urma-mcp 0.1.0-rc.1

package/dist/src/acquisition/storyboard.js

@@ -0,0 +1,324 @@
+import { readdir, readFile, writeFile } from "node:fs/promises";
+import path from "node:path";
+import { UrmaError } from "../core/errors.js";
+import { assertInterval, assertMs } from "../core/coverage.js";
+import { deterministicRequestKey } from "../core/request-key.js";
+import { candidateKeyForSourceFormat } from "../sources/candidates.js";
+import { Ffmpeg } from "../subprocess/ffmpeg.js";
+import { YtDlp } from "../subprocess/ytdlp.js";
+import { collectBinaryVersions } from "../subprocess/versions.js";
+import { BlobStore } from "../store/blob-store.js";
+import { URMA_VERSION } from "../version.js";
+import { assertExpectedRemoteBytes, assertRemoteDirectoryWithinBudget, withRemoteAcquisitionDirectory, } from "./remote-budget.js";
+import { startAcquisition } from "./records.js";
+const MAX_STORYBOARD_CELLS = 100_000;
+function headerEnd(buffer, start) {
+    const crlf = buffer.indexOf(Buffer.from("\r\n\r\n"), start);
+    const lf = buffer.indexOf(Buffer.from("\n\n"), start);
+    if (crlf < 0 && lf < 0)
+        return null;
+    return crlf >= 0 && (lf < 0 || crlf <= lf)
+        ? { index: crlf, length: 4 }
+        : { index: lf, length: 2 };
+}
+export function parseMhtmlImages(buffer) {
+    const root = headerEnd(buffer, 0);
+    if (!root) {
+        throw new UrmaError("MEDIA_INVALID", "Storyboard MHTML root headers are incomplete");
+    }
+    const boundary = buffer
+        .subarray(0, root.index)
+        .toString("utf8")
+        .match(/boundary="?([^";\r\n]+)"?/i)?.[1];
+    if (!boundary) {
+        throw new UrmaError("MEDIA_INVALID", "Storyboard MHTML multipart boundary is missing");
+    }
+    const marker = Buffer.from(`--${boundary}`);
+    const parts = [];
+    let cursor = buffer.indexOf(marker, root.index);
+    while (cursor >= 0) {
+        const start = cursor + marker.length;
+        if (buffer.subarray(start, start + 2).toString("ascii") === "--")
+            break;
+        const end = headerEnd(buffer, start);
+        if (!end)
+            break;
+        const headers = buffer.subarray(start, end.index).toString("utf8");
+        const length = Number(headers.match(/Content-Length:\s*(\d+)/i)?.[1]);
+        const contentType = headers.match(/Content-Type:\s*([^;\r\n]+)/i)?.[1]?.trim() ??
+            "application/octet-stream";
+        const bodyStart = end.index + end.length;
+        if (!Number.isFinite(length) ||
+            length < 0 ||
+            bodyStart + length > buffer.length) {
+            throw new UrmaError("MEDIA_INVALID", "Storyboard MHTML part length is invalid");
+        }
+        if (contentType.startsWith("image/")) {
+            const seconds = Number(headers.match(/X\.yt-dlp\.Duration:\s*([0-9.]+)/i)?.[1]);
+            parts.push({
+                contentType,
+                durationMs: Number.isFinite(seconds) ? Math.round(seconds * 1000) : 0,
+                bytes: buffer.subarray(bodyStart, bodyStart + length),
+            });
+        }
+        cursor = buffer.indexOf(marker, bodyStart + length);
+    }
+    if (parts.length === 0) {
+        throw new UrmaError("MEDIA_INVALID", "Storyboard MHTML contains no image parts");
+    }
+    return parts;
+}
+function layoutFor(format) {
+    const width = format.width;
+    const height = format.height;
+    const fps = format.fps;
+    const columns = format.columns ?? 10;
+    const rows = format.rows ?? 10;
+    if (width === null ||
+        height === null ||
+        fps === null ||
+        !Number.isSafeInteger(width) ||
+        width < 1 ||
+        !Number.isSafeInteger(height) ||
+        height < 1 ||
+        !Number.isFinite(fps) ||
+        fps <= 0 ||
+        !Number.isSafeInteger(columns) ||
+        columns < 1 ||
+        !Number.isSafeInteger(rows) ||
+        rows < 1 ||
+        columns * rows > MAX_STORYBOARD_CELLS) {
+        throw new UrmaError("STORYBOARD_UNAVAILABLE", "Source advertises storyboard geometry that cannot be sampled safely");
+    }
+    return {
+        width,
+        height,
+        fps,
+        columns,
+        rows,
+        intervalMs: Math.max(1, Math.round(1000 / fps)),
+    };
+}
+function storyboardCells(parts, layout, sourceDurationMs) {
+    const cells = [];
+    let startMs = 0;
+    for (const [partIndex, part] of parts.entries()) {
+        if (!Number.isSafeInteger(part.durationMs) || part.durationMs < 1) {
+            throw new UrmaError("STORYBOARD_UNAVAILABLE", "Storyboard sheet has no positive finite duration; the source cannot be mapped safely");
+        }
+        const partEndMs = startMs + part.durationMs;
+        if (!Number.isSafeInteger(partEndMs)) {
+            throw new UrmaError("STORYBOARD_UNAVAILABLE", "Storyboard sheet durations exceed the safe timestamp range");
+        }
+        const count = Math.min(layout.columns * layout.rows, Math.max(1, Math.ceil(part.durationMs / layout.intervalMs)));
+        for (let cell = 0; cell < count; cell += 1) {
+            const timestampMs = startMs + cell * layout.intervalMs;
+            if (timestampMs >= partEndMs || timestampMs >= sourceDurationMs)
+                break;
+            cells.push({ partIndex, cell, timestampMs });
+        }
+        startMs = partEndMs;
+    }
+    return cells;
+}
+function lowerBound(cells, targetMs) {
+    let low = 0;
+    let high = cells.length;
+    while (low < high) {
+        const middle = Math.floor((low + high) / 2);
+        if (cells[middle].timestampMs < targetMs)
+            low = middle + 1;
+        else
+            high = middle;
+    }
+    return low;
+}
+function selectStoryboardCells(cells, requested, scope) {
+    assertInterval(scope.startMs, scope.endMs);
+    for (const point of requested) {
+        const timestampMs = assertMs(point, "storyboard request point");
+        if (timestampMs < scope.startMs || timestampMs >= scope.endMs) {
+            throw new RangeError(`Storyboard request point ${timestampMs} is outside [${scope.startMs},${scope.endMs})`);
+        }
+    }
+    const available = cells.filter((cell) => cell.timestampMs >= scope.startMs && cell.timestampMs < scope.endMs);
+    if (available.length === 0) {
+        throw new UrmaError("STORYBOARD_UNAVAILABLE", `Storyboard has no representative cell inside [${scope.startMs},${scope.endMs}); use another bounded evidence source`);
+    }
+    const selected = new Map();
+    for (const targetMs of [...requested].sort((a, b) => a - b)) {
+        const index = lowerBound(available, targetMs);
+        let candidate = index > 0 ? available[index - 1] : available[0];
+        if (candidate && candidate.timestampMs < scope.startMs) {
+            candidate = available[index];
+        }
+        if (candidate &&
+            candidate.timestampMs >= scope.startMs &&
+            candidate.timestampMs < scope.endMs) {
+            selected.set(`${candidate.partIndex}:${candidate.cell}`, candidate);
+        }
+    }
+    return [...selected.values()].sort((a, b) => a.timestampMs - b.timestampMs ||
+        a.partIndex - b.partIndex ||
+        a.cell - b.cell);
+}
+export class StoryboardAcquirer {
+    config;
+    store;
+    blobs;
+    remoteContext;
+    constructor(config, store, blobs, remoteContext = null) {
+        this.config = config;
+        this.store = store;
+        this.blobs = blobs;
+        this.remoteContext = remoteContext;
+    }
+    async cells(source, investigationRef, requested, working, signal, scope = { startMs: 0, endMs: source.durationMs }) {
+        assertInterval(scope.startMs, scope.endMs, source.durationMs);
+        const format = [...source.formats]
+            .filter((item) => item.ext === "mhtml" || item.protocol === "mhtml")
+            .sort((a, b) => (b.width ?? 0) - (a.width ?? 0))[0];
+        if (!format) {
+            throw new UrmaError("STORYBOARD_UNAVAILABLE", "Source advertises no usable timestamped native storyboard");
+        }
+        const layout = layoutFor(format);
+        const requestKey = deterministicRequestKey(source.revision, "storyboard", {
+            candidateKey: candidateKeyForSourceFormat(source, format),
+            formatId: format.id,
+        }, "storyboard-parser");
+        let artifact = this.store.getArtifactByRequest(requestKey);
+        let bytes;
+        let cacheHit = false;
+        const acquisitionBudget = Math.min(this.config.limits.maxNavigationCopyBytes, this.config.limits.resourceBytes);
+        if (artifact) {
+            try {
+                bytes = await this.blobs.read(artifact.artifactId, artifact.blobPath, acquisitionBudget);
+                cacheHit = true;
+            }
+            catch {
+                artifact = null;
+            }
+        }
+        if (!artifact) {
+            const acquisition = startAcquisition(this.store, {
+                sourceRef: source.sourceRef,
+                sourceRevision: source.revision,
+                investigationRef,
+                operation: "acquire-storyboard",
+                requestKey,
+                method: "yt-dlp-storyboard",
+                debug: this.config.debug,
+            });
+            try {
+                const budget = acquisitionBudget;
+                assertExpectedRemoteBytes(format.estimatedBytes, budget, "storyboard");
+                artifact = await withRemoteAcquisitionDirectory(this.config, "storyboard", budget, signal, async (temporary, remoteSignal) => {
+                    const ytdlp = new YtDlp(this.config, undefined, this.remoteContext);
+                    const exact = await ytdlp.exactFormatSnapshot(source, format, remoteSignal);
+                    const infoPath = path.join(temporary, "snapshot-info.json");
+                    await writeFile(infoPath, JSON.stringify(exact.metadata), "utf8");
+                    await ytdlp.run([
+                        "--load-info-json",
+                        infoPath,
+                        "-f",
+                        exact.formatId,
+                        "--paths",
+                        temporary,
+                        "-o",
+                        "storyboard.%(ext)s",
+                    ], {
+                        signal: remoteSignal,
+                        timeoutMs: this.config.limits.maxRemoteAcquisitionWallMs,
+                        remote: true,
+                    });
+                    await assertRemoteDirectoryWithinBudget(temporary, budget, "storyboard");
+                    const names = (await readdir(temporary)).filter((item) => /^[A-Za-z0-9_.-]{1,200}\.mhtml$/u.test(item));
+                    if (names.length !== 1) {
+                        throw new UrmaError("STORYBOARD_UNAVAILABLE", names.length === 0
+                            ? "yt-dlp completed without the requested MHTML storyboard"
+                            : "yt-dlp completed with ambiguous MHTML storyboard outputs");
+                    }
+                    bytes = await readFile(path.join(temporary, names[0]));
+                    const parts = parseMhtmlImages(bytes);
+                    storyboardCells(parts, layout, source.durationMs);
+                    const blob = await this.blobs.put(bytes);
+                    const versions = await collectBinaryVersions(this.config, ["ytdlp"], remoteSignal);
+                    const stored = {
+                        artifactId: blob.artifactId,
+                        sourceRef: source.sourceRef,
+                        sourceRevision: source.revision,
+                        kind: "storyboard",
+                        role: "locator",
+                        mimeType: "multipart/related",
+                        sha256: blob.sha256,
+                        byteSize: blob.byteSize,
+                        blobPath: blob.relativePath,
+                        startMs: 0,
+                        endMs: source.durationMs,
+                        params: {
+                            candidateKey: candidateKeyForSourceFormat(source, format),
+                            formatId: format.id,
+                        },
+                        producer: {
+                            version: "storyboard-parser",
+                            urmaVersion: URMA_VERSION,
+                            ...versions,
+                            cellWidth: format.width,
+                            cellHeight: format.height,
+                            fps: format.fps,
+                            columns: format.columns,
+                            rows: format.rows,
+                        },
+                        createdAt: new Date().toISOString(),
+                    };
+                    this.store.putArtifact(stored, {
+                        requestKey,
+                        operation: "storyboard",
+                    });
+                    acquisition.succeed({
+                        networkBytes: null,
+                        networkAccountingComplete: false,
+                        metadata: { artifactId: stored.artifactId, ...versions },
+                    });
+                    return stored;
+                });
+            }
+            catch (error) {
+                acquisition.fail(error);
+                throw error;
+            }
+        }
+        const parts = parseMhtmlImages(bytes);
+        const selected = selectStoryboardCells(storyboardCells(parts, layout, source.durationMs), requested, scope);
+        const emittedSprites = new Map();
+        const paths = [];
+        const pointsMs = [];
+        const ffmpeg = new Ffmpeg(this.config);
+        const samples = [];
+        for (const [index, cell] of selected.entries()) {
+            let sprite = emittedSprites.get(cell.partIndex);
+            if (!sprite) {
+                sprite = path.join(working, `sprite-${cell.partIndex}.img`);
+                await writeFile(sprite, parts[cell.partIndex].bytes);
+                emittedSprites.set(cell.partIndex, sprite);
+            }
+            const output = path.join(working, `cell-${index}.jpg`);
+            await ffmpeg.cropJpeg(sprite, {
+                width: layout.width,
+                height: layout.height,
+                x: (cell.cell % layout.columns) * layout.width,
+                y: Math.floor(cell.cell / layout.columns) * layout.height,
+            }, output, signal);
+            await ffmpeg.validateJpeg(output);
+            paths.push(output);
+            pointsMs.push(cell.timestampMs);
+            samples.push({
+                timestampMs: cell.timestampMs,
+                fragmentIndex: cell.partIndex,
+                cellIndex: cell.cell,
+            });
+        }
+        return { paths, pointsMs, samples, artifact, cacheHit };
+    }
+}
+//# sourceMappingURL=storyboard.js.map

package/dist/src/acquisition/storyboard.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"storyboard.js","sourceRoot":"","sources":["../../../src/acquisition/storyboard.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,OAAO,EAAE,QAAQ,EAAE,SAAS,EAAE,MAAM,kBAAkB,CAAC;AAChE,OAAO,IAAI,MAAM,WAAW,CAAC;AAE7B,OAAO,EAAE,SAAS,EAAE,MAAM,mBAAmB,CAAC;AAC9C,OAAO,EAAE,cAAc,EAAE,QAAQ,EAAE,MAAM,qBAAqB,CAAC;AAE/D,OAAO,EAAE,uBAAuB,EAAE,MAAM,wBAAwB,CAAC;AAEjE,OAAO,EAAE,2BAA2B,EAAE,MAAM,0BAA0B,CAAC;AACvE,OAAO,EAAE,MAAM,EAAE,MAAM,yBAAyB,CAAC;AACjD,OAAO,EAAE,KAAK,EAAE,MAAM,wBAAwB,CAAC;AAE/C,OAAO,EAAE,qBAAqB,EAAE,MAAM,2BAA2B,CAAC;AAClE,OAAO,EAAE,SAAS,EAAE,MAAM,wBAAwB,CAAC;AAEnD,OAAO,EAAE,YAAY,EAAE,MAAM,eAAe,CAAC;AAC7C,OAAO,EACL,yBAAyB,EACzB,iCAAiC,EACjC,8BAA8B,GAC/B,MAAM,oBAAoB,CAAC;AAC5B,OAAO,EAAE,gBAAgB,EAAE,MAAM,cAAc,CAAC;AA2BhD,MAAM,oBAAoB,GAAG,OAAO,CAAC;AAErC,SAAS,SAAS,CAChB,MAAc,EACd,KAAa;IAEb,MAAM,IAAI,GAAG,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,IAAI,CAAC,UAAU,CAAC,EAAE,KAAK,CAAC,CAAC;IAC5D,MAAM,EAAE,GAAG,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,KAAK,CAAC,CAAC;IACtD,IAAI,IAAI,GAAG,CAAC,IAAI,EAAE,GAAG,CAAC;QAAE,OAAO,IAAI,CAAC;IACpC,OAAO,IAAI,IAAI,CAAC,IAAI,CAAC,EAAE,GAAG,CAAC,IAAI,IAAI,IAAI,EAAE,CAAC;QACxC,CAAC,CAAC,EAAE,KAAK,EAAE,IAAI,EAAE,MAAM,EAAE,CAAC,EAAE;QAC5B,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE,MAAM,EAAE,CAAC,EAAE,CAAC;AAC/B,CAAC;AAED,MAAM,UAAU,gBAAgB,CAAC,MAAc;IAC7C,MAAM,IAAI,GAAG,SAAS,CAAC,MAAM,EAAE,CAAC,CAAC,CAAC;IAClC,IAAI,CAAC,IAAI,EAAE,CAAC;QACV,MAAM,IAAI,SAAS,CACjB,eAAe,EACf,8CAA8C,CAC/C,CAAC;IACJ,CAAC;IACD,MAAM,QAAQ,GAAG,MAAM;SACpB,QAAQ,CAAC,CAAC,EAAE,IAAI,CAAC,KAAK,CAAC;SACvB,QAAQ,CAAC,MAAM,CAAC;SAChB,KAAK,CAAC,4BAA4B,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;IAC5C,IAAI,CAAC,QAAQ,EAAE,CAAC;QACd,MAAM,IAAI,SAAS,CACjB,eAAe,EACf,gDAAgD,CACjD,CAAC;IACJ,CAAC;IACD,MAAM,MAAM,GAAG,MAAM,CAAC,IAAI,CAAC,KAAK,QAAQ,EAAE,CAAC,CAAC;IAC5C,MAAM,KAAK,GAAqB,EAAE,CAAC;IACnC,IAAI,MAAM,GAAG,MAAM,CAAC,OAAO,CAAC,MAAM,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC;IAChD,OAAO,MAAM,IAAI,CAAC,EAAE,CAAC;QACnB,MAAM,KAAK,GAAG,MAAM,GAAG,MAAM,CAAC,MAAM,CAAC;QACrC,IAAI,MAAM,CAAC,QAAQ,CAAC,KAAK,EAAE,KAAK,GAAG,CAAC,CAAC,CAAC,QAAQ,CAAC,OAAO,CAAC,KAAK,IAAI;YAAE,MAAM;QACxE,MAAM,GAAG,GAAG,SAAS,CAAC,MAAM,EAAE,KAAK,CAAC,CAAC;QACrC,IAAI,CAAC,GAAG;YAAE,MAAM;QAChB,MAAM,OAAO,GAAG,MAAM,CAAC,QAAQ,CAAC,KAAK,EAAE,GAAG,CAAC,KAAK,CAAC,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC;QACnE,MAAM,MAAM,GAAG,MAAM,CAAC,OAAO,CAAC,KAAK,CAAC,0BAA0B,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;QACtE,MAAM,WAAW,GACf,OAAO,CAAC,KAAK,CAAC,8BAA8B,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE;YACxD,0BAA0B,CAAC;QAC/B,MAAM,SAAS,GAAG,GAAG,CAAC,KAAK,GAAG,GAAG,CAAC,MAAM,CAAC;QACzC,IACE,CAAC,MAAM,CAAC,QAAQ,CAAC,MAAM,CAAC;YACxB,MAAM,GAAG,CAAC;YACV,SAAS,GAAG,MAAM,GAAG,MAAM,CAAC,MAAM,EAClC,CAAC;YACD,MAAM,IAAI,SAAS,CACjB,eAAe,EACf,yCAAyC,CAC1C,CAAC;QACJ,CAAC;QACD,IAAI,WAAW,CAAC,UAAU,CAAC,QAAQ,CAAC,EAAE,CAAC;YACrC,MAAM,OAAO,GAAG,MAAM,CACpB,OAAO,CAAC,KAAK,CAAC,mCAAmC,CAAC,EAAE,CAAC,CAAC,CAAC,CACxD,CAAC;YACF,KAAK,CAAC,IAAI,CAAC;gBACT,WAAW;gBACX,UAAU,EAAE,MAAM,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,OAAO,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC;gBACrE,KAAK,EAAE,MAAM,CAAC,QAAQ,CAAC,SAAS,EAAE,SAAS,GAAG,MAAM,CAAC;aACtD,CAAC,CAAC;QACL,CAAC;QACD,MAAM,GAAG,MAAM,CAAC,OAAO,CAAC,MAAM,EAAE,SAAS,GAAG,MAAM,CAAC,CAAC;IACtD,CAAC;IACD,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACvB,MAAM,IAAI,SAAS,CACjB,eAAe,EACf,0CAA0C,CAC3C,CAAC;IACJ,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED,SAAS,SAAS,CAChB,MAAsD;IAEtD,MAAM,KAAK,GAAG,MAAM,CAAC,KAAK,CAAC;IAC3B,MAAM,MAAM,GAAG,MAAM,CAAC,MAAM,CAAC;IAC7B,MAAM,GAAG,GAAG,MAAM,CAAC,GAAG,CAAC;IACvB,MAAM,OAAO,GAAG,MAAM,CAAC,OAAO,IAAI,EAAE,CAAC;IACrC,MAAM,IAAI,GAAG,MAAM,CAAC,IAAI,IAAI,EAAE,CAAC;IAC/B,IACE,KAAK,KAAK,IAAI;QACd,MAAM,KAAK,IAAI;QACf,GAAG,KAAK,IAAI;QACZ,CAAC,MAAM,CAAC,aAAa,CAAC,KAAK,CAAC;QAC5B,KAAK,GAAG,CAAC;QACT,CAAC,MAAM,CAAC,aAAa,CAAC,MAAM,CAAC;QAC7B,MAAM,GAAG,CAAC;QACV,CAAC,MAAM,CAAC,QAAQ,CAAC,GAAG,CAAC;QACrB,GAAG,IAAI,CAAC;QACR,CAAC,MAAM,CAAC,aAAa,CAAC,OAAO,CAAC;QAC9B,OAAO,GAAG,CAAC;QACX,CAAC,MAAM,CAAC,aAAa,CAAC,IAAI,CAAC;QAC3B,IAAI,GAAG,CAAC;QACR,OAAO,GAAG,IAAI,GAAG,oBAAoB,EACrC,CAAC;QACD,MAAM,IAAI,SAAS,CACjB,wBAAwB,EACxB,qEAAqE,CACtE,CAAC;IACJ,CAAC;IACD,OAAO;QACL,KAAK;QACL,MAAM;QACN,GAAG;QACH,OAAO;QACP,IAAI;QACJ,UAAU,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,KAAK,CAAC,IAAI,GAAG,GAAG,CAAC,CAAC;KAChD,CAAC;AACJ,CAAC;AAED,SAAS,eAAe,CACtB,KAAgC,EAChC,MAAwB,EACxB,gBAAwB;IAExB,MAAM,KAAK,GAAqB,EAAE,CAAC;IACnC,IAAI,OAAO,GAAG,CAAC,CAAC;IAChB,KAAK,MAAM,CAAC,SAAS,EAAE,IAAI,CAAC,IAAI,KAAK,CAAC,OAAO,EAAE,EAAE,CAAC;QAChD,IAAI,CAAC,MAAM,CAAC,aAAa,CAAC,IAAI,CAAC,UAAU,CAAC,IAAI,IAAI,CAAC,UAAU,GAAG,CAAC,EAAE,CAAC;YAClE,MAAM,IAAI,SAAS,CACjB,wBAAwB,EACxB,sFAAsF,CACvF,CAAC;QACJ,CAAC;QACD,MAAM,SAAS,GAAG,OAAO,GAAG,IAAI,CAAC,UAAU,CAAC;QAC5C,IAAI,CAAC,MAAM,CAAC,aAAa,CAAC,SAAS,CAAC,EAAE,CAAC;YACrC,MAAM,IAAI,SAAS,CACjB,wBAAwB,EACxB,4DAA4D,CAC7D,CAAC;QACJ,CAAC;QACD,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,CACpB,MAAM,CAAC,OAAO,GAAG,MAAM,CAAC,IAAI,EAC5B,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,UAAU,GAAG,MAAM,CAAC,UAAU,CAAC,CAAC,CAC5D,CAAC;QACF,KAAK,IAAI,IAAI,GAAG,CAAC,EAAE,IAAI,GAAG,KAAK,EAAE,IAAI,IAAI,CAAC,EAAE,CAAC;YAC3C,MAAM,WAAW,GAAG,OAAO,GAAG,IAAI,GAAG,MAAM,CAAC,UAAU,CAAC;YACvD,IAAI,WAAW,IAAI,SAAS,IAAI,WAAW,IAAI,gBAAgB;gBAAE,MAAM;YACvE,KAAK,CAAC,IAAI,CAAC,EAAE,SAAS,EAAE,IAAI,EAAE,WAAW,EAAE,CAAC,CAAC;QAC/C,CAAC;QACD,OAAO,GAAG,SAAS,CAAC;IACtB,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED,SAAS,UAAU,CACjB,KAAgC,EAChC,QAAgB;IAEhB,IAAI,GAAG,GAAG,CAAC,CAAC;IACZ,IAAI,IAAI,GAAG,KAAK,CAAC,MAAM,CAAC;IACxB,OAAO,GAAG,GAAG,IAAI,EAAE,CAAC;QAClB,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC;QAC5C,IAAI,KAAK,CAAC,MAAM,CAAE,CAAC,WAAW,GAAG,QAAQ;YAAE,GAAG,GAAG,MAAM,GAAG,CAAC,CAAC;;YACvD,IAAI,GAAG,MAAM,CAAC;IACrB,CAAC;IACD,OAAO,GAAG,CAAC;AACb,CAAC;AAED,SAAS,qBAAqB,CAC5B,KAAgC,EAChC,SAA4B,EAC5B,KAAsB;IAEtB,cAAc,CAAC,KAAK,CAAC,OAAO,EAAE,KAAK,CAAC,KAAK,CAAC,CAAC;IAC3C,KAAK,MAAM,KAAK,IAAI,SAAS,EAAE,CAAC;QAC9B,MAAM,WAAW,GAAG,QAAQ,CAAC,KAAK,EAAE,0BAA0B,CAAC,CAAC;QAChE,IAAI,WAAW,GAAG,KAAK,CAAC,OAAO,IAAI,WAAW,IAAI,KAAK,CAAC,KAAK,EAAE,CAAC;YAC9D,MAAM,IAAI,UAAU,CAClB,4BAA4B,WAAW,gBAAgB,KAAK,CAAC,OAAO,IAAI,KAAK,CAAC,KAAK,GAAG,CACvF,CAAC;QACJ,CAAC;IACH,CAAC;IACD,MAAM,SAAS,GAAG,KAAK,CAAC,MAAM,CAC5B,CAAC,IAAI,EAAE,EAAE,CACP,IAAI,CAAC,WAAW,IAAI,KAAK,CAAC,OAAO,IAAI,IAAI,CAAC,WAAW,GAAG,KAAK,CAAC,KAAK,CACtE,CAAC;IACF,IAAI,SAAS,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QAC3B,MAAM,IAAI,SAAS,CACjB,wBAAwB,EACxB,iDAAiD,KAAK,CAAC,OAAO,IAAI,KAAK,CAAC,KAAK,wCAAwC,CACtH,CAAC;IACJ,CAAC;IACD,MAAM,QAAQ,GAAG,IAAI,GAAG,EAA0B,CAAC;IACnD,KAAK,MAAM,QAAQ,IAAI,CAAC,GAAG,SAAS,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC;QAC5D,MAAM,KAAK,GAAG,UAAU,CAAC,SAAS,EAAE,QAAQ,CAAC,CAAC;QAC9C,IAAI,SAAS,GAAG,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC;QAChE,IAAI,SAAS,IAAI,SAAS,CAAC,WAAW,GAAG,KAAK,CAAC,OAAO,EAAE,CAAC;YACvD,SAAS,GAAG,SAAS,CAAC,KAAK,CAAC,CAAC;QAC/B,CAAC;QACD,IACE,SAAS;YACT,SAAS,CAAC,WAAW,IAAI,KAAK,CAAC,OAAO;YACtC,SAAS,CAAC,WAAW,GAAG,KAAK,CAAC,KAAK,EACnC,CAAC;YACD,QAAQ,CAAC,GAAG,CAAC,GAAG,SAAS,CAAC,SAAS,IAAI,SAAS,CAAC,IAAI,EAAE,EAAE,SAAS,CAAC,CAAC;QACtE,CAAC;IACH,CAAC;IACD,OAAO,CAAC,GAAG,QAAQ,CAAC,MAAM,EAAE,CAAC,CAAC,IAAI,CAChC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CACP,CAAC,CAAC,WAAW,GAAG,CAAC,CAAC,WAAW;QAC7B,CAAC,CAAC,SAAS,GAAG,CAAC,CAAC,SAAS;QACzB,CAAC,CAAC,IAAI,GAAG,CAAC,CAAC,IAAI,CAClB,CAAC;AACJ,CAAC;AAED,MAAM,OAAO,kBAAkB;IAElB,MAAM;IACN,KAAK;IACL,KAAK;IACL,aAAa;IAJxB,YACW,MAAkB,EAClB,KAAgB,EAChB,KAAgB,EAChB,aAAa,GAAkC,IAAI;sBAHnD,MAAM;qBACN,KAAK;qBACL,KAAK;6BACL,aAAa;IACrB,CAAC;IAEJ,KAAK,CAAC,KAAK,CACT,MAAsB,EACtB,gBAAkC,EAClC,SAA4B,EAC5B,OAAe,EACf,MAAoB,EACpB,KAAK,GAAoB,EAAE,OAAO,EAAE,CAAC,EAAE,KAAK,EAAE,MAAM,CAAC,UAAU,EAAE;QAQjE,cAAc,CAAC,KAAK,CAAC,OAAO,EAAE,KAAK,CAAC,KAAK,EAAE,MAAM,CAAC,UAAU,CAAC,CAAC;QAC9D,MAAM,MAAM,GAAG,CAAC,GAAG,MAAM,CAAC,OAAO,CAAC;aAC/B,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,GAAG,KAAK,OAAO,IAAI,IAAI,CAAC,QAAQ,KAAK,OAAO,CAAC;aACnE,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,KAAK,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,KAAK,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;QACtD,IAAI,CAAC,MAAM,EAAE,CAAC;YACZ,MAAM,IAAI,SAAS,CACjB,wBAAwB,EACxB,2DAA2D,CAC5D,CAAC;QACJ,CAAC;QACD,MAAM,MAAM,GAAG,SAAS,CAAC,MAAM,CAAC,CAAC;QACjC,MAAM,UAAU,GAAG,uBAAuB,CACxC,MAAM,CAAC,QAAQ,EACf,YAAY,EACZ;YACE,YAAY,EAAE,2BAA2B,CAAC,MAAM,EAAE,MAAM,CAAC;YACzD,QAAQ,EAAE,MAAM,CAAC,EAAE;SACpB,EACD,mBAAmB,CACpB,CAAC;QACF,IAAI,QAAQ,GAAG,IAAI,CAAC,KAAK,CAAC,oBAAoB,CAAC,UAAU,CAAC,CAAC;QAC3D,IAAI,KAAa,CAAC;QAClB,IAAI,QAAQ,GAAG,KAAK,CAAC;QACrB,MAAM,iBAAiB,GAAG,IAAI,CAAC,GAAG,CAChC,IAAI,CAAC,MAAM,CAAC,MAAM,CAAC,sBAAsB,EACzC,IAAI,CAAC,MAAM,CAAC,MAAM,CAAC,aAAa,CACjC,CAAC;QACF,IAAI,QAAQ,EAAE,CAAC;YACb,IAAI,CAAC;gBACH,KAAK,GAAG,MAAM,IAAI,CAAC,KAAK,CAAC,IAAI,CAC3B,QAAQ,CAAC,UAAU,EACnB,QAAQ,CAAC,QAAQ,EACjB,iBAAiB,CAClB,CAAC;gBACF,QAAQ,GAAG,IAAI,CAAC;YAClB,CAAC;YAAC,MAAM,CAAC;gBACP,QAAQ,GAAG,IAAI,CAAC;YAClB,CAAC;QACH,CAAC;QACD,IAAI,CAAC,QAAQ,EAAE,CAAC;YACd,MAAM,WAAW,GAAG,gBAAgB,CAAC,IAAI,CAAC,KAAK,EAAE;gBAC/C,SAAS,EAAE,MAAM,CAAC,SAAS;gBAC3B,cAAc,EAAE,MAAM,CAAC,QAAQ;gBAC/B,gBAAgB;gBAChB,SAAS,EAAE,oBAAoB;gBAC/B,UAAU;gBACV,MAAM,EAAE,mBAAmB;gBAC3B,KAAK,EAAE,IAAI,CAAC,MAAM,CAAC,KAAK;aACzB,CAAC,CAAC;YACH,IAAI,CAAC;gBACH,MAAM,MAAM,GAAG,iBAAiB,CAAC;gBACjC,yBAAyB,CAAC,MAAM,CAAC,cAAc,EAAE,MAAM,EAAE,YAAY,CAAC,CAAC;gBACvE,QAAQ,GAAG,MAAM,8BAA8B,CAC7C,IAAI,CAAC,MAAM,EACX,YAAY,EACZ,MAAM,EACN,MAAM,EACN,KAAK,EAAE,SAAS,EAAE,YAAY,EAAE,EAAE;oBAChC,MAAM,KAAK,GAAG,IAAI,KAAK,CAAC,IAAI,CAAC,MAAM,EAAE,SAAS,EAAE,IAAI,CAAC,aAAa,CAAC,CAAC;oBACpE,MAAM,KAAK,GAAG,MAAM,KAAK,CAAC,mBAAmB,CAC3C,MAAM,EACN,MAAM,EACN,YAAY,CACb,CAAC;oBACF,MAAM,QAAQ,GAAG,IAAI,CAAC,IAAI,CAAC,SAAS,EAAE,oBAAoB,CAAC,CAAC;oBAC5D,MAAM,SAAS,CAAC,QAAQ,EAAE,IAAI,CAAC,SAAS,CAAC,KAAK,CAAC,QAAQ,CAAC,EAAE,MAAM,CAAC,CAAC;oBAClE,MAAM,KAAK,CAAC,GAAG,CACb;wBACE,kBAAkB;wBAClB,QAAQ;wBACR,IAAI;wBACJ,KAAK,CAAC,QAAQ;wBACd,SAAS;wBACT,SAAS;wBACT,IAAI;wBACJ,oBAAoB;qBACrB,EACD;wBACE,MAAM,EAAE,YAAY;wBACpB,SAAS,EAAE,IAAI,CAAC,MAAM,CAAC,MAAM,CAAC,0BAA0B;wBACxD,MAAM,EAAE,IAAI;qBACb,CACF,CAAC;oBACF,MAAM,iCAAiC,CACrC,SAAS,EACT,MAAM,EACN,YAAY,CACb,CAAC;oBACF,MAAM,KAAK,GAAG,CAAC,MAAM,OAAO,CAAC,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CACvD,iCAAiC,CAAC,IAAI,CAAC,IAAI,CAAC,CAC7C,CAAC;oBACF,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;wBACvB,MAAM,IAAI,SAAS,CACjB,wBAAwB,EACxB,KAAK,CAAC,MAAM,KAAK,CAAC;4BAChB,CAAC,CAAC,yDAAyD;4BAC3D,CAAC,CAAC,0DAA0D,CAC/D,CAAC;oBACJ,CAAC;oBACD,KAAK,GAAG,MAAM,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,SAAS,EAAE,KAAK,CAAC,CAAC,CAAE,CAAC,CAAC,CAAC;oBACxD,MAAM,KAAK,GAAG,gBAAgB,CAAC,KAAK,CAAC,CAAC;oBACtC,eAAe,CAAC,KAAK,EAAE,MAAM,EAAE,MAAM,CAAC,UAAU,CAAC,CAAC;oBAClD,MAAM,IAAI,GAAG,MAAM,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;oBACzC,MAAM,QAAQ,GAAG,MAAM,qBAAqB,CAC1C,IAAI,CAAC,MAAM,EACX,CAAC,OAAO,CAAC,EACT,YAAY,CACb,CAAC;oBACF,MAAM,MAAM,GAAmB;wBAC7B,UAAU,EAAE,IAAI,CAAC,UAAU;wBAC3B,SAAS,EAAE,MAAM,CAAC,SAAS;wBAC3B,cAAc,EAAE,MAAM,CAAC,QAAQ;wBAC/B,IAAI,EAAE,YAAY;wBAClB,IAAI,EAAE,SAAS;wBACf,QAAQ,EAAE,mBAAmB;wBAC7B,MAAM,EAAE,IAAI,CAAC,MAAM;wBACnB,QAAQ,EAAE,IAAI,CAAC,QAAQ;wBACvB,QAAQ,EAAE,IAAI,CAAC,YAAY;wBAC3B,OAAO,EAAE,CAAC;wBACV,KAAK,EAAE,MAAM,CAAC,UAAU;wBACxB,MAAM,EAAE;4BACN,YAAY,EAAE,2BAA2B,CAAC,MAAM,EAAE,MAAM,CAAC;4BACzD,QAAQ,EAAE,MAAM,CAAC,EAAE;yBACpB;wBACD,QAAQ,EAAE;4BACR,OAAO,EAAE,mBAAmB;4BAC5B,WAAW,EAAE,YAAY;4BACzB,GAAG,QAAQ;4BACX,SAAS,EAAE,MAAM,CAAC,KAAK;4BACvB,UAAU,EAAE,MAAM,CAAC,MAAM;4BACzB,GAAG,EAAE,MAAM,CAAC,GAAG;4BACf,OAAO,EAAE,MAAM,CAAC,OAAO;4BACvB,IAAI,EAAE,MAAM,CAAC,IAAI;yBAClB;wBACD,SAAS,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE;qBACpC,CAAC;oBACF,IAAI,CAAC,KAAK,CAAC,WAAW,CAAC,MAAM,EAAE;wBAC7B,UAAU;wBACV,SAAS,EAAE,YAAY;qBACxB,CAAC,CAAC;oBACH,WAAW,CAAC,OAAO,CAAC;wBAClB,YAAY,EAAE,IAAI;wBAClB,yBAAyB,EAAE,KAAK;wBAChC,QAAQ,EAAE,EAAE,UAAU,EAAE,MAAM,CAAC,UAAU,EAAE,GAAG,QAAQ,EAAE;qBACzD,CAAC,CAAC;oBACH,OAAO,MAAM,CAAC;gBAChB,CAAC,CACF,CAAC;YACJ,CAAC;YAAC,OAAO,KAAK,EAAE,CAAC;gBACf,WAAW,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;gBACxB,MAAM,KAAK,CAAC;YACd,CAAC;QACH,CAAC;QACD,MAAM,KAAK,GAAG,gBAAgB,CAAC,KAAM,CAAC,CAAC;QACvC,MAAM,QAAQ,GAAG,qBAAqB,CACpC,eAAe,CAAC,KAAK,EAAE,MAAM,EAAE,MAAM,CAAC,UAAU,CAAC,EACjD,SAAS,EACT,KAAK,CACN,CAAC;QACF,MAAM,cAAc,GAAG,IAAI,GAAG,EAAkB,CAAC;QACjD,MAAM,KAAK,GAAa,EAAE,CAAC;QAC3B,MAAM,QAAQ,GAAa,EAAE,CAAC;QAC9B,MAAM,MAAM,GAAG,IAAI,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QACvC,MAAM,OAAO,GAAuB,EAAE,CAAC;QACvC,KAAK,MAAM,CAAC,KAAK,EAAE,IAAI,CAAC,IAAI,QAAQ,CAAC,OAAO,EAAE,EAAE,CAAC;YAC/C,IAAI,MAAM,GAAG,cAAc,CAAC,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;YAChD,IAAI,CAAC,MAAM,EAAE,CAAC;gBACZ,MAAM,GAAG,IAAI,CAAC,IAAI,CAAC,OAAO,EAAE,UAAU,IAAI,CAAC,SAAS,MAAM,CAAC,CAAC;gBAC5D,MAAM,SAAS,CAAC,MAAM,EAAE,KAAK,CAAC,IAAI,CAAC,SAAS,CAAE,CAAC,KAAK,CAAC,CAAC;gBACtD,cAAc,CAAC,GAAG,CAAC,IAAI,CAAC,SAAS,EAAE,MAAM,CAAC,CAAC;YAC7C,CAAC;YACD,MAAM,MAAM,GAAG,IAAI,CAAC,IAAI,CAAC,OAAO,EAAE,QAAQ,KAAK,MAAM,CAAC,CAAC;YACvD,MAAM,MAAM,CAAC,QAAQ,CACnB,MAAM,EACN;gBACE,KAAK,EAAE,MAAM,CAAC,KAAK;gBACnB,MAAM,EAAE,MAAM,CAAC,MAAM;gBACrB,CAAC,EAAE,CAAC,IAAI,CAAC,IAAI,GAAG,MAAM,CAAC,OAAO,CAAC,GAAG,MAAM,CAAC,KAAK;gBAC9C,CAAC,EAAE,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,GAAG,MAAM,CAAC,OAAO,CAAC,GAAG,MAAM,CAAC,MAAM;aAC1D,EACD,MAAM,EACN,MAAM,CACP,CAAC;YACF,MAAM,MAAM,CAAC,YAAY,CAAC,MAAM,CAAC,CAAC;YAClC,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;YACnB,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,WAAW,CAAC,CAAC;YAChC,OAAO,CAAC,IAAI,CAAC;gBACX,WAAW,EAAE,IAAI,CAAC,WAAW;gBAC7B,aAAa,EAAE,IAAI,CAAC,SAAS;gBAC7B,SAAS,EAAE,IAAI,CAAC,IAAI;aACrB,CAAC,CAAC;QACL,CAAC;QACD,OAAO,EAAE,KAAK,EAAE,QAAQ,EAAE,OAAO,EAAE,QAAQ,EAAE,QAAQ,EAAE,CAAC;IAC1D,CAAC;CACF"}

package/dist/src/acquisition/transcript.d.ts

@@ -0,0 +1,25 @@
+import type { UrmaConfig } from "../config.js";
+import type { InvestigationRef, TrackRef } from "../core/ids.js";
+import { BlobStore } from "../store/blob-store.js";
+import type { StoredArtifact, StoredSegment, StoredTrack, UrmaStore } from "../store/store.js";
+import { type RemoteOperationContext } from "../remote/worker.js";
+import type { CaptionTrackSummary, ResolvedSource } from "../sources/types.js";
+export declare function parseJson3Captions(value: string, trackId: string): Omit<StoredSegment, "id">[];
+export declare function parseVttCaptions(value: string, trackId: string): Omit<StoredSegment, "id">[];
+export declare function parseSrtCaptions(value: string, trackId: string): Omit<StoredSegment, "id">[];
+export declare function chooseTrack(tracks: readonly CaptionTrackSummary[], preferredLanguage?: string | null): CaptionTrackSummary | null;
+export declare function selectTrack(source: ResolvedSource, requested: string | undefined): CaptionTrackSummary;
+export declare function transcriptTrackId(trackRef: TrackRef): string;
+export declare class TranscriptAcquirer {
+    readonly config: UrmaConfig;
+    readonly store: UrmaStore;
+    readonly blobs: BlobStore;
+    readonly remoteContext: RemoteOperationContext | null;
+    constructor(config: UrmaConfig, store: UrmaStore, blobs: BlobStore, remoteContext?: RemoteOperationContext | null);
+    ensure(source: ResolvedSource, investigationRef: InvestigationRef, requestedTrackRef?: string, signal?: AbortSignal): Promise<{
+        track: StoredTrack;
+        segments: StoredSegment[];
+        artifact: StoredArtifact;
+        cacheHit: boolean;
+    }>;
+}