Skip to main content

Advanced permission management for Datasette

Project description

datasette-acl

PyPI Changelog Tests License

Advanced permission management for Datasette. Highly experimental.

Installation

Install this plugin in the same environment as Datasette. This plugin requires Datasette 1.0a15 or higher.

datasette install datasette-acl

Usage

This plugin is under active development. It currently only supports configuring permissions for individual tables, controlling the following:

  • insert-row
  • delete-row
  • update-row
  • alter-table
  • drop-table

Permissions are saved in the internal database. This means you should run Datasette with the --internal path/to/internal.db option, otherwise your permissions will be reset every time you restart Datasette.

Managing permissions for a table

The interface for configuring table permissions lives at /database-name/table-name/-/acl. It can be accessed from the table actions menu on the table page.

Permission can be granted for each of the above table actions. They can be assigned to both groups and individual users, who can be added using their actor["id"].

An audit log tracks which permissions were added and removed, displayed at the bottom of the table permissions page.

Controlling who can edit permissions

Users with the new datasette-acl permission will have the ability to access a UI for setting permissions for users and groups on a table.

To configure the root user to have this permission, add the following to your Datasette configuration:

permissions:
  datasette-acl:
    id: root

Alternatively you can start Datasette running like this:

datasette mydata.db --root --internal internal.db \
  -s permissions.datasette-acl.id root

User groups

Users can be assigned to groups, and those groups can then be used to quickly assign permissions to all of those users at once.

To manage your groups, visit /-/acl/groups or use the "Manage user groups" item in the Datasette application menu.

Add users to a group by typing in their actor ID. Remove them using the remove user button.

The page for each group includes an audit log showing changes made to that group's list of members.

When you delete a group its members will all be removed and it will be marked as deleted. Creating a group with the same name will reuse that group's record and display its existing audit log, but will not re-add the members that were removed.

Dynamic groups

You may wish to define permission rules against groups of actors based on their actor attributes, without needing to manually add those actors to a group. This can be achieved by defining a dynamic group in the datasette-acl configuration.

Dynamic groups are defined in terms of allow blocks. The following configuration defines two dynamic groups - one called admin that contains all users with "is_admin": true in their attributes, and another called sales that explicitly includes users with "sales" as one of the values in their departments array.

plugins:
  datasette-acl:
    dynamic-groups:
      admin:
        is_admin: true
      sales:
        departments: ["sales"]

Any time an actor has their permissions checked they will be dynamically added to or removed from these groups based on the current value of their actor attributes.

Dynamic groups are displayed in the list of groups, but their members cannot be manually added or removed.

Table creator permissions

If you allow regular users to create tables in Datasette, you may want them to maintain a level of "ownership" over those tables, such that other users are unable to modify those tables without the creator's permission.

The table-creator-permissions plugin setting can be used to automatically configure permissions for the actor who created a table.

Enable that like this:

plugins:
  datasette-acl:
    table-creator-permissions:
    - alter-table
    - drop-table
    - insert-row
    - update-row
    - delete-row

Configuring autocomplete against actor IDs

By default, users of this plugin can assign permissions to any actor ID by entering that ID, whether or not that ID corresponds to a user that exists elsewhere in the current Datasette configuration.

If you are running this plugin in an environment with a fixed, known list of actor IDs you can implement a plugin using the datasette_acl_valid_actors(datasette) plugin hook which returns an iterable sequence of string actor IDs or {"id": "actor-id", "display": "Actor Name"} dictionaries

These will then be used for both validation and autocomplete, ensuring users do not attach actor IDs that are not in that list.

Example plugin implementation:

from datasette import hookimpl

@hookimpl
def datasette_acl_valid_actors(datasette):
    return ["paulo", "rohan", "simon"]

This function can also return an async inner function, for making async calls. This example uses the [{"id": "actor-id", "display": "Actor Name"}] format:

from datasette import hookimpl

@hookimpl
def datasette_acl_valid_actors(datasette):
    async def inner():
        db = datasette.get_internal_database()
        return (await db.execute("select id, username as display from users")).dicts()
    return inner

Development

To set up this plugin locally, first checkout the code. Then create a new virtual environment:

cd datasette-acl
python -m venv venv
source venv/bin/activate

Now install the dependencies and test dependencies:

pip install -e '.[test]'

To run the tests:

python -m pytest

Tips for local development

Here's how to run the plugin with all of its features enabled.

First, grab a test database:

wget https://latest.datasette.io/fixtures.db

Install the datasette-unsafe-actor-debug plugin, so you can use the http://127.0.0.1:8001/-/unsafe-actor page to quickly imitate any actor for testing purposes:

datasette install datasette-unsafe-actor-debug

And datasette-visible-internal-db to make it easy to see what's going on in the internal database:

datasette install datasette-visible-internal-db

Then start Datasette like this:

datasette fixtures.db --internal internal.db \
  -s permissions.datasette-acl.id root \
  -s plugins.datasette-unsafe-actor-debug.enabled 1 \
  -s plugins.datasette-acl.table-creator-permissions '["insert-row", "update-row"]' \
  -s plugins.datasette-acl.dynamic-groups.staff.is_staff true \
  --root \
  --secret 1 \
  --reload

This configures Datasette to provide a URL for you to sign in as root, which will give you access to the permission editing tool.

It ensures that any user who creates a table (which you can test using the /-/api API explorer tool) will be granted initial insert-row and update-row permissions.

It sets up a dynamic group such that any actor with {"is_staff": true} in their JSON will be treated as a member of that group.

--reload means Datasette will reload on any code changes to the plugin, and --secret 1 ensures your Datasette authentication cookies will continue to work across server restarts.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

datasette_acl-0.4a4.tar.gz (46.9 kB view details)

Uploaded Source

Built Distribution

datasette_acl-0.4a4-py3-none-any.whl (43.8 kB view details)

Uploaded Python 3

File details

Details for the file datasette_acl-0.4a4.tar.gz.

File metadata

  • Download URL: datasette_acl-0.4a4.tar.gz
  • Upload date:
  • Size: 46.9 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/5.1.1 CPython/3.12.6

File hashes

Hashes for datasette_acl-0.4a4.tar.gz
Algorithm Hash digest
SHA256 4df4ed797aca031c34ff48c740ef589055e316beba0c97ee957e9579f59e9265
MD5 b4fbc90d22d87cd5f2b277f1ec1063b4
BLAKE2b-256 bb5302d5eabacb892dff792649fea53616356566842cb989ec877ee849160dca

See more details on using hashes here.

File details

Details for the file datasette_acl-0.4a4-py3-none-any.whl.

File metadata

File hashes

Hashes for datasette_acl-0.4a4-py3-none-any.whl
Algorithm Hash digest
SHA256 578ef0d349c7af30ab70baed260ed7c9bc6771826eccfb57b57899020f79c903
MD5 de987e796c416e4f79f83c015b93ee12
BLAKE2b-256 32380845eb0471f58da4d23389ef9661d079b33215e3c4e4b06f128266365a94

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page