feat(subtitles): add local Japanese subtitle generation

- Generate Japanese SRTs through the overlay modal or launcher
- Support managed whisper.cpp models and optional dialogue detection
This commit is contained in:
2026-09-07 21:43:53 -07:00
parent c14c690875
commit 2d623838d5
82 changed files with 4001 additions and 54 deletions
@@ -5,6 +5,7 @@ type ControllerInteractionModalState = {
kikuModalOpen: boolean;
runtimeOptionsModalOpen: boolean;
subsyncModalOpen: boolean;
subtitleGenerationModalOpen?: boolean;
youtubePickerModalOpen: boolean;
sessionHelpModalOpen: boolean;
subtitleSidebarModalOpen: boolean;
@@ -18,6 +19,7 @@ export function isControllerInteractionBlocked(state: ControllerInteractionModal
state.kikuModalOpen ||
state.runtimeOptionsModalOpen ||
state.subsyncModalOpen ||
Boolean(state.subtitleGenerationModalOpen) ||
state.youtubePickerModalOpen ||
state.sessionHelpModalOpen
);
+23
View File
@@ -91,6 +91,7 @@ function createEmptyShortcuts(): ConfiguredShortcuts {
openRuntimeOptions: null,
openJimaku: null,
openTsukihime: null,
openSubtitleGeneration: null,
openSessionHelp: null,
openControllerSelect: null,
openControllerDebug: null,
@@ -1580,6 +1581,28 @@ test('session binding: Ctrl+Alt+S dispatches subsync action locally', async () =
}
});
test('session binding: Ctrl+Shift+G dispatches subtitle generation with the sidebar closed', async () => {
const { handlers, testGlobals } = createKeyboardHandlerHarness();
try {
await handlers.setupMpvInputForwarding();
handlers.updateSessionBindings([
{
sourcePath: 'shortcuts.openSubtitleGeneration',
originalKey: 'Ctrl+Shift+G',
key: { code: 'KeyG', modifiers: ['ctrl', 'shift'] },
actionType: 'session-action',
actionId: 'openSubtitleGeneration',
},
]);
testGlobals.dispatchKeydown({ key: 'G', code: 'KeyG', ctrlKey: true, shiftKey: true });
assert.deepEqual(testGlobals.sessionActions, [
{ actionId: 'openSubtitleGeneration', payload: undefined },
]);
} finally {
testGlobals.restore();
}
});
test('session binding: Ctrl+Shift+J dispatches jimaku action locally', async () => {
const { handlers, testGlobals } = createKeyboardHandlerHarness();
+6
View File
@@ -14,6 +14,7 @@ export function createKeyboardHandlers(
handleRuntimeOptionsKeydown: (e: KeyboardEvent) => boolean;
handleCharacterDictionaryKeydown: (e: KeyboardEvent) => boolean;
handleSubsyncKeydown: (e: KeyboardEvent) => boolean;
handleSubtitleGenerationKeydown?: (e: KeyboardEvent) => boolean;
handleKikuKeydown: (e: KeyboardEvent) => boolean;
handleJimakuKeydown: (e: KeyboardEvent) => boolean;
handleTsukihimeKeydown: (e: KeyboardEvent) => boolean;
@@ -1079,6 +1080,11 @@ export function createKeyboardHandlers(
);
document.addEventListener('keydown', (e: KeyboardEvent) => {
if (ctx.state.subtitleGenerationModalOpen) {
options.handleSubtitleGenerationKeydown?.(e);
return;
}
if (ctx.state.mediaTimingReviewModalOpen) {
options.handleMediaTimingReviewKeydown(e);
return;
+127
View File
@@ -634,6 +634,126 @@
</div>
</div>
</div>
<div
id="subtitleGenerationModal"
class="modal hidden"
aria-hidden="true"
role="dialog"
aria-modal="true"
aria-labelledby="subtitleGenerationTitle"
>
<div class="modal-content subtitle-generation-content">
<div class="modal-header">
<div id="subtitleGenerationTitle" class="modal-title">Generate Japanese subtitles</div>
<button id="subtitleGenerationClose" class="modal-close" type="button">Close</button>
</div>
<div class="modal-body subtitle-generation-body">
<p class="subtitle-generation-intro">
Turn the current audio track into timed Japanese subtitles. Audio stays on this
device.
</p>
<div class="subtitle-generation-detail">
<span class="subtitle-generation-label">Current media</span>
<div id="subtitleGenerationMedia">Checking current media...</div>
</div>
<div class="subtitle-generation-detail">
<span class="subtitle-generation-label">Speech model</span>
<div id="subtitleGenerationModel">Checking local models...</div>
<div
id="subtitleGenerationModelPicker"
class="subtitle-generation-model-picker hidden"
>
<label for="subtitleGenerationModelSelect">Model</label>
<select
id="subtitleGenerationModelSelect"
aria-describedby="subtitleGenerationModelDescription subtitleGenerationModelRecommendation"
></select>
<p id="subtitleGenerationModelDescription" class="subtitle-generation-hint"></p>
<p id="subtitleGenerationModelRecommendation" class="subtitle-generation-hint">
Start with small for a balance of accuracy and CPU time. Larger models need more
memory; speed depends on your hardware. Sizes shown are downloads.
</p>
<p class="subtitle-generation-hint">
This choice lasts for this SubMiner session. Set the default model in Settings.
</p>
</div>
<p class="subtitle-generation-hint">
Use your own model in Settings → Integrations → Japanese Subtitle Generation → Model
Path.
</p>
<button
id="subtitleGenerationDownload"
class="kiku-cancel-button hidden"
type="button"
>
Download model
</button>
</div>
<div class="subtitle-generation-detail">
<label class="subtitle-generation-vad-toggle" for="subtitleGenerationVadEnabled">
<input
id="subtitleGenerationVadEnabled"
type="checkbox"
aria-describedby="subtitleGenerationVadHint"
/>
Prioritize dialogue <span class="subtitle-generation-hint">Optional</span>
</label>
<div id="subtitleGenerationVadModel" class="subtitle-generation-hint"></div>
<p id="subtitleGenerationVadHint" class="subtitle-generation-hint">
Uses a small Silero model to transcribe separate speech passages. May skip songs,
quiet speech, and dialogue under music. Requires whisper.cpp's speech detector.
</p>
<p class="subtitle-generation-hint">
Applies for this session. Set VAD Model Path in Settings to enable it by default.
</p>
<button
id="subtitleGenerationVadDownload"
class="kiku-cancel-button hidden"
type="button"
>
Download Silero
</button>
</div>
<div id="subtitleGenerationActivity" class="subtitle-generation-activity hidden">
<div class="subtitle-generation-progress-heading">
<span id="subtitleGenerationStage">Preparing</span>
<span id="subtitleGenerationPercent"></span>
</div>
<progress
id="subtitleGenerationProgress"
max="100"
aria-label="Subtitle generation progress"
></progress>
</div>
<div
id="subtitleGenerationStatus"
class="runtime-options-status"
role="status"
aria-live="polite"
></div>
<div class="subtitle-generation-actions">
<button id="subtitleGenerationRefresh" class="kiku-cancel-button" type="button">
Check again
</button>
<button id="subtitleGenerationCancel" class="kiku-cancel-button hidden" type="button">
Cancel
</button>
<button
id="subtitleGenerationStart"
class="kiku-confirm-button"
type="button"
disabled
>
Generate subtitles
</button>
</div>
<p class="subtitle-generation-hint">
The generated SRT will be saved locally and loaded into the player. You can close this
window while it runs.
</p>
</div>
</div>
</div>
<div id="subsyncModal" class="modal hidden" aria-hidden="true">
<div class="modal-content subsync-modal-content">
<div class="modal-header">
@@ -734,6 +854,13 @@
<button id="subtitleSidebarClose" class="modal-close" type="button">Close</button>
</div>
<div class="modal-body subtitle-sidebar-body">
<button
id="subtitleGenerationOpen"
class="kiku-cancel-button subtitle-generation-open"
type="button"
>
Generate Japanese subtitles
</button>
<div id="subtitleSidebarStatus" class="runtime-options-status"></div>
<ul id="subtitleSidebarList" class="subtitle-sidebar-list"></ul>
</div>
@@ -225,6 +225,8 @@ function describeSessionAction(
return 'Open jimaku';
case 'openTsukihime':
return 'Open TsukiHime';
case 'openSubtitleGeneration':
return 'Generate Japanese subtitles';
case 'openYoutubePicker':
return 'Open YouTube subtitle picker';
case 'openPlaylistBrowser':
@@ -266,6 +268,7 @@ function sectionForSessionBinding(binding: CompiledSessionBinding): string {
case 'openJimaku':
case 'openTsukihime':
case 'openCharacterDictionaryManager':
case 'openSubtitleGeneration':
case 'openControllerSelect':
case 'openControllerDebug':
case 'openYoutubePicker':
@@ -0,0 +1,75 @@
import assert from 'node:assert/strict';
import test from 'node:test';
import {
describeGenerationModel,
describeGenerationProgress,
describeGenerationVad,
} from './subtitle-generation-view';
test('only a missing managed model offers a download', () => {
assert.deepEqual(describeGenerationModel({ kind: 'missing', path: '/models/small.bin' }), {
ready: false,
download: true,
text: 'Download a speech model to get started.',
});
for (const kind of ['managed', 'external'] as const) {
const model = describeGenerationModel({ kind, path: '/models/ggml-small.bin' });
assert.equal(model.ready, true);
assert.equal(model.download, false);
}
assert.deepEqual(
describeGenerationModel({
kind: 'invalid',
path: '/missing/model.bin',
message: 'Configured model does not exist.',
}),
{
ready: false,
download: false,
text: 'Configured model does not exist.',
},
);
});
test('generation progress distinguishes measured work from indeterminate stages', () => {
assert.deepEqual(
describeGenerationProgress({ stage: 'transcribe', percent: 42.8, message: 'Transcribing' }),
{
stage: 'Recognizing Japanese speech',
percent: 42.8,
label: '42%',
},
);
assert.deepEqual(describeGenerationProgress({ stage: 'extract', message: 'Extracting audio' }), {
stage: 'Preparing audio',
percent: null,
label: 'Working...',
});
assert.equal(
describeGenerationProgress({ stage: 'download', percent: 0, message: '' }).label,
'0%',
);
assert.equal(
describeGenerationProgress({ stage: 'download', percent: Number.NaN, message: '' }).percent,
null,
);
assert.equal(
describeGenerationProgress({ stage: 'write', percent: 120, message: '' }).percent,
100,
);
});
test('optional speech detection only gates generation when selected', () => {
const missing = { kind: 'missing', path: '/vad.bin' } as const;
assert.equal(describeGenerationVad({ enabled: false, model: missing }).ready, true);
assert.equal(describeGenerationVad({ enabled: false, model: missing }).download, false);
assert.equal(describeGenerationVad({ enabled: true, model: missing }).ready, false);
assert.equal(describeGenerationVad({ enabled: true, model: missing }).download, true);
assert.equal(
describeGenerationVad({ enabled: true, model: { kind: 'managed', path: '/vad.bin' } }).ready,
true,
);
const invalid = { kind: 'invalid', path: '/vad.bin', message: 'Cannot read model' } as const;
assert.equal(describeGenerationVad({ enabled: true, model: invalid }).download, false);
assert.equal(describeGenerationVad({ enabled: false, model: invalid }).ready, true);
});
@@ -0,0 +1,57 @@
import type {
SubtitleGenerationModelStatus,
SubtitleGenerationProgress,
} from '../../shared/subtitle-generation';
import type { SubtitleGenerationStatus } from '../../shared/subtitle-generation-ipc';
export function describeGenerationVad(vad: SubtitleGenerationStatus['vad']) {
const model = describeGenerationModel(vad.model);
return {
ready: !vad.enabled || model.ready,
download: vad.enabled && model.download,
text: !vad.enabled
? 'Optional. Generate from the full audio when unchecked.'
: vad.model.kind === 'missing'
? 'Download Silero to prioritize spoken dialogue.'
: vad.model.kind === 'invalid'
? vad.model.message
: vad.model.kind === 'external'
? `Your speech detection model: ${vad.model.path}`
: 'Silero speech detection model installed.',
};
}
export function describeGenerationModel(model: SubtitleGenerationModelStatus) {
switch (model.kind) {
case 'external':
return { ready: true, download: false, text: `Your model: ${model.path}` };
case 'managed':
return { ready: true, download: false, text: `SubMiner model: ${model.path}` };
case 'missing':
return { ready: false, download: true, text: 'Download a speech model to get started.' };
case 'invalid':
return { ready: false, download: false, text: model.message };
default: {
const exhaustive: never = model;
return exhaustive;
}
}
}
const STAGE_LABELS = {
download: 'Downloading speech model',
extract: 'Preparing audio',
transcribe: 'Recognizing Japanese speech',
write: 'Saving subtitles',
} satisfies Record<SubtitleGenerationProgress['stage'], string>;
export function describeGenerationProgress(progress: SubtitleGenerationProgress | null) {
const raw = progress?.percent;
const percent =
raw !== undefined && Number.isFinite(raw) ? Math.max(0, Math.min(100, raw)) : null;
return {
stage: progress ? STAGE_LABELS[progress.stage] : 'Preparing',
percent,
label: percent === null ? 'Working...' : `${Math.floor(percent)}%`,
};
}
+324
View File
@@ -0,0 +1,324 @@
import type { SubtitleGenerationProgress } from '../../shared/subtitle-generation';
import {
SUBTITLE_GENERATION_MODELS,
RECOMMENDED_SUBTITLE_GENERATION_MODEL,
formatSubtitleGenerationModelSize,
getSubtitleGenerationModel,
isSubtitleGenerationModelId,
} from '../../shared/subtitle-generation-model-catalog';
import type {
SubtitleGenerationResult,
SubtitleGenerationStatus,
} from '../../shared/subtitle-generation-ipc';
import type { ModalStateReader, RendererContext } from '../context';
import { syncOverlayMouseIgnoreState } from '../overlay-mouse-ignore';
import { createModalFocusGuard } from './modal-focus-guard';
import {
describeGenerationModel,
describeGenerationProgress,
describeGenerationVad,
} from './subtitle-generation-view';
import { SUBTITLE_GENERATION_VAD_MODEL } from '../../shared/subtitle-generation-vad-model';
function element<T extends HTMLElement>(id: string, constructor: new () => T): T {
const node = document.getElementById(id);
if (!(node instanceof constructor)) throw new Error(`Missing subtitle generation element: ${id}`);
return node;
}
export function createSubtitleGenerationModal(
ctx: RendererContext,
options: {
modalStateReader: Pick<ModalStateReader, 'isAnyModalOpen'>;
syncSettingsModalSubtitleSuppression: () => void;
},
) {
const dom = {
modal: element('subtitleGenerationModal', HTMLDivElement),
close: element('subtitleGenerationClose', HTMLButtonElement),
open: element('subtitleGenerationOpen', HTMLButtonElement),
media: element('subtitleGenerationMedia', HTMLDivElement),
model: element('subtitleGenerationModel', HTMLDivElement),
modelPicker: element('subtitleGenerationModelPicker', HTMLDivElement),
modelSelect: element('subtitleGenerationModelSelect', HTMLSelectElement),
modelDescription: element('subtitleGenerationModelDescription', HTMLParagraphElement),
download: element('subtitleGenerationDownload', HTMLButtonElement),
vadEnabled: element('subtitleGenerationVadEnabled', HTMLInputElement),
vadModel: element('subtitleGenerationVadModel', HTMLDivElement),
vadDownload: element('subtitleGenerationVadDownload', HTMLButtonElement),
activity: element('subtitleGenerationActivity', HTMLDivElement),
stage: element('subtitleGenerationStage', HTMLSpanElement),
percent: element('subtitleGenerationPercent', HTMLSpanElement),
progress: element('subtitleGenerationProgress', HTMLProgressElement),
status: element('subtitleGenerationStatus', HTMLDivElement),
refresh: element('subtitleGenerationRefresh', HTMLButtonElement),
cancel: element('subtitleGenerationCancel', HTMLButtonElement),
start: element('subtitleGenerationStart', HTMLButtonElement),
};
let snapshot: SubtitleGenerationStatus | null = null;
let progress: SubtitleGenerationProgress | null = null;
let result: SubtitleGenerationResult | null = null;
let pending = false;
let cancelling = false;
let checking = false;
let error: string | null = null;
let priorFocus: Element | null = null;
let poll: ReturnType<typeof setTimeout> | null = null;
let unsubscribe: (() => void) | null = null;
for (const model of SUBTITLE_GENERATION_MODELS) {
const option = document.createElement('option');
option.value = model.id;
const recommended = model.id === RECOMMENDED_SUBTITLE_GENERATION_MODEL ? ' (recommended)' : '';
option.textContent = `${model.id}${recommended} · ${formatSubtitleGenerationModelSize(model.size)}`;
dom.modelSelect.append(option);
}
const focus = createModalFocusGuard({
isOpen: () => ctx.state.subtitleGenerationModalOpen,
getModalRoot: () => dom.modal,
getPreferredFocusTargets: () => [dom.modelSelect, dom.download, dom.start],
getFallbackFocusTarget: () => dom.close,
isModalLayer: ctx.platform.isModalLayer,
});
function render(): void {
const busy = pending || Boolean(snapshot?.running);
const model = snapshot ? describeGenerationModel(snapshot.model) : null;
const vad = snapshot ? describeGenerationVad(snapshot.vad) : null;
const readyMessage = !snapshot?.mediaPath
? 'Open local media to generate subtitles.'
: !model?.ready
? 'Set up a speech model to continue.'
: !vad?.ready
? 'Download the speech detection model or uncheck Prioritize dialogue.'
: 'Ready when you are.';
dom.media.textContent = snapshot?.mediaPath ?? 'Open a local media file in the player first.';
dom.model.textContent = model?.text ?? 'Checking local models...';
dom.modelPicker.classList.toggle('hidden', !snapshot || Boolean(snapshot.externalModelPath));
dom.modelSelect.disabled = busy || checking || !snapshot || Boolean(snapshot.externalModelPath);
if (snapshot) {
dom.modelSelect.value = snapshot.managedModel;
dom.modelDescription.textContent = getSubtitleGenerationModel(
snapshot.managedModel,
).description;
}
dom.download.classList.toggle('hidden', !model?.download);
dom.download.textContent = snapshot
? `Download ${snapshot.managedModel} model`
: 'Download model';
dom.download.disabled = busy || checking;
if (!checking) dom.vadEnabled.checked = snapshot?.vad.enabled ?? false;
dom.vadEnabled.disabled = busy || checking || !snapshot;
dom.vadModel.textContent = vad?.text ?? 'Checking speech detection...';
dom.vadDownload.classList.toggle('hidden', !vad?.download);
dom.vadDownload.textContent = `Download Silero · ${formatSubtitleGenerationModelSize(SUBTITLE_GENERATION_VAD_MODEL.size)}`;
dom.vadDownload.disabled = busy || checking;
dom.start.disabled = busy || checking || !model?.ready || !vad?.ready || !snapshot?.mediaPath;
dom.refresh.disabled = busy || checking;
dom.cancel.classList.toggle('hidden', !busy);
dom.cancel.disabled = cancelling;
dom.cancel.textContent = cancelling ? 'Cancelling...' : 'Cancel';
dom.activity.classList.toggle('hidden', !busy);
const activity = describeGenerationProgress(progress);
dom.stage.textContent = activity.stage;
dom.percent.textContent = activity.label;
if (activity.percent === null) dom.progress.removeAttribute('value');
else dom.progress.value = activity.percent;
dom.status.classList.toggle('error', Boolean(error || (!busy && result && !result.ok)));
dom.status.textContent =
error ??
(busy
? cancelling
? 'Stopping the current operation...'
: (progress?.message ?? 'Starting...')
: (result?.message ?? (checking ? 'Checking local setup...' : readyMessage)));
}
function stopPolling(): void {
if (poll !== null) clearTimeout(poll);
poll = null;
}
async function refresh(): Promise<void> {
if (checking) return;
checking = true;
error = null;
render();
try {
snapshot = await window.electronAPI.getSubtitleGenerationStatus();
if (!pending) {
progress = snapshot.progress;
result = snapshot.lastResult ?? result;
}
} catch (cause) {
error = cause instanceof Error ? cause.message : 'Could not check subtitle generation setup.';
} finally {
checking = false;
render();
stopPolling();
if (ctx.state.subtitleGenerationModalOpen && snapshot?.running && !pending) {
poll = setTimeout(() => void refresh(), 1500);
}
}
}
async function run(action: 'download' | 'download-vad' | 'generate'): Promise<void> {
if (pending || snapshot?.running || checking || !snapshot) return;
const model = describeGenerationModel(snapshot.model);
const vad = describeGenerationVad(snapshot.vad);
if (
action === 'download'
? !model.download
: action === 'download-vad'
? !vad.download
: !model.ready || !vad.ready || !snapshot.mediaPath
)
return;
pending = true;
result = null;
error = null;
progress = { stage: action === 'generate' ? 'extract' : 'download', message: 'Starting...' };
render();
try {
result = await (action === 'download'
? window.electronAPI.downloadSubtitleGenerationModel()
: action === 'download-vad'
? window.electronAPI.downloadSubtitleGenerationVadModel()
: window.electronAPI.startSubtitleGeneration());
} catch (cause) {
result = {
ok: false,
message: cause instanceof Error ? cause.message : 'Subtitle generation failed.',
};
} finally {
pending = false;
cancelling = false;
// Recheck model availability after downloads and recover controls after errors.
await refresh();
}
}
async function selectModel(): Promise<void> {
const model = dom.modelSelect.value;
if (pending || snapshot?.running || checking || !isSubtitleGenerationModelId(model)) return;
checking = true;
error = null;
render();
try {
snapshot = await window.electronAPI.selectSubtitleGenerationModel(model);
result = snapshot.lastResult;
progress = snapshot.progress;
} catch (cause) {
error = cause instanceof Error ? cause.message : 'Could not select the model.';
} finally {
checking = false;
render();
}
}
async function selectVad(): Promise<void> {
if (pending || snapshot?.running || checking || !snapshot) return;
const enabled = dom.vadEnabled.checked;
checking = true;
error = null;
render();
try {
snapshot = await window.electronAPI.setSubtitleGenerationVadEnabled(enabled);
result = snapshot.lastResult;
progress = snapshot.progress;
} catch (cause) {
error = cause instanceof Error ? cause.message : 'Could not change speech detection.';
} finally {
checking = false;
render();
}
}
async function cancel(): Promise<void> {
if (cancelling) return;
cancelling = true;
render();
try {
await window.electronAPI.cancelSubtitleGeneration();
await refresh();
} catch (cause) {
error = cause instanceof Error ? cause.message : 'Could not cancel the operation.';
} finally {
cancelling = false;
render();
}
}
function open(): void {
if (ctx.state.subtitleGenerationModalOpen) return;
priorFocus = document.activeElement;
ctx.state.subtitleGenerationModalOpen = true;
options.syncSettingsModalSubtitleSuppression();
dom.modal.classList.remove('hidden');
dom.modal.setAttribute('aria-hidden', 'false');
syncOverlayMouseIgnoreState(ctx);
focus.attach();
focus.focusFallbackTarget();
window.electronAPI.notifyOverlayModalOpened('subtitle-generation');
void refresh();
}
function close(): void {
if (!ctx.state.subtitleGenerationModalOpen) return;
ctx.state.subtitleGenerationModalOpen = false;
options.syncSettingsModalSubtitleSuppression();
dom.modal.classList.add('hidden');
dom.modal.setAttribute('aria-hidden', 'true');
focus.detach();
stopPolling();
window.electronAPI.notifyOverlayModalClosed('subtitle-generation');
if (priorFocus instanceof HTMLElement && priorFocus.isConnected)
priorFocus.focus({ preventScroll: true });
if (!options.modalStateReader.isAnyModalOpen()) syncOverlayMouseIgnoreState(ctx);
}
function handleKeydown(event: KeyboardEvent): boolean {
if (!ctx.state.subtitleGenerationModalOpen) return false;
if (event.key === 'Escape') {
event.preventDefault();
close();
}
return true;
}
function wireDomEvents(): void {
dom.close.addEventListener('click', close);
dom.open.addEventListener('click', () => {
void window.electronAPI
.requestSubtitleGenerationOpen()
.then((opened) => {
if (!opened)
ctx.dom.subtitleSidebarStatus.textContent = 'Could not open subtitle generation.';
})
.catch((cause: unknown) => {
ctx.dom.subtitleSidebarStatus.textContent =
cause instanceof Error ? cause.message : 'Could not open subtitle generation.';
});
});
dom.download.addEventListener('click', () => void run('download'));
dom.vadDownload.addEventListener('click', () => void run('download-vad'));
dom.vadEnabled.addEventListener('change', () => void selectVad());
dom.modelSelect.addEventListener('change', () => void selectModel());
dom.start.addEventListener('click', () => void run('generate'));
dom.cancel.addEventListener('click', () => void cancel());
dom.refresh.addEventListener('click', () => void refresh());
unsubscribe = window.electronAPI.onSubtitleGenerationProgress((update) => {
progress = update;
render();
});
}
function dispose(): void {
stopPolling();
focus.detach();
unsubscribe?.();
unsubscribe = null;
}
return { open, close, handleKeydown, wireDomEvents, dispose };
}
@@ -113,6 +113,20 @@ test('findActiveSubtitleCueIndex prefers current subtitle timing over near-futur
assert.equal(findActiveSubtitleCueIndex(cues, { text: 'previous', startTime: 231 }, 233, 0), 0);
});
test('findActiveSubtitleCueIndex follows playback through empty subtitle gaps', () => {
const cues = [
{ startTime: 0, endTime: 2, text: 'first' },
{ startTime: 100, endTime: 102, text: 'later' },
{ startTime: 105, endTime: 107, text: 'next' },
];
assert.equal(findActiveSubtitleCueIndex(cues, { text: 'later', startTime: 100 }, 101, 1), 1);
assert.equal(findActiveSubtitleCueIndex(cues, { text: '', startTime: 0 }, 103, 1), 2);
assert.equal(findActiveSubtitleCueIndex(cues, { text: 'next', startTime: 105 }, 105, 2), 2);
assert.equal(findActiveSubtitleCueIndex(cues, { text: '', startTime: 0 }, 108, 2), -1);
assert.equal(findActiveSubtitleCueIndex(cues, { text: 'first', startTime: 0 }, 0, 2), 0);
});
test('subtitle sidebar mining context resolves selected row cue timing', () => {
const globals = globalThis as typeof globalThis & {
Element?: unknown;
+5 -1
View File
@@ -116,8 +116,12 @@ export function findActiveSubtitleCueIndex(
return -1;
}
// The mpv client maps cleared sub-start to zero. Empty text has no active cue timing.
const hasCurrentTiming =
typeof current?.startTime === 'number' && Number.isFinite(current.startTime);
current !== null &&
normalizeCueText(current.text).length > 0 &&
typeof current.startTime === 'number' &&
Number.isFinite(current.startTime);
if (hasCurrentTiming) {
const timingMatch = cues.findIndex(
+1
View File
@@ -11,6 +11,7 @@ function isBlockingOverlayModalOpen(state: RendererState): boolean {
state.kikuModalOpen ||
state.runtimeOptionsModalOpen ||
state.subsyncModalOpen ||
state.subtitleGenerationModalOpen ||
state.sessionHelpModalOpen,
);
}
+17
View File
@@ -44,6 +44,7 @@ import { isControllerInteractionBlocked } from './controller-interaction-blockin
import { createCharacterDictionaryModal } from './modals/character-dictionary.js';
import { createRuntimeOptionsModal } from './modals/runtime-options.js';
import { createSubsyncModal } from './modals/subsync.js';
import { createSubtitleGenerationModal } from './modals/subtitle-generation.js';
import { createYoutubeTrackPickerModal } from './modals/youtube-track-picker.js';
import { createMediaTimingReviewModal } from './modals/media-timing-review.js';
import { createPositioningController } from './positioning.js';
@@ -79,6 +80,12 @@ const ctx = {
};
const modalDescriptors = [
{
id: 'subtitle-generation',
isOpen: () => ctx.state.subtitleGenerationModalOpen,
close: () => subtitleGenerationModal.close(),
suppressesSubtitles: true,
},
{
id: 'controller-select',
isOpen: () => ctx.state.controllerSelectModalOpen,
@@ -212,6 +219,10 @@ const subsyncModal = createSubsyncModal(ctx, {
modalStateReader: { isAnyModalOpen },
syncSettingsModalSubtitleSuppression,
});
const subtitleGenerationModal = createSubtitleGenerationModal(ctx, {
modalStateReader: { isAnyModalOpen },
syncSettingsModalSubtitleSuppression,
});
const controllerSelectModal = createControllerSelectModal(ctx, {
modalStateReader: { isAnyModalOpen },
syncSettingsModalSubtitleSuppression,
@@ -280,6 +291,7 @@ const keyboardHandlers = createKeyboardHandlers(ctx, {
handleRuntimeOptionsKeydown: runtimeOptionsModal.handleRuntimeOptionsKeydown,
handleCharacterDictionaryKeydown: characterDictionaryModal.handleCharacterDictionaryKeydown,
handleSubsyncKeydown: subsyncModal.handleSubsyncKeydown,
handleSubtitleGenerationKeydown: subtitleGenerationModal.handleKeydown,
handleKikuKeydown: kikuModal.handleKikuKeydown,
handleJimakuKeydown: jimakuModal.handleJimakuKeydown,
handleTsukihimeKeydown: tsukihimeModal.handleTsukihimeKeydown,
@@ -532,6 +544,9 @@ const recovery = createRendererRecoveryController({
registerRendererGlobalErrorHandlers(window, recovery);
function registerModalOpenHandlers(): void {
window.electronAPI.onSubtitleGenerationOpen(() => {
runGuarded('subtitle-generation:open', () => subtitleGenerationModal.open());
});
window.electronAPI.onOpenRuntimeOptions(() => {
runGuarded('runtime-options:open', () => {
runtimeOptionsModal.openRuntimeOptionsModal();
@@ -832,6 +847,7 @@ async function init(): Promise<void> {
kikuModal.wireDomEvents();
runtimeOptionsModal.wireDomEvents();
subsyncModal.wireDomEvents();
subtitleGenerationModal.wireDomEvents();
controllerSelectModal.wireDomEvents();
controllerDebugModal.wireDomEvents();
sessionHelpModal.wireDomEvents();
@@ -839,6 +855,7 @@ async function init(): Promise<void> {
subtitleSidebarModal.wireDomEvents();
characterDictionaryModal.wireDomEvents();
window.addEventListener('beforeunload', () => {
subtitleGenerationModal.dispose();
subtitleSidebarModal.disposeDomEvents();
});
+2
View File
@@ -87,6 +87,7 @@ export type RendererState = {
characterDictionaryStatus: string;
subsyncModalOpen: boolean;
subtitleGenerationModalOpen: boolean;
subsyncSubtitleTracks: SubsyncSubtitleTrack[];
subsyncSubmitting: boolean;
@@ -219,6 +220,7 @@ export function createRendererState(): RendererState {
characterDictionaryStatus: '',
subsyncModalOpen: false,
subtitleGenerationModalOpen: false,
subsyncSubtitleTracks: [],
subsyncSubmitting: false,
+132
View File
@@ -2976,6 +2976,138 @@ iframe[id^='yomitan-popup'],
width: min(560px, 92%);
}
.subtitle-generation-content {
width: min(580px, 92%);
max-height: 92%;
border-top: 3px solid var(--ctp-green);
}
.subtitle-generation-body {
display: flex;
flex-direction: column;
gap: 18px;
overflow-y: auto;
scrollbar-width: thin;
scrollbar-color: var(--ctp-surface1) transparent;
}
.subtitle-generation-body > * {
flex-shrink: 0;
}
.subtitle-generation-intro {
margin: 0;
color: var(--ctp-subtext1);
font-size: 14px;
line-height: 1.65;
}
.subtitle-generation-detail {
padding: 12px 14px;
border-left: 2px solid var(--ctp-surface1);
background: rgba(24, 25, 38, 0.35);
overflow-wrap: anywhere;
font-size: 13px;
}
.subtitle-generation-label {
display: block;
margin-bottom: 6px;
color: var(--ctp-green);
font-size: 11px;
letter-spacing: 0.06em;
text-transform: uppercase;
}
.subtitle-generation-hint {
margin: 8px 0 0;
color: var(--ctp-subtext0);
font-size: 12px;
line-height: 1.6;
}
.subtitle-generation-model-picker {
margin-top: 14px;
}
.subtitle-generation-model-picker label {
display: block;
margin-bottom: 6px;
color: var(--ctp-subtext1);
font-size: 12px;
}
.subtitle-generation-model-picker select {
width: 100%;
padding: 8px 10px;
border: 1px solid var(--ctp-surface1);
border-radius: 6px;
background: var(--ctp-mantle);
color: var(--ctp-text);
font: inherit;
}
.subtitle-generation-model-picker select:focus-visible {
outline: 2px solid var(--ctp-green);
outline-offset: 2px;
}
.subtitle-generation-model-picker select:disabled {
opacity: 0.5;
}
.subtitle-generation-detail button {
margin-top: 12px;
}
.subtitle-generation-vad-toggle {
display: flex;
align-items: center;
gap: 8px;
cursor: pointer;
}
.subtitle-generation-vad-toggle input {
accent-color: var(--ctp-green);
}
.subtitle-generation-vad-toggle .subtitle-generation-hint {
margin: 0 0 0 auto;
}
.subtitle-generation-progress-heading {
display: flex;
justify-content: space-between;
margin-bottom: 9px;
font-size: 13px;
}
.subtitle-generation-activity progress {
display: block;
width: 100%;
height: 8px;
accent-color: var(--ctp-green);
}
.subtitle-generation-actions {
display: flex;
flex-wrap: wrap;
justify-content: flex-end;
gap: 8px;
}
.subtitle-generation-actions button:disabled,
.subtitle-generation-detail button:disabled {
opacity: 0.5;
cursor: default;
}
.subtitle-generation-open {
align-self: flex-start;
margin: 0 16px 8px;
font-size: 12px;
}
.subsync-form {
display: flex;
flex-direction: column;