixed a bug where
sqlite_timelimit()
occasionally failed to clean up after itself (bac4e01) - We no longer load additional plugins when executing pytest (#438)
- Homepage now links to database views if there are less than five tables in a database (#373)
- The
--cors
option is now respected by error pages (#453) datasette publish heroku
now uses the--include-vcs-ignore
option, which means it works under Travis CI (#407)datasette publish heroku
now publishes using Python 3.6.8 (666c374)- Renamed
datasette publish now
todatasette publish nowv1
(#472) datasette publish nowv1
now accepts multiple--alias
parameters (09ef305)- Removed the
datasette skeleton
command (#476) - The documentation on how to build the documentation now recommends
sphinx-autobuild
0.27.1 (2019-05-09)¶
- Tiny bugfix release: don’t install
tests/
in the wrong place. Thanks, Veit Heller.
0.27 (2019-01-31)¶
- New command:
datasette plugins
(documentation) shows you the currently installed list of plugins. - Datasette can now output newline-delimited JSON using the new
?_shape=array&_nl=on
querystring option. - Added documentation on The Datasette Ecosystem.
- Now using Python 3.7.2 as the base for the official Datasette Docker image.
0.26.1 (2019-01-10)¶
/-/versions
now includes SQLitecompile_options
(#396)- datasetteproject/datasette Docker image now uses SQLite 3.26.0 (#397)
- Cleaned up some deprecation warnings under Python 3.7
0.26 (2019-01-02)¶
datasette serve --reload
now restarts Datasette if a database file changes on disk.datasette publish now
now takes an optional--alias mysite.now.sh
argument. This will attempt to set an alias after the deploy completes.- Fixed a bug where the advanced CSV export form failed to include the currently selected filters (#393)
0.25.2 (2018-12-16)¶
datasette publish heroku
now uses thepython-3.6.7
runtime- Added documentation on how to build the documentation
- Added documentation covering our release process
- Upgraded to pytest 4.0.2
0.25.1 (2018-11-04)¶
Documentation improvements plus a fix for publishing to Zeit Now.
datasette publish now
now uses Zeit’s v1 platform, to work around the new 100MB image limit. Thanks, @slygent - closes #366.
0.25 (2018-09-19)¶
New plugin hooks, improved database view support and an easier way to use more recent versions of SQLite.
- New
publish_subcommand
plugin hook. A plugin can now add additionaldatasette publish
publishers in addition to the defaultnow
andheroku
, both of which have been refactored into default plugins. publish_subcommand documentation. Closes #349 - New
render_cell
plugin hook. Plugins can now customize how values are displayed in the HTML tables produced by Datasette’s browseable interface. datasette-json-html and datasette-render-images are two new plugins that use this hook. render_cell documentation. Closes #352 - New
extra_body_script
plugin hook, enabling plugins to provide additional JavaScript that should be added to the page footer. extra_body_script documentation. extra_css_urls
andextra_js_urls
hooks now take additional optional parameters, allowing them to be more selective about which pages they apply to. Documentation.- You can now use the sortable_columns metadata setting to explicitly enable sort-by-column in the interface for database views, as well as for specific tables.
- The new
fts_table
andfts_pk
metadata settings can now be used to explicitly configure full-text search for a table or a view, even if that table is not directly coupled to the SQLite FTS feature in the database schema itself. - Datasette will now use pysqlite3 in place of the standard library
sqlite3
module if it has been installed in the current environment. This makes it much easier to run Datasette against a more recent version of SQLite, including the just-released SQLite 3.25.0 which adds window function support. More details on how to use this in #360 - New mechanism that allows plugin configuration options to be set using
metadata.json
.
0.24 (2018-07-23)¶
A number of small new features:
datasette publish heroku
now supports--extra-options
, fixes #334- Custom error message if SpatiaLite is needed for specified database, closes #331
- New config option:
truncate_cells_html
for truncating long cell values in HTML view - closes #330 - Documentation for datasette publish and datasette package, closes #337
- Fixed compatibility with Python 3.7
datasette publish heroku
now supports app names via the-n
option, which can also be used to overwrite an existing application [Russ Garrett]- Title and description metadata can now be set for canned SQL queries, closes #342
- New
force_https_on
config option, fixeshttps://
API URLs when deploying to Zeit Now - closes #333 ?_json_infinity=1
querystring argument for handling Infinity/-Infinity values in JSON, closes #332- URLs displayed in the results of custom SQL queries are now URLified, closes #298
0.23.2 (2018-07-07)¶
Minor bugfix and documentation release.
- CSV export now respects
--cors
, fixes #326 - Installation instructions, including docker image - closes #328
- Fix for row pages for tables with / in, closes #325
0.23.1 (2018-06-21)¶
Minor bugfix release.
- Correctly display empty strings in HTML table, closes #314
- Allow “.” in database filenames, closes #302
- 404s ending in slash redirect to remove that slash, closes #309
- Fixed incorrect display of compound primary keys with foreign key references. Closes #319
- Docs + example of canned SQL query using || concatenation. Closes #321
- Correctly display facets with value of 0 - closes #318
- Default ‘expand labels’ to checked in CSV advanced export
0.23 (2018-06-18)¶
This release features CSV export, improved options for foreign key expansions, new configuration settings and improved support for SpatiaLite.
See datasette/compare/0.22.1…0.23 for a full list of commits added since the last release.
CSV export¶
Any Datasette table, view or custom SQL query can now be exported as CSV.
Check out the CSV export documentation for more details, or try the feature out on https://fivethirtyeight.datasettes.com/fivethirtyeight/bechdel%2Fmovies
If your table has more than max_returned_rows (default 1,000) Datasette provides the option to stream all rows. This option takes advantage of async Python and Datasette’s efficient pagination to iterate through the entire matching result set and stream it back as a downloadable CSV file.
Foreign key expansions¶
When Datasette detects a foreign key reference it attempts to resolve a label for that reference (automatically or using the Specifying the label column for a table metadata option) so it can display a link to the associated row.
This expansion is now also available for JSON and CSV representations of the
table, using the new _labels=on
querystring option. See
Expanding foreign key references for more details.
New configuration settings¶
Datasette’s Configuration now also supports boolean settings. A number of new configuration options have been added:
num_sql_threads
- the number of threads used to execute SQLite queries. Defaults to 3.allow_facet
- enable or disable custom Facets using the _facet= parameter. Defaults to on.suggest_facets
- should Datasette suggest facets? Defaults to on.allow_download
- should users be allowed to download the entire SQLite database? Defaults to on.allow_sql
- should users be allowed to execute custom SQL queries? Defaults to on.default_cache_ttl
- Default HTTP caching max-age header in seconds. Defaults to 365 days - caching can be disabled entirely by settings this to 0.cache_size_kb
- Set the amount of memory SQLite uses for its per-connection cache, in KB.allow_csv_stream
- allow users to stream entire result sets as a single CSV file. Defaults to on.max_csv_mb
- maximum size of a returned CSV file in MB. Defaults to 100MB, set to 0 to disable this limit.
Control HTTP caching with ?_ttl=¶
You can now customize the HTTP max-age header that is sent on a per-URL basis, using the new ?_ttl=
querystring parameter.
You can set this to any value in seconds, or you can set it to 0 to disable HTTP caching entirely.
Consider for example this query which returns a randomly selected member of the Avengers:
select * from [avengers/avengers] order by random() limit 1
If you hit the following page repeatedly you will get the same result, due to HTTP caching:
/fivethirtyeight?sql=select+*+from+%5Bavengers%2Favengers%5D+order+by+random%28%29+limit+1
By adding ?_ttl=0 to the zero you can ensure the page will not be cached and get back a different super hero every time:
/fivethirtyeight?sql=select+*+from+%5Bavengers%2Favengers%5D+order+by+random%28%29+limit+1&_ttl=0
Improved support for SpatiaLite¶
The SpatiaLite module for SQLite adds robust geospatial features to the database.
Getting SpatiaLite working can be tricky, especially if you want to use the most recent alpha version (with support for K-nearest neighbor).
Datasette now includes extensive documentation on SpatiaLite, and thanks to Ravi Kotecha our GitHub repo includes a Dockerfile that can build the latest SpatiaLite and configure it for use with Datasette.
The datasette publish
and datasette package
commands now accept a new
--spatialite
argument which causes them to install and configure SpatiaLite
as part of the container they deploy.
latest.datasette.io¶
Every commit to Datasette master is now automatically deployed by Travis CI to https://latest.datasette.io/ - ensuring there is always a live demo of the latest version of the software.
The demo uses the fixtures from our unit tests, ensuring it demonstrates the same range of functionality that is covered by the tests.
You can see how the deployment mechanism works in our .travis.yml file.
0.22.1 (2018-05-23)¶
Bugfix release, plus we now use versioneer for our version numbers.
Faceting no longer breaks pagination, fixes #282
Add
__version_info__
derived from __version__ [Robert Gieseke]This might be tuple of more than two values (major and minor version) if commits have been made after a release.
Add version number support with Versioneer. [Robert Gieseke]
Versioneer Licence: Public Domain (CC0-1.0)
Closes #273
Refactor inspect logic [Russ Garrett]
0.22 (2018-05-20)¶
The big new feature in this release is Facets. Datasette can now apply faceted browse to any column in any table. It will also suggest possible facets. See the Datasette Facets announcement post for more details.
In addition to the work on facets:
New
--config
option, added--help-config
, closes #274Removed the
--page_size=
argument todatasette serve
in favour of:datasette serve --config default_page_size:50 mydb.db
Added new help section:
$ datasette --help-config Config options: default_page_size Default page size for the table view (default=100) max_returned_rows Maximum rows that can be returned from a table or custom query (default=1000) sql_time_limit_ms Time limit for a SQL query in milliseconds (default=1000) default_facet_size Number of values to return for requested facets (default=30) facet_time_limit_ms Time limit for calculating a requested facet (default=200) facet_suggest_time_limit_ms Time limit for calculating a suggested facet (default=50)
Only apply responsive table styles to
.rows-and-column
Otherwise they interfere with tables in the description, e.g. on https://fivethirtyeight.datasettes.com/fivethirtyeight/nba-elo%2Fnbaallelo
Refactored views into new
views/
modules, refs #256Documentation for SQLite full-text search support, closes #253
/-/versions
now includes SQLitefts_versions
, closes #252
0.21 (2018-05-05)¶
New JSON _shape=
options, the ability to set table _size=
and a mechanism for searching within specific columns.
Default tests to using a longer timelimit
Every now and then a test will fail in Travis CI on Python 3.5 because it hit the default 20ms SQL time limit.
Test fixtures now default to a 200ms time limit, and we only use the 20ms time limit for the specific test that tests query interruption. This should make our tests on Python 3.5 in Travis much more stable.
Support
_search_COLUMN=text
searches, closes #237Show version on
/-/plugins
page, closes #248?_size=max
option, closes #249Added
/-/versions
and/-/versions.json
, closes #244Sample output:
{ "python": { "version": "3.6.3", "full": "3.6.3 (default, Oct 4 2017, 06:09:38) \n[GCC 4.2.1 Compatible Apple LLVM 9.0.0 (clang-900.0.37)]" }, "datasette": { "version": "0.20" }, "sqlite": { "version": "3.23.1", "extensions": { "json1": null, "spatialite": "4.3.0a" } } }
Renamed
?_sql_time_limit_ms=
to?_timelimit
, closes #242New
?_shape=array
option + tweaks to_shape
, closes #245- Default is now
?_shape=arrays
(renamed fromlists
) - New
?_shape=array
returns an array of objects as the root object - Changed
?_shape=object
to return the object as the root - Updated docs
- Default is now
FTS tables now detected by
inspect()
, closes #240New
?_size=XXX
querystring parameter for table view, closes #229Also added documentation for all of the
_special
arguments.Plus deleted some duplicate logic implementing
_group_count
.If
max_returned_rows==page_size
, incrementmax_returned_rows
- fixes #230New
hidden: True
option for table metadata, closes #239Hide
idx_*
tables if spatialite detected, closes #228Added
class=rows-and-columns
to custom query results tableAdded CSS class
rows-and-columns
to main tablelabel_column
option inmetadata.json
- closes #234
0.20 (2018-04-20)¶
Mostly new work on the Plugins mechanism: plugins can now bundle static assets and custom templates, and datasette publish
has a new --install=name-of-plugin
option.
Add col-X classes to HTML table on custom query page
Fixed out-dated template in documentation
Plugins can now bundle custom templates, #224
Added /-/metadata /-/plugins /-/inspect, #225
Documentation for –install option, refs #223
Datasette publish/package –install option, #223
Fix for plugins in Python 3.5, #222
New plugin hooks: extra_css_urls() and extra_js_urls(), #214
/-/static-plugins/PLUGIN_NAME/ now serves static/ from plugins
<th> now gets class=”col-X” - plus added col-X documentation
Use to_css_class for table cell column classes
This ensures that columns with spaces in the name will still generate usable CSS class names. Refs #209
Add column name classes to <td>s, make PK bold [Russ Garrett]
Don’t duplicate simple primary keys in the link column [Russ Garrett]
When there’s a simple (single-column) primary key, it looks weird to duplicate it in the link column.
This change removes the second PK column and treats the link column as if it were the PK column from a header/sorting perspective.
Correct escaping for HTML display of row links [Russ Garrett]
Longer time limit for test_paginate_compound_keys
It was failing intermittently in Travis - see #209
Use application/octet-stream for downloadable databases
Updated PyPI classifiers
Updated PyPI link to pypi.org
0.19 (2018-04-16)¶
This is the first preview of the new Datasette plugins mechanism. Only two plugin hooks are available so far - for custom SQL functions and custom template filters. There’s plenty more to come - read the documentation and get involved in the tracking ticket if you have feedback on the direction so far.
Fix for
_sort_desc=sortable_with_nulls
test, refs #216Fixed #216 - paginate correctly when sorting by nullable column
Initial documentation for plugins, closes #213
New
--plugins-dir=plugins/
option (#212)New option causing Datasette to load and evaluate all of the Python files in the specified directory and register any plugins that are defined in those files.
This new option is available for the following commands:
datasette serve mydb.db --plugins-dir=plugins/ datasette publish now/heroku mydb.db --plugins-dir=plugins/ datasette package mydb.db --plugins-dir=plugins/
Start of the plugin system, based on pluggy (#210)
Uses https://pluggy.readthedocs.io/ originally created for the py.test project
We’re starting with two plugin hooks:
prepare_connection(conn)
This is called when a new SQLite connection is created. It can be used to register custom SQL functions.
prepare_jinja2_environment(env)
This is called with the Jinja2 environment. It can be used to register custom template tags and filters.
An example plugin which uses these two hooks can be found at https://github.com/simonw/datasette-plugin-demos or installed using
pip install datasette-plugin-demos
Refs #14
Return HTTP 405 on InvalidUsage rather than 500. [Russ Garrett]
This also stops it filling up the logs. This happens for HEAD requests at the moment - which perhaps should be handled better, but that’s a different issue.
0.18 (2018-04-14)¶
This release introduces support for units,
contributed by Russ Garrett (#203).
You can now optionally specify the units for specific columns using metadata.json
.
Once specified, units will be displayed in the HTML view of your table. They also become
available for use in filters - if a column is configured with a unit of distance, you can
request all rows where that column is less than 50 meters or more than 20 feet for example.
Link foreign keys which don’t have labels. [Russ Garrett]
This renders unlabeled FKs as simple links.
Also includes bonus fixes for two minor issues:
- In foreign key link hrefs the primary key was escaped using HTML escaping rather than URL escaping. This broke some non-integer PKs.
- Print tracebacks to console when handling 500 errors.
Fix SQLite error when loading rows with no incoming FKs. [Russ Garrett]
This fixes
ERROR: conn=<sqlite3.Connection object at 0x10bbb9f10>, sql = 'select ', params = {'id': '1'}
caused by an invalid query when loading incoming FKs.The error was ignored due to async but it still got printed to the console.
Allow custom units to be registered with Pint. [Russ Garrett]
Support units in filters. [Russ Garrett]
Tidy up units support. [Russ Garrett]
- Add units to exported JSON
- Units key in metadata skeleton
- Docs
Initial units support. [Russ Garrett]
Add support for specifying units for a column in
metadata.json
and rendering them on display using pint
0.17 (2018-04-13)¶
- Release 0.17 to fix issues with PyPI
0.16 (2018-04-13)¶
Better mechanism for handling errors; 404s for missing table/database
New error mechanism closes #193
404s for missing tables/databases closes #184
long_description in markdown for the new PyPI
Hide Spatialite system tables. [Russ Garrett]
Allow
explain select
/explain query plan select
#201Datasette inspect now finds primary_keys #195
Ability to sort using form fields (for mobile portrait mode) #199
We now display sort options as a select box plus a descending checkbox, which means you can apply sort orders even in portrait mode on a mobile phone where the column headers are hidden.
0.15 (2018-04-09)¶
The biggest new feature in this release is the ability to sort by column. On the
table page the column headers can now be clicked to apply sort (or descending
sort), or you can specify ?_sort=column
or ?_sort_desc=column
directly
in the URL.
table_rows
=>table_rows_count
,filtered_table_rows
=>filtered_table_rows_count
Renamed properties. Closes #194
New
sortable_columns
option inmetadata.json
to control sort options.You can now explicitly set which columns in a table can be used for sorting using the
_sort
and_sort_desc
arguments usingmetadata.json
:{ "databases": { "database1": { "tables": { "example_table": { "sortable_columns": [ "height", "weight" ] } } } } }
Refs #189
Column headers now link to sort/desc sort - refs #189
_sort
and_sort_desc
parameters for table viewsAllows for paginated sorted results based on a specified column.
Refs #189
Total row count now correct even if
_next
appliedUse .custom_sql() for _group_count implementation (refs #150)
Make HTML title more readable in query template (#180) [Ryan Pitts]
New
?_shape=objects/object/lists
param for JSON API (#192)New
_shape=
parameter replacing old.jsono
extensionNow instead of this:
/database/table.jsono
We use the
_shape
parameter like this:/database/table.json?_shape=objects
Also introduced a new
_shape
calledobject
which looks like this:/database/table.json?_shape=object
Returning an object for the rows key:
... "rows": { "pk1": { ... }, "pk2": { ... } }
Refs #122
Utility for writing test database fixtures to a .db file
python tests/fixtures.py /tmp/hello.db
This is useful for making a SQLite database of the test fixtures for interactive exploration.
Compound primary key
_next=
now plays well with extra filtersCloses #190
Fixed bug with keyset pagination over compound primary keys
Refs #190
Database/Table views inherit
source/license/source_url/license_url
metadataIf you set the
source_url/license_url/source/license
fields in your root metadata those values will now be inherited all the way down to the database and table templates.The
title/description
are NOT inherited.Also added unit tests for the HTML generated by the metadata.
Refs #185
Add metadata, if it exists, to heroku temp dir (#178) [Tony Hirst]
Initial documentation for pagination
Broke up test_app into test_api and test_html
Fixed bug with .json path regular expression
I had a table called
geojson
and it caused an exception because the regex was matching.json
and not\.json
Deploy to Heroku with Python 3.6.3
0.14 (2017-12-09)¶
The theme of this release is customization: Datasette now allows every aspect of its presentation to be customized either using additional CSS or by providing entirely new templates.
Datasette’s metadata.json format
has also been expanded, to allow per-database and per-table metadata. A new
datasette skeleton
command can be used to generate a skeleton JSON file
ready to be filled in with per-database and per-table details.
The metadata.json
file can also be used to define
canned queries,
as a more powerful alternative to SQL views.
extra_css_urls
/extra_js_urls
in metadataA mechanism in the
metadata.json
format for adding custom CSS and JS urls.Create a
metadata.json
file that looks like this:{ "extra_css_urls": [ "https://simonwillison.net/static/css/all.bf8cd891642c.css" ], "extra_js_urls": [ "https://code.jquery.com/jquery-3.2.1.slim.min.js" ] }
Then start datasette like this:
datasette mydb.db --metadata=metadata.json
The CSS and JavaScript files will be linked in the
<head>
of every page.You can also specify a SRI (subresource integrity hash) for these assets:
{ "extra_css_urls": [ { "url": "https://simonwillison.net/static/css/all.bf8cd891642c.css", "sri": "sha384-9qIZekWUyjCyDIf2YK1FRoKiPJq4PHt6tp/ulnuuyRBvazd0hG7pWbE99zvwSznI" } ], "extra_js_urls": [ { "url": "https://code.jquery.com/jquery-3.2.1.slim.min.js", "sri": "sha256-k2WSCIexGzOj3Euiig+TlR8gA0EmPjuc79OEeY5L45g=" } ] }
Modern browsers will only execute the stylesheet or JavaScript if the SRI hash matches the content served. You can generate hashes using https://www.srihash.org/
Auto-link column values that look like URLs (#153)
CSS styling hooks as classes on the body (#153)
Every template now gets CSS classes in the body designed to support custom styling.
The index template (the top level page at
/
) gets this:<body class="index">
The database template (
/dbname/
) gets this:<body class="db db-dbname">
The table template (
/dbname/tablename
) gets:<body class="table db-dbname table-tablename">
The row template (
/dbname/tablename/rowid
) gets:<body class="row db-dbname table-tablename">
The
db-x
andtable-x
classes use the database or table names themselves IF they are valid CSS identifiers. If they aren’t, we strip any invalid characters out and append a 6 character md5 digest of the original name, in order to ensure that multiple tables which resolve to the same stripped character version still have different CSS classes.Some examples (extracted from the unit tests):
"simple" => "simple" "MixedCase" => "MixedCase" "-no-leading-hyphens" => "no-leading-hyphens-65bea6" "_no-leading-underscores" => "no-leading-underscores-b921bc" "no spaces" => "no-spaces-7088d7" "-" => "336d5e" "no $ characters" => "no--characters-59e024"
datasette --template-dir=mytemplates/
argumentYou can now pass an additional argument specifying a directory to look for custom templates in.
Datasette will fall back on the default templates if a template is not found in that directory.
Ability to over-ride templates for individual tables/databases.
It is now possible to over-ride templates on a per-database / per-row or per- table basis.
When you access e.g.
/mydatabase/mytable
Datasette will look for the following:- table-mydatabase-mytable.html - table.html
If you provided a
--template-dir
argument to datasette serve it will look in that directory first.The lookup rules are as follows:
Index page (/): index.html Database page (/mydatabase): database-mydatabase.html database.html Table page (/mydatabase/mytable): table-mydatabase-mytable.html table.html Row page (/mydatabase/mytable/id): row-mydatabase-mytable.html row.html
If a table name has spaces or other unexpected characters in it, the template filename will follow the same rules as our custom
<body>
CSS classes - for example, a table called “Food Trucks” will attempt to load the following templates:table-mydatabase-Food-Trucks-399138.html table.html
It is possible to extend the default templates using Jinja template inheritance. If you want to customize EVERY row template with some additional content you can do so by creating a row.html template like this:
{% extends "default:row.html" %} {% block content %} <h1>EXTRA HTML AT THE TOP OF THE CONTENT BLOCK</h1> <p>This line renders the original block:</p> {{ super() }} {% endblock %}
--static
option for datasette serve (#160)You can now tell Datasette to serve static files from a specific location at a specific mountpoint.
For example:
datasette serve mydb.db --static extra-css:/tmp/static/css
Now if you visit this URL:
http://localhost:8001/extra-css/blah.css
The following file will be served:
/tmp/static/css/blah.css
Canned query support.
Named canned queries can now be defined in
metadata.json
like this:{ "databases": { "timezones": { "queries": { "timezone_for_point": "select tzid from timezones ..." } } } }
These will be shown in a new “Queries” section beneath “Views” on the database page.
New
datasette skeleton
command for generatingmetadata.json
(#164)metadata.json
support for per-table/per-database metadata (#165)Also added support for descriptions and HTML descriptions.
Here’s an example metadata.json file illustrating custom per-database and per- table metadata:
{ "title": "Overall datasette title", "description_html": "This is a <em>description with HTML</em>.", "databases": { "db1": { "title": "First database", "description": "This is a string description & has no HTML", "license_url": "http://example.com/", "license": "The example license", "queries": { "canned_query": "select * from table1 limit 3;" }, "tables": { "table1": { "title": "Custom title for table1", "description": "Tables can have descriptions too", "source": "This has a custom source", "source_url": "http://example.com/" } } } } }
Renamed
datasette build
command todatasette inspect
(#130)Upgrade to Sanic 0.7.0 (#168)
Package and publish commands now accept
--static
and--template-dir
Example usage:
datasette package --static css:extra-css/ --static js:extra-js/ \ sf-trees.db --template-dir templates/ --tag sf-trees --branch master
This creates a local Docker image that includes copies of the templates/, extra-css/ and extra-js/ directories. You can then run it like this:
docker run -p 8001:8001 sf-trees
For publishing to Zeit now:
datasette publish now --static css:extra-css/ --static js:extra-js/ \ sf-trees.db --template-dir templates/ --name sf-trees --branch master
HTML comment showing which templates were considered for a page (#171)
0.13 (2017-11-24)¶
Search now applies to current filters.
Combined search into the same form as filters.
Closes #133
Much tidier design for table view header.
Closes #147
Added
?column__not=blah
filter.Closes #148
Row page now resolves foreign keys.
Closes #132
Further tweaks to select/input filter styling.
Refs #86 - thanks for the help, @natbat!
Show linked foreign key in table cells.
Added UI for editing table filters.
Refs #86
Hide FTS-created tables on index pages.
Closes #129
Add publish to heroku support [Jacob Kaplan-Moss]
datasette publish heroku mydb.db
Pull request #104
Initial implementation of
?_group_count=column
.URL shortcut for counting rows grouped by one or more columns.
?_group_count=column1&_group_count=column2
works as well.SQL generated looks like this:
select "qSpecies", count(*) as "count" from Street_Tree_List group by "qSpecies" order by "count" desc limit 100
Or for two columns like this:
select "qSpecies", "qSiteInfo", count(*) as "count" from Street_Tree_List group by "qSpecies", "qSiteInfo" order by "count" desc limit 100
Refs #44
Added
--build=master
option to datasette publish and package.The
datasette publish
anddatasette package
commands both now accept an optional--build
argument. If provided, this can be used to specify a branch published to GitHub that should be built into the container.This makes it easier to test code that has not yet been officially released to PyPI, e.g.:
datasette publish now mydb.db --branch=master
Implemented
?_search=XXX
+ UI if a FTS table is detected.Closes #131
Added
datasette --version
support.Table views now show expanded foreign key references, if possible.
If a table has foreign key columns, and those foreign key tables have
label_columns
, the TableView will now query those other tables for the corresponding values and display those values as links in the corresponding table cells.label_columns are currently detected by the
inspect()
function, which looks for any table that has just two columns - an ID column and one other - and sets thelabel_column
to be that second non-ID column.Don’t prevent tabbing to “Run SQL” button (#117) [Robert Gieseke]
See comment in #115
Add keyboard shortcut to execute SQL query (#115) [Robert Gieseke]
Allow
--load-extension
to be set via environment variable.Add support for
?field__isnull=1
(#107) [Ray N]Add spatialite, switch to debian and local build (#114) [Ariel Núñez]
Added
--load-extension
argument to datasette serve.Allows loading of SQLite extensions. Refs #110.
0.12 (2017-11-16)¶
Added
__version__
, now displayed as tooltip in page footer (#108).Added initial docs, including a changelog (#99).
Turned on auto-escaping in Jinja.
Added a UI for editing named parameters (#96).
You can now construct a custom SQL statement using SQLite named parameters (e.g.
:name
) and datasette will display form fields for editing those parameters. Here’s an example which lets you see the most popular names for dogs of different species registered through various dog registration schemes in Australia.
Pin to specific Jinja version. (#100).
Default to 127.0.0.1 not 0.0.0.0. (#98).
Added extra metadata options to publish and package commands. (#92).
You can now run these commands like so:
datasette now publish mydb.db \ --title="My Title" \ --source="Source" \ --source_url="http://www.example.com/" \ --license="CC0" \ --license_url="https://creativecommons.org/publicdomain/zero/1.0/"
This will write those values into the metadata.json that is packaged with the app. If you also pass
--metadata=metadata.json
that file will be updated with the extra values before being written into the Docker image.Added simple production-ready Dockerfile (#94) [Andrew Cutler]
New
?_sql_time_limit_ms=10
argument to database and table page (#95)SQL syntax highlighting with Codemirror (#89) [Tom Dyson]
0.11 (2017-11-14)¶
Added
datasette publish now --force
option.This calls
now
with--force
- useful as it means you get a fresh copy of datasette even if Now has already cached that docker layer.Enable
--cors
by default when running in a container.
0.10 (2017-11-14)¶
Fixed #83 - 500 error on individual row pages.
Stop using sqlite WITH RECURSIVE in our tests.
The version of Python 3 running in Travis CI doesn’t support this.
0.9 (2017-11-13)¶
Added
--sql_time_limit_ms
and--extra-options
.The serve command now accepts
--sql_time_limit_ms
for customizing the SQL time limit.The publish and package commands now accept
--extra-options
which can be used to specify additional options to be passed to the datasite serve command when it executes inside the resulting Docker containers.
0.8 (2017-11-13)¶
V0.8 - added PyPI metadata, ready to ship.
Implemented offset/limit pagination for views (#70).
Improved pagination. (#78)
Limit on max rows returned, controlled by
--max_returned_rows
option. (#69)If someone executes ‘select * from table’ against a table with a million rows in it, we could run into problems: just serializing that much data as JSON is likely to lock up the server.
Solution: we now have a hard limit on the maximum number of rows that can be returned by a query. If that limit is exceeded, the server will return a
"truncated": true
field in the JSON.This limit can be optionally controlled by the new
--max_returned_rows
option. Setting that option to 0 disables the limit entirely.