Selecting & editing
With the pointer tool active, people click an annotation to select it. A box with handles appears
around it: drag the annotation to move it, a handle to resize it, and the round handle above it to
turn it. Your code reads and changes the same selection, and you build the menu and the style
panel around it.
How people select#
- Click an annotation to select it. Shift-click adds it to the selection, or takes it out.
- Drag on empty space to draw a box: everything it touches is selected. It has to touch what an annotation paints: its lines, or the inside of a filled shape. A box drawn inside an empty rectangle doesn’t select the rectangle.
- Click empty space to clear the selection.
- Double-click a text box to type in it.
While moving, the selection snaps to the edges and centers of other annotations and of the page, and a turn snaps to 0°, 90°, 180° and 270°. Hold Shift to move or turn freely. See Snapping to change it.
Select with code#
annotation.selection.set([ref]); // select these
annotation.selection.add([otherRef]); // add to the selection
annotation.selection.selectAll(page.ref); // everything on a page (or leave out the page)
annotation.selection.selectInRect(page.ref, { x: 50, y: 50, width: 200, height: 100 });
annotation.selection.clear();To show the selection in a component, use useAnnotationState(). Your component
updates when the selection changes:
<script lang="ts">
import { useAnnotationState } from '@embedpdf/svelte/annotation';
const state = useAnnotationState(); // state.selected: the selected annotations
</script>
<p>{state.selected.length} selected</p>Loading live preview…
<script lang="ts" module>
import { DocumentGate, Viewer, type OpenInput } from '@embedpdf/svelte/runtime';
import { Stage, stagePlugin } from '@embedpdf/svelte/stage';
import { RenderLayer, renderPlugin } from '@embedpdf/svelte/render';
import { interactionPlugin } from '@embedpdf/svelte/interaction';
import { AnnotationLayer, annotationPlugin } from '@embedpdf/svelte/annotation';
import { localEngine } from '@embedpdf/engine';
import AddShapes from './AddShapes.svelte';
import SelectionToolbar from './SelectionToolbar.svelte';
import '../select-code.css';
const engine = localEngine();
const plugins = [stagePlugin(), renderPlugin(), interactionPlugin(), annotationPlugin()];
const ebook = async (): Promise<OpenInput> => {
const response = await fetch('https://snippet.embedpdf.com/ebook.pdf');
return { kind: 'bytes', id: 'ebook', bytes: new Uint8Array(await response.arrayBuffer()) };
};
</script>
<Viewer {engine} {plugins} initialDocuments={[{ source: ebook }]}>
<DocumentGate>
{#snippet fallback()}
<p class="loading">Loading…</p>
{/snippet}
<AddShapes />
<SelectionToolbar />
<Stage class="stage">
<RenderLayer />
<AnnotationLayer />
</Stage>
</DocumentGate>
</Viewer>Try it: The top half selects the two shapes the box touches. Shift-click a shape to add it, or take it out.
Every field is listed under State.
Add a menu#
<AnnotationMenu> places your own UI next to the selection. Mount it in the
Stage’s overlay snippet,
and it stays attached while the user scrolls and zooms:
<script lang="ts">
import {
AnnotationLayer,
AnnotationMenu,
useAnnotation,
useAnnotationState,
} from '@embedpdf/svelte/annotation';
import { RenderLayer } from '@embedpdf/svelte/render';
import { Stage } from '@embedpdf/svelte/stage';
const annotation = useAnnotation();
const state = useAnnotationState();
const allText = $derived(state.selected.every((a) => a.subtype === 'free-text'));
</script>
<Stage>
<RenderLayer />
<AnnotationLayer />
{#snippet overlay()}
<AnnotationMenu placement="bottom">
<div class="menu">
{#if allText}
<button onclick={() => annotation.text.toggleFormat('bold')}>Bold</button>
{/if}
<button onclick={() => annotation.selection.update({ color: '#dc143c' })}>Red</button>
<button onclick={() => annotation.selection.delete()}>Delete</button>
</div>
</AnnotationMenu>
{/snippet}
</Stage>Loading live preview…
<script lang="ts" module>
import { DocumentGate, Viewer, type OpenInput } from '@embedpdf/svelte/runtime';
import { Stage, stagePlugin } from '@embedpdf/svelte/stage';
import { RenderLayer, renderPlugin } from '@embedpdf/svelte/render';
import { interactionPlugin } from '@embedpdf/svelte/interaction';
import { AnnotationLayer, AnnotationMenu, annotationPlugin } from '@embedpdf/svelte/annotation';
import { localEngine } from '@embedpdf/engine';
import AddAnnotations from './AddAnnotations.svelte';
import SelectionActions from './SelectionActions.svelte';
import '../menu.css';
const engine = localEngine();
const plugins = [stagePlugin(), renderPlugin(), interactionPlugin(), annotationPlugin()];
const ebook = async (): Promise<OpenInput> => {
const response = await fetch('https://snippet.embedpdf.com/ebook.pdf');
return { kind: 'bytes', id: 'ebook', bytes: new Uint8Array(await response.arrayBuffer()) };
};
</script>
<Viewer {engine} {plugins} initialDocuments={[{ source: ebook }]}>
<DocumentGate>
{#snippet fallback()}
<p class="loading">Loading…</p>
{/snippet}
<AddAnnotations />
<Stage class="stage">
<RenderLayer />
<AnnotationLayer />
{#snippet overlay()}
<AnnotationMenu placement="bottom">
<SelectionActions />
</AnnotationMenu>
{/snippet}
</Stage>
</DocumentGate>
</Viewer>Try it: select the rectangle: the menu loses Bold and Edit text. Drag it, and the menu hides until you let go.
- The menu only places things. What’s in it is yours: build it from the selection, as above, to show buttons that fit what’s selected.
- It moves out of the way. It hides while the user drags, and it never covers the rotation handle.
placementis'top'(the default),'bottom','left'or'right', andgapis the space between the selection and the menu, in pixels.
A style panel#
A style panel shows the properties of what’s selected, such as color, width, opacity and font, with
their current values. useAnnotationProperties() gives you everything to draw it:
propertiesare what the selection can change, in display order. Each one says which control fits: a color, a number with its range, a choice from a list, or an on/off flag.valuesare their current values.mixedlists the properties where the selected annotations differ. Show those as “mixed”.
When different kinds are selected, you get only the properties they all have.
Each property’s key is the annotation’s own field name, so most read and write with no mapping.
Two group several fields: borderStyle is the style, the dashes and the cloudy edge together, and
lineEndings reads as a { start, end } pair while its options are single endings. The panel below
shows both:
<script lang="ts">
import { useAnnotation, useAnnotationProperties } from '@embedpdf/svelte/annotation';
import Control from './Control.svelte'; // your own control
const annotation = useAnnotation();
const panel = useAnnotationProperties(); // panel.current: { properties, values, mixed }
</script>
{#each panel.current.properties as property (property.key)}
<Control
{property}
value={panel.current.values[property.key]}
mixed={panel.current.mixed.includes(property.key)}
onchange={(value: unknown) => annotation.selection.update({ [property.key]: value })}
/>
{/each}Loading live preview…
<script lang="ts" module>
import { DocumentGate, Viewer, type OpenInput } from '@embedpdf/svelte/runtime';
import { Stage, stagePlugin } from '@embedpdf/svelte/stage';
import { RenderLayer, renderPlugin } from '@embedpdf/svelte/render';
import { interactionPlugin } from '@embedpdf/svelte/interaction';
import { AnnotationLayer, annotationPlugin } from '@embedpdf/svelte/annotation';
import { localEngine } from '@embedpdf/engine';
import AddAnnotations from './AddAnnotations.svelte';
import StylePanel from './StylePanel.svelte';
import '../style-panel.css';
const engine = localEngine();
const plugins = [stagePlugin(), renderPlugin(), interactionPlugin(), annotationPlugin()];
const ebook = async (): Promise<OpenInput> => {
const response = await fetch('https://snippet.embedpdf.com/ebook.pdf');
return { kind: 'bytes', id: 'ebook', bytes: new Uint8Array(await response.arrayBuffer()) };
};
</script>
<Viewer {engine} {plugins} initialDocuments={[{ source: ebook }]}>
<DocumentGate>
{#snippet fallback()}
<p class="loading">Loading…</p>
{/snippet}
<AddAnnotations />
<div class="viewer">
<Stage class="stage">
<RenderLayer />
<AnnotationLayer />
</Stage>
<StylePanel />
</div>
</DocumentGate>
</Viewer>Try it: change the rectangle’s colors and width, then select the arrow or the text box: the panel shows what each one has. Shift-click two to see what they share.
property.control | What to show | Also on the property |
|---|---|---|
'color' | A color picker. null means no color. | |
'number' | A slider or a number input. | min, max, step |
'choice' | A list, such as fonts or line endings. Each option is { value, label, labelKey }: write value, show label. | options |
'flag' | An on/off switch, such as locked or print. | |
'text' | A text input, such as a redaction’s label. | |
'textFormat' | A bold, italic or underline toggle for text boxes. Write it with annotation.text.toggleFormat(property.format). | format |
'link' | Where the annotation links to. Write it with annotation.selection.updateLink(target), or null to remove it. |
selection.update() gives each selected annotation the properties its kind has, and skips the rest.
To change each annotation relative to itself, pass a function. For example, to set only the end
arrow of lines that keep their own start:
await annotation.selection.update((member) =>
member.subtype === 'line' ? { lineEndings: { ...member.lineEndings, end: 'open-arrow' } } : {},
);The same properties drive a panel for a tool’s defaults; see Let people change the defaults.
In the user’s language#
Each property has a label in English and a labelKey to translate it, and so does each option:
annotation.style.stroke, annotation.flags.noView, annotation.lineEndings.openArrow. With the
i18n plugin, show t(property.labelKey, { fallback: property.label }), and
give your languages strings for the keys; without a string for a key, the English shows. A key names
the wording, not the field: color is annotation.style.stroke on a rectangle and
annotation.style.outline on a redaction mark, so each kind keeps its own word.
Change the handles#
By default the selection has an outline in the accent color, square handles on its corners and sides (or on each point of a line or a polygon), and a round rotation handle above it, on a short line called its stalk. Everything about them can change.
Set them when you register the plugin. One color changes every part; the rest is optional:
annotationPlugin({
chrome: {
accent: '#e91e63',
outline: { style: 'dashed', width: 1 },
handles: { shape: 'circle', size: 10 },
rotationHandle: { offset: 30, stalk: false },
},
});| Setting | What it changes | From CSS |
|---|---|---|
accent | The color of every part below, unless you set one of its own. Defaults to the viewer’s accent. | --epdf-annotation-accent |
outline | The box around the selection: style ('solid', 'dashed'), width, color. | --epdf-annotation-outline, --epdf-annotation-outline-width, --epdf-annotation-outline-dash |
handles | The resize and point handles: shape ('square', 'circle'), size, hitSize, fill, stroke. | --epdf-annotation-handle-fill, --epdf-annotation-handle-stroke |
rotationHandle | The turn handle: enabled, size, hitSize, offset (distance from the box), stalk (false hides the line to it), fill, stroke. | --epdf-annotation-rotation-handle-fill, --epdf-annotation-rotation-handle-stroke |
guides | The lines shown while snapping and turning: enabled, color, rotationColor, style, width. | --epdf-annotation-guide, --epdf-annotation-rotation-guide, --epdf-annotation-guide-width, --epdf-annotation-guide-dash |
marquee | The box you drag to select several annotations: fill, stroke. | --epdf-annotation-marquee-fill, --epdf-annotation-marquee-stroke |
textOutline | The outline of a text box while you type in it. | --epdf-annotation-text-outline |
readout | The angle while you turn: enabled, background, color. | --epdf-readout-background, --epdf-readout-color |
Every size is in screen pixels, so handles stay the same size at every zoom. hitSize is the area
you can grab; keep it at 24 or more for touch.
Override the colors from CSS, for example for dark mode. A CSS variable wins over the setting:
@media (prefers-color-scheme: dark) {
.pdf-viewer {
--epdf-annotation-accent: #ff80ab;
}
}Colors, widths and dashes can come from CSS. Sizes and the offset can’t, because they decide where you can grab. Every variable is on Colors & theming.
Change them while the app runs, for example when the user picks another accent:
annotation.updateSettings({ chrome: { accent: '#0f6e56' } });
annotation.resetSettings(); // back to what the plugin was registered withDraw them yourself. Give the layer handle and rotationHandle
snippets, and it draws them in place of its handles:
<script lang="ts">
import { AnnotationLayer } from '@embedpdf/svelte/annotation';
</script>
<AnnotationLayer>
{#snippet handle({ at, size, rotation })}
<div
class="my-handle"
style:left="{at.x - size / 2}px"
style:top="{at.y - size / 2}px"
style:width="{size}px"
style:height="{size}px"
style:rotate="{rotation}deg"
></div>
{/snippet}
<!-- a rotationHandle snippet works the same way -->
</AnnotationLayer>| Prop | What it is |
|---|---|
at | The handle’s center, in pixels on the page. |
size | The size from the settings. |
rotation | The selection’s rotation, so a square handle can turn with it. |
kind | 'corner', 'side' or 'point' (a point of a line or a polygon). |
active | True while it’s being dragged. |
The rotation handle gets the same props except kind, plus from, where its stalk starts on the box. You only draw:
the viewer still decides where a handle can be grabbed, so what you draw can’t break dragging.
Loading live preview…
<script lang="ts" module>
import { DocumentGate, Viewer, type OpenInput } from '@embedpdf/svelte/runtime';
import { Stage, stagePlugin } from '@embedpdf/svelte/stage';
import { RenderLayer, renderPlugin } from '@embedpdf/svelte/render';
import { interactionPlugin } from '@embedpdf/svelte/interaction';
import {
AnnotationLayer,
annotationPlugin,
type HandleProps,
type RotationHandleProps,
} from '@embedpdf/svelte/annotation';
import { localEngine } from '@embedpdf/engine';
import AddRectangle from './AddRectangle.svelte';
import ChromeControls from './ChromeControls.svelte';
import '../handles.css';
const engine = localEngine();
const plugins = [stagePlugin(), renderPlugin(), interactionPlugin(), annotationPlugin()];
const ebook = async (): Promise<OpenInput> => {
const response = await fetch('https://snippet.embedpdf.com/ebook.pdf');
return { kind: 'bytes', id: 'ebook', bytes: new Uint8Array(await response.arrayBuffer()) };
};
</script>
<script lang="ts">
let own = $state(false);
</script>
<!-- Your own handles: the layer places them, and still decides where they can be grabbed. -->
{#snippet handle({ at, size, rotation, kind, active }: HandleProps)}
<div
class={['handle', `handle--${kind}`, active && 'handle--active']}
style:left="{at.x - size / 2}px"
style:top="{at.y - size / 2}px"
style:width="{size}px"
style:height="{size}px"
style:rotate="{rotation}deg"
></div>
{/snippet}
{#snippet rotationHandle({ at, size, active }: RotationHandleProps)}
<div
class={['turn', active && 'turn--active']}
style:left="{at.x - size}px"
style:top="{at.y - size}px"
style:width="{size * 2}px"
style:height="{size * 2}px"
>
↻
</div>
{/snippet}
<Viewer {engine} {plugins} initialDocuments={[{ source: ebook }]}>
<DocumentGate>
{#snippet fallback()}
<p class="loading">Loading…</p>
{/snippet}
<AddRectangle />
<ChromeControls bind:own />
<Stage class="stage">
<RenderLayer />
<AnnotationLayer
handle={own ? handle : undefined}
rotationHandle={own ? rotationHandle : undefined}
/>
</Stage>
</DocumentGate>
</Viewer>Try it: pick an accent and round handles, then Draw them myself. Reset goes back to what the plugin was registered with.
Snapping#
annotationPlugin({
snap: {
alignment: true, // snap to other annotations and the page while moving
alignmentThreshold: 6, // how close, in screen pixels, before it snaps
rotation: true,
rotationAngles: [0, 45, 90, 135, 180, 225, 270, 315],
rotationThreshold: 4, // degrees
},
});
annotation.updateSettings({ snap: { alignment: false } }); // while the app runsThe threshold is in screen pixels, so snapping feels the same at every zoom. The lines that show
what an annotation snaps to are --epdf-annotation-guide.
Loading live preview…
<script lang="ts" module>
import { DocumentGate, Viewer, type OpenInput } from '@embedpdf/svelte/runtime';
import { Stage, stagePlugin } from '@embedpdf/svelte/stage';
import { RenderLayer, renderPlugin } from '@embedpdf/svelte/render';
import { interactionPlugin } from '@embedpdf/svelte/interaction';
import { AnnotationLayer, annotationPlugin } from '@embedpdf/svelte/annotation';
import { localEngine } from '@embedpdf/engine';
import AddRectangles from './AddRectangles.svelte';
import SnapControls from './SnapControls.svelte';
import '../snapping.css';
const engine = localEngine();
const plugins = [stagePlugin(), renderPlugin(), interactionPlugin(), annotationPlugin()];
const ebook = async (): Promise<OpenInput> => {
const response = await fetch('https://snippet.embedpdf.com/ebook.pdf');
return { kind: 'bytes', id: 'ebook', bytes: new Uint8Array(await response.arrayBuffer()) };
};
</script>
<Viewer {engine} {plugins} initialDocuments={[{ source: ebook }]}>
<DocumentGate>
{#snippet fallback()}
<p class="loading">Loading…</p>
{/snippet}
<AddRectangles />
<SnapControls />
<Stage class="stage">
<RenderLayer />
<AnnotationLayer />
</Stage>
</DocumentGate>
</Viewer>Try it: drag the red rectangle until its edge lines up with the blue one, then turn it with the round handle.
Turn and group#
annotation.selection.rotateBy(90); // a quarter turn clockwise; -90 the other way
annotation.selection.resetRotation(); // upright againGrouping makes several annotations select and move as one. The group is saved in the PDF the way Acrobat saves it, so it stays a group there too:
if (annotation.selection.canGroup()) await annotation.selection.group();
if (annotation.selection.canUngroup()) await annotation.selection.ungroup();Loading live preview…
<script lang="ts" module>
import { DocumentGate, Viewer, type OpenInput } from '@embedpdf/svelte/runtime';
import { Stage, stagePlugin } from '@embedpdf/svelte/stage';
import { RenderLayer, renderPlugin } from '@embedpdf/svelte/render';
import { interactionPlugin } from '@embedpdf/svelte/interaction';
import { AnnotationLayer, annotationPlugin } from '@embedpdf/svelte/annotation';
import { localEngine } from '@embedpdf/engine';
import AddShapes from './AddShapes.svelte';
import TurnAndGroup from './TurnAndGroup.svelte';
import '../turn-group.css';
const engine = localEngine();
const plugins = [stagePlugin(), renderPlugin(), interactionPlugin(), annotationPlugin()];
const ebook = async (): Promise<OpenInput> => {
const response = await fetch('https://snippet.embedpdf.com/ebook.pdf');
return { kind: 'bytes', id: 'ebook', bytes: new Uint8Array(await response.arrayBuffer()) };
};
</script>
<Viewer {engine} {plugins} initialDocuments={[{ source: ebook }]}>
<DocumentGate>
{#snippet fallback()}
<p class="loading">Loading…</p>
{/snippet}
<AddShapes />
<TurnAndGroup />
<Stage class="stage">
<RenderLayer />
<AnnotationLayer />
</Stage>
</DocumentGate>
</Viewer>Try it: Group the two shapes, click empty space, then click one: both are selected.
To bring an annotation to the front or send it to the back, see Change the drawing order.
Edit text#
A double-click on a text box starts typing in it; a click outside ends it. From code:
annotation.text.begin(ref);
await annotation.text.toggleFormat('bold'); // the selected words, or the whole box
await annotation.text.end();annotation.text.getEditing() is the box being typed in, or null. How fonts and formatting work
is on Rich text.
What people may change#
editable says which annotations people may change by hand. One it leaves out can still be
selected, hovered, read and commented on, but it has no handles, doesn’t move, turn or take typing,
the style panel has nothing for it, and selection.update() and selection.delete() leave it as
it is:
annotationPlugin({ editable: ['highlight', 'ink'] }); // only highlights and drawings change
annotation.updateSettings({ editable: false }); // nothing changes by hand, while the app runs
annotation.updateSettings({ editable: true }); // everything again, the defaultThe list names kinds by their subtype. A measurement goes by its intent instead
('line-dimension', 'polyline-dimension' or 'polygon-dimension'), so ['line'] keeps
measurements still and ['line-dimension'] lets only them change. annotation.isEditable(ref) says
whether people may change one.
It’s only about people: your code still changes any annotation with update() and delete(), and
canUpdate(ref) and canDelete(ref) stay the same. To stop your code too, leave the permission out
of the scope.
Shapes with many points#
A polygon or a polyline takes a click per point, and a double-click ends it. While it’s being drawn,
<AnnotationDraftMenu> can show buttons to finish or cancel it, which helps on touch screens:
<script lang="ts">
import { AnnotationDraftMenu, useAnnotation } from '@embedpdf/svelte/annotation';
const annotation = useAnnotation();
</script>
<AnnotationDraftMenu>
{#snippet children(draft)}
<button disabled={!draft.canFinish} onclick={() => annotation.draft.finish()}>Done</button>
<button onclick={() => annotation.draft.cancel()}>Cancel</button>
{/snippet}
</AnnotationDraftMenu>Keyboard#
The headless plugin doesn’t take over any keys, since your app may already use them. Delete and Escape are a few lines:
<script lang="ts">
import { useAnnotation } from '@embedpdf/svelte/annotation';
import { useInteraction } from '@embedpdf/svelte/interaction';
const annotation = useAnnotation();
const interaction = useInteraction();
function onKey(event: KeyboardEvent) {
if (event.target instanceof HTMLInputElement || annotation.text.getEditing()) return;
if (event.key === 'Delete' || event.key === 'Backspace') void annotation.selection.delete();
if (event.key === 'Escape') {
annotation.cancel(); // a drag or a polygon in progress
interaction.activateDefaultTool();
}
}
</script>
<svelte:window onkeydown={onKey} />Loading live preview…
<script lang="ts" module>
import { DocumentGate, Viewer, type OpenInput } from '@embedpdf/svelte/runtime';
import { Stage, stagePlugin } from '@embedpdf/svelte/stage';
import { RenderLayer, renderPlugin } from '@embedpdf/svelte/render';
import { interactionPlugin } from '@embedpdf/svelte/interaction';
import { AnnotationLayer, annotationPlugin } from '@embedpdf/svelte/annotation';
import { localEngine } from '@embedpdf/engine';
import AnnotationKeys from './AnnotationKeys.svelte';
import DraftMenu from './DraftMenu.svelte';
import Toolbar from './Toolbar.svelte';
import '../polygon.css';
const engine = localEngine();
const plugins = [stagePlugin(), renderPlugin(), interactionPlugin(), annotationPlugin()];
const ebook = async (): Promise<OpenInput> => {
const response = await fetch('https://snippet.embedpdf.com/ebook.pdf');
return { kind: 'bytes', id: 'ebook', bytes: new Uint8Array(await response.arrayBuffer()) };
};
</script>
<Viewer {engine} {plugins} initialDocuments={[{ source: ebook }]}>
<DocumentGate>
{#snippet fallback()}
<p class="loading">Loading…</p>
{/snippet}
<AnnotationKeys />
<Toolbar />
<Stage class="stage">
<RenderLayer />
<AnnotationLayer />
{#snippet overlay()}
<DraftMenu />
{/snippet}
</Stage>
</DocumentGate>
</Viewer>Try it: click a few points on the page, then Done, or double-click. Select a shape and press Delete; Escape stops a shape you’re drawing.
Methods#
What useAnnotation() gives you for selecting, drawing and editing text: its selection, draft and text parts. The rest is on
Annotations in code.
| Method | What it does |
|---|---|
selection.set(refs), selection.add(refs) | Selects these annotations, or adds them to the selection. An annotation in a group brings its group. Select with code |
selection.selectAll(page?) | Selects everything that can be selected, on one page or in the whole document |
selection.selectInRect(page, rect) | Selects everything a box on a page touches of what it paints, with the rest of each one’s group, as dragging on empty space does |
selection.clear() | Clears the selection |
selection.update(changes) | Changes the selected annotations: each takes the fields its kind has. Pass a function to change each one relative to itself. A style panel |
selection.updateLink(target) | Links the selected annotations to a page or a website, or removes their links with null. A style panel |
selection.getProperties() | What a style panel shows for the selection: { properties, values, mixed }. In a component, use useAnnotationProperties(). |
selection.delete() | Deletes every selected annotation |
selection.rotateBy(degrees), selection.resetRotation() | Turns the selection a quarter turn (90 clockwise, -90 the other way), or back upright. Turn and group |
selection.group(), selection.ungroup() | Makes the selected annotations one group, saved the way Acrobat saves it, or splits the selected group up again |
selection.canGroup(), selection.canUngroup() | Whether the selection can be grouped, or is a group that can be split |
selection.getAnchor() | Where a menu attaches: { page, bounds } around the selection, or null. <AnnotationMenu> uses it. Add a menu |
selection.getRotationAnchor() | While a selection is being turned: { page, at, angle }, the pointer’s point and the angle so far in degrees clockwise, or null. <AnnotationRotationBadge> uses it to show the angle next to the pointer. |
draft.get() | The polygon or polyline being drawn ({ page, bounds, pointCount, canFinish }), or null. <AnnotationDraftMenu> uses it. Shapes with many points |
draft.finish(), draft.cancel() | Ends the shape and creates it, as a double-click does, or drops it |
text.begin(ref), text.end() | Starts typing in a text box, as a double-click does, or stops and saves what was typed. Edit text |
text.toggleFormat(format) | Turns 'bold', 'italic' or 'underline' on or off, for the selected words or the whole box |
isEditable(ref) | Whether people may change this annotation by hand, as the editable setting says. What people may change |
cancel() | Stops a drag, or a polygon in progress. Keyboard |
Every method that returns a promise also takes a signal in its options, to cancel it.
State#
Everything useAnnotationState() gives you, as one reactive object: read its fields, and your markup updates when they change. The method in the last column reads the same value once.
| Field | What it is | Read once |
|---|---|---|
status | 'loading' until every annotation is in, then 'ready'; 'forbidden' or 'error' when they can’t load | getStatus() |
selected | The selected annotations | selection.list() |
hovered | The annotation under the pointer, or null | getHovered() |
editing | The text box being typed in, or null | text.getEditing() |
The annotations themselves come from useAnnotationList(filter), a tool’s defaults from
useAnnotationDefaults(id), and a style panel’s properties from useAnnotationProperties().
Events#
<script lang="ts">
import { useAnnotationEvent } from '@embedpdf/svelte/annotation';
let sidebarOpen = $state(false);
useAnnotationEvent(
(annotation) => annotation.onSelectionChanged,
({ refs }) => (sidebarOpen = refs.length > 0),
);
</script>
{#if sidebarOpen}
<aside><!-- … --></aside>
{/if}| Event | Fires when | Carries |
|---|---|---|
onSelectionChanged | The selection changed | refs, previousRefs |
onDraftChanged | A polygon or polyline is being drawn, or ended | draft |
onEditingChanged | Typing in a text box started or ended | ref |
onHoverChanged | The pointer moved onto an annotation, or off it | ref |
Changes to the annotations themselves are on Events, saving & sharing.
Good to know#
Locked annotations#
People can select a locked annotation, but not move, resize, turn or delete it, and
lockedContents stops them from changing its text. A hidden annotation can’t be selected at all.
The editable setting does what locked does without writing the flag into the PDF: see
What people may change.
Touch#
A tap selects. The handles of a selected annotation are easier to grab with a finger: their grab area is twice as large on touch. One finger on the selected annotation moves it; anywhere else, it scrolls the page.
Your feedback goes directly to the documentation team.