first version

This commit is contained in:
2026-04-25 16:23:09 +05:30
commit 7d3c36050b
73 changed files with 19033 additions and 0 deletions

View File

@@ -0,0 +1,648 @@
import { useCallback, useEffect, useMemo, useRef, useState } from "react";
import { open as openDialog } from "@tauri-apps/plugin-dialog";
import { listen } from "@tauri-apps/api/event";
import { api } from "../ipc";
import type { CocoShape, ImportedCoco, VideoInfo } from "../types";
import { FrameCanvas } from "../components/extract/FrameCanvas";
import { drawOverlay } from "../components/extract/overlay";
const SPEEDS = [0.25, 0.5, 1, 1.5, 2, 3, 4];
interface Props {
username: string;
}
interface BulkProgress {
done: number;
total: number;
phase: "extract" | "undo";
}
export function ExtractMode({ username }: Props) {
const [video, setVideo] = useState<VideoInfo | null>(null);
const [frameIdx, setFrameIdx] = useState(0);
const [playing, setPlaying] = useState(false);
const [speed, setSpeed] = useState(1);
const [reviewMode, setReviewMode] = useState(false);
const [overlayVisible, setOverlayVisible] = useState(true);
const [loadError, setLoadError] = useState<string | null>(null);
const [coco, setCoco] = useState<ImportedCoco | null>(null);
// Extraction state
const [outputFolder, setOutputFolder] = useState<string | null>(null);
const [extractedFrames, setExtractedFrames] = useState<Set<number>>(new Set());
const [bulkStart, setBulkStart] = useState<number | null>(null);
const [bulkEnd, setBulkEnd] = useState<number | null>(null);
const [bulkProgress, setBulkProgress] = useState<BulkProgress | null>(null);
const [bulkError, setBulkError] = useState<string | null>(null);
const [infoCollapsed, setInfoCollapsed] = useState(false);
const canvasRef = useRef<HTMLCanvasElement>(null);
const frameMap = useMemo(() => {
const m = new Map<number, CocoShape[]>();
if (!coco) return m;
for (const [k, v] of Object.entries(coco.frames)) m.set(Number(k), v);
return m;
}, [coco]);
// Refs mirroring state for effects that shouldn't depend on fast-changing values.
const frameIdxRef = useRef(0);
const frameMapRef = useRef<Map<number, CocoShape[]>>(new Map());
const overlayVisibleRef = useRef(overlayVisible);
const annotatedFrames = useRef<Set<number>>(new Set());
// Playback clock anchor — set when playback starts, rebased on seeks so a
// slider drag during play doesn't get overwritten by the tick loop.
const playAnchorRef = useRef<{ time: number; idx: number } | null>(null);
useEffect(() => { frameIdxRef.current = frameIdx; }, [frameIdx]);
useEffect(() => {
frameMapRef.current = frameMap;
annotatedFrames.current = new Set(frameMap.keys());
}, [frameMap]);
useEffect(() => { overlayVisibleRef.current = overlayVisible; }, [overlayVisible]);
// --- decode pump (latest-wins coalescing) ----------------------------
const requestedIdx = useRef(0);
const renderedIdx = useRef(-1);
const decodingInFlight = useRef(false);
const renderFrame = useCallback(
async (bytes: ArrayBuffer, atFrameIdx: number) => {
const canvas = canvasRef.current;
if (!canvas) return;
const blob = new Blob([bytes], { type: "image/jpeg" });
let bitmap: ImageBitmap;
try { bitmap = await createImageBitmap(blob); }
catch (e) { console.error("createImageBitmap failed", e); return; }
const ctx = canvas.getContext("2d");
if (!ctx) { bitmap.close(); return; }
ctx.clearRect(0, 0, canvas.width, canvas.height);
ctx.drawImage(bitmap, 0, 0, canvas.width, canvas.height);
bitmap.close();
if (overlayVisibleRef.current) {
const shapes = frameMapRef.current.get(atFrameIdx);
if (shapes && shapes.length > 0) {
drawOverlay(ctx, shapes, canvas.width, canvas.height);
}
}
},
[]
);
const pump = useCallback(async () => {
if (decodingInFlight.current) return;
decodingInFlight.current = true;
try {
while (requestedIdx.current !== renderedIdx.current) {
const want = requestedIdx.current;
try {
const bytes = await api.decodeFrame(want);
if (requestedIdx.current !== want) continue;
await renderFrame(bytes, want);
renderedIdx.current = want;
} catch (e) {
if (requestedIdx.current !== want) continue;
console.error("decode_frame failed", e);
break;
}
}
} finally {
decodingInFlight.current = false;
}
}, [renderFrame]);
useEffect(() => {
if (!video) return;
requestedIdx.current = frameIdx;
pump();
}, [frameIdx, pump, video]);
useEffect(() => {
renderedIdx.current = -1;
pump();
}, [overlayVisible, frameMap, pump]);
// --- playback loop ---------------------------------------------------
useEffect(() => {
if (!playing || !video) return;
playAnchorRef.current = {
time: performance.now(),
idx: frameIdxRef.current,
};
let raf = 0;
const tick = () => {
const anchor = playAnchorRef.current;
if (!anchor) return;
const elapsed = (performance.now() - anchor.time) / 1000;
const want = Math.min(
video.total_frames - 1,
anchor.idx + Math.floor(elapsed * video.fps * speed)
);
if (reviewMode && annotatedFrames.current.has(want)) {
setFrameIdx(want);
setPlaying(false);
return;
}
setFrameIdx(want);
if (want >= video.total_frames - 1) { setPlaying(false); return; }
raf = requestAnimationFrame(tick);
};
raf = requestAnimationFrame(tick);
return () => {
cancelAnimationFrame(raf);
playAnchorRef.current = null;
};
}, [playing, video, speed, reviewMode]);
// --- navigation helpers ----------------------------------------------
const seekToFrame = useCallback((idx: number) => {
if (!video) return;
const clamped = Math.max(0, Math.min(video.total_frames - 1, idx));
setFrameIdx(clamped);
// If a playback tick is active, rebase its anchor so the loop continues
// from the new position instead of snapping back.
if (playAnchorRef.current) {
playAnchorRef.current = { time: performance.now(), idx: clamped };
}
}, [video]);
const stepFrames = useCallback((delta: number) => {
if (!video) return;
setPlaying(false);
setFrameIdx((i) => Math.max(0, Math.min(video.total_frames - 1, i + delta)));
}, [video]);
const togglePlay = useCallback(() => {
if (!video) return;
setPlaying((p) => !p);
}, [video]);
const bumpSpeed = useCallback((dir: 1 | -1) => {
setSpeed((s) => {
const i = SPEEDS.indexOf(s);
const next = Math.max(0, Math.min(SPEEDS.length - 1, (i < 0 ? 2 : i) + dir));
return SPEEDS[next];
});
}, []);
// --- extraction helpers ----------------------------------------------
const applyOutputFolder = useCallback(async (path: string) => {
setOutputFolder(path);
try {
const existing = await api.listExtractedFrames(path);
setExtractedFrames(new Set(existing));
} catch (e) {
// Folder may not exist yet — that's fine, we'll create it on extract.
setExtractedFrames(new Set());
console.warn("listExtractedFrames failed (non-fatal):", e);
}
if (coco && coco.categories.length > 0) {
try { await api.writeClassesTxt(path, coco.categories); } catch (e) { console.error(e); }
}
}, [coco]);
async function pickOutputFolder() {
const selected = await openDialog({ directory: true, multiple: false });
if (!selected || typeof selected !== "string") return;
await applyOutputFolder(selected);
}
const extractCurrentFrame = useCallback(async () => {
if (!outputFolder || !video) return;
setBulkError(null);
try {
const shapes = frameMap.get(frameIdx) ?? [];
await api.extractFrame(frameIdx, outputFolder, shapes, username);
setExtractedFrames((prev) => new Set(prev).add(frameIdx));
} catch (e) {
setBulkError(String(e));
}
}, [outputFolder, video, frameIdx, frameMap, username]);
const undoCurrentFrame = useCallback(async () => {
if (!outputFolder) return;
setBulkError(null);
try {
await api.undoExtract(frameIdx, outputFolder);
setExtractedFrames((prev) => {
const n = new Set(prev);
n.delete(frameIdx);
return n;
});
} catch (e) {
setBulkError(String(e));
}
}, [outputFolder, frameIdx]);
const setRangeStart = useCallback(() => setBulkStart(frameIdx), [frameIdx]);
const setRangeEnd = useCallback(() => setBulkEnd(frameIdx), [frameIdx]);
const clearRange = useCallback(() => { setBulkStart(null); setBulkEnd(null); }, []);
const bulkRange = useMemo(() => {
if (bulkStart == null || bulkEnd == null) return null;
return {
start: Math.min(bulkStart, bulkEnd),
end: Math.max(bulkStart, bulkEnd),
};
}, [bulkStart, bulkEnd]);
async function runBulkExtract() {
if (!outputFolder || !bulkRange) return;
setBulkError(null);
setBulkProgress({ done: 0, total: bulkRange.end - bulkRange.start + 1, phase: "extract" });
const unlisten = await listen<{ done: number; total: number }>(
"bulk-extract-progress",
(e) => setBulkProgress({ ...e.payload, phase: "extract" })
);
try {
await api.bulkExtract(
bulkRange.start,
bulkRange.end,
outputFolder,
coco?.frames ?? {},
username
);
const existing = await api.listExtractedFrames(outputFolder);
setExtractedFrames(new Set(existing));
} catch (e) {
setBulkError(String(e));
} finally {
unlisten();
setBulkProgress(null);
}
}
async function runBulkUndo() {
if (!outputFolder || !bulkRange) return;
setBulkError(null);
setBulkProgress({ done: 0, total: bulkRange.end - bulkRange.start + 1, phase: "undo" });
const unlisten = await listen<{ done: number; total: number }>(
"bulk-undo-progress",
(e) => setBulkProgress({ ...e.payload, phase: "undo" })
);
try {
await api.bulkUndo(bulkRange.start, bulkRange.end, outputFolder);
const existing = await api.listExtractedFrames(outputFolder);
setExtractedFrames(new Set(existing));
} catch (e) {
setBulkError(String(e));
} finally {
unlisten();
setBulkProgress(null);
}
}
// --- keyboard --------------------------------------------------------
useEffect(() => {
function onKey(e: KeyboardEvent) {
const tag = (e.target as HTMLElement | null)?.tagName;
if (tag === "INPUT" || tag === "TEXTAREA") return;
const step = e.ctrlKey || e.metaKey ? 10 : 1;
switch (e.key) {
case " ":
e.preventDefault(); togglePlay(); break;
case "ArrowRight":
e.preventDefault(); stepFrames(step); break;
case "ArrowLeft":
e.preventDefault(); stepFrames(-step); break;
case "ArrowUp":
e.preventDefault(); bumpSpeed(1); break;
case "ArrowDown":
e.preventDefault(); bumpSpeed(-1); break;
case "h": case "H":
e.preventDefault(); setOverlayVisible((v) => !v); break;
case "e": case "E":
e.preventDefault(); extractCurrentFrame(); break;
case "u": case "U":
e.preventDefault(); undoCurrentFrame(); break;
case "[":
e.preventDefault(); setRangeStart(); break;
case "]":
e.preventDefault(); setRangeEnd(); break;
}
}
window.addEventListener("keydown", onKey);
return () => window.removeEventListener("keydown", onKey);
}, [togglePlay, stepFrames, bumpSpeed, extractCurrentFrame, undoCurrentFrame, setRangeStart, setRangeEnd]);
// --- video/coco open -------------------------------------------------
useEffect(() => {
api.currentVideo().then((v) => {
if (!v) return;
setVideo(v);
// Re-adopt the default folder on mount too, so switching modes and
// coming back doesn't drop the extracted-frames overlay.
applyOutputFolder(v.default_output_folder);
});
// eslint-disable-next-line react-hooks/exhaustive-deps
}, []);
async function chooseVideo() {
setLoadError(null);
const selected = await openDialog({
multiple: false,
filters: [
{
name: "Video",
extensions: [
"mp4","MP4","mov","MOV","avi","AVI","mkv","MKV","webm","WEBM","flv","FLV","wmv","WMV",
],
},
{ name: "All files", extensions: ["*"] },
],
});
if (!selected || typeof selected !== "string") return;
try {
const info = await api.openVideo(selected);
setVideo(info);
setFrameIdx(0);
setPlaying(false);
setSpeed(1);
renderedIdx.current = -1;
// Reset per-video extraction context, then auto-adopt the per-video
// default folder. If the user already extracted frames into it in a
// prior session, listExtractedFrames rehydrates them immediately.
setBulkStart(null);
setBulkEnd(null);
await applyOutputFolder(info.default_output_folder);
} catch (e) {
setLoadError(String(e));
}
}
async function chooseCoco() {
const selected = await openDialog({
multiple: false,
filters: [{ name: "Annotations JSON", extensions: ["json"] }, { name: "All files", extensions: ["*"] }],
});
if (!selected || typeof selected !== "string") return;
try {
const imported = await api.importCoco(selected);
setCoco(imported);
// If a folder is already picked, write (or refresh) classes.txt.
if (outputFolder && imported.categories.length > 0) {
try { await api.writeClassesTxt(outputFolder, imported.categories); } catch (e) { console.error(e); }
}
} catch (e) {
setLoadError(String(e));
}
}
function clearCoco() { setCoco(null); }
// --- render ----------------------------------------------------------
if (!video) {
return (
<div className="upload-pane">
<button className="primary" onClick={chooseVideo}>Upload Video</button>
{loadError && <p className="error">{loadError}</p>}
<p className="dim">
Supported: mp4, mov, avi, mkv, webm, flv, wmv. Decoded via <code>ffmpeg</code>.
</p>
</div>
);
}
const t = frameIdx / video.fps;
const currentShapes = frameMap.get(frameIdx);
const isCurrentExtracted = extractedFrames.has(frameIdx);
const pct = bulkProgress ? Math.round((bulkProgress.done / bulkProgress.total) * 100) : 0;
return (
<div className="extract-mode">
{!infoCollapsed && (
<>
<div className="video-info">
<span className="path" title={video.path}>{video.path.split("/").pop()}</span>
<span className="dim">
{video.width}×{video.height} · {video.fps.toFixed(2)} fps · {video.total_frames} frames · {video.duration_s.toFixed(1)}s
</span>
<button className="link" onClick={chooseVideo}>change</button>
</div>
<div className="coco-info">
{coco ? (
<>
<span>
<span className="dim">{coco.stats.source_format === "sam" ? "SAM" : "COCO"}:</span>{" "}
<strong>{coco.stats.total_annotations}</strong> annotations ·{" "}
<strong>{coco.stats.frame_count}</strong> frames ·{" "}
<strong>{coco.stats.categories}</strong> categories
{coco.stats.skipped_rle > 0 && <span className="dim"> · {coco.stats.skipped_rle} RLE skipped</span>}
{coco.stats.skipped_no_image > 0 && <span className="dim"> · {coco.stats.skipped_no_image} unmappable</span>}
</span>
<button className="link" onClick={chooseCoco}>replace</button>
<button className="link" onClick={clearCoco}>clear</button>
</>
) : (
<>
<span className="dim">no annotations loaded</span>
<button className="link" onClick={chooseCoco}>import annotations</button>
</>
)}
</div>
</>
)}
<div className={`extract-info${infoCollapsed ? " collapsed" : ""}`}>
{outputFolder ? (
<>
<span>Output: <strong title={outputFolder}>{outputFolder.split("/").pop()}</strong></span>
<span className="dim"> · {extractedFrames.size} extracted</span>
<button className="link" onClick={pickOutputFolder}>change</button>
</>
) : (
<button className="link" onClick={pickOutputFolder}>choose output folder</button>
)}
<div className="extract-actions">
<button
className={isCurrentExtracted ? "extracted" : "primary-ghost"}
onClick={extractCurrentFrame}
disabled={!outputFolder}
title="E"
>
{isCurrentExtracted ? "✓ re-extract (E)" : "Extract (E)"}
</button>
<button
onClick={undoCurrentFrame}
disabled={!outputFolder || !isCurrentExtracted}
title="U"
>
Undo (U)
</button>
<div className="range-picker">
<button onClick={setRangeStart} title="[">
Start [ {bulkStart ?? ""}
</button>
<button onClick={setRangeEnd} title="]">
End ] {bulkEnd ?? ""}
</button>
{(bulkStart != null || bulkEnd != null) && (
<button className="link" onClick={clearRange}>clear</button>
)}
</div>
<button
onClick={runBulkExtract}
disabled={!outputFolder || !bulkRange || !!bulkProgress}
>
Bulk Extract{bulkRange && ` (${bulkRange.end - bulkRange.start + 1})`}
</button>
<button
onClick={runBulkUndo}
disabled={!outputFolder || !bulkRange || !!bulkProgress}
>
Bulk Undo
</button>
<button
className="collapse-toggle"
onClick={() => setInfoCollapsed((v) => !v)}
title={infoCollapsed ? "show info" : "hide info"}
>
{infoCollapsed ? "v" : "^"}
</button>
</div>
{bulkProgress && (
<div className="progress">
<div className="progress-bar" style={{ width: `${pct}%` }} />
<span>
{bulkProgress.phase === "extract" ? "extracting" : "undoing"} ·{" "}
{bulkProgress.done} / {bulkProgress.total}
</span>
</div>
)}
{bulkError && <p className="error">{bulkError}</p>}
</div>
<FrameCanvas ref={canvasRef} video={video} extracted={isCurrentExtracted} />
<div className="timeline">
<div className="timeline-track-wrap">
<input
type="range"
min={0}
max={video.total_frames - 1}
value={frameIdx}
onChange={(e) => seekToFrame(Number(e.target.value))}
style={
{
"--val": `${(frameIdx / Math.max(1, video.total_frames - 1)) * 100}%`,
} as React.CSSProperties
}
/>
{extractedFrames.size > 0 && (
<ExtractedTicks
extracted={extractedFrames}
totalFrames={video.total_frames}
/>
)}
</div>
<div className="timeline-readout">
frame <strong>{frameIdx}</strong> / {video.total_frames - 1}
<span className="dim"> · {formatTime(t)}</span>
{currentShapes && currentShapes.length > 0 && (
<span className="badge">
{currentShapes.length} annotation{currentShapes.length === 1 ? "" : "s"}
</span>
)}
{isCurrentExtracted && <span className="badge extracted"> extracted</span>}
</div>
</div>
<div className="controls">
<button className={playing ? "active" : ""} onClick={togglePlay}>
{playing ? "❚❚ pause" : "▶ play"}
</button>
<button onClick={() => stepFrames(-1)} title="←"> 1</button>
<button onClick={() => stepFrames(1)} title="→">1 </button>
<button onClick={() => stepFrames(-10)} title="Ctrl+←"> 10</button>
<button onClick={() => stepFrames(10)} title="Ctrl+→">10 </button>
<div className="speed">
<button onClick={() => bumpSpeed(-1)} title="↓"></button>
<span className="speed-readout">{speed}×</span>
<button onClick={() => bumpSpeed(1)} title="↑">+</button>
</div>
<label className={`review-toggle ${reviewMode ? "on" : ""}`}>
<input
type="checkbox"
checked={reviewMode}
onChange={(e) => setReviewMode(e.target.checked)}
/>
Review mode
</label>
<button
className={`overlay-toggle ${overlayVisible ? "on" : ""}`}
onClick={() => setOverlayVisible((v) => !v)}
title="H"
>
{overlayVisible ? "👁 overlay on" : "◌ overlay off"}
</button>
<span className="dim shortcuts">space · · ctrl+ · · H · E · U · [ ]</span>
</div>
</div>
);
}
function formatTime(s: number) {
const m = Math.floor(s / 60);
const r = s - m * 60;
return `${m}:${r.toFixed(2).padStart(5, "0")}`;
}
/**
* Canvas overlay that draws one green tick per pixel-column of the timeline
* track where at least one extracted frame falls. Canvas avoids the DOM
* blow-up of 1-div-per-frame on long videos, and collapses dense neighboring
* frames into a single visible stripe at narrow timeline widths.
*/
function ExtractedTicks({
extracted,
totalFrames,
}: {
extracted: Set<number>;
totalFrames: number;
}) {
const ref = useRef<HTMLCanvasElement>(null);
useEffect(() => {
const canvas = ref.current;
if (!canvas) return;
let raf = 0;
const draw = () => {
raf = 0;
const rect = canvas.getBoundingClientRect();
const dpr = window.devicePixelRatio || 1;
const w = Math.max(1, Math.floor(rect.width * dpr));
const h = Math.max(1, Math.floor(rect.height * dpr));
if (canvas.width !== w) canvas.width = w;
if (canvas.height !== h) canvas.height = h;
const ctx = canvas.getContext("2d");
if (!ctx) return;
ctx.clearRect(0, 0, w, h);
if (extracted.size === 0 || totalFrames <= 0) return;
ctx.fillStyle = "rgba(92, 207, 117, 0.85)";
const max = Math.max(1, totalFrames - 1);
for (const f of extracted) {
const x = Math.round((f / max) * (w - 1));
ctx.fillRect(x, 0, Math.max(1, Math.floor(dpr)), h);
}
};
const schedule = () => {
if (raf) return;
raf = requestAnimationFrame(draw);
};
schedule();
const ro = new ResizeObserver(schedule);
ro.observe(canvas);
return () => {
if (raf) cancelAnimationFrame(raf);
ro.disconnect();
};
}, [extracted, totalFrames]);
return <canvas ref={ref} className="timeline-ticks" aria-hidden="true" />;
}