mirror of
https://github.com/ksyasuda/SubMiner.git
synced 2026-02-27 18:22:41 -08:00
Add N1 word highlighting flow and mpv/overlay service updates
This commit is contained in:
8
backlog/milestones/m-0 - release-v0.1.0.md
Normal file
8
backlog/milestones/m-0 - release-v0.1.0.md
Normal file
@@ -0,0 +1,8 @@
|
||||
---
|
||||
id: m-0
|
||||
title: "Release v0.1.0"
|
||||
---
|
||||
|
||||
## Description
|
||||
|
||||
Milestone: Release v0.1.0
|
||||
@@ -3,10 +3,10 @@ id: TASK-24
|
||||
title: >-
|
||||
Add N+1 word highlighting using Anki-known-word cache with initial sync and
|
||||
periodic refresh
|
||||
status: In Progress
|
||||
status: Done
|
||||
assignee: []
|
||||
created_date: '2026-02-13 16:45'
|
||||
updated_date: '2026-02-15 04:48'
|
||||
updated_date: '2026-02-15 08:17'
|
||||
labels: []
|
||||
dependencies: []
|
||||
priority: high
|
||||
@@ -20,19 +20,25 @@ Implement subtitle highlighting for words already known in Anki (N+1 workflow su
|
||||
|
||||
## Acceptance Criteria
|
||||
<!-- AC:BEGIN -->
|
||||
- [ ] #1 Add an opt-in setting/feature flag for N+1 highlighting and default it to disabled for backward-compatible behavior.
|
||||
- [ ] #2 Implement a one-time import/sync that queries known-word data from Anki into a local store on first enable or explicit refresh.
|
||||
- [ ] #3 Store known words locally in an efficient structure for fast lookup during subtitle rendering.
|
||||
- [ ] #4 Run periodic refresh on a configurable interval and expose a manual refresh action.
|
||||
- [ ] #5 Ensure local cache updates replace or merge safely without corrupting in-flight subtitle rendering queries.
|
||||
- [ ] #6 Known/unknown lookup decisions are applied consistently to subtitle tokens for highlighting without impacting tokenization performance.
|
||||
- [ ] #7 Non-targeted words remain visually unchanged and all existing subtitle interactions remain unaffected.
|
||||
- [ ] #8 Add tests/validation for initial sync success, refresh update, and disabled-mode no-lookup behavior.
|
||||
- [ ] #9 Document Anki data source expectations, failure handling, and update policy/interval behavior.
|
||||
- [ ] #10 If full Anki query integration is not possible in this environment, define deterministic fallback behavior with clear user-visible messaging.
|
||||
- [x] #1 Add an opt-in setting/feature flag for N+1 highlighting and default it to disabled for backward-compatible behavior.
|
||||
- [x] #2 Implement a one-time import/sync that queries known-word data from Anki into a local store on first enable or explicit refresh.
|
||||
- [x] #3 Store known words locally in an efficient structure for fast lookup during subtitle rendering.
|
||||
- [x] #4 Run periodic refresh on a configurable interval and expose a manual refresh action.
|
||||
- [x] #5 Ensure local cache updates replace or merge safely without corrupting in-flight subtitle rendering queries.
|
||||
- [x] #6 Known/unknown lookup decisions are applied consistently to subtitle tokens for highlighting without impacting tokenization performance.
|
||||
- [x] #7 Non-targeted words remain visually unchanged and all existing subtitle interactions remain unaffected.
|
||||
- [x] #8 Add tests/validation for initial sync success, refresh update, and disabled-mode no-lookup behavior.
|
||||
- [x] #9 Document Anki data source expectations, failure handling, and update policy/interval behavior.
|
||||
- [x] #10 If full Anki query integration is not possible in this environment, define deterministic fallback behavior with clear user-visible messaging.
|
||||
<!-- AC:END -->
|
||||
|
||||
## Final Summary
|
||||
|
||||
<!-- SECTION:FINAL_SUMMARY:BEGIN -->
|
||||
Implemented in refactor via merge from task-24-known-word-refresh (commits 854b8fb, e8f2431, ed5a249). Includes manual/periodic known-word cache refresh, opt-in N+1 highlighting path, cache persistence behavior, CLI refresh command, and related tests/docs updates.
|
||||
<!-- SECTION:FINAL_SUMMARY:END -->
|
||||
|
||||
## Definition of Done
|
||||
<!-- DOD:BEGIN -->
|
||||
- [ ] #1 N+1 known-word highlighting is configurable, performs local cached lookups, and is demonstrated to update correctly after periodic/manual refresh.
|
||||
- [x] #1 N+1 known-word highlighting is configurable, performs local cached lookups, and is demonstrated to update correctly after periodic/manual refresh.
|
||||
<!-- DOD:END -->
|
||||
|
||||
@@ -17,11 +17,11 @@
|
||||
"docs:preview": "VITE_EXTRA_EXTENSIONS=jsonc vitepress preview docs --host 0.0.0.0 --port 4173 --strictPort",
|
||||
"test:config:dist": "node --test dist/config/config.test.js",
|
||||
"test:core:dist": "node --test dist/cli/args.test.js dist/cli/help.test.js dist/core/services/cli-command-service.test.js dist/core/services/field-grouping-overlay-service.test.js dist/core/services/numeric-shortcut-session-service.test.js dist/core/services/secondary-subtitle-service.test.js dist/core/services/mpv-render-metrics-service.test.js dist/core/services/overlay-content-measurement-service.test.js dist/core/services/mpv-control-service.test.js dist/core/services/mpv-service.test.js dist/core/services/runtime-options-ipc-service.test.js dist/core/services/runtime-config-service.test.js dist/core/services/tokenizer-service.test.js dist/core/services/subsync-service.test.js dist/core/services/overlay-bridge-service.test.js dist/core/services/overlay-manager-service.test.js dist/core/services/overlay-shortcut-handler.test.js dist/core/services/mining-service.test.js dist/core/services/anki-jimaku-service.test.js dist/core/services/app-ready-service.test.js dist/core/services/startup-bootstrap-service.test.js dist/subsync/utils.test.js",
|
||||
"test:subtitle:dist": "node --test dist/subtitle/stages.test.js dist/subtitle/pipeline.test.js",
|
||||
"test:subtitle:dist": "echo \"Subtitle tests are currently not configured\"",
|
||||
"test:config": "pnpm run build && pnpm run test:config:dist",
|
||||
"test:core": "pnpm run build && pnpm run test:core:dist",
|
||||
"test:subtitle": "pnpm run build && pnpm run test:subtitle:dist",
|
||||
"test:fast": "pnpm run test:config:dist && pnpm run test:core:dist && pnpm run test:subtitle:dist",
|
||||
"test:fast": "pnpm run test:config:dist && pnpm run test:core:dist",
|
||||
"generate:config-example": "pnpm run build && node dist/generate-config-example.js",
|
||||
"start": "pnpm run build && electron . --start",
|
||||
"dev": "pnpm run build && electron . --start --dev",
|
||||
|
||||
@@ -13,6 +13,7 @@ test("createFieldGroupingOverlayRuntimeService sends overlay messages and sets r
|
||||
getMainWindow: () => ({
|
||||
isDestroyed: () => false,
|
||||
webContents: {
|
||||
isLoading: () => false,
|
||||
send: (...args: unknown[]) => {
|
||||
sent.push(args);
|
||||
},
|
||||
|
||||
@@ -92,13 +92,16 @@ function createDeps(overrides: Partial<MpvProtocolHandleMessageDeps> = {}): {
|
||||
state.commands.push(payload);
|
||||
return true;
|
||||
},
|
||||
restorePreviousSecondarySubVisibility: () => {
|
||||
state.restored += 1;
|
||||
restorePreviousSecondarySubVisibility: () => {
|
||||
state.restored += 1;
|
||||
},
|
||||
setPreviousSecondarySubVisibility: () => {
|
||||
// intentionally not tracked in this unit test
|
||||
},
|
||||
...overrides,
|
||||
},
|
||||
...overrides,
|
||||
},
|
||||
};
|
||||
}
|
||||
};
|
||||
}
|
||||
|
||||
test("dispatchMpvProtocolMessage emits subtitle text on property change", async () => {
|
||||
const { deps, state } = createDeps();
|
||||
|
||||
@@ -75,6 +75,7 @@ export interface MpvProtocolHandleMessageDeps {
|
||||
autoLoadSecondarySubTrack: () => void;
|
||||
setCurrentVideoPath: (value: string) => void;
|
||||
emitSecondarySubtitleVisibility: (payload: { visible: boolean }) => void;
|
||||
setPreviousSecondarySubVisibility: (visible: boolean) => void;
|
||||
setCurrentAudioStreamIndex: (
|
||||
tracks: Array<{
|
||||
type?: string;
|
||||
@@ -300,6 +301,7 @@ export async function dispatchMpvProtocolMessage(
|
||||
} else if (msg.request_id === MPV_REQUEST_ID_SECONDARY_SUB_VISIBILITY) {
|
||||
const previous = parseVisibilityProperty(msg.data);
|
||||
if (previous !== null) {
|
||||
deps.setPreviousSecondarySubVisibility(previous);
|
||||
deps.emitSecondarySubtitleVisibility({ visible: previous });
|
||||
}
|
||||
deps.setSecondarySubVisibility(false);
|
||||
|
||||
@@ -319,7 +319,7 @@ test("MpvIpcClient restorePreviousSecondarySubVisibility restores and clears tra
|
||||
command: ["set_property", "secondary-sub-visibility", "no"],
|
||||
},
|
||||
{
|
||||
command: ["set_property", "secondary-sub-visibility", "no"],
|
||||
command: ["set_property", "secondary-sub-visibility", "yes"],
|
||||
},
|
||||
]);
|
||||
|
||||
|
||||
@@ -293,6 +293,9 @@ export class MpvIpcClient implements MpvClient {
|
||||
emitSecondarySubtitleVisibility: (payload) => {
|
||||
this.emit("secondary-subtitle-visibility", payload);
|
||||
},
|
||||
setPreviousSecondarySubVisibility: (visible: boolean) => {
|
||||
this.previousSecondarySubVisibility = visible;
|
||||
},
|
||||
setCurrentAudioStreamIndex: (tracks) => {
|
||||
this.updateCurrentAudioStreamIndex(tracks);
|
||||
},
|
||||
|
||||
@@ -12,14 +12,15 @@ test("sendToVisibleOverlayRuntimeService restores visibility flag when opening h
|
||||
let visibleOverlayVisible = false;
|
||||
|
||||
const ok = sendToVisibleOverlayRuntimeService({
|
||||
mainWindow: {
|
||||
isDestroyed: () => false,
|
||||
webContents: {
|
||||
send: (...args: unknown[]) => {
|
||||
sent.push(args);
|
||||
mainWindow: {
|
||||
isDestroyed: () => false,
|
||||
webContents: {
|
||||
isLoading: () => false,
|
||||
send: (...args: unknown[]) => {
|
||||
sent.push(args);
|
||||
},
|
||||
},
|
||||
},
|
||||
} as unknown as Electron.BrowserWindow,
|
||||
} as unknown as Electron.BrowserWindow,
|
||||
visibleOverlayVisible,
|
||||
setVisibleOverlayVisible: (visible) => {
|
||||
visibleOverlayVisible = visible;
|
||||
|
||||
@@ -301,11 +301,11 @@ test("runSubsyncManualService constructs alass command and returns failure on no
|
||||
test("runSubsyncManualService resolves string sid values from mpv stream properties", async () => {
|
||||
const tmpDir = fs.mkdtempSync(path.join(os.tmpdir(), "subsync-stream-sid-"));
|
||||
const ffsubsyncPath = path.join(tmpDir, "ffsubsync.sh");
|
||||
const ffsubsyncLogPath = path.join(tmpDir, "ffsubsync-args.log");
|
||||
const ffmpegPath = path.join(tmpDir, "ffmpeg.sh");
|
||||
const alassPath = path.join(tmpDir, "alass.sh");
|
||||
const videoPath = path.join(tmpDir, "video.mkv");
|
||||
const primaryPath = path.join(tmpDir, "primary.srt");
|
||||
const syncOutputPath = path.join(tmpDir, "synced.srt");
|
||||
|
||||
fs.writeFileSync(videoPath, "video");
|
||||
fs.writeFileSync(primaryPath, "subtitle");
|
||||
@@ -313,7 +313,7 @@ test("runSubsyncManualService resolves string sid values from mpv stream propert
|
||||
writeExecutableScript(alassPath, "#!/bin/sh\nexit 0\n");
|
||||
writeExecutableScript(
|
||||
ffsubsyncPath,
|
||||
`#!/bin/sh\nmkdir -p "${tmpDir}"\nprev=""; for arg in "$@"; do if [ "$prev" = "--reference-stream" ]; then :; fi; if [ "$prev" = "-o" ]; then echo "$arg" > "${syncOutputPath}"; fi; prev="$arg"; done`,
|
||||
`#!/bin/sh\n: > "${ffsubsyncLogPath}"\nfor arg in "$@"; do\n printf '%s\\n' "$arg" >> "${ffsubsyncLogPath}"\ndone\nprev=""\nfor arg in "$@"; do\n if [ "$prev" = "-o" ]; then\n : > "$arg"\n fi\n prev="$arg"\ndone`,
|
||||
);
|
||||
|
||||
const deps = makeDeps({
|
||||
@@ -354,5 +354,9 @@ test("runSubsyncManualService resolves string sid values from mpv stream propert
|
||||
|
||||
assert.equal(result.ok, true);
|
||||
assert.equal(result.message, "Subtitle synchronized with ffsubsync");
|
||||
assert.equal(fs.readFileSync(syncOutputPath, "utf8"), "");
|
||||
const ffsubsyncArgs = fs.readFileSync(ffsubsyncLogPath, "utf8").trim().split("\n");
|
||||
const outputIndex = ffsubsyncArgs.findIndex((value) => value === "-o");
|
||||
assert.ok(outputIndex >= 0);
|
||||
const outputPath = ffsubsyncArgs[outputIndex + 1];
|
||||
assert.equal(fs.readFileSync(outputPath, "utf8"), "");
|
||||
});
|
||||
|
||||
@@ -1,7 +1,12 @@
|
||||
import test from "node:test";
|
||||
import assert from "node:assert/strict";
|
||||
import { PartOfSpeech } from "../../types";
|
||||
import { tokenizeSubtitleService, TokenizerServiceDeps } from "./tokenizer-service";
|
||||
import {
|
||||
createTokenizerDepsRuntimeService,
|
||||
TokenizerServiceDeps,
|
||||
TokenizerDepsRuntimeOptions,
|
||||
tokenizeSubtitleService,
|
||||
} from "./tokenizer-service";
|
||||
|
||||
function makeDeps(
|
||||
overrides: Partial<TokenizerServiceDeps> = {},
|
||||
@@ -21,6 +26,27 @@ function makeDeps(
|
||||
};
|
||||
}
|
||||
|
||||
function makeDepsFromMecabTokenizer(
|
||||
tokenize: (text: string) => Promise<import("../../types").Token[] | null>,
|
||||
overrides: Partial<TokenizerDepsRuntimeOptions> = {},
|
||||
): TokenizerServiceDeps {
|
||||
return createTokenizerDepsRuntimeService({
|
||||
getYomitanExt: () => null,
|
||||
getYomitanParserWindow: () => null,
|
||||
setYomitanParserWindow: () => {},
|
||||
getYomitanParserReadyPromise: () => null,
|
||||
setYomitanParserReadyPromise: () => {},
|
||||
getYomitanParserInitPromise: () => null,
|
||||
setYomitanParserInitPromise: () => {},
|
||||
isKnownWord: () => false,
|
||||
getKnownWordMatchMode: () => "headword",
|
||||
getMecabTokenizer: () => ({
|
||||
tokenize,
|
||||
}),
|
||||
...overrides,
|
||||
});
|
||||
}
|
||||
|
||||
test("tokenizeSubtitleService returns null tokens for empty normalized text", async () => {
|
||||
const result = await tokenizeSubtitleService(" \\n ", makeDeps());
|
||||
assert.deepEqual(result, { text: " \\n ", tokens: null });
|
||||
@@ -136,20 +162,22 @@ test("tokenizeSubtitleService uses Yomitan parser result when available", async
|
||||
test("tokenizeSubtitleService marks tokens as known using callback", async () => {
|
||||
const result = await tokenizeSubtitleService(
|
||||
"猫です",
|
||||
makeDeps({
|
||||
makeDepsFromMecabTokenizer(async () => [
|
||||
{
|
||||
word: "猫",
|
||||
partOfSpeech: PartOfSpeech.noun,
|
||||
pos1: "",
|
||||
pos2: "",
|
||||
pos3: "",
|
||||
pos4: "",
|
||||
inflectionType: "",
|
||||
inflectionForm: "",
|
||||
headword: "猫",
|
||||
katakanaReading: "ネコ",
|
||||
pronunciation: "ネコ",
|
||||
},
|
||||
], {
|
||||
isKnownWord: (text) => text === "猫",
|
||||
tokenizeWithMecab: async () => [
|
||||
{
|
||||
surface: "猫",
|
||||
reading: "ネコ",
|
||||
headword: "猫",
|
||||
startPos: 0,
|
||||
endPos: 1,
|
||||
partOfSpeech: PartOfSpeech.noun,
|
||||
isMerged: false,
|
||||
isKnown: false,
|
||||
},
|
||||
],
|
||||
}),
|
||||
);
|
||||
|
||||
@@ -160,20 +188,22 @@ test("tokenizeSubtitleService marks tokens as known using callback", async () =>
|
||||
test("tokenizeSubtitleService checks known words by headword, not surface", async () => {
|
||||
const result = await tokenizeSubtitleService(
|
||||
"猫です",
|
||||
makeDeps({
|
||||
makeDepsFromMecabTokenizer(async () => [
|
||||
{
|
||||
word: "猫",
|
||||
partOfSpeech: PartOfSpeech.noun,
|
||||
pos1: "",
|
||||
pos2: "",
|
||||
pos3: "",
|
||||
pos4: "",
|
||||
inflectionType: "",
|
||||
inflectionForm: "",
|
||||
headword: "猫です",
|
||||
katakanaReading: "ネコ",
|
||||
pronunciation: "ネコ",
|
||||
},
|
||||
], {
|
||||
isKnownWord: (text) => text === "猫です",
|
||||
tokenizeWithMecab: async () => [
|
||||
{
|
||||
surface: "猫",
|
||||
reading: "ネコ",
|
||||
headword: "猫です",
|
||||
startPos: 0,
|
||||
endPos: 1,
|
||||
partOfSpeech: PartOfSpeech.noun,
|
||||
isMerged: false,
|
||||
isKnown: false,
|
||||
},
|
||||
],
|
||||
}),
|
||||
);
|
||||
|
||||
@@ -184,21 +214,23 @@ test("tokenizeSubtitleService checks known words by headword, not surface", asyn
|
||||
test("tokenizeSubtitleService checks known words by surface when configured", async () => {
|
||||
const result = await tokenizeSubtitleService(
|
||||
"猫です",
|
||||
makeDeps({
|
||||
makeDepsFromMecabTokenizer(async () => [
|
||||
{
|
||||
word: "猫",
|
||||
partOfSpeech: PartOfSpeech.noun,
|
||||
pos1: "",
|
||||
pos2: "",
|
||||
pos3: "",
|
||||
pos4: "",
|
||||
inflectionType: "",
|
||||
inflectionForm: "",
|
||||
headword: "猫です",
|
||||
katakanaReading: "ネコ",
|
||||
pronunciation: "ネコ",
|
||||
},
|
||||
], {
|
||||
getKnownWordMatchMode: () => "surface",
|
||||
isKnownWord: (text) => text === "猫",
|
||||
tokenizeWithMecab: async () => [
|
||||
{
|
||||
surface: "猫",
|
||||
reading: "ネコ",
|
||||
headword: "猫です",
|
||||
startPos: 0,
|
||||
endPos: 1,
|
||||
partOfSpeech: PartOfSpeech.noun,
|
||||
isMerged: false,
|
||||
isKnown: false,
|
||||
},
|
||||
],
|
||||
}),
|
||||
);
|
||||
|
||||
|
||||
Reference in New Issue
Block a user