@agent-native/core 0.103.0 → 0.104.0

package/dist/ingestion/media.js

@@ -0,0 +1,178 @@
+import { createHash } from "node:crypto";
+export function fingerprintMedia(data, mimeType) {
+    return {
+        sha256: createHash("sha256").update(data).digest("hex"),
+        byteLength: data.byteLength,
+        ...(mimeType ? { mimeType } : {}),
+    };
+}
+export function extractCssColors(value) {
+    const colors = new Set();
+    const pattern = /(?:#[0-9a-f]{3,8}\b|(?:rgb|hsl)a?\([^)]*\)|\b(?:transparent|currentColor)\b)/gi;
+    for (const match of value.match(pattern) ?? [])
+        colors.add(match.trim());
+    return [...colors];
+}
+export function rankColorSamples(samples, limit = 24) {
+    const counts = new Map();
+    for (const sample of samples) {
+        const normalized = sample.trim().replace(/\s+/g, " ").toLowerCase();
+        if (!normalized ||
+            normalized === "rgba(0, 0, 0, 0)" ||
+            normalized === "transparent")
+            continue;
+        counts.set(normalized, (counts.get(normalized) ?? 0) + 1);
+    }
+    return [...counts.entries()]
+        .sort((a, b) => b[1] - a[1] || a[0].localeCompare(b[0]))
+        .slice(0, Math.max(1, limit))
+        .map(([color]) => color);
+}
+export async function extractDominantColors(data, limit = 6) {
+    const sharp = await loadSharp();
+    if (!sharp) {
+        throw new Error("Image palette extraction requires the optional sharp dependency.");
+    }
+    const output = await sharp(Buffer.from(data), { failOn: "none" })
+        .resize(64, 64, { fit: "inside" })
+        .removeAlpha()
+        .raw()
+        .toBuffer({ resolveWithObject: true });
+    const buckets = new Map();
+    for (let index = 0; index < output.data.length; index += 3) {
+        const key = [
+            output.data[index],
+            output.data[index + 1],
+            output.data[index + 2],
+        ]
+            .map((channel) => Math.round(channel / 32) * 32)
+            .map((channel) => Math.max(0, Math.min(255, channel)).toString(16).padStart(2, "0"))
+            .join("");
+        buckets.set(key, (buckets.get(key) ?? 0) + 1);
+    }
+    return [...buckets.entries()]
+        .sort((a, b) => b[1] - a[1])
+        .slice(0, Math.max(1, limit))
+        .map(([hex]) => `#${hex.toUpperCase()}`);
+}
+export async function compareRasterImages(input) {
+    const maxInputBytes = input.maxInputBytes ?? 20 * 1024 * 1024;
+    if (input.source.byteLength > maxInputBytes ||
+        input.rendered.byteLength > maxInputBytes) {
+        throw new Error(`Image comparison input exceeds ${maxInputBytes} bytes.`);
+    }
+    const sharp = await loadSharp();
+    if (!sharp) {
+        throw new Error("Image comparison requires the optional sharp dependency.");
+    }
+    const source = sharp(Buffer.from(input.source), { failOn: "none" });
+    const rendered = sharp(Buffer.from(input.rendered), { failOn: "none" });
+    const [sourceMetadata, renderedMetadata] = await Promise.all([
+        source.metadata(),
+        rendered.metadata(),
+    ]);
+    if (!sourceMetadata.width ||
+        !sourceMetadata.height ||
+        sourceMetadata.width !== renderedMetadata.width ||
+        sourceMetadata.height !== renderedMetadata.height ||
+        sourceMetadata.channels !== renderedMetadata.channels) {
+        throw new Error("Image comparison requires identical dimensions and channels.");
+    }
+    const [sourcePixels, renderedPixels] = await Promise.all([
+        source.raw().toBuffer(),
+        rendered.raw().toBuffer(),
+    ]);
+    if (sourcePixels.byteLength !== renderedPixels.byteLength) {
+        throw new Error("Image comparison buffers have different lengths.");
+    }
+    let difference = 0;
+    for (let index = 0; index < sourcePixels.length; index++) {
+        difference += Math.abs(sourcePixels[index] - renderedPixels[index]);
+    }
+    return {
+        meanAbsoluteDifference: difference / sourcePixels.length,
+        width: sourceMetadata.width,
+        height: sourceMetadata.height,
+    };
+}
+/** Crops a bounded raster region without retaining the source image. */
+export async function cropImageRegion(input) {
+    const maxInputBytes = input.maxInputBytes ?? 20 * 1024 * 1024;
+    const maxOutputPixels = input.maxOutputPixels ?? 16_000_000;
+    if (input.data.byteLength > maxInputBytes) {
+        throw new Error(`Image crop input exceeds ${maxInputBytes} bytes.`);
+    }
+    const region = Object.fromEntries(["left", "top", "width", "height"].map((key) => [
+        key,
+        Math.round(input[key]),
+    ]));
+    if (region.left < 0 ||
+        region.top < 0 ||
+        region.width <= 0 ||
+        region.height <= 0 ||
+        region.width * region.height > maxOutputPixels) {
+        throw new Error("Image crop region is invalid or exceeds the pixel limit.");
+    }
+    const sharp = await loadSharp();
+    if (!sharp) {
+        throw new Error("Image cropping requires the optional sharp dependency.");
+    }
+    const pipeline = sharp(Buffer.from(input.data), { failOn: "none" });
+    const metadata = await pipeline.metadata();
+    if (!metadata.width ||
+        !metadata.height ||
+        region.left + region.width > metadata.width ||
+        region.top + region.height > metadata.height) {
+        throw new Error("Image crop region falls outside the source image.");
+    }
+    const data = await pipeline.extract(region).png().toBuffer();
+    return {
+        data: new Uint8Array(data),
+        mimeType: "image/png",
+        width: region.width,
+        height: region.height,
+    };
+}
+export async function readBoundedResponseBytes(response, maxBytes) {
+    if (!Number.isInteger(maxBytes) || maxBytes <= 0) {
+        throw new Error("maxBytes must be a positive integer.");
+    }
+    const contentLength = Number(response.headers.get("content-length"));
+    if (Number.isFinite(contentLength) && contentLength > maxBytes) {
+        throw new Error(`Remote artifact exceeds ${maxBytes} bytes.`);
+    }
+    if (!response.body)
+        return new Uint8Array(await response.arrayBuffer());
+    const reader = response.body.getReader();
+    const chunks = [];
+    let total = 0;
+    while (true) {
+        const { done, value } = await reader.read();
+        if (done)
+            break;
+        total += value.byteLength;
+        if (total > maxBytes) {
+            await reader.cancel();
+            throw new Error(`Remote artifact exceeds ${maxBytes} bytes.`);
+        }
+        chunks.push(value);
+    }
+    const joined = new Uint8Array(total);
+    let offset = 0;
+    for (const chunk of chunks) {
+        joined.set(chunk, offset);
+        offset += chunk.byteLength;
+    }
+    return joined;
+}
+async function loadSharp() {
+    const specifier = "sharp";
+    try {
+        const module = (await import(/* @vite-ignore */ specifier));
+        return module.default ?? module;
+    }
+    catch {
+        return null;
+    }
+}
+//# sourceMappingURL=media.js.map

package/dist/ingestion/media.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"media.js","sourceRoot":"","sources":["../../src/ingestion/media.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,UAAU,EAAE,MAAM,aAAa,CAAC;AAQzC,MAAM,UAAU,gBAAgB,CAC9B,IAAgB,EAChB,QAAiB;IAEjB,OAAO;QACL,MAAM,EAAE,UAAU,CAAC,QAAQ,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC;QACvD,UAAU,EAAE,IAAI,CAAC,UAAU;QAC3B,GAAG,CAAC,QAAQ,CAAC,CAAC,CAAC,EAAE,QAAQ,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;KAClC,CAAC;AACJ,CAAC;AAED,MAAM,UAAU,gBAAgB,CAAC,KAAa;IAC5C,MAAM,MAAM,GAAG,IAAI,GAAG,EAAU,CAAC;IACjC,MAAM,OAAO,GACX,gFAAgF,CAAC;IACnF,KAAK,MAAM,KAAK,IAAI,KAAK,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,EAAE;QAAE,MAAM,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC,CAAC;IACzE,OAAO,CAAC,GAAG,MAAM,CAAC,CAAC;AACrB,CAAC;AAED,MAAM,UAAU,gBAAgB,CAC9B,OAA0B,EAC1B,KAAK,GAAG,EAAE;IAEV,MAAM,MAAM,GAAG,IAAI,GAAG,EAAkB,CAAC;IACzC,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;QAC7B,MAAM,UAAU,GAAG,MAAM,CAAC,IAAI,EAAE,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,WAAW,EAAE,CAAC;QACpE,IACE,CAAC,UAAU;YACX,UAAU,KAAK,kBAAkB;YACjC,UAAU,KAAK,aAAa;YAE5B,SAAS;QACX,MAAM,CAAC,GAAG,CAAC,UAAU,EAAE,CAAC,MAAM,CAAC,GAAG,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;IAC5D,CAAC;IACD,OAAO,CAAC,GAAG,MAAM,CAAC,OAAO,EAAE,CAAC;SACzB,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,aAAa,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;SACvD,KAAK,CAAC,CAAC,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC;SAC5B,GAAG,CAAC,CAAC,CAAC,KAAK,CAAC,EAAE,EAAE,CAAC,KAAK,CAAC,CAAC;AAC7B,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,qBAAqB,CACzC,IAAgB,EAChB,KAAK,GAAG,CAAC;IAET,MAAM,KAAK,GAAG,MAAM,SAAS,EAAE,CAAC;IAChC,IAAI,CAAC,KAAK,EAAE,CAAC;QACX,MAAM,IAAI,KAAK,CACb,kEAAkE,CACnE,CAAC;IACJ,CAAC;IACD,MAAM,MAAM,GAAG,MAAM,KAAK,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,EAAE,MAAM,EAAE,MAAM,EAAE,CAAC;SAC9D,MAAM,CAAC,EAAE,EAAE,EAAE,EAAE,EAAE,GAAG,EAAE,QAAQ,EAAE,CAAC;SACjC,WAAW,EAAE;SACb,GAAG,EAAE;SACL,QAAQ,CAAC,EAAE,iBAAiB,EAAE,IAAI,EAAE,CAAC,CAAC;IACzC,MAAM,OAAO,GAAG,IAAI,GAAG,EAAkB,CAAC;IAC1C,KAAK,IAAI,KAAK,GAAG,CAAC,EAAE,KAAK,GAAG,MAAM,CAAC,IAAI,CAAC,MAAM,EAAE,KAAK,IAAI,CAAC,EAAE,CAAC;QAC3D,MAAM,GAAG,GAAG;YACV,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC;YAClB,MAAM,CAAC,IAAI,CAAC,KAAK,GAAG,CAAC,CAAC;YACtB,MAAM,CAAC,IAAI,CAAC,KAAK,GAAG,CAAC,CAAC;SACvB;aACE,GAAG,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,IAAI,CAAC,KAAK,CAAC,OAAO,GAAG,EAAE,CAAC,GAAG,EAAE,CAAC;aAC/C,GAAG,CAAC,CAAC,OAAO,EAAE,EAAE,CACf,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,GAAG,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC,CAAC,QAAQ,CAAC,EAAE,CAAC,CAAC,QAAQ,CAAC,CAAC,EAAE,GAAG,CAAC,CAClE;aACA,IAAI,CAAC,EAAE,CAAC,CAAC;QACZ,OAAO,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,OAAO,CAAC,GAAG,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;IAChD,CAAC;IACD,OAAO,CAAC,GAAG,OAAO,CAAC,OAAO,EAAE,CAAC;SAC1B,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;SAC3B,KAAK,CAAC,CAAC,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC;SAC5B,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,EAAE,EAAE,CAAC,IAAI,GAAG,CAAC,WAAW,EAAE,EAAE,CAAC,CAAC;AAC7C,CAAC;AASD,MAAM,CAAC,KAAK,UAAU,mBAAmB,CAAC,KAIzC;IACC,MAAM,aAAa,GAAG,KAAK,CAAC,aAAa,IAAI,EAAE,GAAG,IAAI,GAAG,IAAI,CAAC;IAC9D,IACE,KAAK,CAAC,MAAM,CAAC,UAAU,GAAG,aAAa;QACvC,KAAK,CAAC,QAAQ,CAAC,UAAU,GAAG,aAAa,EACzC,CAAC;QACD,MAAM,IAAI,KAAK,CAAC,kCAAkC,aAAa,SAAS,CAAC,CAAC;IAC5E,CAAC;IACD,MAAM,KAAK,GAAG,MAAM,SAAS,EAAE,CAAC;IAChC,IAAI,CAAC,KAAK,EAAE,CAAC;QACX,MAAM,IAAI,KAAK,CAAC,0DAA0D,CAAC,CAAC;IAC9E,CAAC;IACD,MAAM,MAAM,GAAG,KAAK,CAAC,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,EAAE,MAAM,EAAE,MAAM,EAAE,CAAC,CAAC;IACpE,MAAM,QAAQ,GAAG,KAAK,CAAC,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,QAAQ,CAAC,EAAE,EAAE,MAAM,EAAE,MAAM,EAAE,CAAC,CAAC;IACxE,MAAM,CAAC,cAAc,EAAE,gBAAgB,CAAC,GAAG,MAAM,OAAO,CAAC,GAAG,CAAC;QAC3D,MAAM,CAAC,QAAQ,EAAE;QACjB,QAAQ,CAAC,QAAQ,EAAE;KACpB,CAAC,CAAC;IACH,IACE,CAAC,cAAc,CAAC,KAAK;QACrB,CAAC,cAAc,CAAC,MAAM;QACtB,cAAc,CAAC,KAAK,KAAK,gBAAgB,CAAC,KAAK;QAC/C,cAAc,CAAC,MAAM,KAAK,gBAAgB,CAAC,MAAM;QACjD,cAAc,CAAC,QAAQ,KAAK,gBAAgB,CAAC,QAAQ,EACrD,CAAC;QACD,MAAM,IAAI,KAAK,CACb,8DAA8D,CAC/D,CAAC;IACJ,CAAC;IACD,MAAM,CAAC,YAAY,EAAE,cAAc,CAAC,GAAG,MAAM,OAAO,CAAC,GAAG,CAAC;QACvD,MAAM,CAAC,GAAG,EAAE,CAAC,QAAQ,EAAE;QACvB,QAAQ,CAAC,GAAG,EAAE,CAAC,QAAQ,EAAE;KAC1B,CAAC,CAAC;IACH,IAAI,YAAY,CAAC,UAAU,KAAK,cAAc,CAAC,UAAU,EAAE,CAAC;QAC1D,MAAM,IAAI,KAAK,CAAC,kDAAkD,CAAC,CAAC;IACtE,CAAC;IACD,IAAI,UAAU,GAAG,CAAC,CAAC;IACnB,KAAK,IAAI,KAAK,GAAG,CAAC,EAAE,KAAK,GAAG,YAAY,CAAC,MAAM,EAAE,KAAK,EAAE,EAAE,CAAC;QACzD,UAAU,IAAI,IAAI,CAAC,GAAG,CAAC,YAAY,CAAC,KAAK,CAAE,GAAG,cAAc,CAAC,KAAK,CAAE,CAAC,CAAC;IACxE,CAAC;IACD,OAAO;QACL,sBAAsB,EAAE,UAAU,GAAG,YAAY,CAAC,MAAM;QACxD,KAAK,EAAE,cAAc,CAAC,KAAK;QAC3B,MAAM,EAAE,cAAc,CAAC,MAAM;KAC9B,CAAC;AACJ,CAAC;AAED,wEAAwE;AACxE,MAAM,CAAC,KAAK,UAAU,eAAe,CAAC,KAQrC;IACC,MAAM,aAAa,GAAG,KAAK,CAAC,aAAa,IAAI,EAAE,GAAG,IAAI,GAAG,IAAI,CAAC;IAC9D,MAAM,eAAe,GAAG,KAAK,CAAC,eAAe,IAAI,UAAU,CAAC;IAC5D,IAAI,KAAK,CAAC,IAAI,CAAC,UAAU,GAAG,aAAa,EAAE,CAAC;QAC1C,MAAM,IAAI,KAAK,CAAC,4BAA4B,aAAa,SAAS,CAAC,CAAC;IACtE,CAAC;IACD,MAAM,MAAM,GAAG,MAAM,CAAC,WAAW,CAC9B,CAAC,MAAM,EAAE,KAAK,EAAE,OAAO,EAAE,QAAQ,CAAW,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC;QACzD,GAAG;QACH,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;KACvB,CAAC,CAC6D,CAAC;IAClE,IACE,MAAM,CAAC,IAAI,GAAG,CAAC;QACf,MAAM,CAAC,GAAG,GAAG,CAAC;QACd,MAAM,CAAC,KAAK,IAAI,CAAC;QACjB,MAAM,CAAC,MAAM,IAAI,CAAC;QAClB,MAAM,CAAC,KAAK,GAAG,MAAM,CAAC,MAAM,GAAG,eAAe,EAC9C,CAAC;QACD,MAAM,IAAI,KAAK,CAAC,0DAA0D,CAAC,CAAC;IAC9E,CAAC;IACD,MAAM,KAAK,GAAG,MAAM,SAAS,EAAE,CAAC;IAChC,IAAI,CAAC,KAAK,EAAE,CAAC;QACX,MAAM,IAAI,KAAK,CAAC,wDAAwD,CAAC,CAAC;IAC5E,CAAC;IACD,MAAM,QAAQ,GAAG,KAAK,CAAC,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,EAAE,EAAE,MAAM,EAAE,MAAM,EAAE,CAAC,CAAC;IACpE,MAAM,QAAQ,GAAG,MAAM,QAAQ,CAAC,QAAQ,EAAE,CAAC;IAC3C,IACE,CAAC,QAAQ,CAAC,KAAK;QACf,CAAC,QAAQ,CAAC,MAAM;QAChB,MAAM,CAAC,IAAI,GAAG,MAAM,CAAC,KAAK,GAAG,QAAQ,CAAC,KAAK;QAC3C,MAAM,CAAC,GAAG,GAAG,MAAM,CAAC,MAAM,GAAG,QAAQ,CAAC,MAAM,EAC5C,CAAC;QACD,MAAM,IAAI,KAAK,CAAC,mDAAmD,CAAC,CAAC;IACvE,CAAC;IACD,MAAM,IAAI,GAAG,MAAM,QAAQ,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,CAAC,QAAQ,EAAE,CAAC;IAC7D,OAAO;QACL,IAAI,EAAE,IAAI,UAAU,CAAC,IAAI,CAAC;QAC1B,QAAQ,EAAE,WAAW;QACrB,KAAK,EAAE,MAAM,CAAC,KAAK;QACnB,MAAM,EAAE,MAAM,CAAC,MAAM;KACtB,CAAC;AACJ,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,wBAAwB,CAC5C,QAAkB,EAClB,QAAgB;IAEhB,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,QAAQ,CAAC,IAAI,QAAQ,IAAI,CAAC,EAAE,CAAC;QACjD,MAAM,IAAI,KAAK,CAAC,sCAAsC,CAAC,CAAC;IAC1D,CAAC;IACD,MAAM,aAAa,GAAG,MAAM,CAAC,QAAQ,CAAC,OAAO,CAAC,GAAG,CAAC,gBAAgB,CAAC,CAAC,CAAC;IACrE,IAAI,MAAM,CAAC,QAAQ,CAAC,aAAa,CAAC,IAAI,aAAa,GAAG,QAAQ,EAAE,CAAC;QAC/D,MAAM,IAAI,KAAK,CAAC,2BAA2B,QAAQ,SAAS,CAAC,CAAC;IAChE,CAAC;IACD,IAAI,CAAC,QAAQ,CAAC,IAAI;QAAE,OAAO,IAAI,UAAU,CAAC,MAAM,QAAQ,CAAC,WAAW,EAAE,CAAC,CAAC;IACxE,MAAM,MAAM,GAAG,QAAQ,CAAC,IAAI,CAAC,SAAS,EAAE,CAAC;IACzC,MAAM,MAAM,GAAiB,EAAE,CAAC;IAChC,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,OAAO,IAAI,EAAE,CAAC;QACZ,MAAM,EAAE,IAAI,EAAE,KAAK,EAAE,GAAG,MAAM,MAAM,CAAC,IAAI,EAAE,CAAC;QAC5C,IAAI,IAAI;YAAE,MAAM;QAChB,KAAK,IAAI,KAAK,CAAC,UAAU,CAAC;QAC1B,IAAI,KAAK,GAAG,QAAQ,EAAE,CAAC;YACrB,MAAM,MAAM,CAAC,MAAM,EAAE,CAAC;YACtB,MAAM,IAAI,KAAK,CAAC,2BAA2B,QAAQ,SAAS,CAAC,CAAC;QAChE,CAAC;QACD,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;IACrB,CAAC;IACD,MAAM,MAAM,GAAG,IAAI,UAAU,CAAC,KAAK,CAAC,CAAC;IACrC,IAAI,MAAM,GAAG,CAAC,CAAC;IACf,KAAK,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;QAC3B,MAAM,CAAC,GAAG,CAAC,KAAK,EAAE,MAAM,CAAC,CAAC;QAC1B,MAAM,IAAI,KAAK,CAAC,UAAU,CAAC;IAC7B,CAAC;IACD,OAAO,MAAM,CAAC;AAChB,CAAC;AA2BD,KAAK,UAAU,SAAS;IACtB,MAAM,SAAS,GAAG,OAAO,CAAC;IAC1B,IAAI,CAAC;QACH,MAAM,MAAM,GAAG,CAAC,MAAM,MAAM,CAAC,kBAAkB,CAAC,SAAS,CAAC,CAEzD,CAAC;QACF,OAAO,MAAM,CAAC,OAAO,IAAK,MAAkC,CAAC;IAC/D,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,IAAI,CAAC;IACd,CAAC;AACH,CAAC","sourcesContent":["import { createHash } from \"node:crypto\";\n\nexport interface MediaFingerprint {\n  sha256: string;\n  byteLength: number;\n  mimeType?: string;\n}\n\nexport function fingerprintMedia(\n  data: Uint8Array,\n  mimeType?: string,\n): MediaFingerprint {\n  return {\n    sha256: createHash(\"sha256\").update(data).digest(\"hex\"),\n    byteLength: data.byteLength,\n    ...(mimeType ? { mimeType } : {}),\n  };\n}\n\nexport function extractCssColors(value: string): string[] {\n  const colors = new Set<string>();\n  const pattern =\n    /(?:#[0-9a-f]{3,8}\\b|(?:rgb|hsl)a?\\([^)]*\\)|\\b(?:transparent|currentColor)\\b)/gi;\n  for (const match of value.match(pattern) ?? []) colors.add(match.trim());\n  return [...colors];\n}\n\nexport function rankColorSamples(\n  samples: readonly string[],\n  limit = 24,\n): string[] {\n  const counts = new Map<string, number>();\n  for (const sample of samples) {\n    const normalized = sample.trim().replace(/\\s+/g, \" \").toLowerCase();\n    if (\n      !normalized ||\n      normalized === \"rgba(0, 0, 0, 0)\" ||\n      normalized === \"transparent\"\n    )\n      continue;\n    counts.set(normalized, (counts.get(normalized) ?? 0) + 1);\n  }\n  return [...counts.entries()]\n    .sort((a, b) => b[1] - a[1] || a[0].localeCompare(b[0]))\n    .slice(0, Math.max(1, limit))\n    .map(([color]) => color);\n}\n\nexport async function extractDominantColors(\n  data: Uint8Array,\n  limit = 6,\n): Promise<string[]> {\n  const sharp = await loadSharp();\n  if (!sharp) {\n    throw new Error(\n      \"Image palette extraction requires the optional sharp dependency.\",\n    );\n  }\n  const output = await sharp(Buffer.from(data), { failOn: \"none\" })\n    .resize(64, 64, { fit: \"inside\" })\n    .removeAlpha()\n    .raw()\n    .toBuffer({ resolveWithObject: true });\n  const buckets = new Map<string, number>();\n  for (let index = 0; index < output.data.length; index += 3) {\n    const key = [\n      output.data[index],\n      output.data[index + 1],\n      output.data[index + 2],\n    ]\n      .map((channel) => Math.round(channel / 32) * 32)\n      .map((channel) =>\n        Math.max(0, Math.min(255, channel)).toString(16).padStart(2, \"0\"),\n      )\n      .join(\"\");\n    buckets.set(key, (buckets.get(key) ?? 0) + 1);\n  }\n  return [...buckets.entries()]\n    .sort((a, b) => b[1] - a[1])\n    .slice(0, Math.max(1, limit))\n    .map(([hex]) => `#${hex.toUpperCase()}`);\n}\n\nexport interface CroppedImageRegion {\n  data: Uint8Array;\n  mimeType: \"image/png\";\n  width: number;\n  height: number;\n}\n\nexport async function compareRasterImages(input: {\n  source: Uint8Array;\n  rendered: Uint8Array;\n  maxInputBytes?: number;\n}): Promise<{ meanAbsoluteDifference: number; width: number; height: number }> {\n  const maxInputBytes = input.maxInputBytes ?? 20 * 1024 * 1024;\n  if (\n    input.source.byteLength > maxInputBytes ||\n    input.rendered.byteLength > maxInputBytes\n  ) {\n    throw new Error(`Image comparison input exceeds ${maxInputBytes} bytes.`);\n  }\n  const sharp = await loadSharp();\n  if (!sharp) {\n    throw new Error(\"Image comparison requires the optional sharp dependency.\");\n  }\n  const source = sharp(Buffer.from(input.source), { failOn: \"none\" });\n  const rendered = sharp(Buffer.from(input.rendered), { failOn: \"none\" });\n  const [sourceMetadata, renderedMetadata] = await Promise.all([\n    source.metadata(),\n    rendered.metadata(),\n  ]);\n  if (\n    !sourceMetadata.width ||\n    !sourceMetadata.height ||\n    sourceMetadata.width !== renderedMetadata.width ||\n    sourceMetadata.height !== renderedMetadata.height ||\n    sourceMetadata.channels !== renderedMetadata.channels\n  ) {\n    throw new Error(\n      \"Image comparison requires identical dimensions and channels.\",\n    );\n  }\n  const [sourcePixels, renderedPixels] = await Promise.all([\n    source.raw().toBuffer(),\n    rendered.raw().toBuffer(),\n  ]);\n  if (sourcePixels.byteLength !== renderedPixels.byteLength) {\n    throw new Error(\"Image comparison buffers have different lengths.\");\n  }\n  let difference = 0;\n  for (let index = 0; index < sourcePixels.length; index++) {\n    difference += Math.abs(sourcePixels[index]! - renderedPixels[index]!);\n  }\n  return {\n    meanAbsoluteDifference: difference / sourcePixels.length,\n    width: sourceMetadata.width,\n    height: sourceMetadata.height,\n  };\n}\n\n/** Crops a bounded raster region without retaining the source image. */\nexport async function cropImageRegion(input: {\n  data: Uint8Array;\n  left: number;\n  top: number;\n  width: number;\n  height: number;\n  maxInputBytes?: number;\n  maxOutputPixels?: number;\n}): Promise<CroppedImageRegion> {\n  const maxInputBytes = input.maxInputBytes ?? 20 * 1024 * 1024;\n  const maxOutputPixels = input.maxOutputPixels ?? 16_000_000;\n  if (input.data.byteLength > maxInputBytes) {\n    throw new Error(`Image crop input exceeds ${maxInputBytes} bytes.`);\n  }\n  const region = Object.fromEntries(\n    ([\"left\", \"top\", \"width\", \"height\"] as const).map((key) => [\n      key,\n      Math.round(input[key]),\n    ]),\n  ) as { left: number; top: number; width: number; height: number };\n  if (\n    region.left < 0 ||\n    region.top < 0 ||\n    region.width <= 0 ||\n    region.height <= 0 ||\n    region.width * region.height > maxOutputPixels\n  ) {\n    throw new Error(\"Image crop region is invalid or exceeds the pixel limit.\");\n  }\n  const sharp = await loadSharp();\n  if (!sharp) {\n    throw new Error(\"Image cropping requires the optional sharp dependency.\");\n  }\n  const pipeline = sharp(Buffer.from(input.data), { failOn: \"none\" });\n  const metadata = await pipeline.metadata();\n  if (\n    !metadata.width ||\n    !metadata.height ||\n    region.left + region.width > metadata.width ||\n    region.top + region.height > metadata.height\n  ) {\n    throw new Error(\"Image crop region falls outside the source image.\");\n  }\n  const data = await pipeline.extract(region).png().toBuffer();\n  return {\n    data: new Uint8Array(data),\n    mimeType: \"image/png\",\n    width: region.width,\n    height: region.height,\n  };\n}\n\nexport async function readBoundedResponseBytes(\n  response: Response,\n  maxBytes: number,\n): Promise<Uint8Array> {\n  if (!Number.isInteger(maxBytes) || maxBytes <= 0) {\n    throw new Error(\"maxBytes must be a positive integer.\");\n  }\n  const contentLength = Number(response.headers.get(\"content-length\"));\n  if (Number.isFinite(contentLength) && contentLength > maxBytes) {\n    throw new Error(`Remote artifact exceeds ${maxBytes} bytes.`);\n  }\n  if (!response.body) return new Uint8Array(await response.arrayBuffer());\n  const reader = response.body.getReader();\n  const chunks: Uint8Array[] = [];\n  let total = 0;\n  while (true) {\n    const { done, value } = await reader.read();\n    if (done) break;\n    total += value.byteLength;\n    if (total > maxBytes) {\n      await reader.cancel();\n      throw new Error(`Remote artifact exceeds ${maxBytes} bytes.`);\n    }\n    chunks.push(value);\n  }\n  const joined = new Uint8Array(total);\n  let offset = 0;\n  for (const chunk of chunks) {\n    joined.set(chunk, offset);\n    offset += chunk.byteLength;\n  }\n  return joined;\n}\n\ninterface SharpPipeline {\n  resize(\n    width: number,\n    height: number,\n    options: { fit: \"inside\" },\n  ): SharpPipeline;\n  removeAlpha(): SharpPipeline;\n  metadata(): Promise<{ width?: number; height?: number; channels?: number }>;\n  extract(region: {\n    left: number;\n    top: number;\n    width: number;\n    height: number;\n  }): SharpPipeline;\n  png(): SharpPipeline;\n  toBuffer(): Promise<Buffer>;\n  raw(): SharpPipeline;\n  toBuffer(options: { resolveWithObject: true }): Promise<{ data: Uint8Array }>;\n}\n\ntype SharpFactory = (\n  data: Buffer,\n  options: { failOn: \"none\" },\n) => SharpPipeline;\n\nasync function loadSharp(): Promise<SharpFactory | null> {\n  const specifier = \"sharp\";\n  try {\n    const module = (await import(/* @vite-ignore */ specifier)) as {\n      default?: SharpFactory;\n    };\n    return module.default ?? (module as unknown as SharpFactory);\n  } catch {\n    return null;\n  }\n}\n"]}

package/dist/ingestion/notion.d.ts

@@ -0,0 +1,12 @@
+export interface NotionMarkdownSection {
+    heading: string | null;
+    headingLevel: number | null;
+    markdown: string;
+}
+export interface NotionMarkdownResult {
+    markdown: string;
+    sections: NotionMarkdownSection[];
+}
+export declare function notionBlocksToMarkdown(blocks: readonly unknown[]): NotionMarkdownResult;
+export declare function notionBlockToMarkdown(block: Record<string, unknown>, type?: string): string;
+//# sourceMappingURL=notion.d.ts.map

package/dist/ingestion/notion.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"notion.d.ts","sourceRoot":"","sources":["../../src/ingestion/notion.ts"],"names":[],"mappings":"AAAA,MAAM,WAAW,qBAAqB;IACpC,OAAO,EAAE,MAAM,GAAG,IAAI,CAAC;IACvB,YAAY,EAAE,MAAM,GAAG,IAAI,CAAC;IAC5B,QAAQ,EAAE,MAAM,CAAC;CAClB;AAED,MAAM,WAAW,oBAAoB;IACnC,QAAQ,EAAE,MAAM,CAAC;IACjB,QAAQ,EAAE,qBAAqB,EAAE,CAAC;CACnC;AAED,wBAAgB,sBAAsB,CACpC,MAAM,EAAE,SAAS,OAAO,EAAE,GACzB,oBAAoB,CAmCtB;AAED,wBAAgB,qBAAqB,CACnC,KAAK,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,EAC9B,IAAI,SAAgC,GACnC,MAAM,CA2CR"}

package/dist/ingestion/notion.js

@@ -0,0 +1,104 @@
+export function notionBlocksToMarkdown(blocks) {
+    const sections = [];
+    let current = {
+        heading: null,
+        headingLevel: null,
+        markdown: "",
+    };
+    const flush = () => {
+        const markdown = current.markdown.trim();
+        if (markdown)
+            sections.push({ ...current, markdown });
+    };
+    for (const value of blocks) {
+        const block = asRecord(value);
+        const type = stringValue(block?.type);
+        if (!block || !type)
+            continue;
+        const line = notionBlockToMarkdown(block, type);
+        if (!line)
+            continue;
+        const heading = type.match(/^heading_([1-3])$/);
+        if (heading) {
+            flush();
+            const headingLevel = Number(heading[1]);
+            current = {
+                heading: richText(asRecord(block[type])?.rich_text),
+                headingLevel,
+                markdown: line,
+            };
+        }
+        else {
+            current.markdown += `${current.markdown ? "\n\n" : ""}${line}`;
+        }
+    }
+    flush();
+    return {
+        markdown: sections.map((section) => section.markdown).join("\n\n"),
+        sections,
+    };
+}
+export function notionBlockToMarkdown(block, type = stringValue(block.type) ?? "") {
+    const payload = asRecord(block[type]) ?? {};
+    const text = richText(payload.rich_text);
+    switch (type) {
+        case "heading_1":
+        case "heading_2":
+        case "heading_3":
+            return `${"#".repeat(Number(type.slice(-1)))} ${text}`.trim();
+        case "bulleted_list_item":
+            return `- ${text}`.trim();
+        case "numbered_list_item":
+            return `1. ${text}`.trim();
+        case "to_do":
+            return `- [${payload.checked === true ? "x" : " "}] ${text}`.trim();
+        case "quote":
+            return text
+                .split("\n")
+                .map((line) => `> ${line}`)
+                .join("\n");
+        case "code":
+            return `\`\`\`${stringValue(payload.language) ?? ""}\n${text}\n\`\`\``;
+        case "divider":
+            return "---";
+        case "bookmark":
+        case "embed": {
+            const url = stringValue(payload.url);
+            return url ? `[${url}](${url})` : "";
+        }
+        case "child_page":
+            return `## ${stringValue(payload.title) ?? "Untitled page"}`;
+        case "image":
+        case "video":
+        case "audio":
+        case "file":
+        case "pdf": {
+            const file = asRecord(payload.file) ?? asRecord(payload.external);
+            const url = stringValue(file?.url);
+            const caption = richText(payload.caption) || type;
+            return url ? `![${caption}](${url})` : caption;
+        }
+        default:
+            return text;
+    }
+}
+function richText(value) {
+    return (Array.isArray(value) ? value : [])
+        .map((part) => {
+        const record = asRecord(part);
+        return (stringValue(record?.plain_text) ??
+            stringValue(asRecord(record?.text)?.content) ??
+            "");
+    })
+        .join("")
+        .trim();
+}
+function asRecord(value) {
+    return value && typeof value === "object" && !Array.isArray(value)
+        ? value
+        : null;
+}
+function stringValue(value) {
+    return typeof value === "string" && value.trim() ? value.trim() : undefined;
+}
+//# sourceMappingURL=notion.js.map

package/dist/ingestion/notion.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"notion.js","sourceRoot":"","sources":["../../src/ingestion/notion.ts"],"names":[],"mappings":"AAWA,MAAM,UAAU,sBAAsB,CACpC,MAA0B;IAE1B,MAAM,QAAQ,GAA4B,EAAE,CAAC;IAC7C,IAAI,OAAO,GAA0B;QACnC,OAAO,EAAE,IAAI;QACb,YAAY,EAAE,IAAI;QAClB,QAAQ,EAAE,EAAE;KACb,CAAC;IACF,MAAM,KAAK,GAAG,GAAG,EAAE;QACjB,MAAM,QAAQ,GAAG,OAAO,CAAC,QAAQ,CAAC,IAAI,EAAE,CAAC;QACzC,IAAI,QAAQ;YAAE,QAAQ,CAAC,IAAI,CAAC,EAAE,GAAG,OAAO,EAAE,QAAQ,EAAE,CAAC,CAAC;IACxD,CAAC,CAAC;IACF,KAAK,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;QAC3B,MAAM,KAAK,GAAG,QAAQ,CAAC,KAAK,CAAC,CAAC;QAC9B,MAAM,IAAI,GAAG,WAAW,CAAC,KAAK,EAAE,IAAI,CAAC,CAAC;QACtC,IAAI,CAAC,KAAK,IAAI,CAAC,IAAI;YAAE,SAAS;QAC9B,MAAM,IAAI,GAAG,qBAAqB,CAAC,KAAK,EAAE,IAAI,CAAC,CAAC;QAChD,IAAI,CAAC,IAAI;YAAE,SAAS;QACpB,MAAM,OAAO,GAAG,IAAI,CAAC,KAAK,CAAC,mBAAmB,CAAC,CAAC;QAChD,IAAI,OAAO,EAAE,CAAC;YACZ,KAAK,EAAE,CAAC;YACR,MAAM,YAAY,GAAG,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,CAAC;YACxC,OAAO,GAAG;gBACR,OAAO,EAAE,QAAQ,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,EAAE,SAAS,CAAC;gBACnD,YAAY;gBACZ,QAAQ,EAAE,IAAI;aACf,CAAC;QACJ,CAAC;aAAM,CAAC;YACN,OAAO,CAAC,QAAQ,IAAI,GAAG,OAAO,CAAC,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,GAAG,IAAI,EAAE,CAAC;QACjE,CAAC;IACH,CAAC;IACD,KAAK,EAAE,CAAC;IACR,OAAO;QACL,QAAQ,EAAE,QAAQ,CAAC,GAAG,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC;QAClE,QAAQ;KACT,CAAC;AACJ,CAAC;AAED,MAAM,UAAU,qBAAqB,CACnC,KAA8B,EAC9B,IAAI,GAAG,WAAW,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,EAAE;IAEpC,MAAM,OAAO,GAAG,QAAQ,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,CAAC;IAC5C,MAAM,IAAI,GAAG,QAAQ,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC;IACzC,QAAQ,IAAI,EAAE,CAAC;QACb,KAAK,WAAW,CAAC;QACjB,KAAK,WAAW,CAAC;QACjB,KAAK,WAAW;YACd,OAAO,GAAG,GAAG,CAAC,MAAM,CAAC,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,IAAI,EAAE,CAAC,IAAI,EAAE,CAAC;QAChE,KAAK,oBAAoB;YACvB,OAAO,KAAK,IAAI,EAAE,CAAC,IAAI,EAAE,CAAC;QAC5B,KAAK,oBAAoB;YACvB,OAAO,MAAM,IAAI,EAAE,CAAC,IAAI,EAAE,CAAC;QAC7B,KAAK,OAAO;YACV,OAAO,MAAM,OAAO,CAAC,OAAO,KAAK,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,KAAK,IAAI,EAAE,CAAC,IAAI,EAAE,CAAC;QACtE,KAAK,OAAO;YACV,OAAO,IAAI;iBACR,KAAK,CAAC,IAAI,CAAC;iBACX,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,KAAK,IAAI,EAAE,CAAC;iBAC1B,IAAI,CAAC,IAAI,CAAC,CAAC;QAChB,KAAK,MAAM;YACT,OAAO,SAAS,WAAW,CAAC,OAAO,CAAC,QAAQ,CAAC,IAAI,EAAE,KAAK,IAAI,UAAU,CAAC;QACzE,KAAK,SAAS;YACZ,OAAO,KAAK,CAAC;QACf,KAAK,UAAU,CAAC;QAChB,KAAK,OAAO,EAAE,CAAC;YACb,MAAM,GAAG,GAAG,WAAW,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC;YACrC,OAAO,GAAG,CAAC,CAAC,CAAC,IAAI,GAAG,KAAK,GAAG,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC;QACvC,CAAC;QACD,KAAK,YAAY;YACf,OAAO,MAAM,WAAW,CAAC,OAAO,CAAC,KAAK,CAAC,IAAI,eAAe,EAAE,CAAC;QAC/D,KAAK,OAAO,CAAC;QACb,KAAK,OAAO,CAAC;QACb,KAAK,OAAO,CAAC;QACb,KAAK,MAAM,CAAC;QACZ,KAAK,KAAK,EAAE,CAAC;YACX,MAAM,IAAI,GAAG,QAAQ,CAAC,OAAO,CAAC,IAAI,CAAC,IAAI,QAAQ,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC;YAClE,MAAM,GAAG,GAAG,WAAW,CAAC,IAAI,EAAE,GAAG,CAAC,CAAC;YACnC,MAAM,OAAO,GAAG,QAAQ,CAAC,OAAO,CAAC,OAAO,CAAC,IAAI,IAAI,CAAC;YAClD,OAAO,GAAG,CAAC,CAAC,CAAC,KAAK,OAAO,KAAK,GAAG,GAAG,CAAC,CAAC,CAAC,OAAO,CAAC;QACjD,CAAC;QACD;YACE,OAAO,IAAI,CAAC;IAChB,CAAC;AACH,CAAC;AAED,SAAS,QAAQ,CAAC,KAAc;IAC9B,OAAO,CAAC,KAAK,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,CAAC;SACvC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE;QACZ,MAAM,MAAM,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAC;QAC9B,OAAO,CACL,WAAW,CAAC,MAAM,EAAE,UAAU,CAAC;YAC/B,WAAW,CAAC,QAAQ,CAAC,MAAM,EAAE,IAAI,CAAC,EAAE,OAAO,CAAC;YAC5C,EAAE,CACH,CAAC;IACJ,CAAC,CAAC;SACD,IAAI,CAAC,EAAE,CAAC;SACR,IAAI,EAAE,CAAC;AACZ,CAAC;AAED,SAAS,QAAQ,CAAC,KAAc;IAC9B,OAAO,KAAK,IAAI,OAAO,KAAK,KAAK,QAAQ,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,KAAK,CAAC;QAChE,CAAC,CAAE,KAAiC;QACpC,CAAC,CAAC,IAAI,CAAC;AACX,CAAC;AAED,SAAS,WAAW,CAAC,KAAc;IACjC,OAAO,OAAO,KAAK,KAAK,QAAQ,IAAI,KAAK,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,SAAS,CAAC;AAC9E,CAAC","sourcesContent":["export interface NotionMarkdownSection {\n  heading: string | null;\n  headingLevel: number | null;\n  markdown: string;\n}\n\nexport interface NotionMarkdownResult {\n  markdown: string;\n  sections: NotionMarkdownSection[];\n}\n\nexport function notionBlocksToMarkdown(\n  blocks: readonly unknown[],\n): NotionMarkdownResult {\n  const sections: NotionMarkdownSection[] = [];\n  let current: NotionMarkdownSection = {\n    heading: null,\n    headingLevel: null,\n    markdown: \"\",\n  };\n  const flush = () => {\n    const markdown = current.markdown.trim();\n    if (markdown) sections.push({ ...current, markdown });\n  };\n  for (const value of blocks) {\n    const block = asRecord(value);\n    const type = stringValue(block?.type);\n    if (!block || !type) continue;\n    const line = notionBlockToMarkdown(block, type);\n    if (!line) continue;\n    const heading = type.match(/^heading_([1-3])$/);\n    if (heading) {\n      flush();\n      const headingLevel = Number(heading[1]);\n      current = {\n        heading: richText(asRecord(block[type])?.rich_text),\n        headingLevel,\n        markdown: line,\n      };\n    } else {\n      current.markdown += `${current.markdown ? \"\\n\\n\" : \"\"}${line}`;\n    }\n  }\n  flush();\n  return {\n    markdown: sections.map((section) => section.markdown).join(\"\\n\\n\"),\n    sections,\n  };\n}\n\nexport function notionBlockToMarkdown(\n  block: Record<string, unknown>,\n  type = stringValue(block.type) ?? \"\",\n): string {\n  const payload = asRecord(block[type]) ?? {};\n  const text = richText(payload.rich_text);\n  switch (type) {\n    case \"heading_1\":\n    case \"heading_2\":\n    case \"heading_3\":\n      return `${\"#\".repeat(Number(type.slice(-1)))} ${text}`.trim();\n    case \"bulleted_list_item\":\n      return `- ${text}`.trim();\n    case \"numbered_list_item\":\n      return `1. ${text}`.trim();\n    case \"to_do\":\n      return `- [${payload.checked === true ? \"x\" : \" \"}] ${text}`.trim();\n    case \"quote\":\n      return text\n        .split(\"\\n\")\n        .map((line) => `> ${line}`)\n        .join(\"\\n\");\n    case \"code\":\n      return `\\`\\`\\`${stringValue(payload.language) ?? \"\"}\\n${text}\\n\\`\\`\\``;\n    case \"divider\":\n      return \"---\";\n    case \"bookmark\":\n    case \"embed\": {\n      const url = stringValue(payload.url);\n      return url ? `[${url}](${url})` : \"\";\n    }\n    case \"child_page\":\n      return `## ${stringValue(payload.title) ?? \"Untitled page\"}`;\n    case \"image\":\n    case \"video\":\n    case \"audio\":\n    case \"file\":\n    case \"pdf\": {\n      const file = asRecord(payload.file) ?? asRecord(payload.external);\n      const url = stringValue(file?.url);\n      const caption = richText(payload.caption) || type;\n      return url ? `![${caption}](${url})` : caption;\n    }\n    default:\n      return text;\n  }\n}\n\nfunction richText(value: unknown): string {\n  return (Array.isArray(value) ? value : [])\n    .map((part) => {\n      const record = asRecord(part);\n      return (\n        stringValue(record?.plain_text) ??\n        stringValue(asRecord(record?.text)?.content) ??\n        \"\"\n      );\n    })\n    .join(\"\")\n    .trim();\n}\n\nfunction asRecord(value: unknown): Record<string, unknown> | null {\n  return value && typeof value === \"object\" && !Array.isArray(value)\n    ? (value as Record<string, unknown>)\n    : null;\n}\n\nfunction stringValue(value: unknown): string | undefined {\n  return typeof value === \"string\" && value.trim() ? value.trim() : undefined;\n}\n"]}

package/dist/ingestion/office.d.ts

@@ -0,0 +1,45 @@
+export type OfficeDocumentType = "pptx" | "docx" | "pdf" | "txt" | "md" | "csv";
+export interface ParseOfficeDocumentInput {
+    data: Uint8Array;
+    fileName: string;
+    mimeType?: string;
+    signal?: AbortSignal;
+}
+export interface ParsedOfficeDocument {
+    text: string;
+    fileType: OfficeDocumentType;
+    title: string;
+    parser: "structured-pptx" | "mammoth-docx" | "officeparser-pdf" | "plain-text";
+    parts: OfficeDocumentPart[];
+    metadata: Record<string, unknown>;
+    warnings: string[];
+}
+export interface OfficeDocumentPart {
+    kind: "slide" | "section" | "page" | "document";
+    index: number;
+    title: string;
+    text: string;
+    notes?: string;
+    textRuns?: ParsedPptxTextRun[];
+    images?: ParsedPptxImage[];
+    metadata?: Record<string, unknown>;
+}
+export interface OfficeDocumentAst {
+    type?: string;
+    metadata?: Record<string, unknown>;
+    warnings?: unknown[];
+    toText(): string | Promise<string>;
+}
+export interface OfficeDocumentParser {
+    parseOffice(data: Uint8Array, options: {
+        fileType: OfficeDocumentType;
+        abortSignal?: AbortSignal;
+        extractAttachments: boolean;
+        ignoreNotes: boolean;
+    }): Promise<OfficeDocumentAst>;
+}
+export declare function parseOfficeDocument(input: ParseOfficeDocumentInput, parser?: OfficeDocumentParser): Promise<ParsedOfficeDocument>;
+export declare function detectOfficeDocumentType(fileName: string, mimeType?: string): OfficeDocumentType;
+export declare function normalizeDocumentText(value: string): string;
+import { type ParsedPptxImage, type ParsedPptxTextRun } from "./pptx.js";
+//# sourceMappingURL=office.d.ts.map

package/dist/ingestion/office.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"office.d.ts","sourceRoot":"","sources":["../../src/ingestion/office.ts"],"names":[],"mappings":"AAAA,MAAM,MAAM,kBAAkB,GAAG,MAAM,GAAG,MAAM,GAAG,KAAK,GAAG,KAAK,GAAG,IAAI,GAAG,KAAK,CAAC;AAEhF,MAAM,WAAW,wBAAwB;IACvC,IAAI,EAAE,UAAU,CAAC;IACjB,QAAQ,EAAE,MAAM,CAAC;IACjB,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,MAAM,CAAC,EAAE,WAAW,CAAC;CACtB;AAED,MAAM,WAAW,oBAAoB;IACnC,IAAI,EAAE,MAAM,CAAC;IACb,QAAQ,EAAE,kBAAkB,CAAC;IAC7B,KAAK,EAAE,MAAM,CAAC;IACd,MAAM,EACF,iBAAiB,GACjB,cAAc,GACd,kBAAkB,GAClB,YAAY,CAAC;IACjB,KAAK,EAAE,kBAAkB,EAAE,CAAC;IAC5B,QAAQ,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;IAClC,QAAQ,EAAE,MAAM,EAAE,CAAC;CACpB;AAED,MAAM,WAAW,kBAAkB;IACjC,IAAI,EAAE,OAAO,GAAG,SAAS,GAAG,MAAM,GAAG,UAAU,CAAC;IAChD,KAAK,EAAE,MAAM,CAAC;IACd,KAAK,EAAE,MAAM,CAAC;IACd,IAAI,EAAE,MAAM,CAAC;IACb,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,QAAQ,CAAC,EAAE,iBAAiB,EAAE,CAAC;IAC/B,MAAM,CAAC,EAAE,eAAe,EAAE,CAAC;IAC3B,QAAQ,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;CACpC;AAED,MAAM,WAAW,iBAAiB;IAChC,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,QAAQ,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;IACnC,QAAQ,CAAC,EAAE,OAAO,EAAE,CAAC;IACrB,MAAM,IAAI,MAAM,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC;CACpC;AAED,MAAM,WAAW,oBAAoB;IACnC,WAAW,CACT,IAAI,EAAE,UAAU,EAChB,OAAO,EAAE;QACP,QAAQ,EAAE,kBAAkB,CAAC;QAC7B,WAAW,CAAC,EAAE,WAAW,CAAC;QAC1B,kBAAkB,EAAE,OAAO,CAAC;QAC5B,WAAW,EAAE,OAAO,CAAC;KACtB,GACA,OAAO,CAAC,iBAAiB,CAAC,CAAC;CAC/B;AAED,wBAAsB,mBAAmB,CACvC,KAAK,EAAE,wBAAwB,EAC/B,MAAM,CAAC,EAAE,oBAAoB,GAC5B,OAAO,CAAC,oBAAoB,CAAC,CAyG/B;AAED,wBAAgB,wBAAwB,CACtC,QAAQ,EAAE,MAAM,EAChB,QAAQ,CAAC,EAAE,MAAM,GAChB,kBAAkB,CAapB;AAED,wBAAgB,qBAAqB,CAAC,KAAK,EAAE,MAAM,GAAG,MAAM,CAO3D;AAqBD,OAAO,EAEL,KAAK,eAAe,EACpB,KAAK,iBAAiB,EACvB,MAAM,WAAW,CAAC"}

package/dist/ingestion/office.js

@@ -0,0 +1,141 @@
+export async function parseOfficeDocument(input, parser) {
+    const fileType = detectOfficeDocumentType(input.fileName, input.mimeType);
+    if (fileType === "txt" || fileType === "md" || fileType === "csv") {
+        const text = normalizeDocumentText(new TextDecoder().decode(input.data));
+        return {
+            text,
+            fileType,
+            title: input.fileName,
+            parser: "plain-text",
+            parts: [{ kind: "document", index: 0, title: input.fileName, text }],
+            metadata: {},
+            warnings: [],
+        };
+    }
+    if (fileType === "pptx") {
+        const presentation = await parsePptxPresentation(input.data);
+        const parts = presentation.slides.map((slide, index) => {
+            const body = normalizeDocumentText(slide.texts.map((run) => run.content).join("\n"));
+            return {
+                kind: "slide",
+                index,
+                title: slide.texts
+                    .find((run) => run.content.trim())
+                    ?.content.trim()
+                    .slice(0, 160) ?? `Slide ${index + 1}`,
+                text: normalizeDocumentText([body, slide.notes ? `Speaker notes\n${slide.notes}` : ""]
+                    .filter(Boolean)
+                    .join("\n\n")),
+                notes: slide.notes,
+                textRuns: slide.texts,
+                images: slide.images,
+                metadata: {
+                    layoutHint: slide.layoutHint,
+                    theme: presentation.theme,
+                },
+            };
+        });
+        return {
+            text: parts.map((part) => part.text).join("\n\n"),
+            fileType,
+            title: presentation.title,
+            parser: "structured-pptx",
+            parts,
+            metadata: { theme: presentation.theme, slideCount: parts.length },
+            warnings: [],
+        };
+    }
+    if (fileType === "docx") {
+        const document = await parseDocxDocument(input.data);
+        const parts = document.sections.map((section, index) => {
+            const html = sanitizeInertDocumentHtml(section.html);
+            return {
+                kind: "section",
+                index,
+                title: section.heading || `Section ${index + 1}`,
+                text: normalizeDocumentText([section.heading, section.text].filter(Boolean).join("\n\n")),
+                metadata: { html },
+            };
+        });
+        return {
+            text: normalizeDocumentText(document.text),
+            fileType,
+            title: document.title,
+            parser: "mammoth-docx",
+            parts,
+            metadata: { sectionCount: parts.length },
+            warnings: [],
+        };
+    }
+    const resolvedParser = parser ?? (await loadOfficeParser());
+    if (!resolvedParser) {
+        throw new Error(`Parsing ${fileType.toUpperCase()} uploads requires the optional officeparser dependency.`);
+    }
+    const ast = await resolvedParser.parseOffice(input.data, {
+        fileType,
+        abortSignal: input.signal,
+        extractAttachments: false,
+        ignoreNotes: false,
+    });
+    const text = normalizeDocumentText(await ast.toText());
+    const pages = text.split(/\f+/).map(normalizeDocumentText).filter(Boolean);
+    const parts = (pages.length > 0 ? pages : [text]).map((page, index) => ({
+        kind: "page",
+        index,
+        title: `Page ${index + 1}`,
+        text: page,
+    }));
+    return {
+        text,
+        fileType: isOfficeDocumentType(ast.type) ? ast.type : fileType,
+        title: input.fileName,
+        parser: "officeparser-pdf",
+        parts,
+        metadata: ast.metadata ?? {},
+        warnings: (ast.warnings ?? []).map(String),
+    };
+}
+export function detectOfficeDocumentType(fileName, mimeType) {
+    const mime = mimeType?.toLowerCase() ?? "";
+    if (mime.includes("presentationml"))
+        return "pptx";
+    if (mime.includes("wordprocessingml"))
+        return "docx";
+    if (mime.includes("pdf"))
+        return "pdf";
+    if (mime.includes("markdown"))
+        return "md";
+    if (mime.includes("csv"))
+        return "csv";
+    if (mime.startsWith("text/"))
+        return "txt";
+    const extension = fileName.toLowerCase().match(/\.([a-z0-9]+)$/)?.[1];
+    if (isOfficeDocumentType(extension))
+        return extension;
+    throw new Error(`Unsupported upload type ${mimeType ?? extension ?? "unknown"}; supported formats are PPTX, DOCX, PDF, and text.`);
+}
+export function normalizeDocumentText(value) {
+    return value
+        .replace(/\r\n?/g, "\n")
+        .replace(/[\t\f\v ]+/g, " ")
+        .replace(/ *\n */g, "\n")
+        .replace(/\n{3,}/g, "\n\n")
+        .trim();
+}
+async function loadOfficeParser() {
+    const specifier = "officeparser";
+    try {
+        const module = (await import(specifier));
+        return module.OfficeParser ?? null;
+    }
+    catch {
+        return null;
+    }
+}
+function isOfficeDocumentType(value) {
+    return (typeof value === "string" &&
+        ["pptx", "docx", "pdf", "txt", "md", "csv"].includes(value));
+}
+import { parseDocxDocument, sanitizeInertDocumentHtml } from "./docx.js";
+import { parsePptxPresentation, } from "./pptx.js";
+//# sourceMappingURL=office.js.map

package/dist/ingestion/office.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"office.js","sourceRoot":"","sources":["../../src/ingestion/office.ts"],"names":[],"mappings":"AAqDA,MAAM,CAAC,KAAK,UAAU,mBAAmB,CACvC,KAA+B,EAC/B,MAA6B;IAE7B,MAAM,QAAQ,GAAG,wBAAwB,CAAC,KAAK,CAAC,QAAQ,EAAE,KAAK,CAAC,QAAQ,CAAC,CAAC;IAC1E,IAAI,QAAQ,KAAK,KAAK,IAAI,QAAQ,KAAK,IAAI,IAAI,QAAQ,KAAK,KAAK,EAAE,CAAC;QAClE,MAAM,IAAI,GAAG,qBAAqB,CAAC,IAAI,WAAW,EAAE,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC;QACzE,OAAO;YACL,IAAI;YACJ,QAAQ;YACR,KAAK,EAAE,KAAK,CAAC,QAAQ;YACrB,MAAM,EAAE,YAAY;YACpB,KAAK,EAAE,CAAC,EAAE,IAAI,EAAE,UAAU,EAAE,KAAK,EAAE,CAAC,EAAE,KAAK,EAAE,KAAK,CAAC,QAAQ,EAAE,IAAI,EAAE,CAAC;YACpE,QAAQ,EAAE,EAAE;YACZ,QAAQ,EAAE,EAAE;SACb,CAAC;IACJ,CAAC;IACD,IAAI,QAAQ,KAAK,MAAM,EAAE,CAAC;QACxB,MAAM,YAAY,GAAG,MAAM,qBAAqB,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QAC7D,MAAM,KAAK,GAAG,YAAY,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,KAAK,EAAE,EAAE;YACrD,MAAM,IAAI,GAAG,qBAAqB,CAChC,KAAK,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CACjD,CAAC;YACF,OAAO;gBACL,IAAI,EAAE,OAAgB;gBACtB,KAAK;gBACL,KAAK,EACH,KAAK,CAAC,KAAK;qBACR,IAAI,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,OAAO,CAAC,IAAI,EAAE,CAAC;oBAClC,EAAE,OAAO,CAAC,IAAI,EAAE;qBACf,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,IAAI,SAAS,KAAK,GAAG,CAAC,EAAE;gBAC1C,IAAI,EAAE,qBAAqB,CACzB,CAAC,IAAI,EAAE,KAAK,CAAC,KAAK,CAAC,CAAC,CAAC,kBAAkB,KAAK,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;qBACvD,MAAM,CAAC,OAAO,CAAC;qBACf,IAAI,CAAC,MAAM,CAAC,CAChB;gBACD,KAAK,EAAE,KAAK,CAAC,KAAK;gBAClB,QAAQ,EAAE,KAAK,CAAC,KAAK;gBACrB,MAAM,EAAE,KAAK,CAAC,MAAM;gBACpB,QAAQ,EAAE;oBACR,UAAU,EAAE,KAAK,CAAC,UAAU;oBAC5B,KAAK,EAAE,YAAY,CAAC,KAAK;iBAC1B;aACF,CAAC;QACJ,CAAC,CAAC,CAAC;QACH,OAAO;YACL,IAAI,EAAE,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC;YACjD,QAAQ;YACR,KAAK,EAAE,YAAY,CAAC,KAAK;YACzB,MAAM,EAAE,iBAAiB;YACzB,KAAK;YACL,QAAQ,EAAE,EAAE,KAAK,EAAE,YAAY,CAAC,KAAK,EAAE,UAAU,EAAE,KAAK,CAAC,MAAM,EAAE;YACjE,QAAQ,EAAE,EAAE;SACb,CAAC;IACJ,CAAC;IACD,IAAI,QAAQ,KAAK,MAAM,EAAE,CAAC;QACxB,MAAM,QAAQ,GAAG,MAAM,iBAAiB,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QACrD,MAAM,KAAK,GAAG,QAAQ,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,OAAO,EAAE,KAAK,EAAE,EAAE;YACrD,MAAM,IAAI,GAAG,yBAAyB,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;YACrD,OAAO;gBACL,IAAI,EAAE,SAAkB;gBACxB,KAAK;gBACL,KAAK,EAAE,OAAO,CAAC,OAAO,IAAI,WAAW,KAAK,GAAG,CAAC,EAAE;gBAChD,IAAI,EAAE,qBAAqB,CACzB,CAAC,OAAO,CAAC,OAAO,EAAE,OAAO,CAAC,IAAI,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,CAC7D;gBACD,QAAQ,EAAE,EAAE,IAAI,EAAE;aACnB,CAAC;QACJ,CAAC,CAAC,CAAC;QACH,OAAO;YACL,IAAI,EAAE,qBAAqB,CAAC,QAAQ,CAAC,IAAI,CAAC;YAC1C,QAAQ;YACR,KAAK,EAAE,QAAQ,CAAC,KAAK;YACrB,MAAM,EAAE,cAAc;YACtB,KAAK;YACL,QAAQ,EAAE,EAAE,YAAY,EAAE,KAAK,CAAC,MAAM,EAAE;YACxC,QAAQ,EAAE,EAAE;SACb,CAAC;IACJ,CAAC;IACD,MAAM,cAAc,GAAG,MAAM,IAAI,CAAC,MAAM,gBAAgB,EAAE,CAAC,CAAC;IAC5D,IAAI,CAAC,cAAc,EAAE,CAAC;QACpB,MAAM,IAAI,KAAK,CACb,WAAW,QAAQ,CAAC,WAAW,EAAE,yDAAyD,CAC3F,CAAC;IACJ,CAAC;IACD,MAAM,GAAG,GAAG,MAAM,cAAc,CAAC,WAAW,CAAC,KAAK,CAAC,IAAI,EAAE;QACvD,QAAQ;QACR,WAAW,EAAE,KAAK,CAAC,MAAM;QACzB,kBAAkB,EAAE,KAAK;QACzB,WAAW,EAAE,KAAK;KACnB,CAAC,CAAC;IACH,MAAM,IAAI,GAAG,qBAAqB,CAAC,MAAM,GAAG,CAAC,MAAM,EAAE,CAAC,CAAC;IACvD,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,qBAAqB,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC;IAC3E,MAAM,KAAK,GAAG,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC;QACtE,IAAI,EAAE,MAAe;QACrB,KAAK;QACL,KAAK,EAAE,QAAQ,KAAK,GAAG,CAAC,EAAE;QAC1B,IAAI,EAAE,IAAI;KACX,CAAC,CAAC,CAAC;IACJ,OAAO;QACL,IAAI;QACJ,QAAQ,EAAE,oBAAoB,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,QAAQ;QAC9D,KAAK,EAAE,KAAK,CAAC,QAAQ;QACrB,MAAM,EAAE,kBAAkB;QAC1B,KAAK;QACL,QAAQ,EAAE,GAAG,CAAC,QAAQ,IAAI,EAAE;QAC5B,QAAQ,EAAE,CAAC,GAAG,CAAC,QAAQ,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,MAAM,CAAC;KAC3C,CAAC;AACJ,CAAC;AAED,MAAM,UAAU,wBAAwB,CACtC,QAAgB,EAChB,QAAiB;IAEjB,MAAM,IAAI,GAAG,QAAQ,EAAE,WAAW,EAAE,IAAI,EAAE,CAAC;IAC3C,IAAI,IAAI,CAAC,QAAQ,CAAC,gBAAgB,CAAC;QAAE,OAAO,MAAM,CAAC;IACnD,IAAI,IAAI,CAAC,QAAQ,CAAC,kBAAkB,CAAC;QAAE,OAAO,MAAM,CAAC;IACrD,IAAI,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC;QAAE,OAAO,KAAK,CAAC;IACvC,IAAI,IAAI,CAAC,QAAQ,CAAC,UAAU,CAAC;QAAE,OAAO,IAAI,CAAC;IAC3C,IAAI,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC;QAAE,OAAO,KAAK,CAAC;IACvC,IAAI,IAAI,CAAC,UAAU,CAAC,OAAO,CAAC;QAAE,OAAO,KAAK,CAAC;IAC3C,MAAM,SAAS,GAAG,QAAQ,CAAC,WAAW,EAAE,CAAC,KAAK,CAAC,gBAAgB,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;IACtE,IAAI,oBAAoB,CAAC,SAAS,CAAC;QAAE,OAAO,SAAS,CAAC;IACtD,MAAM,IAAI,KAAK,CACb,2BAA2B,QAAQ,IAAI,SAAS,IAAI,SAAS,oDAAoD,CAClH,CAAC;AACJ,CAAC;AAED,MAAM,UAAU,qBAAqB,CAAC,KAAa;IACjD,OAAO,KAAK;SACT,OAAO,CAAC,QAAQ,EAAE,IAAI,CAAC;SACvB,OAAO,CAAC,aAAa,EAAE,GAAG,CAAC;SAC3B,OAAO,CAAC,SAAS,EAAE,IAAI,CAAC;SACxB,OAAO,CAAC,SAAS,EAAE,MAAM,CAAC;SAC1B,IAAI,EAAE,CAAC;AACZ,CAAC;AAED,KAAK,UAAU,gBAAgB;IAC7B,MAAM,SAAS,GAAG,cAAc,CAAC;IACjC,IAAI,CAAC;QACH,MAAM,MAAM,GAAG,CAAC,MAAM,MAAM,CAAC,SAAS,CAAC,CAEtC,CAAC;QACF,OAAO,MAAM,CAAC,YAAY,IAAI,IAAI,CAAC;IACrC,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,IAAI,CAAC;IACd,CAAC;AACH,CAAC;AAED,SAAS,oBAAoB,CAAC,KAAc;IAC1C,OAAO,CACL,OAAO,KAAK,KAAK,QAAQ;QACxB,CAAC,MAAM,EAAE,MAAM,EAAE,KAAK,EAAE,KAAK,EAAE,IAAI,EAAE,KAAK,CAAc,CAAC,QAAQ,CAAC,KAAK,CAAC,CAC1E,CAAC;AACJ,CAAC;AACD,OAAO,EAAE,iBAAiB,EAAE,yBAAyB,EAAE,MAAM,WAAW,CAAC;AACzE,OAAO,EACL,qBAAqB,GAGtB,MAAM,WAAW,CAAC","sourcesContent":["export type OfficeDocumentType = \"pptx\" | \"docx\" | \"pdf\" | \"txt\" | \"md\" | \"csv\";\n\nexport interface ParseOfficeDocumentInput {\n  data: Uint8Array;\n  fileName: string;\n  mimeType?: string;\n  signal?: AbortSignal;\n}\n\nexport interface ParsedOfficeDocument {\n  text: string;\n  fileType: OfficeDocumentType;\n  title: string;\n  parser:\n    | \"structured-pptx\"\n    | \"mammoth-docx\"\n    | \"officeparser-pdf\"\n    | \"plain-text\";\n  parts: OfficeDocumentPart[];\n  metadata: Record<string, unknown>;\n  warnings: string[];\n}\n\nexport interface OfficeDocumentPart {\n  kind: \"slide\" | \"section\" | \"page\" | \"document\";\n  index: number;\n  title: string;\n  text: string;\n  notes?: string;\n  textRuns?: ParsedPptxTextRun[];\n  images?: ParsedPptxImage[];\n  metadata?: Record<string, unknown>;\n}\n\nexport interface OfficeDocumentAst {\n  type?: string;\n  metadata?: Record<string, unknown>;\n  warnings?: unknown[];\n  toText(): string | Promise<string>;\n}\n\nexport interface OfficeDocumentParser {\n  parseOffice(\n    data: Uint8Array,\n    options: {\n      fileType: OfficeDocumentType;\n      abortSignal?: AbortSignal;\n      extractAttachments: boolean;\n      ignoreNotes: boolean;\n    },\n  ): Promise<OfficeDocumentAst>;\n}\n\nexport async function parseOfficeDocument(\n  input: ParseOfficeDocumentInput,\n  parser?: OfficeDocumentParser,\n): Promise<ParsedOfficeDocument> {\n  const fileType = detectOfficeDocumentType(input.fileName, input.mimeType);\n  if (fileType === \"txt\" || fileType === \"md\" || fileType === \"csv\") {\n    const text = normalizeDocumentText(new TextDecoder().decode(input.data));\n    return {\n      text,\n      fileType,\n      title: input.fileName,\n      parser: \"plain-text\",\n      parts: [{ kind: \"document\", index: 0, title: input.fileName, text }],\n      metadata: {},\n      warnings: [],\n    };\n  }\n  if (fileType === \"pptx\") {\n    const presentation = await parsePptxPresentation(input.data);\n    const parts = presentation.slides.map((slide, index) => {\n      const body = normalizeDocumentText(\n        slide.texts.map((run) => run.content).join(\"\\n\"),\n      );\n      return {\n        kind: \"slide\" as const,\n        index,\n        title:\n          slide.texts\n            .find((run) => run.content.trim())\n            ?.content.trim()\n            .slice(0, 160) ?? `Slide ${index + 1}`,\n        text: normalizeDocumentText(\n          [body, slide.notes ? `Speaker notes\\n${slide.notes}` : \"\"]\n            .filter(Boolean)\n            .join(\"\\n\\n\"),\n        ),\n        notes: slide.notes,\n        textRuns: slide.texts,\n        images: slide.images,\n        metadata: {\n          layoutHint: slide.layoutHint,\n          theme: presentation.theme,\n        },\n      };\n    });\n    return {\n      text: parts.map((part) => part.text).join(\"\\n\\n\"),\n      fileType,\n      title: presentation.title,\n      parser: \"structured-pptx\",\n      parts,\n      metadata: { theme: presentation.theme, slideCount: parts.length },\n      warnings: [],\n    };\n  }\n  if (fileType === \"docx\") {\n    const document = await parseDocxDocument(input.data);\n    const parts = document.sections.map((section, index) => {\n      const html = sanitizeInertDocumentHtml(section.html);\n      return {\n        kind: \"section\" as const,\n        index,\n        title: section.heading || `Section ${index + 1}`,\n        text: normalizeDocumentText(\n          [section.heading, section.text].filter(Boolean).join(\"\\n\\n\"),\n        ),\n        metadata: { html },\n      };\n    });\n    return {\n      text: normalizeDocumentText(document.text),\n      fileType,\n      title: document.title,\n      parser: \"mammoth-docx\",\n      parts,\n      metadata: { sectionCount: parts.length },\n      warnings: [],\n    };\n  }\n  const resolvedParser = parser ?? (await loadOfficeParser());\n  if (!resolvedParser) {\n    throw new Error(\n      `Parsing ${fileType.toUpperCase()} uploads requires the optional officeparser dependency.`,\n    );\n  }\n  const ast = await resolvedParser.parseOffice(input.data, {\n    fileType,\n    abortSignal: input.signal,\n    extractAttachments: false,\n    ignoreNotes: false,\n  });\n  const text = normalizeDocumentText(await ast.toText());\n  const pages = text.split(/\\f+/).map(normalizeDocumentText).filter(Boolean);\n  const parts = (pages.length > 0 ? pages : [text]).map((page, index) => ({\n    kind: \"page\" as const,\n    index,\n    title: `Page ${index + 1}`,\n    text: page,\n  }));\n  return {\n    text,\n    fileType: isOfficeDocumentType(ast.type) ? ast.type : fileType,\n    title: input.fileName,\n    parser: \"officeparser-pdf\",\n    parts,\n    metadata: ast.metadata ?? {},\n    warnings: (ast.warnings ?? []).map(String),\n  };\n}\n\nexport function detectOfficeDocumentType(\n  fileName: string,\n  mimeType?: string,\n): OfficeDocumentType {\n  const mime = mimeType?.toLowerCase() ?? \"\";\n  if (mime.includes(\"presentationml\")) return \"pptx\";\n  if (mime.includes(\"wordprocessingml\")) return \"docx\";\n  if (mime.includes(\"pdf\")) return \"pdf\";\n  if (mime.includes(\"markdown\")) return \"md\";\n  if (mime.includes(\"csv\")) return \"csv\";\n  if (mime.startsWith(\"text/\")) return \"txt\";\n  const extension = fileName.toLowerCase().match(/\\.([a-z0-9]+)$/)?.[1];\n  if (isOfficeDocumentType(extension)) return extension;\n  throw new Error(\n    `Unsupported upload type ${mimeType ?? extension ?? \"unknown\"}; supported formats are PPTX, DOCX, PDF, and text.`,\n  );\n}\n\nexport function normalizeDocumentText(value: string): string {\n  return value\n    .replace(/\\r\\n?/g, \"\\n\")\n    .replace(/[\\t\\f\\v ]+/g, \" \")\n    .replace(/ *\\n */g, \"\\n\")\n    .replace(/\\n{3,}/g, \"\\n\\n\")\n    .trim();\n}\n\nasync function loadOfficeParser(): Promise<OfficeDocumentParser | null> {\n  const specifier = \"officeparser\";\n  try {\n    const module = (await import(specifier)) as {\n      OfficeParser?: OfficeDocumentParser;\n    };\n    return module.OfficeParser ?? null;\n  } catch {\n    return null;\n  }\n}\n\nfunction isOfficeDocumentType(value: unknown): value is OfficeDocumentType {\n  return (\n    typeof value === \"string\" &&\n    ([\"pptx\", \"docx\", \"pdf\", \"txt\", \"md\", \"csv\"] as string[]).includes(value)\n  );\n}\nimport { parseDocxDocument, sanitizeInertDocumentHtml } from \"./docx.js\";\nimport {\n  parsePptxPresentation,\n  type ParsedPptxImage,\n  type ParsedPptxTextRun,\n} from \"./pptx.js\";\n"]}

package/dist/ingestion/orchestration.d.ts

@@ -0,0 +1,52 @@
+export type IngestionPhase = "inventory" | "fetch" | "completed";
+export interface IngestionCheckpoint {
+    phase: IngestionPhase;
+    cursor: string | null;
+    inventoryComplete: boolean;
+    itemOffset: number;
+    itemsInventoried: number;
+    itemsProcessed: number;
+    itemsFailed: number;
+}
+export interface IngestionLease {
+    owner: string;
+    token: string;
+    acquiredAt: number;
+    expiresAt: number;
+}
+export interface AcquireIngestionLeaseInput {
+    owner: string;
+    token: string;
+    ttlMs: number;
+    now?: number;
+}
+export type AcquireIngestionLeaseResult = {
+    acquired: true;
+    lease: IngestionLease;
+} | {
+    acquired: false;
+    lease: IngestionLease;
+};
+export interface IngestionBudgetLimits {
+    runtimeMs: number;
+    itemBudget: number;
+    batchBudget: number;
+}
+export interface IngestionBudgetState {
+    startedAt: number;
+    items: number;
+    batches: number;
+}
+export type IngestionBudgetStopReason = "runtime" | "items" | "batches" | null;
+export declare function createIngestionCheckpoint(): IngestionCheckpoint;
+export declare function acquireIngestionLease(current: IngestionLease | null | undefined, input: AcquireIngestionLeaseInput): AcquireIngestionLeaseResult;
+export declare function renewIngestionLease(current: IngestionLease, input: Pick<AcquireIngestionLeaseInput, "token" | "ttlMs" | "now">): IngestionLease | null;
+export declare function releaseIngestionLease(current: IngestionLease | null | undefined, token: string): null | IngestionLease;
+export declare function createIngestionBudgetState(now?: number): IngestionBudgetState;
+export declare function consumeIngestionBudget(state: IngestionBudgetState, input: {
+    items?: number;
+    batches?: number;
+}): IngestionBudgetState;
+export declare function ingestionBudgetStopReason(limits: IngestionBudgetLimits, state: IngestionBudgetState, now?: number): IngestionBudgetStopReason;
+export declare function assertInventoryComplete(checkpoint: Pick<IngestionCheckpoint, "inventoryComplete">): void;
+//# sourceMappingURL=orchestration.d.ts.map

package/dist/ingestion/orchestration.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"orchestration.d.ts","sourceRoot":"","sources":["../../src/ingestion/orchestration.ts"],"names":[],"mappings":"AAAA,MAAM,MAAM,cAAc,GAAG,WAAW,GAAG,OAAO,GAAG,WAAW,CAAC;AAEjE,MAAM,WAAW,mBAAmB;IAClC,KAAK,EAAE,cAAc,CAAC;IACtB,MAAM,EAAE,MAAM,GAAG,IAAI,CAAC;IACtB,iBAAiB,EAAE,OAAO,CAAC;IAC3B,UAAU,EAAE,MAAM,CAAC;IACnB,gBAAgB,EAAE,MAAM,CAAC;IACzB,cAAc,EAAE,MAAM,CAAC;IACvB,WAAW,EAAE,MAAM,CAAC;CACrB;AAED,MAAM,WAAW,cAAc;IAC7B,KAAK,EAAE,MAAM,CAAC;IACd,KAAK,EAAE,MAAM,CAAC;IACd,UAAU,EAAE,MAAM,CAAC;IACnB,SAAS,EAAE,MAAM,CAAC;CACnB;AAED,MAAM,WAAW,0BAA0B;IACzC,KAAK,EAAE,MAAM,CAAC;IACd,KAAK,EAAE,MAAM,CAAC;IACd,KAAK,EAAE,MAAM,CAAC;IACd,GAAG,CAAC,EAAE,MAAM,CAAC;CACd;AAED,MAAM,MAAM,2BAA2B,GACnC;IAAE,QAAQ,EAAE,IAAI,CAAC;IAAC,KAAK,EAAE,cAAc,CAAA;CAAE,GACzC;IAAE,QAAQ,EAAE,KAAK,CAAC;IAAC,KAAK,EAAE,cAAc,CAAA;CAAE,CAAC;AAE/C,MAAM,WAAW,qBAAqB;IACpC,SAAS,EAAE,MAAM,CAAC;IAClB,UAAU,EAAE,MAAM,CAAC;IACnB,WAAW,EAAE,MAAM,CAAC;CACrB;AAED,MAAM,WAAW,oBAAoB;IACnC,SAAS,EAAE,MAAM,CAAC;IAClB,KAAK,EAAE,MAAM,CAAC;IACd,OAAO,EAAE,MAAM,CAAC;CACjB;AAED,MAAM,MAAM,yBAAyB,GAAG,SAAS,GAAG,OAAO,GAAG,SAAS,GAAG,IAAI,CAAC;AAE/E,wBAAgB,yBAAyB,IAAI,mBAAmB,CAU/D;AAED,wBAAgB,qBAAqB,CACnC,OAAO,EAAE,cAAc,GAAG,IAAI,GAAG,SAAS,EAC1C,KAAK,EAAE,0BAA0B,GAChC,2BAA2B,CAe7B;AAED,wBAAgB,mBAAmB,CACjC,OAAO,EAAE,cAAc,EACvB,KAAK,EAAE,IAAI,CAAC,0BAA0B,EAAE,OAAO,GAAG,OAAO,GAAG,KAAK,CAAC,GACjE,cAAc,GAAG,IAAI,CAQvB;AAED,wBAAgB,qBAAqB,CACnC,OAAO,EAAE,cAAc,GAAG,IAAI,GAAG,SAAS,EAC1C,KAAK,EAAE,MAAM,GACZ,IAAI,GAAG,cAAc,CAGvB;AAED,wBAAgB,0BAA0B,CACxC,GAAG,SAAa,GACf,oBAAoB,CAEtB;AAED,wBAAgB,sBAAsB,CACpC,KAAK,EAAE,oBAAoB,EAC3B,KAAK,EAAE;IAAE,KAAK,CAAC,EAAE,MAAM,CAAC;IAAC,OAAO,CAAC,EAAE,MAAM,CAAA;CAAE,GAC1C,oBAAoB,CAMtB;AAED,wBAAgB,yBAAyB,CACvC,MAAM,EAAE,qBAAqB,EAC7B,KAAK,EAAE,oBAAoB,EAC3B,GAAG,SAAa,GACf,yBAAyB,CAW3B;AAED,wBAAgB,uBAAuB,CACrC,UAAU,EAAE,IAAI,CAAC,mBAAmB,EAAE,mBAAmB,CAAC,GACzD,IAAI,CAMN"}