textual-autocomplete is a Python library for creating dropdown autocompletion menus in Textual applications, allowing users to quickly select from a list of suggestions as they type. textual-autocomplete supports Textual version 0.14.0 and above.
Video example
textual-autocomplete-styles.mov
Warning
This README is relevant to version 2.x of textual-autocomplete
. Alpha versions of version 3 are available on PyPI now, but they are unstable and undocumented.
For best results, pin textual-autocomplete
to version 2.
Simply wrap a Textual Input
widget as follows:
from textual.app import ComposeResult
from textual.widgets import Input
from textual_autocomplete import AutoComplete, Dropdown, DropdownItem
def compose(self) -> ComposeResult:
yield AutoComplete(
Input(placeholder="Type to search..."),
Dropdown(items=[
DropdownItem("Glasgow"),
DropdownItem("Edinburgh"),
DropdownItem("Aberdeen"),
DropdownItem("Dundee"),
]),
)
There are more complete examples here.
textual-autocomplete
can be installed from PyPI using your favourite dependency
manager.
As shown in the quickstart, you can wrap the Textual builtin Input
widget with
AutoComplete
, and supply a Dropdown
.
The AutoComplete
manages communication between the Input
and the Dropdown
.
The Dropdown
is the widget you see on screen, as you type into the input.
The DropdownItem
s contain up to 3 columns. All must contain a "main" column, which
is the central column used in the filtering. They can also optionally contain a left and right metadata
column.
You can supply the data for the dropdown via a list or a callback function.
The easiest way to use textual-autocomplete is to pass in a list of DropdownItem
s,
as shown in the quickstart.
Instead of passing a list of DropdownItems
, you can supply a callback function
which will be called with the current input state. From this function, you should
return the list of DropdownItems
you wish to be displayed.
See here for a usage example.
- Press the Up/Down arrow keys to navigate.
- Press Enter to select the item in the dropdown and fill it in.
- Press Tab to fill in the selected item, and move focus.
- Press Esc to hide the dropdown.
- Press the Up/Down arrow keys to force the dropdown to appear.
The Dropdown
itself can be styled using Textual CSS.
For more fine-grained control over styling, you can target the following CSS classes:
.autocomplete--highlight-match
: the highlighted portion of a matching item.autocomplete--selection-cursor
: the item the selection cursor is on.autocomplete--left-column
: the left metadata column, if it exists.autocomplete--right-column
: the right metadata column, if it exists
Since the 3 columns in DropdownItem
support Rich Text
objects, they can be styled dynamically.
The custom_meta.py file is an example of this, showing how the rightmost column is coloured dynamically based on the city population.
The examples directory contains multiple examples of custom styling.
When you select an item in the dropdown, an AutoComplete.Selected
event is emitted.
You can declare a handler for this event on_auto_complete_selected(self, event)
to respond
to an item being selected.
An item is selected when it's highlighted in the dropdown, and you press Enter or Tab.
Pressing Enter simply fills the value in the dropdown, whilst Tab fills the value and then shifts focus from the input.
- textual-autocomplete will create a new layer at runtime on the
Screen
that theAutoComplete
is on. TheDropdown
will be rendered on this layer. - The position of the dropdown is currently fixed below the value entered into the
Input
. This means if yourInput
is at the bottom of the screen, it's probably not going to be much use for now. I'm happy to discuss or look at PRs that offer a flag for having it float above. - There's currently no special handling for when the dropdown meets the right-hand side of the screen.
- Do not apply
margin
to theDropdown
. The position of the dropdown is updated by applying margin to the top/left of it. - There's currently no debouncing support, but I'm happy to discuss or look at PRs for this.
- There are a few known issues/TODOs in the code, which will later be transferred to GitHub.
- Test coverage is currently non-existent - sorry!