fix(web): Create Person face preview not working for video assets

FaceEditor previously required an HTMLImageElement | HTMLVideoElement prop to
compute layout metrics and generate the face crop preview. This was unavailable
for video assets, so the preview thumbnail in the Create Person modal was always
missing, and face positions could be NaN during image load (naturalWidth is 0
before the image decodes).

Replace the DOM element prop with assetSize: Size and containerSize: Size, using
asset metadata dimensions that are always available from the API response.
computeContentMetrics() is extracted as a pure utility alongside
mapContentRectToNatural() for converting face rect coordinates back to original
image space.

For videos, VideoNativeViewer now captures the current frame to canvas when face
edit mode opens and sets assetViewerManager.imgRef, giving FaceEditor the same
image-based preview path as photo assets.

Change-Id: I0e9da549e3af40211abad4ab2c0270706a6a6964

.github/FUNDING.yml

@@ -1 +0,0 @@
-custom: ['https://buy.immich.app', 'https://immich.store']

CODE_OF_CONDUCT.md

@@ -1,134 +0,0 @@
-# Contributor Covenant Code of Conduct
-
-## Our Pledge
-
-We as members, contributors, and leaders pledge to make participation
-in our community a harassment-free experience for everyone, regardless
-of age, body size, visible or invisible disability, ethnicity, sex
-characteristics, gender identity and expression, level of experience,
-education, socio-economic status, nationality, personal appearance,
-race, religion, or sexual identity and orientation.
-
-We pledge to act and interact in ways that contribute to an open,
-welcoming, diverse, inclusive, and healthy community.
-
-## Our Standards
-
-Examples of behavior that contributes to a positive environment for
-our community include:
-
-- Demonstrating empathy and kindness toward other people
-- Being respectful of differing opinions, viewpoints, and experiences
-- Giving and gracefully accepting constructive feedback
-- Accepting responsibility and apologizing to those affected by our
-  mistakes, and learning from the experience
-- Focusing on what is best not just for us as individuals, but for the
-  overall community
-
-Examples of unacceptable behavior include:
-
-- The use of sexualized language or imagery, and sexual attention or
-  advances of any kind
-- Trolling, insulting or derogatory comments, and personal or
-  political attacks
-- Public or private harassment
-- Publishing others' private information, such as a physical or email
-  address, without their explicit permission
-- Other conduct which could reasonably be considered inappropriate in
-  a professional setting
-
-## Enforcement Responsibilities
-
-Community leaders are responsible for clarifying and enforcing our
-standards of acceptable behavior and will take appropriate and fair
-corrective action in response to any behavior that they deem
-inappropriate, threatening, offensive, or harmful.
-
-Community leaders have the right and responsibility to remove, edit,
-or reject comments, commits, code, wiki edits, issues, and other
-contributions that are not aligned to this Code of Conduct, and will
-communicate reasons for moderation decisions when appropriate.
-
-## Scope
-
-This Code of Conduct applies within all community spaces, and also
-applies when an individual is officially representing the community in
-public spaces. Examples of representing our community include using an
-official e-mail address, posting via an official social media account,
-or acting as an appointed representative at an online or offline
-event.
-
-## Enforcement
-
-Instances of abusive, harassing, or otherwise unacceptable behavior
-may be reported to the community leaders responsible for enforcement
-at our Discord channel. All complaints
-will be reviewed and investigated promptly and fairly.
-
-All community leaders are obligated to respect the privacy and
-security of the reporter of any incident.
-
-## Enforcement Guidelines
-
-Community leaders will follow these Community Impact Guidelines in
-determining the consequences for any action they deem in violation of
-this Code of Conduct:
-
-### 1. Correction
-
-**Community Impact**: Use of inappropriate language or other behavior
-deemed unprofessional or unwelcome in the community.
-
-**Consequence**: A private, written warning from community leaders,
-providing clarity around the nature of the violation and an
-explanation of why the behavior was inappropriate. A public apology
-may be requested.
-
-### 2. Warning
-
-**Community Impact**: A violation through a single incident or series
-of actions.
-
-**Consequence**: A warning with consequences for continued
-behavior. No interaction with the people involved, including
-unsolicited interaction with those enforcing the Code of Conduct, for
-a specified period of time. This includes avoiding interactions in
-community spaces as well as external channels like social
-media. Violating these terms may lead to a temporary or permanent ban.
-
-### 3. Temporary Ban
-
-**Community Impact**: A serious violation of community standards,
-including sustained inappropriate behavior.
-
-**Consequence**: A temporary ban from any sort of interaction or
-public communication with the community for a specified period of
-time. No public or private interaction with the people involved,
-including unsolicited interaction with those enforcing the Code of
-Conduct, is allowed during this period. Violating these terms may lead
-to a permanent ban.
-
-### 4. Permanent Ban
-
-**Community Impact**: Demonstrating a pattern of violation of
-community standards, including sustained inappropriate behavior,
-harassment of an individual, or aggression toward or disparagement of
-classes of individuals.
-
-**Consequence**: A permanent ban from any sort of public interaction
-within the community.
-
-## Attribution
-
-This Code of Conduct is adapted from the [Contributor
-Covenant][homepage], version 2.0, available at
-https://www.contributor-covenant.org/version/2/0/code_of_conduct.html.
-
-Community Impact Guidelines were inspired by [Mozilla's code of
-conduct enforcement ladder](https://github.com/mozilla/diversity).
-
-[homepage]: https://www.contributor-covenant.org
-
-For answers to common questions about this code of conduct, see the
-FAQ at https://www.contributor-covenant.org/faq. Translations are
-available at https://www.contributor-covenant.org/translations.

SECURITY.md

@@ -1,5 +0,0 @@
-# Security Policy
-
-## Reporting a Vulnerability
-
-Please report security issues to `security@immich.app`

machine-learning/Dockerfile

@@ -1,8 +1,8 @@
 ARG DEVICE=cpu
 
-FROM python:3.11-bookworm@sha256:970c99f886b839fc8829289040c1845dadaf2cae46b37acc7710333158ec29b4 AS builder-cpu
+FROM python:3.11-bookworm@sha256:121d86b6d08752968a7dddbc708849e5f3a839bbff47f32212b46d2a1d842bab AS builder-cpu
 
-FROM python:3.13-slim-trixie@sha256:d168b8d9eb761f4d3fe305ebd04aeb7e7f2de0297cec5fb2f8f6403244621664 AS builder-openvino
+FROM python:3.13-slim-trixie@sha256:b04b5d7233d2ad9c379e22ea8927cd1378cd15c60d4ef876c065b25ea8fb3bf3 AS builder-openvino
 
 FROM builder-cpu AS builder-cuda
 
@@ -39,12 +39,12 @@ RUN --mount=type=cache,target=/root/.cache/uv \
     --mount=type=bind,source=pyproject.toml,target=pyproject.toml \
     uv sync --frozen --extra ${DEVICE} --no-dev --no-editable --no-install-project --compile-bytecode --no-progress --active --link-mode copy
 
-FROM python:3.11-slim-bookworm@sha256:9c6f90801e6b68e772b7c0ca74260cbf7af9f320acec894e26fccdaccfbe3b47 AS prod-cpu
+FROM python:3.11-slim-bookworm@sha256:8dca233de9f3d9bb410665f00a4da6dd06f331083137e0e98ccf227236fcc438 AS prod-cpu
 
 ENV LD_PRELOAD=/usr/lib/libmimalloc.so.2 \
     MACHINE_LEARNING_MODEL_ARENA=false
 
-FROM python:3.13-slim-trixie@sha256:d168b8d9eb761f4d3fe305ebd04aeb7e7f2de0297cec5fb2f8f6403244621664 AS prod-openvino
+FROM python:3.13-slim-trixie@sha256:b04b5d7233d2ad9c379e22ea8927cd1378cd15c60d4ef876c065b25ea8fb3bf3 AS prod-openvino
 
 RUN apt-get update && \
     apt-get install --no-install-recommends -yqq ocl-icd-libopencl1 wget && \

machine-learning/immich_ml/__main__.py

@@ -49,6 +49,7 @@ try:
             str(settings.http_keepalive_timeout_s),
             "--graceful-timeout",
             "10",
+            "--no-control-socket",
         ],
     ) as cmd:
         cmd.wait()

machine-learning/immich_ml/main.py

@@ -12,7 +12,7 @@ from zipfile import BadZipFile
 
 import orjson
 from fastapi import Depends, FastAPI, File, Form, HTTPException
-from fastapi.responses import ORJSONResponse, PlainTextResponse
+from fastapi.responses import PlainTextResponse
 from onnxruntime.capi.onnxruntime_pybind11_state import InvalidProtobuf, NoSuchFile
 from PIL.Image import Image
 from pydantic import ValidationError
@@ -32,6 +32,7 @@ from .schemas import (
     ModelIdentity,
     ModelTask,
     ModelType,
+    ORJSONResponse,
     PipelineRequest,
     T,
 )

machine-learning/immich_ml/models/clip/textual.py

@@ -89,7 +89,9 @@ class OpenClipTextualEncoder(BaseCLIPTextualEncoder):
 
         tokenizer: Tokenizer = Tokenizer.from_file(self.tokenizer_file_path.as_posix())
 
-        pad_id: int = tokenizer.token_to_id(pad_token)
+        pad_id = tokenizer.token_to_id(pad_token)
+        if pad_id is None:
+            raise ValueError(f"Pad token '{pad_token}' not found in tokenizer vocab")
         tokenizer.enable_padding(length=context_length, pad_token=pad_token, pad_id=pad_id)
         tokenizer.enable_truncation(max_length=context_length)
 

machine-learning/immich_ml/schemas.py

@@ -3,9 +3,16 @@ from typing import Any, Literal, Protocol, TypeGuard, TypeVar
 
 import numpy as np
 import numpy.typing as npt
+import orjson
+from fastapi.responses import JSONResponse
 from typing_extensions import TypedDict
 
 
+class ORJSONResponse(JSONResponse):
+    def render(self, content: Any) -> bytes:
+        return orjson.dumps(content, option=orjson.OPT_SERIALIZE_NUMPY)
+
+
 class StrEnum(str, Enum):
     value: str
 

packages/plugin-sdk/package.json

@@ -31,7 +31,7 @@
     "@types/node": "^24.12.4",
     "esbuild": "^0.28.0",
     "tsc-alias": "^1.8.16",
-    "typescript": "^6.0.0"
+    "typescript": "^5.9.3"
   },
   "peerDependencies": {
     "@extism/js-pdk": "^1.1.1"

pnpm-lock.yaml

@@ -348,8 +348,8 @@ importers:
         specifier: ^1.8.16
         version: 1.8.17
       typescript:
-        specifier: ^6.0.0
-        version: 6.0.3
+        specifier: ^5.9.3
+        version: 5.9.3
 
   packages/sdk:
     dependencies:

server/Dockerfile

@@ -1,4 +1,4 @@
-FROM ghcr.io/immich-app/base-server-dev:202605121138@sha256:127cc323590e1d64d765016492e1de1e9355da8f658f078aef7be6bede0fdd0f AS builder
+FROM ghcr.io/immich-app/base-server-dev:202606021219@sha256:63fa91aa011f6f2921dd32fe6d1be8d637e9bd7f3e3dd0c8e446afb31b282af4 AS builder
 ENV COREPACK_ENABLE_DOWNLOAD_PROMPT=0 \
   CI=1 \
   COREPACK_HOME=/tmp \
@@ -80,7 +80,7 @@ RUN --mount=type=cache,id=pnpm-packages,target=/buildcache/pnpm-store \
   --mount=type=cache,id=mise-tools-${TARGETPLATFORM},target=/buildcache/mise \
   mise //:plugins
 
-FROM ghcr.io/immich-app/base-server-prod:202605121138@sha256:b346d42d86f42799fede6b6c9f6feed0018c5ee46e4ac31d1165f50737cee842
+FROM ghcr.io/immich-app/base-server-prod:202606021219@sha256:6ef9ef5859492149af770a6c884b5e2ddbaeef99f8885ea5f2d9f73625a3d9ec
 
 WORKDIR /usr/src/app
 ENV NODE_ENV=production \

server/Dockerfile.dev

@@ -1,5 +1,5 @@
 # dev build
-FROM ghcr.io/immich-app/base-server-dev:202605121138@sha256:127cc323590e1d64d765016492e1de1e9355da8f658f078aef7be6bede0fdd0f AS dev
+FROM ghcr.io/immich-app/base-server-dev:202606021219@sha256:63fa91aa011f6f2921dd32fe6d1be8d637e9bd7f3e3dd0c8e446afb31b282af4 AS dev
 
 
 COPY --from=ghcr.io/jdx/mise:2026.5.18@sha256:5bb3311994fa78cef307ca3077cdb18f9551da0886371fc26ea91ab56220ffc5 /usr/local/bin/mise /usr/local/bin/mise

server/src/schema/migrations/1780435471692-DeleteMismatchedAssetFaces.ts

@@ -0,0 +1,16 @@
+import { Kysely, sql } from 'kysely';
+
+export async function up(db: Kysely<any>): Promise<void> {
+  // Delete unauthorized cross-owner asset faces
+  await sql`
+    DELETE FROM asset_face
+    USING person, asset
+    WHERE asset_face."personId" = person.id
+      AND asset_face."assetId" = asset.id
+      AND person."ownerId" != asset."ownerId"
+  `.execute(db);
+}
+
+export async function down(): Promise<void> {
+  // Not implemented: the deleted rows were unauthorized cross-owner entries
+}

server/src/services/person.service.spec.ts

@@ -452,6 +452,30 @@ describe(PersonService.name, () => {
       expect(mocks.person.update).not.toHaveBeenCalled();
       expect(mocks.job.queueAll).not.toHaveBeenCalled();
     });
+
+    it('should reject creating a face on an asset the user does not own', async () => {
+      const auth = AuthFactory.create();
+      const asset = AssetFactory.create();
+      const person = PersonFactory.create({ faceAssetId: null });
+
+      mocks.access.asset.checkOwnerAccess.mockResolvedValue(new Set());
+      mocks.access.person.checkOwnerAccess.mockResolvedValue(new Set([person.id]));
+
+      await expect(
+        sut.createFace(auth, {
+          assetId: asset.id,
+          personId: person.id,
+          imageHeight: 500,
+          imageWidth: 400,
+          x: 10,
+          y: 20,
+          width: 100,
+          height: 110,
+        }),
+      ).rejects.toBeInstanceOf(BadRequestException);
+
+      expect(mocks.person.createAssetFace).not.toHaveBeenCalled();
+    });
   });
 
   describe('createNewFeaturePhoto', () => {

server/src/services/person.service.ts

@@ -625,7 +625,7 @@ export class PersonService extends BaseService {
   // TODO return a asset face response
   async createFace(auth: AuthDto, dto: AssetFaceCreateDto): Promise<void> {
     await Promise.all([
-      this.requireAccess({ auth, permission: Permission.AssetRead, ids: [dto.assetId] }),
+      this.requireAccess({ auth, permission: Permission.AssetUpdate, ids: [dto.assetId] }),
       this.requireAccess({ auth, permission: Permission.PersonRead, ids: [dto.personId] }),
     ]);
 

web/src/lib/components/asset-viewer/PhotoViewer.svelte

@@ -286,7 +286,11 @@
     {/snippet}
   </AdaptiveImage>
 
-  {#if assetViewerManager.isFaceEditMode && assetViewerManager.imgRef}
-    <FaceEditor htmlElement={assetViewerManager.imgRef} {containerWidth} {containerHeight} assetId={asset.id} />
+  {#if assetViewerManager.isFaceEditMode && assetViewerManager.imgRef && asset.width && asset.height}
+    <FaceEditor
+      imageSize={{ width: asset.width, height: asset.height }}
+      containerSize={{ width: containerWidth, height: containerHeight }}
+      assetId={asset.id}
+    />
   {/if}
 </div>

web/src/lib/components/asset-viewer/VideoNativeViewer.svelte

@@ -308,9 +308,40 @@
   let containerHeight = $state(0);
 
   $effect(() => {
-    if (assetViewerManager.isFaceEditMode) {
-      videoPlayer?.pause();
+    if (!assetViewerManager.isFaceEditMode || !videoPlayer) {
+      return;
     }
+    videoPlayer.pause();
+
+    const { videoWidth, videoHeight } = videoPlayer;
+    if (videoWidth === 0 || videoHeight === 0) {
+      return;
+    }
+
+    const canvas = document.createElement('canvas');
+    canvas.width = videoWidth;
+    canvas.height = videoHeight;
+    const context = canvas.getContext('2d');
+    if (!context) {
+      return;
+    }
+
+    context.drawImage(videoPlayer, 0, 0);
+    const dataUrl = canvas.toDataURL('image/png');
+    canvas.width = 0;
+
+    const img = new Image();
+    const onLoad = () => {
+      assetViewerManager.imgRef = img;
+    };
+    img.addEventListener('load', onLoad);
+    img.src = dataUrl;
+
+    return () => {
+      img.removeEventListener('load', onLoad);
+      img.src = '';
+      assetViewerManager.imgRef = undefined;
+    };
   });
 
   // The time is only refreshed on HLS fragment decode by default,
@@ -454,8 +485,12 @@
         </div>
       {/if}
 
-      {#if assetViewerManager.isFaceEditMode && videoPlayer}
-        <FaceEditor htmlElement={videoPlayer} {containerWidth} {containerHeight} {assetId} />
+      {#if assetViewerManager.isFaceEditMode}
+        <FaceEditor
+          imageSize={{ width: asset.width ?? 0, height: asset.height ?? 0 }}
+          containerSize={{ width: containerWidth, height: containerHeight }}
+          {assetId}
+        />
       {/if}
     {/if}
   </div>

web/src/lib/components/asset-viewer/face-editor/FaceEditor.svelte

@@ -4,7 +4,7 @@
   import { assetViewerManager } from '$lib/managers/asset-viewer-manager.svelte';
   import FaceCreateTagModal from '$lib/modals/CreateFaceModal.svelte';
   import { getPeopleThumbnailUrl } from '$lib/utils';
-  import { getNaturalSize, scaleToFit } from '$lib/utils/container-utils';
+  import { computeContentMetrics, mapContentRectToNatural, type Size } from '$lib/utils/container-utils';
   import { handleError } from '$lib/utils/handle-error';
   import { createFace, getAllPeople, type PersonResponseDto } from '@immich/sdk';
   import { Button, Input, modalManager, toastManager } from '@immich/ui';
@@ -14,13 +14,12 @@
   import { t } from 'svelte-i18n';
 
   type Props = {
-    htmlElement: HTMLImageElement | HTMLVideoElement;
-    containerWidth: number;
-    containerHeight: number;
+    imageSize: Size;
+    containerSize: Size;
     assetId: string;
   };
 
-  let { htmlElement, containerWidth, containerHeight, assetId }: Props = $props();
+  let { imageSize, containerSize, assetId }: Props = $props();
 
   let canvasEl: HTMLCanvasElement | undefined = $state();
   let canvas: Canvas | undefined = $state();
@@ -54,7 +53,7 @@
   };
 
   const setupCanvas = () => {
-    if (!canvasEl || !htmlElement) {
+    if (!canvasEl) {
       return;
     }
 
@@ -86,24 +85,14 @@
     searchInputEl?.focus();
   });
 
-  const imageContentMetrics = $derived.by(() => {
-    const natural = getNaturalSize(htmlElement);
-    const container = { width: containerWidth, height: containerHeight };
-    const { width: contentWidth, height: contentHeight } = scaleToFit(natural, container);
-    return {
-      contentWidth,
-      contentHeight,
-      offsetX: (containerWidth - contentWidth) / 2,
-      offsetY: (containerHeight - contentHeight) / 2,
-    };
-  });
+  const imageContentMetrics = $derived(computeContentMetrics(imageSize, containerSize));
 
   const setDefaultFaceRectanglePosition = (faceRect: Rect) => {
-    const { offsetX, offsetY } = imageContentMetrics;
+    const { offsetX, offsetY, contentWidth, contentHeight } = imageContentMetrics;
 
     faceRect.set({
-      top: offsetY + 200,
-      left: offsetX + 200,
+      top: offsetY + contentHeight / 2 - 56,
+      left: offsetX + contentWidth / 2 - 56,
     });
 
     faceRect.setCoords();
@@ -116,8 +105,8 @@
     }
 
     canvas.setDimensions({
-      width: containerWidth,
-      height: containerHeight,
+      width: containerSize.width,
+      height: containerSize.height,
     });
 
     if (!faceRect) {
@@ -167,6 +156,9 @@
     const gap = 15;
     const padding = faceRect.padding ?? 0;
     const rawBox = faceRect.getBoundingRect();
+    if (Number.isNaN(rawBox.left) || Number.isNaN(rawBox.width)) {
+      return;
+    }
     const faceBox = {
       left: rawBox.left - padding,
       top: rawBox.top - padding,
@@ -175,11 +167,11 @@
     };
     const selectorWidth = faceSelectorEl.offsetWidth;
     const chromeHeight = faceSelectorEl.offsetHeight - scrollableListEl.offsetHeight;
-    const listHeight = Math.min(MAX_LIST_HEIGHT, containerHeight - gap * 2 - chromeHeight);
+    const listHeight = Math.min(MAX_LIST_HEIGHT, containerSize.height - gap * 2 - chromeHeight);
     const selectorHeight = listHeight + chromeHeight;
 
-    const clampTop = (top: number) => clamp(top, gap, containerHeight - selectorHeight - gap);
-    const clampLeft = (left: number) => clamp(left, gap, containerWidth - selectorWidth - gap);
+    const clampTop = (top: number) => clamp(top, gap, containerSize.height - selectorHeight - gap);
+    const clampLeft = (left: number) => clamp(left, gap, containerSize.width - selectorWidth - gap);
 
     const overlapArea = (position: { top: number; left: number }) => {
       const selectorRight = position.left + selectorWidth;
@@ -238,45 +230,37 @@
   });
 
   const getFaceCroppedCoordinates = () => {
-    if (!faceRect || !htmlElement) {
+    if (!faceRect || imageContentMetrics.contentWidth === 0) {
       return;
     }
 
-    const { left, top, width, height } = faceRect.getBoundingRect();
-    const { offsetX, offsetY, contentWidth, contentHeight } = imageContentMetrics;
-    const natural = getNaturalSize(htmlElement);
-
-    const scaleX = natural.width / contentWidth;
-    const scaleY = natural.height / contentHeight;
-    const imageX = (left - offsetX) * scaleX;
-    const imageY = (top - offsetY) * scaleY;
+    const imageRect = mapContentRectToNatural(faceRect.getBoundingRect(), imageContentMetrics, imageSize);
 
     return {
-      imageWidth: natural.width,
-      imageHeight: natural.height,
-      x: Math.floor(imageX),
-      y: Math.floor(imageY),
-      width: Math.floor(width * scaleX),
-      height: Math.floor(height * scaleY),
+      imageWidth: imageSize.width,
+      imageHeight: imageSize.height,
+      x: Math.floor(imageRect.left),
+      y: Math.floor(imageRect.top),
+      width: Math.floor(imageRect.width),
+      height: Math.floor(imageRect.height),
     };
   };
 
   type FaceCoordinates = NonNullable<ReturnType<typeof getFaceCroppedCoordinates>>;
 
   const getFacePreviewUrl = (data: FaceCoordinates) => {
-    if (!htmlElement) {
+    const imgRef = assetViewerManager.imgRef;
+    if (!imgRef || imageContentMetrics.contentWidth === 0) {
       return;
     }
 
-    const natural = getNaturalSize(htmlElement);
-    if (natural.width <= 0 || natural.height <= 0) {
-      return;
-    }
+    const scaleX = imgRef.naturalWidth / imageSize.width;
+    const scaleY = imgRef.naturalHeight / imageSize.height;
 
-    const x = clamp(data.x, 0, natural.width - 1);
-    const y = clamp(data.y, 0, natural.height - 1);
-    const width = clamp(data.width, 1, natural.width - x);
-    const height = clamp(data.height, 1, natural.height - y);
+    const x = clamp(Math.floor(data.x * scaleX), 0, imgRef.naturalWidth - 1);
+    const y = clamp(Math.floor(data.y * scaleY), 0, imgRef.naturalHeight - 1);
+    const width = clamp(Math.floor(data.width * scaleX), 1, imgRef.naturalWidth - x);
+    const height = clamp(Math.floor(data.height * scaleY), 1, imgRef.naturalHeight - y);
 
     if (width <= 0 || height <= 0) {
       return;
@@ -292,7 +276,7 @@
     }
 
     try {
-      context.drawImage(htmlElement, x, y, width, height, 0, 0, width, height);
+      context.drawImage(imgRef, x, y, width, height, 0, 0, width, height);
       return canvas.toDataURL('image/png');
     } catch {
       return;

web/src/lib/utils/container-utils.spec.ts

@@ -1,18 +1,15 @@
 import {
-  getContentMetrics,
+  computeContentMetrics,
   getNaturalSize,
+  mapContentRectToNatural,
   mapNormalizedRectToContent,
   mapNormalizedToContent,
   scaleToCover,
   scaleToFit,
 } from '$lib/utils/container-utils';
 
-const mockImage = (props: {
-  naturalWidth: number;
-  naturalHeight: number;
-  width: number;
-  height: number;
-}): HTMLImageElement => props as unknown as HTMLImageElement;
+const mockImage = (props: { naturalWidth: number; naturalHeight: number }): HTMLImageElement =>
+  props as unknown as HTMLImageElement;
 
 const mockVideo = (props: {
   videoWidth: number;
@@ -49,48 +46,85 @@ describe('scaleToFit', () => {
   });
 });
 
-describe('getContentMetrics', () => {
-  it('should compute zero offsets when aspect ratios match', () => {
-    const img = mockImage({ naturalWidth: 1600, naturalHeight: 900, width: 800, height: 450 });
-    expect(getContentMetrics(img)).toEqual({
+describe('computeContentMetrics', () => {
+  it('should return zero metrics for zero-width content', () => {
+    expect(computeContentMetrics({ width: 0, height: 1080 }, { width: 800, height: 600 })).toEqual({
+      contentWidth: 0,
+      contentHeight: 0,
+      offsetX: 0,
+      offsetY: 0,
+    });
+  });
+
+  it('should return zero metrics for zero-height content', () => {
+    expect(computeContentMetrics({ width: 1920, height: 0 }, { width: 800, height: 600 })).toEqual({
+      contentWidth: 0,
+      contentHeight: 0,
+      offsetX: 0,
+      offsetY: 0,
+    });
+  });
+
+  it('should center wide content vertically', () => {
+    expect(computeContentMetrics({ width: 2000, height: 1000 }, { width: 800, height: 600 })).toEqual({
+      contentWidth: 800,
+      contentHeight: 400,
+      offsetX: 0,
+      offsetY: 100,
+    });
+  });
+
+  it('should center tall content horizontally', () => {
+    expect(computeContentMetrics({ width: 1000, height: 2000 }, { width: 800, height: 600 })).toEqual({
+      contentWidth: 300,
+      contentHeight: 600,
+      offsetX: 250,
+      offsetY: 0,
+    });
+  });
+
+  it('should produce zero offsets when aspect ratios match', () => {
+    expect(computeContentMetrics({ width: 1600, height: 900 }, { width: 800, height: 450 })).toEqual({
       contentWidth: 800,
       contentHeight: 450,
       offsetX: 0,
       offsetY: 0,
     });
   });
+});
 
-  it('should compute horizontal letterbox offsets for tall image', () => {
-    const img = mockImage({ naturalWidth: 1000, naturalHeight: 2000, width: 800, height: 600 });
-    const metrics = getContentMetrics(img);
-    expect(metrics.contentWidth).toBe(300);
-    expect(metrics.contentHeight).toBe(600);
-    expect(metrics.offsetX).toBe(250);
-    expect(metrics.offsetY).toBe(0);
+describe('mapContentRectToNatural', () => {
+  it('should map a full-content rect back to natural size', () => {
+    const metrics = { contentWidth: 800, contentHeight: 400, offsetX: 0, offsetY: 100 };
+    const rect = mapContentRectToNatural({ left: 0, top: 100, width: 800, height: 400 }, metrics, {
+      width: 2000,
+      height: 1000,
+    });
+    expect(rect).toEqual({ left: 0, top: 0, width: 2000, height: 1000 });
   });
 
-  it('should compute vertical letterbox offsets for wide image', () => {
-    const img = mockImage({ naturalWidth: 2000, naturalHeight: 1000, width: 800, height: 600 });
-    const metrics = getContentMetrics(img);
-    expect(metrics.contentWidth).toBe(800);
-    expect(metrics.contentHeight).toBe(400);
-    expect(metrics.offsetX).toBe(0);
-    expect(metrics.offsetY).toBe(100);
+  it('should map a centered sub-rect to natural coordinates', () => {
+    const metrics = { contentWidth: 800, contentHeight: 400, offsetX: 0, offsetY: 100 };
+    const rect = mapContentRectToNatural({ left: 200, top: 200, width: 400, height: 200 }, metrics, {
+      width: 2000,
+      height: 1000,
+    });
+    expect(rect).toEqual({ left: 500, top: 250, width: 1000, height: 500 });
   });
 
-  it('should use clientWidth/clientHeight for video elements', () => {
-    const video = mockVideo({ videoWidth: 1920, videoHeight: 1080, clientWidth: 800, clientHeight: 600 });
-    const metrics = getContentMetrics(video);
-    expect(metrics.contentWidth).toBe(800);
-    expect(metrics.contentHeight).toBe(450);
-    expect(metrics.offsetX).toBe(0);
-    expect(metrics.offsetY).toBe(75);
+  it('should handle letterboxed content with horizontal offset', () => {
+    const metrics = { contentWidth: 300, contentHeight: 600, offsetX: 250, offsetY: 0 };
+    const rect = mapContentRectToNatural({ left: 250, top: 0, width: 300, height: 600 }, metrics, {
+      width: 1000,
+      height: 2000,
+    });
+    expect(rect).toEqual({ left: 0, top: 0, width: 1000, height: 2000 });
   });
 });
 
 describe('getNaturalSize', () => {
   it('should return naturalWidth/naturalHeight for images', () => {
-    const img = mockImage({ naturalWidth: 4000, naturalHeight: 3000, width: 800, height: 600 });
+    const img = mockImage({ naturalWidth: 4000, naturalHeight: 3000 });
     expect(getNaturalSize(img)).toEqual({ width: 4000, height: 3000 });
   });
 

web/src/lib/utils/container-utils.ts

@@ -49,13 +49,6 @@ export const scaleToFit = (dimensions: Size, container: Size): Size => {
   };
 };
 
-const getElementSize = (element: HTMLImageElement | HTMLVideoElement): Size => {
-  if (element instanceof HTMLVideoElement) {
-    return { width: element.clientWidth, height: element.clientHeight };
-  }
-  return { width: element.width, height: element.height };
-};
-
 export const getNaturalSize = (element: HTMLImageElement | HTMLVideoElement): Size => {
   if (element instanceof HTMLVideoElement) {
     return { width: element.videoWidth, height: element.videoHeight };
@@ -63,17 +56,18 @@ export const getNaturalSize = (element: HTMLImageElement | HTMLVideoElement): Si
   return { width: element.naturalWidth, height: element.naturalHeight };
 };
 
-export const getContentMetrics = (element: HTMLImageElement | HTMLVideoElement): ContentMetrics => {
-  const natural = getNaturalSize(element);
-  const client = getElementSize(element);
-  const { width: contentWidth, height: contentHeight } = scaleToFit(natural, client);
+export function computeContentMetrics(content: Size, container: Size): ContentMetrics {
+  if (content.width === 0 || content.height === 0) {
+    return { contentWidth: 0, contentHeight: 0, offsetX: 0, offsetY: 0 };
+  }
+  const { width: contentWidth, height: contentHeight } = scaleToFit(content, container);
   return {
     contentWidth,
     contentHeight,
-    offsetX: (client.width - contentWidth) / 2,
-    offsetY: (client.height - contentHeight) / 2,
+    offsetX: (container.width - contentWidth) / 2,
+    offsetY: (container.height - contentHeight) / 2,
   };
-};
+}
 
 export function mapNormalizedToContent(point: Point, sizeOrMetrics: Size | ContentMetrics): Point {
   if ('contentWidth' in sizeOrMetrics) {
@@ -109,3 +103,25 @@ export function mapNormalizedRectToContent(
     height: br.y - tl.y,
   };
 }
+
+function mapContentToNatural(point: Point, metrics: ContentMetrics, naturalSize: Size): Point {
+  return {
+    x: ((point.x - metrics.offsetX) / metrics.contentWidth) * naturalSize.width,
+    y: ((point.y - metrics.offsetY) / metrics.contentHeight) * naturalSize.height,
+  };
+}
+
+export function mapContentRectToNatural(rect: Rect, metrics: ContentMetrics, naturalSize: Size): Rect {
+  const topLeft = mapContentToNatural({ x: rect.left, y: rect.top }, metrics, naturalSize);
+  const bottomRight = mapContentToNatural(
+    { x: rect.left + rect.width, y: rect.top + rect.height },
+    metrics,
+    naturalSize,
+  );
+  return {
+    top: topLeft.y,
+    left: topLeft.x,
+    width: bottomRight.x - topLeft.x,
+    height: bottomRight.y - topLeft.y,
+  };
+}