@agent-native/core 0.103.1 → 0.105.0

package/dist/ingestion/orchestration.js

@@ -0,0 +1,89 @@
+export function createIngestionCheckpoint() {
+    return {
+        phase: "inventory",
+        cursor: null,
+        inventoryComplete: false,
+        itemOffset: 0,
+        itemsInventoried: 0,
+        itemsProcessed: 0,
+        itemsFailed: 0,
+    };
+}
+export function acquireIngestionLease(current, input) {
+    const now = input.now ?? Date.now();
+    if (current && current.expiresAt > now && current.owner !== input.owner) {
+        return { acquired: false, lease: current };
+    }
+    const ttlMs = positiveInteger(input.ttlMs, "lease ttlMs");
+    return {
+        acquired: true,
+        lease: {
+            owner: nonEmpty(input.owner, "lease owner"),
+            token: nonEmpty(input.token, "lease token"),
+            acquiredAt: now,
+            expiresAt: now + ttlMs,
+        },
+    };
+}
+export function renewIngestionLease(current, input) {
+    if (current.token !== input.token)
+        return null;
+    const now = input.now ?? Date.now();
+    if (current.expiresAt <= now)
+        return null;
+    return {
+        ...current,
+        expiresAt: now + positiveInteger(input.ttlMs, "lease ttlMs"),
+    };
+}
+export function releaseIngestionLease(current, token) {
+    if (!current || current.token === token)
+        return null;
+    return current;
+}
+export function createIngestionBudgetState(now = Date.now()) {
+    return { startedAt: now, items: 0, batches: 0 };
+}
+export function consumeIngestionBudget(state, input) {
+    return {
+        ...state,
+        items: state.items + nonNegativeInteger(input.items ?? 0, "items"),
+        batches: state.batches + nonNegativeInteger(input.batches ?? 0, "batches"),
+    };
+}
+export function ingestionBudgetStopReason(limits, state, now = Date.now()) {
+    if (now - state.startedAt >= positiveInteger(limits.runtimeMs, "runtimeMs")) {
+        return "runtime";
+    }
+    if (state.items >= positiveInteger(limits.itemBudget, "itemBudget")) {
+        return "items";
+    }
+    if (state.batches >= positiveInteger(limits.batchBudget, "batchBudget")) {
+        return "batches";
+    }
+    return null;
+}
+export function assertInventoryComplete(checkpoint) {
+    if (!checkpoint.inventoryComplete) {
+        throw new Error("Context import fetch cannot start before inventory is complete.");
+    }
+}
+function nonEmpty(value, name) {
+    const normalized = value.trim();
+    if (!normalized)
+        throw new Error(`${name} is required.`);
+    return normalized;
+}
+function positiveInteger(value, name) {
+    if (!Number.isInteger(value) || value <= 0) {
+        throw new Error(`${name} must be a positive integer.`);
+    }
+    return value;
+}
+function nonNegativeInteger(value, name) {
+    if (!Number.isInteger(value) || value < 0) {
+        throw new Error(`${name} must be a non-negative integer.`);
+    }
+    return value;
+}
+//# sourceMappingURL=orchestration.js.map

package/dist/ingestion/orchestration.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"orchestration.js","sourceRoot":"","sources":["../../src/ingestion/orchestration.ts"],"names":[],"mappings":"AA4CA,MAAM,UAAU,yBAAyB;IACvC,OAAO;QACL,KAAK,EAAE,WAAW;QAClB,MAAM,EAAE,IAAI;QACZ,iBAAiB,EAAE,KAAK;QACxB,UAAU,EAAE,CAAC;QACb,gBAAgB,EAAE,CAAC;QACnB,cAAc,EAAE,CAAC;QACjB,WAAW,EAAE,CAAC;KACf,CAAC;AACJ,CAAC;AAED,MAAM,UAAU,qBAAqB,CACnC,OAA0C,EAC1C,KAAiC;IAEjC,MAAM,GAAG,GAAG,KAAK,CAAC,GAAG,IAAI,IAAI,CAAC,GAAG,EAAE,CAAC;IACpC,IAAI,OAAO,IAAI,OAAO,CAAC,SAAS,GAAG,GAAG,IAAI,OAAO,CAAC,KAAK,KAAK,KAAK,CAAC,KAAK,EAAE,CAAC;QACxE,OAAO,EAAE,QAAQ,EAAE,KAAK,EAAE,KAAK,EAAE,OAAO,EAAE,CAAC;IAC7C,CAAC;IACD,MAAM,KAAK,GAAG,eAAe,CAAC,KAAK,CAAC,KAAK,EAAE,aAAa,CAAC,CAAC;IAC1D,OAAO;QACL,QAAQ,EAAE,IAAI;QACd,KAAK,EAAE;YACL,KAAK,EAAE,QAAQ,CAAC,KAAK,CAAC,KAAK,EAAE,aAAa,CAAC;YAC3C,KAAK,EAAE,QAAQ,CAAC,KAAK,CAAC,KAAK,EAAE,aAAa,CAAC;YAC3C,UAAU,EAAE,GAAG;YACf,SAAS,EAAE,GAAG,GAAG,KAAK;SACvB;KACF,CAAC;AACJ,CAAC;AAED,MAAM,UAAU,mBAAmB,CACjC,OAAuB,EACvB,KAAkE;IAElE,IAAI,OAAO,CAAC,KAAK,KAAK,KAAK,CAAC,KAAK;QAAE,OAAO,IAAI,CAAC;IAC/C,MAAM,GAAG,GAAG,KAAK,CAAC,GAAG,IAAI,IAAI,CAAC,GAAG,EAAE,CAAC;IACpC,IAAI,OAAO,CAAC,SAAS,IAAI,GAAG;QAAE,OAAO,IAAI,CAAC;IAC1C,OAAO;QACL,GAAG,OAAO;QACV,SAAS,EAAE,GAAG,GAAG,eAAe,CAAC,KAAK,CAAC,KAAK,EAAE,aAAa,CAAC;KAC7D,CAAC;AACJ,CAAC;AAED,MAAM,UAAU,qBAAqB,CACnC,OAA0C,EAC1C,KAAa;IAEb,IAAI,CAAC,OAAO,IAAI,OAAO,CAAC,KAAK,KAAK,KAAK;QAAE,OAAO,IAAI,CAAC;IACrD,OAAO,OAAO,CAAC;AACjB,CAAC;AAED,MAAM,UAAU,0BAA0B,CACxC,GAAG,GAAG,IAAI,CAAC,GAAG,EAAE;IAEhB,OAAO,EAAE,SAAS,EAAE,GAAG,EAAE,KAAK,EAAE,CAAC,EAAE,OAAO,EAAE,CAAC,EAAE,CAAC;AAClD,CAAC;AAED,MAAM,UAAU,sBAAsB,CACpC,KAA2B,EAC3B,KAA2C;IAE3C,OAAO;QACL,GAAG,KAAK;QACR,KAAK,EAAE,KAAK,CAAC,KAAK,GAAG,kBAAkB,CAAC,KAAK,CAAC,KAAK,IAAI,CAAC,EAAE,OAAO,CAAC;QAClE,OAAO,EAAE,KAAK,CAAC,OAAO,GAAG,kBAAkB,CAAC,KAAK,CAAC,OAAO,IAAI,CAAC,EAAE,SAAS,CAAC;KAC3E,CAAC;AACJ,CAAC;AAED,MAAM,UAAU,yBAAyB,CACvC,MAA6B,EAC7B,KAA2B,EAC3B,GAAG,GAAG,IAAI,CAAC,GAAG,EAAE;IAEhB,IAAI,GAAG,GAAG,KAAK,CAAC,SAAS,IAAI,eAAe,CAAC,MAAM,CAAC,SAAS,EAAE,WAAW,CAAC,EAAE,CAAC;QAC5E,OAAO,SAAS,CAAC;IACnB,CAAC;IACD,IAAI,KAAK,CAAC,KAAK,IAAI,eAAe,CAAC,MAAM,CAAC,UAAU,EAAE,YAAY,CAAC,EAAE,CAAC;QACpE,OAAO,OAAO,CAAC;IACjB,CAAC;IACD,IAAI,KAAK,CAAC,OAAO,IAAI,eAAe,CAAC,MAAM,CAAC,WAAW,EAAE,aAAa,CAAC,EAAE,CAAC;QACxE,OAAO,SAAS,CAAC;IACnB,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC;AAED,MAAM,UAAU,uBAAuB,CACrC,UAA0D;IAE1D,IAAI,CAAC,UAAU,CAAC,iBAAiB,EAAE,CAAC;QAClC,MAAM,IAAI,KAAK,CACb,iEAAiE,CAClE,CAAC;IACJ,CAAC;AACH,CAAC;AAED,SAAS,QAAQ,CAAC,KAAa,EAAE,IAAY;IAC3C,MAAM,UAAU,GAAG,KAAK,CAAC,IAAI,EAAE,CAAC;IAChC,IAAI,CAAC,UAAU;QAAE,MAAM,IAAI,KAAK,CAAC,GAAG,IAAI,eAAe,CAAC,CAAC;IACzD,OAAO,UAAU,CAAC;AACpB,CAAC;AAED,SAAS,eAAe,CAAC,KAAa,EAAE,IAAY;IAClD,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,KAAK,CAAC,IAAI,KAAK,IAAI,CAAC,EAAE,CAAC;QAC3C,MAAM,IAAI,KAAK,CAAC,GAAG,IAAI,8BAA8B,CAAC,CAAC;IACzD,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED,SAAS,kBAAkB,CAAC,KAAa,EAAE,IAAY;IACrD,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,KAAK,CAAC,IAAI,KAAK,GAAG,CAAC,EAAE,CAAC;QAC1C,MAAM,IAAI,KAAK,CAAC,GAAG,IAAI,kCAAkC,CAAC,CAAC;IAC7D,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC","sourcesContent":["export type IngestionPhase = \"inventory\" | \"fetch\" | \"completed\";\n\nexport interface IngestionCheckpoint {\n  phase: IngestionPhase;\n  cursor: string | null;\n  inventoryComplete: boolean;\n  itemOffset: number;\n  itemsInventoried: number;\n  itemsProcessed: number;\n  itemsFailed: number;\n}\n\nexport interface IngestionLease {\n  owner: string;\n  token: string;\n  acquiredAt: number;\n  expiresAt: number;\n}\n\nexport interface AcquireIngestionLeaseInput {\n  owner: string;\n  token: string;\n  ttlMs: number;\n  now?: number;\n}\n\nexport type AcquireIngestionLeaseResult =\n  | { acquired: true; lease: IngestionLease }\n  | { acquired: false; lease: IngestionLease };\n\nexport interface IngestionBudgetLimits {\n  runtimeMs: number;\n  itemBudget: number;\n  batchBudget: number;\n}\n\nexport interface IngestionBudgetState {\n  startedAt: number;\n  items: number;\n  batches: number;\n}\n\nexport type IngestionBudgetStopReason = \"runtime\" | \"items\" | \"batches\" | null;\n\nexport function createIngestionCheckpoint(): IngestionCheckpoint {\n  return {\n    phase: \"inventory\",\n    cursor: null,\n    inventoryComplete: false,\n    itemOffset: 0,\n    itemsInventoried: 0,\n    itemsProcessed: 0,\n    itemsFailed: 0,\n  };\n}\n\nexport function acquireIngestionLease(\n  current: IngestionLease | null | undefined,\n  input: AcquireIngestionLeaseInput,\n): AcquireIngestionLeaseResult {\n  const now = input.now ?? Date.now();\n  if (current && current.expiresAt > now && current.owner !== input.owner) {\n    return { acquired: false, lease: current };\n  }\n  const ttlMs = positiveInteger(input.ttlMs, \"lease ttlMs\");\n  return {\n    acquired: true,\n    lease: {\n      owner: nonEmpty(input.owner, \"lease owner\"),\n      token: nonEmpty(input.token, \"lease token\"),\n      acquiredAt: now,\n      expiresAt: now + ttlMs,\n    },\n  };\n}\n\nexport function renewIngestionLease(\n  current: IngestionLease,\n  input: Pick<AcquireIngestionLeaseInput, \"token\" | \"ttlMs\" | \"now\">,\n): IngestionLease | null {\n  if (current.token !== input.token) return null;\n  const now = input.now ?? Date.now();\n  if (current.expiresAt <= now) return null;\n  return {\n    ...current,\n    expiresAt: now + positiveInteger(input.ttlMs, \"lease ttlMs\"),\n  };\n}\n\nexport function releaseIngestionLease(\n  current: IngestionLease | null | undefined,\n  token: string,\n): null | IngestionLease {\n  if (!current || current.token === token) return null;\n  return current;\n}\n\nexport function createIngestionBudgetState(\n  now = Date.now(),\n): IngestionBudgetState {\n  return { startedAt: now, items: 0, batches: 0 };\n}\n\nexport function consumeIngestionBudget(\n  state: IngestionBudgetState,\n  input: { items?: number; batches?: number },\n): IngestionBudgetState {\n  return {\n    ...state,\n    items: state.items + nonNegativeInteger(input.items ?? 0, \"items\"),\n    batches: state.batches + nonNegativeInteger(input.batches ?? 0, \"batches\"),\n  };\n}\n\nexport function ingestionBudgetStopReason(\n  limits: IngestionBudgetLimits,\n  state: IngestionBudgetState,\n  now = Date.now(),\n): IngestionBudgetStopReason {\n  if (now - state.startedAt >= positiveInteger(limits.runtimeMs, \"runtimeMs\")) {\n    return \"runtime\";\n  }\n  if (state.items >= positiveInteger(limits.itemBudget, \"itemBudget\")) {\n    return \"items\";\n  }\n  if (state.batches >= positiveInteger(limits.batchBudget, \"batchBudget\")) {\n    return \"batches\";\n  }\n  return null;\n}\n\nexport function assertInventoryComplete(\n  checkpoint: Pick<IngestionCheckpoint, \"inventoryComplete\">,\n): void {\n  if (!checkpoint.inventoryComplete) {\n    throw new Error(\n      \"Context import fetch cannot start before inventory is complete.\",\n    );\n  }\n}\n\nfunction nonEmpty(value: string, name: string): string {\n  const normalized = value.trim();\n  if (!normalized) throw new Error(`${name} is required.`);\n  return normalized;\n}\n\nfunction positiveInteger(value: number, name: string): number {\n  if (!Number.isInteger(value) || value <= 0) {\n    throw new Error(`${name} must be a positive integer.`);\n  }\n  return value;\n}\n\nfunction nonNegativeInteger(value: number, name: string): number {\n  if (!Number.isInteger(value) || value < 0) {\n    throw new Error(`${name} must be a non-negative integer.`);\n  }\n  return value;\n}\n"]}

package/dist/ingestion/pptx.d.ts

@@ -0,0 +1,28 @@
+export interface ParsedPptxTextRun {
+    content: string;
+    bold?: boolean;
+    italic?: boolean;
+    fontSize?: number;
+    color?: string;
+}
+export interface ParsedPptxImage {
+    data: Uint8Array;
+    mimeType: string;
+    name: string;
+}
+export interface ParsedPptxSlide {
+    texts: ParsedPptxTextRun[];
+    images: ParsedPptxImage[];
+    notes?: string;
+    layoutHint?: string;
+}
+export interface ParsedPptxPresentation {
+    title: string;
+    slides: ParsedPptxSlide[];
+    theme?: {
+        colors: string[];
+        fonts: string[];
+    };
+}
+export declare function parsePptxPresentation(fileBuffer: Uint8Array): Promise<ParsedPptxPresentation>;
+//# sourceMappingURL=pptx.d.ts.map

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

@@ -0,0 +1 @@
+{"version":3,"file":"pptx.d.ts","sourceRoot":"","sources":["../../src/ingestion/pptx.ts"],"names":[],"mappings":"AAAA,MAAM,WAAW,iBAAiB;IAChC,OAAO,EAAE,MAAM,CAAC;IAChB,IAAI,CAAC,EAAE,OAAO,CAAC;IACf,MAAM,CAAC,EAAE,OAAO,CAAC;IACjB,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,KAAK,CAAC,EAAE,MAAM,CAAC;CAChB;AAED,MAAM,WAAW,eAAe;IAC9B,IAAI,EAAE,UAAU,CAAC;IACjB,QAAQ,EAAE,MAAM,CAAC;IACjB,IAAI,EAAE,MAAM,CAAC;CACd;AAED,MAAM,WAAW,eAAe;IAC9B,KAAK,EAAE,iBAAiB,EAAE,CAAC;IAC3B,MAAM,EAAE,eAAe,EAAE,CAAC;IAC1B,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,UAAU,CAAC,EAAE,MAAM,CAAC;CACrB;AAED,MAAM,WAAW,sBAAsB;IACrC,KAAK,EAAE,MAAM,CAAC;IACd,MAAM,EAAE,eAAe,EAAE,CAAC;IAC1B,KAAK,CAAC,EAAE;QAAE,MAAM,EAAE,MAAM,EAAE,CAAC;QAAC,KAAK,EAAE,MAAM,EAAE,CAAA;KAAE,CAAC;CAC/C;AAYD,wBAAsB,qBAAqB,CACzC,UAAU,EAAE,UAAU,GACrB,OAAO,CAAC,sBAAsB,CAAC,CAgHjC"}

package/dist/ingestion/pptx.js

@@ -0,0 +1,255 @@
+export async function parsePptxPresentation(fileBuffer) {
+    const { loadZip, parseXml } = await loadPptxDependencies();
+    const zip = await loadZip(fileBuffer);
+    const presentationXml = await zip
+        .file("ppt/presentation.xml")
+        ?.async("string");
+    if (!presentationXml)
+        throw new Error("Invalid PPTX: missing ppt/presentation.xml");
+    const presentation = parseXml(presentationXml);
+    const slideIds = asArray(record(record(record(presentation)?.["p:presentation"])?.["p:sldIdLst"])?.["p:sldId"]).map((entry) => stringValue(record(entry)?.["@_r:id"]) ?? "");
+    const relationshipsXml = await zip
+        .file("ppt/_rels/presentation.xml.rels")
+        ?.async("string");
+    const relationships = relationshipsXml
+        ? parseRelationships(parseXml(relationshipsXml))
+        : new Map();
+    const slidePaths = slideIds.flatMap((id) => {
+        const relationship = relationships.get(id);
+        if (!relationship)
+            return [];
+        return [
+            relationship.target.startsWith("/")
+                ? relationship.target.slice(1)
+                : `ppt/${relationship.target}`,
+        ];
+    });
+    if (slidePaths.length === 0) {
+        slidePaths.push(...Object.keys(zip.files)
+            .filter((path) => /^ppt\/slides\/slide\d+\.xml$/.test(path))
+            .sort((a, b) => slideNumber(a) - slideNumber(b)));
+    }
+    const theme = await parseTheme(zip, parseXml);
+    const slides = [];
+    for (const slidePath of slidePaths) {
+        const xml = await zip.file(slidePath)?.async("string");
+        if (!xml)
+            continue;
+        let slide;
+        try {
+            slide = parseXml(xml);
+        }
+        catch {
+            continue;
+        }
+        const texts = [];
+        collectTextRuns(slide, texts);
+        const images = [];
+        const relationshipPath = slidePath.replace(/slides\/(slide\d+\.xml)/, "slides/_rels/$1.rels");
+        const slideRelationshipsXml = await zip
+            .file(relationshipPath)
+            ?.async("string");
+        if (slideRelationshipsXml) {
+            for (const relationship of parseRelationships(parseXml(slideRelationshipsXml)).values()) {
+                if (!relationship.type.includes("/image") &&
+                    !/\.(png|jpe?g|gif|svg|webp|bmp|tiff?|emf|wmf)$/i.test(relationship.target)) {
+                    continue;
+                }
+                const imagePath = relationship.target.startsWith("/")
+                    ? relationship.target.slice(1)
+                    : relationship.target.startsWith("../")
+                        ? `ppt/${relationship.target.replace(/^\.\.\//, "")}`
+                        : `ppt/slides/${relationship.target}`;
+                const image = zip.file(imagePath);
+                if (!image)
+                    continue;
+                const name = imagePath.split("/").at(-1) ?? "image";
+                images.push({
+                    data: new Uint8Array(await image.async("nodebuffer")),
+                    mimeType: imageMimeType(name),
+                    name,
+                });
+            }
+        }
+        const number = slideNumber(slidePath);
+        const notesXml = await zip
+            .file(`ppt/notesSlides/notesSlide${number}.xml`)
+            ?.async("string");
+        let notes;
+        if (notesXml) {
+            const runs = [];
+            collectTextRuns(parseXml(notesXml), runs);
+            const value = runs
+                .map((run) => run.content)
+                .join(" ")
+                .trim();
+            if (value.length > 1)
+                notes = value;
+        }
+        slides.push({
+            texts,
+            images,
+            notes,
+            layoutHint: guessLayoutHint(texts, images.length > 0),
+        });
+    }
+    const firstSlide = slides[0]?.texts ?? [];
+    const title = [...firstSlide]
+        .sort((a, b) => (b.fontSize ?? 0) - (a.fontSize ?? 0))[0]
+        ?.content.trim()
+        .slice(0, 200) || "Imported Presentation";
+    return { title, slides, theme };
+}
+async function parseTheme(zip, parseXml) {
+    const xml = await zip.file("ppt/theme/theme1.xml")?.async("string");
+    if (!xml)
+        return undefined;
+    const root = record(parseXml(xml));
+    const elements = record(record(root?.["a:theme"])?.["a:themeElements"]);
+    const scheme = record(elements?.["a:clrScheme"]);
+    const colors = [];
+    for (const [key, value] of Object.entries(scheme ?? {})) {
+        if (key.startsWith("@_"))
+            continue;
+        const color = record(value);
+        const rgb = stringValue(record(color?.["a:srgbClr"])?.["@_val"]);
+        const system = stringValue(record(color?.["a:sysClr"])?.["@_lastClr"]);
+        if (rgb || system)
+            colors.push(`#${rgb ?? system}`);
+    }
+    const fontScheme = record(elements?.["a:fontScheme"]);
+    const fonts = ["a:majorFont", "a:minorFont"].flatMap((key) => {
+        const value = stringValue(record(record(fontScheme?.[key])?.["a:latin"])?.["@_typeface"]);
+        return value ? [value] : [];
+    });
+    return colors.length || fonts.length ? { colors, fonts } : undefined;
+}
+function collectTextRuns(value, runs, inherited = {}) {
+    const node = record(value);
+    if (!node)
+        return;
+    for (const raw of asArray(node["a:r"])) {
+        const run = record(raw);
+        const content = innerText(run?.["a:t"]);
+        if (content)
+            runs.push({
+                content,
+                ...runProperties(record(run?.["a:rPr"]), inherited),
+            });
+    }
+    if (node["a:t"] !== undefined && node["a:r"] === undefined) {
+        const content = innerText(node["a:t"]);
+        if (content)
+            runs.push({ content, ...inherited });
+    }
+    for (const [key, child] of Object.entries(node)) {
+        if (key.startsWith("@_") || key === "a:r" || key === "a:t")
+            continue;
+        for (const item of asArray(child))
+            collectTextRuns(item, runs, inherited);
+    }
+}
+function runProperties(value, inherited) {
+    if (!value)
+        return inherited;
+    const size = Number(value["@_sz"]);
+    const rgb = stringValue(record(record(value["a:solidFill"])?.["a:srgbClr"])?.["@_val"]);
+    return {
+        ...inherited,
+        ...(value["@_b"] === "1" || value["@_b"] === 1 || value["@_b"] === true
+            ? { bold: true }
+            : {}),
+        ...(value["@_i"] === "1" || value["@_i"] === 1 || value["@_i"] === true
+            ? { italic: true }
+            : {}),
+        ...(Number.isFinite(size) && size > 0 ? { fontSize: size / 100 } : {}),
+        ...(rgb ? { color: `#${rgb}` } : {}),
+    };
+}
+function parseRelationships(value) {
+    const output = new Map();
+    for (const raw of asArray(record(record(value)?.Relationships)?.Relationship)) {
+        const relationship = record(raw);
+        const id = stringValue(relationship?.["@_Id"]);
+        const target = stringValue(relationship?.["@_Target"]);
+        if (id && target) {
+            output.set(id, {
+                target,
+                type: stringValue(relationship?.["@_Type"]) ?? "",
+            });
+        }
+    }
+    return output;
+}
+function guessLayoutHint(texts, hasImages) {
+    if (hasImages)
+        return "image";
+    const maxSize = Math.max(...texts.map((text) => text.fontSize ?? 0), 0);
+    const length = texts.reduce((total, text) => total + text.content.length, 0);
+    if (texts.length <= 3 && length < 200 && maxSize >= 28)
+        return "title";
+    if (texts.length <= 2 && length < 100)
+        return "section";
+    return "content";
+}
+function imageMimeType(name) {
+    const extension = name.split(".").at(-1)?.toLowerCase();
+    return ({
+        png: "image/png",
+        jpg: "image/jpeg",
+        jpeg: "image/jpeg",
+        gif: "image/gif",
+        svg: "image/svg+xml",
+        webp: "image/webp",
+        bmp: "image/bmp",
+        tiff: "image/tiff",
+        tif: "image/tiff",
+        emf: "image/emf",
+        wmf: "image/wmf",
+    }[extension ?? ""] ?? "application/octet-stream");
+}
+async function loadPptxDependencies() {
+    const zipModuleName = "jszip";
+    const xmlModuleName = "fast-xml-parser";
+    try {
+        const [zipModule, xmlModule] = await Promise.all([
+            import(zipModuleName),
+            import(xmlModuleName),
+        ]);
+        const parser = new xmlModule.XMLParser({
+            ignoreAttributes: false,
+            attributeNamePrefix: "@_",
+        });
+        return {
+            loadZip: (data) => zipModule.default.loadAsync(data),
+            parseXml: (xml) => parser.parse(xml),
+        };
+    }
+    catch {
+        throw new Error("Structured PPTX parsing requires the optional jszip and fast-xml-parser dependencies.");
+    }
+}
+function slideNumber(value) {
+    return Number(value.match(/slide(\d+)/)?.[1] ?? 0);
+}
+function innerText(value) {
+    if (typeof value === "string" || typeof value === "number")
+        return String(value);
+    return String(record(value)?.["#text"] ?? "");
+}
+function asArray(value) {
+    return value === undefined || value === null
+        ? []
+        : Array.isArray(value)
+            ? value
+            : [value];
+}
+function record(value) {
+    return value && typeof value === "object" && !Array.isArray(value)
+        ? value
+        : null;
+}
+function stringValue(value) {
+    return typeof value === "string" && value.trim() ? value.trim() : undefined;
+}
+//# sourceMappingURL=pptx.js.map

package/dist/ingestion/pptx.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"pptx.js","sourceRoot":"","sources":["../../src/ingestion/pptx.ts"],"names":[],"mappings":"AAqCA,MAAM,CAAC,KAAK,UAAU,qBAAqB,CACzC,UAAsB;IAEtB,MAAM,EAAE,OAAO,EAAE,QAAQ,EAAE,GAAG,MAAM,oBAAoB,EAAE,CAAC;IAC3D,MAAM,GAAG,GAAG,MAAM,OAAO,CAAC,UAAU,CAAC,CAAC;IACtC,MAAM,eAAe,GAAG,MAAM,GAAG;SAC9B,IAAI,CAAC,sBAAsB,CAAC;QAC7B,EAAE,KAAK,CAAC,QAAQ,CAAC,CAAC;IACpB,IAAI,CAAC,eAAe;QAClB,MAAM,IAAI,KAAK,CAAC,4CAA4C,CAAC,CAAC;IAChE,MAAM,YAAY,GAAG,QAAQ,CAAC,eAAe,CAAC,CAAC;IAC/C,MAAM,QAAQ,GAAG,OAAO,CACtB,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,YAAY,CAAC,EAAE,CAAC,gBAAgB,CAAC,CAAC,EAAE,CAAC,YAAY,CAAC,CAAC,EAAE,CACxE,SAAS,CACV,CACF,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,WAAW,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,CAAC,QAAQ,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;IAC/D,MAAM,gBAAgB,GAAG,MAAM,GAAG;SAC/B,IAAI,CAAC,iCAAiC,CAAC;QACxC,EAAE,KAAK,CAAC,QAAQ,CAAC,CAAC;IACpB,MAAM,aAAa,GAAG,gBAAgB;QACpC,CAAC,CAAC,kBAAkB,CAAC,QAAQ,CAAC,gBAAgB,CAAC,CAAC;QAChD,CAAC,CAAC,IAAI,GAAG,EAA4C,CAAC;IACxD,MAAM,UAAU,GAAG,QAAQ,CAAC,OAAO,CAAC,CAAC,EAAE,EAAE,EAAE;QACzC,MAAM,YAAY,GAAG,aAAa,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;QAC3C,IAAI,CAAC,YAAY;YAAE,OAAO,EAAE,CAAC;QAC7B,OAAO;YACL,YAAY,CAAC,MAAM,CAAC,UAAU,CAAC,GAAG,CAAC;gBACjC,CAAC,CAAC,YAAY,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC;gBAC9B,CAAC,CAAC,OAAO,YAAY,CAAC,MAAM,EAAE;SACjC,CAAC;IACJ,CAAC,CAAC,CAAC;IACH,IAAI,UAAU,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QAC5B,UAAU,CAAC,IAAI,CACb,GAAG,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC;aACtB,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,8BAA8B,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;aAC3D,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,WAAW,CAAC,CAAC,CAAC,GAAG,WAAW,CAAC,CAAC,CAAC,CAAC,CACnD,CAAC;IACJ,CAAC;IACD,MAAM,KAAK,GAAG,MAAM,UAAU,CAAC,GAAG,EAAE,QAAQ,CAAC,CAAC;IAC9C,MAAM,MAAM,GAAsB,EAAE,CAAC;IACrC,KAAK,MAAM,SAAS,IAAI,UAAU,EAAE,CAAC;QACnC,MAAM,GAAG,GAAG,MAAM,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,EAAE,KAAK,CAAC,QAAQ,CAAC,CAAC;QACvD,IAAI,CAAC,GAAG;YAAE,SAAS;QACnB,IAAI,KAAc,CAAC;QACnB,IAAI,CAAC;YACH,KAAK,GAAG,QAAQ,CAAC,GAAG,CAAC,CAAC;QACxB,CAAC;QAAC,MAAM,CAAC;YACP,SAAS;QACX,CAAC;QACD,MAAM,KAAK,GAAwB,EAAE,CAAC;QACtC,eAAe,CAAC,KAAK,EAAE,KAAK,CAAC,CAAC;QAC9B,MAAM,MAAM,GAAsB,EAAE,CAAC;QACrC,MAAM,gBAAgB,GAAG,SAAS,CAAC,OAAO,CACxC,yBAAyB,EACzB,sBAAsB,CACvB,CAAC;QACF,MAAM,qBAAqB,GAAG,MAAM,GAAG;aACpC,IAAI,CAAC,gBAAgB,CAAC;YACvB,EAAE,KAAK,CAAC,QAAQ,CAAC,CAAC;QACpB,IAAI,qBAAqB,EAAE,CAAC;YAC1B,KAAK,MAAM,YAAY,IAAI,kBAAkB,CAC3C,QAAQ,CAAC,qBAAqB,CAAC,CAChC,CAAC,MAAM,EAAE,EAAE,CAAC;gBACX,IACE,CAAC,YAAY,CAAC,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC;oBACrC,CAAC,gDAAgD,CAAC,IAAI,CACpD,YAAY,CAAC,MAAM,CACpB,EACD,CAAC;oBACD,SAAS;gBACX,CAAC;gBACD,MAAM,SAAS,GAAG,YAAY,CAAC,MAAM,CAAC,UAAU,CAAC,GAAG,CAAC;oBACnD,CAAC,CAAC,YAAY,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC;oBAC9B,CAAC,CAAC,YAAY,CAAC,MAAM,CAAC,UAAU,CAAC,KAAK,CAAC;wBACrC,CAAC,CAAC,OAAO,YAAY,CAAC,MAAM,CAAC,OAAO,CAAC,SAAS,EAAE,EAAE,CAAC,EAAE;wBACrD,CAAC,CAAC,cAAc,YAAY,CAAC,MAAM,EAAE,CAAC;gBAC1C,MAAM,KAAK,GAAG,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;gBAClC,IAAI,CAAC,KAAK;oBAAE,SAAS;gBACrB,MAAM,IAAI,GAAG,SAAS,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,IAAI,OAAO,CAAC;gBACpD,MAAM,CAAC,IAAI,CAAC;oBACV,IAAI,EAAE,IAAI,UAAU,CAAC,MAAM,KAAK,CAAC,KAAK,CAAC,YAAY,CAAC,CAAC;oBACrD,QAAQ,EAAE,aAAa,CAAC,IAAI,CAAC;oBAC7B,IAAI;iBACL,CAAC,CAAC;YACL,CAAC;QACH,CAAC;QACD,MAAM,MAAM,GAAG,WAAW,CAAC,SAAS,CAAC,CAAC;QACtC,MAAM,QAAQ,GAAG,MAAM,GAAG;aACvB,IAAI,CAAC,6BAA6B,MAAM,MAAM,CAAC;YAChD,EAAE,KAAK,CAAC,QAAQ,CAAC,CAAC;QACpB,IAAI,KAAyB,CAAC;QAC9B,IAAI,QAAQ,EAAE,CAAC;YACb,MAAM,IAAI,GAAwB,EAAE,CAAC;YACrC,eAAe,CAAC,QAAQ,CAAC,QAAQ,CAAC,EAAE,IAAI,CAAC,CAAC;YAC1C,MAAM,KAAK,GAAG,IAAI;iBACf,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,OAAO,CAAC;iBACzB,IAAI,CAAC,GAAG,CAAC;iBACT,IAAI,EAAE,CAAC;YACV,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC;gBAAE,KAAK,GAAG,KAAK,CAAC;QACtC,CAAC;QACD,MAAM,CAAC,IAAI,CAAC;YACV,KAAK;YACL,MAAM;YACN,KAAK;YACL,UAAU,EAAE,eAAe,CAAC,KAAK,EAAE,MAAM,CAAC,MAAM,GAAG,CAAC,CAAC;SACtD,CAAC,CAAC;IACL,CAAC;IACD,MAAM,UAAU,GAAG,MAAM,CAAC,CAAC,CAAC,EAAE,KAAK,IAAI,EAAE,CAAC;IAC1C,MAAM,KAAK,GACT,CAAC,GAAG,UAAU,CAAC;SACZ,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,QAAQ,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,QAAQ,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;QACzD,EAAE,OAAO,CAAC,IAAI,EAAE;SACf,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,IAAI,uBAAuB,CAAC;IAC9C,OAAO,EAAE,KAAK,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC;AAClC,CAAC;AAED,KAAK,UAAU,UAAU,CACvB,GAAe,EACf,QAAkC;IAElC,MAAM,GAAG,GAAG,MAAM,GAAG,CAAC,IAAI,CAAC,sBAAsB,CAAC,EAAE,KAAK,CAAC,QAAQ,CAAC,CAAC;IACpE,IAAI,CAAC,GAAG;QAAE,OAAO,SAAS,CAAC;IAC3B,MAAM,IAAI,GAAG,MAAM,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC;IACnC,MAAM,QAAQ,GAAG,MAAM,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC,SAAS,CAAC,CAAC,EAAE,CAAC,iBAAiB,CAAC,CAAC,CAAC;IACxE,MAAM,MAAM,GAAG,MAAM,CAAC,QAAQ,EAAE,CAAC,aAAa,CAAC,CAAC,CAAC;IACjD,MAAM,MAAM,GAAa,EAAE,CAAC;IAC5B,KAAK,MAAM,CAAC,GAAG,EAAE,KAAK,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,MAAM,IAAI,EAAE,CAAC,EAAE,CAAC;QACxD,IAAI,GAAG,CAAC,UAAU,CAAC,IAAI,CAAC;YAAE,SAAS;QACnC,MAAM,KAAK,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC;QAC5B,MAAM,GAAG,GAAG,WAAW,CAAC,MAAM,CAAC,KAAK,EAAE,CAAC,WAAW,CAAC,CAAC,EAAE,CAAC,OAAO,CAAC,CAAC,CAAC;QACjE,MAAM,MAAM,GAAG,WAAW,CAAC,MAAM,CAAC,KAAK,EAAE,CAAC,UAAU,CAAC,CAAC,EAAE,CAAC,WAAW,CAAC,CAAC,CAAC;QACvE,IAAI,GAAG,IAAI,MAAM;YAAE,MAAM,CAAC,IAAI,CAAC,IAAI,GAAG,IAAI,MAAM,EAAE,CAAC,CAAC;IACtD,CAAC;IACD,MAAM,UAAU,GAAG,MAAM,CAAC,QAAQ,EAAE,CAAC,cAAc,CAAC,CAAC,CAAC;IACtD,MAAM,KAAK,GAAG,CAAC,aAAa,EAAE,aAAa,CAAC,CAAC,OAAO,CAAC,CAAC,GAAG,EAAE,EAAE;QAC3D,MAAM,KAAK,GAAG,WAAW,CACvB,MAAM,CAAC,MAAM,CAAC,UAAU,EAAE,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC,SAAS,CAAC,CAAC,EAAE,CAAC,YAAY,CAAC,CAC/D,CAAC;QACF,OAAO,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;IAC9B,CAAC,CAAC,CAAC;IACH,OAAO,MAAM,CAAC,MAAM,IAAI,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC,CAAC,CAAC,SAAS,CAAC;AACvE,CAAC;AAED,SAAS,eAAe,CACtB,KAAc,EACd,IAAyB,EACzB,SAAS,GAAuC,EAAE;IAElD,MAAM,IAAI,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC;IAC3B,IAAI,CAAC,IAAI;QAAE,OAAO;IAClB,KAAK,MAAM,GAAG,IAAI,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC;QACvC,MAAM,GAAG,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC;QACxB,MAAM,OAAO,GAAG,SAAS,CAAC,GAAG,EAAE,CAAC,KAAK,CAAC,CAAC,CAAC;QACxC,IAAI,OAAO;YACT,IAAI,CAAC,IAAI,CAAC;gBACR,OAAO;gBACP,GAAG,aAAa,CAAC,MAAM,CAAC,GAAG,EAAE,CAAC,OAAO,CAAC,CAAC,EAAE,SAAS,CAAC;aACpD,CAAC,CAAC;IACP,CAAC;IACD,IAAI,IAAI,CAAC,KAAK,CAAC,KAAK,SAAS,IAAI,IAAI,CAAC,KAAK,CAAC,KAAK,SAAS,EAAE,CAAC;QAC3D,MAAM,OAAO,GAAG,SAAS,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC;QACvC,IAAI,OAAO;YAAE,IAAI,CAAC,IAAI,CAAC,EAAE,OAAO,EAAE,GAAG,SAAS,EAAE,CAAC,CAAC;IACpD,CAAC;IACD,KAAK,MAAM,CAAC,GAAG,EAAE,KAAK,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,IAAI,CAAC,EAAE,CAAC;QAChD,IAAI,GAAG,CAAC,UAAU,CAAC,IAAI,CAAC,IAAI,GAAG,KAAK,KAAK,IAAI,GAAG,KAAK,KAAK;YAAE,SAAS;QACrE,KAAK,MAAM,IAAI,IAAI,OAAO,CAAC,KAAK,CAAC;YAAE,eAAe,CAAC,IAAI,EAAE,IAAI,EAAE,SAAS,CAAC,CAAC;IAC5E,CAAC;AACH,CAAC;AAED,SAAS,aAAa,CACpB,KAAqC,EACrC,SAA6C;IAE7C,IAAI,CAAC,KAAK;QAAE,OAAO,SAAS,CAAC;IAC7B,MAAM,IAAI,GAAG,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC;IACnC,MAAM,GAAG,GAAG,WAAW,CACrB,MAAM,CAAC,MAAM,CAAC,KAAK,CAAC,aAAa,CAAC,CAAC,EAAE,CAAC,WAAW,CAAC,CAAC,EAAE,CAAC,OAAO,CAAC,CAC/D,CAAC;IACF,OAAO;QACL,GAAG,SAAS;QACZ,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,KAAK,GAAG,IAAI,KAAK,CAAC,KAAK,CAAC,KAAK,CAAC,IAAI,KAAK,CAAC,KAAK,CAAC,KAAK,IAAI;YACrE,CAAC,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE;YAChB,CAAC,CAAC,EAAE,CAAC;QACP,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,KAAK,GAAG,IAAI,KAAK,CAAC,KAAK,CAAC,KAAK,CAAC,IAAI,KAAK,CAAC,KAAK,CAAC,KAAK,IAAI;YACrE,CAAC,CAAC,EAAE,MAAM,EAAE,IAAI,EAAE;YAClB,CAAC,CAAC,EAAE,CAAC;QACP,GAAG,CAAC,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,IAAI,GAAG,CAAC,CAAC,CAAC,CAAC,EAAE,QAAQ,EAAE,IAAI,GAAG,GAAG,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;QACtE,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,IAAI,GAAG,EAAE,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;KACrC,CAAC;AACJ,CAAC;AAED,SAAS,kBAAkB,CAAC,KAAc;IACxC,MAAM,MAAM,GAAG,IAAI,GAAG,EAA4C,CAAC;IACnE,KAAK,MAAM,GAAG,IAAI,OAAO,CACvB,MAAM,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,aAAa,CAAC,EAAE,YAAY,CACnD,EAAE,CAAC;QACF,MAAM,YAAY,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC;QACjC,MAAM,EAAE,GAAG,WAAW,CAAC,YAAY,EAAE,CAAC,MAAM,CAAC,CAAC,CAAC;QAC/C,MAAM,MAAM,GAAG,WAAW,CAAC,YAAY,EAAE,CAAC,UAAU,CAAC,CAAC,CAAC;QACvD,IAAI,EAAE,IAAI,MAAM,EAAE,CAAC;YACjB,MAAM,CAAC,GAAG,CAAC,EAAE,EAAE;gBACb,MAAM;gBACN,IAAI,EAAE,WAAW,CAAC,YAAY,EAAE,CAAC,QAAQ,CAAC,CAAC,IAAI,EAAE;aAClD,CAAC,CAAC;QACL,CAAC;IACH,CAAC;IACD,OAAO,MAAM,CAAC;AAChB,CAAC;AAED,SAAS,eAAe,CAAC,KAA0B,EAAE,SAAkB;IACrE,IAAI,SAAS;QAAE,OAAO,OAAO,CAAC;IAC9B,MAAM,OAAO,GAAG,IAAI,CAAC,GAAG,CAAC,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,QAAQ,IAAI,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;IACxE,MAAM,MAAM,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,IAAI,EAAE,EAAE,CAAC,KAAK,GAAG,IAAI,CAAC,OAAO,CAAC,MAAM,EAAE,CAAC,CAAC,CAAC;IAC7E,IAAI,KAAK,CAAC,MAAM,IAAI,CAAC,IAAI,MAAM,GAAG,GAAG,IAAI,OAAO,IAAI,EAAE;QAAE,OAAO,OAAO,CAAC;IACvE,IAAI,KAAK,CAAC,MAAM,IAAI,CAAC,IAAI,MAAM,GAAG,GAAG;QAAE,OAAO,SAAS,CAAC;IACxD,OAAO,SAAS,CAAC;AACnB,CAAC;AAED,SAAS,aAAa,CAAC,IAAY;IACjC,MAAM,SAAS,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,WAAW,EAAE,CAAC;IACxD,OAAO,CACL;QACE,GAAG,EAAE,WAAW;QAChB,GAAG,EAAE,YAAY;QACjB,IAAI,EAAE,YAAY;QAClB,GAAG,EAAE,WAAW;QAChB,GAAG,EAAE,eAAe;QACpB,IAAI,EAAE,YAAY;QAClB,GAAG,EAAE,WAAW;QAChB,IAAI,EAAE,YAAY;QAClB,GAAG,EAAE,YAAY;QACjB,GAAG,EAAE,WAAW;QAChB,GAAG,EAAE,WAAW;KACjB,CAAC,SAAS,IAAI,EAAE,CAAC,IAAI,0BAA0B,CACjD,CAAC;AACJ,CAAC;AAED,KAAK,UAAU,oBAAoB;IAIjC,MAAM,aAAa,GAAG,OAAO,CAAC;IAC9B,MAAM,aAAa,GAAG,iBAAiB,CAAC;IACxC,IAAI,CAAC;QACH,MAAM,CAAC,SAAS,EAAE,SAAS,CAAC,GAAG,MAAM,OAAO,CAAC,GAAG,CAAC;YAC/C,MAAM,CAAC,aAAa,CAElB;YACF,MAAM,CAAC,aAAa,CAIlB;SACH,CAAC,CAAC;QACH,MAAM,MAAM,GAAG,IAAI,SAAS,CAAC,SAAS,CAAC;YACrC,gBAAgB,EAAE,KAAK;YACvB,mBAAmB,EAAE,IAAI;SAC1B,CAAC,CAAC;QACH,OAAO;YACL,OAAO,EAAE,CAAC,IAAI,EAAE,EAAE,CAAC,SAAS,CAAC,OAAO,CAAC,SAAS,CAAC,IAAI,CAAC;YACpD,QAAQ,EAAE,CAAC,GAAG,EAAE,EAAE,CAAC,MAAM,CAAC,KAAK,CAAC,GAAG,CAAC;SACrC,CAAC;IACJ,CAAC;IAAC,MAAM,CAAC;QACP,MAAM,IAAI,KAAK,CACb,uFAAuF,CACxF,CAAC;IACJ,CAAC;AACH,CAAC;AAED,SAAS,WAAW,CAAC,KAAa;IAChC,OAAO,MAAM,CAAC,KAAK,CAAC,KAAK,CAAC,YAAY,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;AACrD,CAAC;AAED,SAAS,SAAS,CAAC,KAAc;IAC/B,IAAI,OAAO,KAAK,KAAK,QAAQ,IAAI,OAAO,KAAK,KAAK,QAAQ;QACxD,OAAO,MAAM,CAAC,KAAK,CAAC,CAAC;IACvB,OAAO,MAAM,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,CAAC,OAAO,CAAC,IAAI,EAAE,CAAC,CAAC;AAChD,CAAC;AAED,SAAS,OAAO,CAAC,KAAc;IAC7B,OAAO,KAAK,KAAK,SAAS,IAAI,KAAK,KAAK,IAAI;QAC1C,CAAC,CAAC,EAAE;QACJ,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,KAAK,CAAC;YACpB,CAAC,CAAC,KAAK;YACP,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC;AAChB,CAAC;AAED,SAAS,MAAM,CAAC,KAAc;IAC5B,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 ParsedPptxTextRun {\n  content: string;\n  bold?: boolean;\n  italic?: boolean;\n  fontSize?: number;\n  color?: string;\n}\n\nexport interface ParsedPptxImage {\n  data: Uint8Array;\n  mimeType: string;\n  name: string;\n}\n\nexport interface ParsedPptxSlide {\n  texts: ParsedPptxTextRun[];\n  images: ParsedPptxImage[];\n  notes?: string;\n  layoutHint?: string;\n}\n\nexport interface ParsedPptxPresentation {\n  title: string;\n  slides: ParsedPptxSlide[];\n  theme?: { colors: string[]; fonts: string[] };\n}\n\ninterface ZipFile {\n  async(type: \"string\"): Promise<string>;\n  async(type: \"nodebuffer\"): Promise<Buffer>;\n}\n\ninterface ZipArchive {\n  files: Record<string, unknown>;\n  file(path: string): ZipFile | null;\n}\n\nexport async function parsePptxPresentation(\n  fileBuffer: Uint8Array,\n): Promise<ParsedPptxPresentation> {\n  const { loadZip, parseXml } = await loadPptxDependencies();\n  const zip = await loadZip(fileBuffer);\n  const presentationXml = await zip\n    .file(\"ppt/presentation.xml\")\n    ?.async(\"string\");\n  if (!presentationXml)\n    throw new Error(\"Invalid PPTX: missing ppt/presentation.xml\");\n  const presentation = parseXml(presentationXml);\n  const slideIds = asArray(\n    record(record(record(presentation)?.[\"p:presentation\"])?.[\"p:sldIdLst\"])?.[\n      \"p:sldId\"\n    ],\n  ).map((entry) => stringValue(record(entry)?.[\"@_r:id\"]) ?? \"\");\n  const relationshipsXml = await zip\n    .file(\"ppt/_rels/presentation.xml.rels\")\n    ?.async(\"string\");\n  const relationships = relationshipsXml\n    ? parseRelationships(parseXml(relationshipsXml))\n    : new Map<string, { target: string; type: string }>();\n  const slidePaths = slideIds.flatMap((id) => {\n    const relationship = relationships.get(id);\n    if (!relationship) return [];\n    return [\n      relationship.target.startsWith(\"/\")\n        ? relationship.target.slice(1)\n        : `ppt/${relationship.target}`,\n    ];\n  });\n  if (slidePaths.length === 0) {\n    slidePaths.push(\n      ...Object.keys(zip.files)\n        .filter((path) => /^ppt\\/slides\\/slide\\d+\\.xml$/.test(path))\n        .sort((a, b) => slideNumber(a) - slideNumber(b)),\n    );\n  }\n  const theme = await parseTheme(zip, parseXml);\n  const slides: ParsedPptxSlide[] = [];\n  for (const slidePath of slidePaths) {\n    const xml = await zip.file(slidePath)?.async(\"string\");\n    if (!xml) continue;\n    let slide: unknown;\n    try {\n      slide = parseXml(xml);\n    } catch {\n      continue;\n    }\n    const texts: ParsedPptxTextRun[] = [];\n    collectTextRuns(slide, texts);\n    const images: ParsedPptxImage[] = [];\n    const relationshipPath = slidePath.replace(\n      /slides\\/(slide\\d+\\.xml)/,\n      \"slides/_rels/$1.rels\",\n    );\n    const slideRelationshipsXml = await zip\n      .file(relationshipPath)\n      ?.async(\"string\");\n    if (slideRelationshipsXml) {\n      for (const relationship of parseRelationships(\n        parseXml(slideRelationshipsXml),\n      ).values()) {\n        if (\n          !relationship.type.includes(\"/image\") &&\n          !/\\.(png|jpe?g|gif|svg|webp|bmp|tiff?|emf|wmf)$/i.test(\n            relationship.target,\n          )\n        ) {\n          continue;\n        }\n        const imagePath = relationship.target.startsWith(\"/\")\n          ? relationship.target.slice(1)\n          : relationship.target.startsWith(\"../\")\n            ? `ppt/${relationship.target.replace(/^\\.\\.\\//, \"\")}`\n            : `ppt/slides/${relationship.target}`;\n        const image = zip.file(imagePath);\n        if (!image) continue;\n        const name = imagePath.split(\"/\").at(-1) ?? \"image\";\n        images.push({\n          data: new Uint8Array(await image.async(\"nodebuffer\")),\n          mimeType: imageMimeType(name),\n          name,\n        });\n      }\n    }\n    const number = slideNumber(slidePath);\n    const notesXml = await zip\n      .file(`ppt/notesSlides/notesSlide${number}.xml`)\n      ?.async(\"string\");\n    let notes: string | undefined;\n    if (notesXml) {\n      const runs: ParsedPptxTextRun[] = [];\n      collectTextRuns(parseXml(notesXml), runs);\n      const value = runs\n        .map((run) => run.content)\n        .join(\" \")\n        .trim();\n      if (value.length > 1) notes = value;\n    }\n    slides.push({\n      texts,\n      images,\n      notes,\n      layoutHint: guessLayoutHint(texts, images.length > 0),\n    });\n  }\n  const firstSlide = slides[0]?.texts ?? [];\n  const title =\n    [...firstSlide]\n      .sort((a, b) => (b.fontSize ?? 0) - (a.fontSize ?? 0))[0]\n      ?.content.trim()\n      .slice(0, 200) || \"Imported Presentation\";\n  return { title, slides, theme };\n}\n\nasync function parseTheme(\n  zip: ZipArchive,\n  parseXml: (xml: string) => unknown,\n): Promise<ParsedPptxPresentation[\"theme\"]> {\n  const xml = await zip.file(\"ppt/theme/theme1.xml\")?.async(\"string\");\n  if (!xml) return undefined;\n  const root = record(parseXml(xml));\n  const elements = record(record(root?.[\"a:theme\"])?.[\"a:themeElements\"]);\n  const scheme = record(elements?.[\"a:clrScheme\"]);\n  const colors: string[] = [];\n  for (const [key, value] of Object.entries(scheme ?? {})) {\n    if (key.startsWith(\"@_\")) continue;\n    const color = record(value);\n    const rgb = stringValue(record(color?.[\"a:srgbClr\"])?.[\"@_val\"]);\n    const system = stringValue(record(color?.[\"a:sysClr\"])?.[\"@_lastClr\"]);\n    if (rgb || system) colors.push(`#${rgb ?? system}`);\n  }\n  const fontScheme = record(elements?.[\"a:fontScheme\"]);\n  const fonts = [\"a:majorFont\", \"a:minorFont\"].flatMap((key) => {\n    const value = stringValue(\n      record(record(fontScheme?.[key])?.[\"a:latin\"])?.[\"@_typeface\"],\n    );\n    return value ? [value] : [];\n  });\n  return colors.length || fonts.length ? { colors, fonts } : undefined;\n}\n\nfunction collectTextRuns(\n  value: unknown,\n  runs: ParsedPptxTextRun[],\n  inherited: Omit<ParsedPptxTextRun, \"content\"> = {},\n): void {\n  const node = record(value);\n  if (!node) return;\n  for (const raw of asArray(node[\"a:r\"])) {\n    const run = record(raw);\n    const content = innerText(run?.[\"a:t\"]);\n    if (content)\n      runs.push({\n        content,\n        ...runProperties(record(run?.[\"a:rPr\"]), inherited),\n      });\n  }\n  if (node[\"a:t\"] !== undefined && node[\"a:r\"] === undefined) {\n    const content = innerText(node[\"a:t\"]);\n    if (content) runs.push({ content, ...inherited });\n  }\n  for (const [key, child] of Object.entries(node)) {\n    if (key.startsWith(\"@_\") || key === \"a:r\" || key === \"a:t\") continue;\n    for (const item of asArray(child)) collectTextRuns(item, runs, inherited);\n  }\n}\n\nfunction runProperties(\n  value: Record<string, unknown> | null,\n  inherited: Omit<ParsedPptxTextRun, \"content\">,\n): Omit<ParsedPptxTextRun, \"content\"> {\n  if (!value) return inherited;\n  const size = Number(value[\"@_sz\"]);\n  const rgb = stringValue(\n    record(record(value[\"a:solidFill\"])?.[\"a:srgbClr\"])?.[\"@_val\"],\n  );\n  return {\n    ...inherited,\n    ...(value[\"@_b\"] === \"1\" || value[\"@_b\"] === 1 || value[\"@_b\"] === true\n      ? { bold: true }\n      : {}),\n    ...(value[\"@_i\"] === \"1\" || value[\"@_i\"] === 1 || value[\"@_i\"] === true\n      ? { italic: true }\n      : {}),\n    ...(Number.isFinite(size) && size > 0 ? { fontSize: size / 100 } : {}),\n    ...(rgb ? { color: `#${rgb}` } : {}),\n  };\n}\n\nfunction parseRelationships(value: unknown) {\n  const output = new Map<string, { target: string; type: string }>();\n  for (const raw of asArray(\n    record(record(value)?.Relationships)?.Relationship,\n  )) {\n    const relationship = record(raw);\n    const id = stringValue(relationship?.[\"@_Id\"]);\n    const target = stringValue(relationship?.[\"@_Target\"]);\n    if (id && target) {\n      output.set(id, {\n        target,\n        type: stringValue(relationship?.[\"@_Type\"]) ?? \"\",\n      });\n    }\n  }\n  return output;\n}\n\nfunction guessLayoutHint(texts: ParsedPptxTextRun[], hasImages: boolean) {\n  if (hasImages) return \"image\";\n  const maxSize = Math.max(...texts.map((text) => text.fontSize ?? 0), 0);\n  const length = texts.reduce((total, text) => total + text.content.length, 0);\n  if (texts.length <= 3 && length < 200 && maxSize >= 28) return \"title\";\n  if (texts.length <= 2 && length < 100) return \"section\";\n  return \"content\";\n}\n\nfunction imageMimeType(name: string): string {\n  const extension = name.split(\".\").at(-1)?.toLowerCase();\n  return (\n    {\n      png: \"image/png\",\n      jpg: \"image/jpeg\",\n      jpeg: \"image/jpeg\",\n      gif: \"image/gif\",\n      svg: \"image/svg+xml\",\n      webp: \"image/webp\",\n      bmp: \"image/bmp\",\n      tiff: \"image/tiff\",\n      tif: \"image/tiff\",\n      emf: \"image/emf\",\n      wmf: \"image/wmf\",\n    }[extension ?? \"\"] ?? \"application/octet-stream\"\n  );\n}\n\nasync function loadPptxDependencies(): Promise<{\n  loadZip(data: Uint8Array): Promise<ZipArchive>;\n  parseXml(xml: string): unknown;\n}> {\n  const zipModuleName = \"jszip\";\n  const xmlModuleName = \"fast-xml-parser\";\n  try {\n    const [zipModule, xmlModule] = await Promise.all([\n      import(zipModuleName) as Promise<{\n        default: { loadAsync(data: Uint8Array): Promise<ZipArchive> };\n      }>,\n      import(xmlModuleName) as Promise<{\n        XMLParser: new (options: Record<string, unknown>) => {\n          parse(xml: string): unknown;\n        };\n      }>,\n    ]);\n    const parser = new xmlModule.XMLParser({\n      ignoreAttributes: false,\n      attributeNamePrefix: \"@_\",\n    });\n    return {\n      loadZip: (data) => zipModule.default.loadAsync(data),\n      parseXml: (xml) => parser.parse(xml),\n    };\n  } catch {\n    throw new Error(\n      \"Structured PPTX parsing requires the optional jszip and fast-xml-parser dependencies.\",\n    );\n  }\n}\n\nfunction slideNumber(value: string): number {\n  return Number(value.match(/slide(\\d+)/)?.[1] ?? 0);\n}\n\nfunction innerText(value: unknown): string {\n  if (typeof value === \"string\" || typeof value === \"number\")\n    return String(value);\n  return String(record(value)?.[\"#text\"] ?? \"\");\n}\n\nfunction asArray(value: unknown): unknown[] {\n  return value === undefined || value === null\n    ? []\n    : Array.isArray(value)\n      ? value\n      : [value];\n}\n\nfunction record(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/selection.d.ts

@@ -0,0 +1,20 @@
+export interface IngestionInventoryCandidate {
+    externalId: string;
+    sourceModifiedAt?: string;
+    metadata?: Record<string, unknown>;
+}
+export interface IngestionHydrationPolicy {
+    selectedExternalIds?: readonly string[];
+    canonicalExternalIds?: readonly string[];
+    pinnedExternalIds?: readonly string[];
+    recentWindowMonths?: number;
+    hydrateAll?: boolean;
+    now?: Date;
+}
+export interface IngestionHydrationSelection<T> {
+    selected: T[];
+    deferred: T[];
+    reasons: Record<string, "confirmed" | "canonical" | "pinned" | "recent" | "undated" | "all">;
+}
+export declare function selectInventoryForHydration<T extends IngestionInventoryCandidate>(inventory: readonly T[], policy?: IngestionHydrationPolicy): IngestionHydrationSelection<T>;
+//# sourceMappingURL=selection.d.ts.map

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

@@ -0,0 +1 @@
+{"version":3,"file":"selection.d.ts","sourceRoot":"","sources":["../../src/ingestion/selection.ts"],"names":[],"mappings":"AAAA,MAAM,WAAW,2BAA2B;IAC1C,UAAU,EAAE,MAAM,CAAC;IACnB,gBAAgB,CAAC,EAAE,MAAM,CAAC;IAC1B,QAAQ,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;CACpC;AAED,MAAM,WAAW,wBAAwB;IACvC,mBAAmB,CAAC,EAAE,SAAS,MAAM,EAAE,CAAC;IACxC,oBAAoB,CAAC,EAAE,SAAS,MAAM,EAAE,CAAC;IACzC,iBAAiB,CAAC,EAAE,SAAS,MAAM,EAAE,CAAC;IACtC,kBAAkB,CAAC,EAAE,MAAM,CAAC;IAC5B,UAAU,CAAC,EAAE,OAAO,CAAC;IACrB,GAAG,CAAC,EAAE,IAAI,CAAC;CACZ;AAED,MAAM,WAAW,2BAA2B,CAAC,CAAC;IAC5C,QAAQ,EAAE,CAAC,EAAE,CAAC;IACd,QAAQ,EAAE,CAAC,EAAE,CAAC;IACd,OAAO,EAAE,MAAM,CACb,MAAM,EACN,WAAW,GAAG,WAAW,GAAG,QAAQ,GAAG,QAAQ,GAAG,SAAS,GAAG,KAAK,CACpE,CAAC;CACH;AAED,wBAAgB,2BAA2B,CACzC,CAAC,SAAS,2BAA2B,EAErC,SAAS,EAAE,SAAS,CAAC,EAAE,EACvB,MAAM,GAAE,wBAA6B,GACpC,2BAA2B,CAAC,CAAC,CAAC,CAqChC"}

package/dist/ingestion/selection.js

@@ -0,0 +1,52 @@
+export function selectInventoryForHydration(inventory, policy = {}) {
+    const confirmed = stringSet(policy.selectedExternalIds);
+    const canonical = stringSet(policy.canonicalExternalIds);
+    const pinned = stringSet(policy.pinnedExternalIds);
+    const hasConfirmedSelection = policy.selectedExternalIds !== undefined;
+    const now = policy.now ?? new Date();
+    const cutoff = new Date(now);
+    cutoff.setUTCMonth(cutoff.getUTCMonth() - positiveInteger(policy.recentWindowMonths, 12));
+    const selected = [];
+    const deferred = [];
+    const reasons = {};
+    for (const item of inventory) {
+        let reason;
+        if (policy.hydrateAll)
+            reason = "all";
+        else if (confirmed.has(item.externalId))
+            reason = "confirmed";
+        else if (canonical.has(item.externalId) ||
+            item.metadata?.canonical === true)
+            reason = "canonical";
+        else if (pinned.has(item.externalId) || item.metadata?.pinned === true)
+            reason = "pinned";
+        else if (!hasConfirmedSelection) {
+            const modifiedAt = parsedDate(item.sourceModifiedAt);
+            if (!modifiedAt)
+                reason = "undated";
+            else if (modifiedAt >= cutoff)
+                reason = "recent";
+        }
+        if (reason) {
+            selected.push(item);
+            reasons[item.externalId] = reason;
+        }
+        else {
+            deferred.push(item);
+        }
+    }
+    return { selected, deferred, reasons };
+}
+function stringSet(values) {
+    return new Set((values ?? []).map((value) => value.trim()).filter(Boolean));
+}
+function parsedDate(value) {
+    if (!value)
+        return null;
+    const date = new Date(value);
+    return Number.isNaN(date.getTime()) ? null : date;
+}
+function positiveInteger(value, fallback) {
+    return Number.isInteger(value) && value > 0 ? value : fallback;
+}
+//# sourceMappingURL=selection.js.map

package/dist/ingestion/selection.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"selection.js","sourceRoot":"","sources":["../../src/ingestion/selection.ts"],"names":[],"mappings":"AAwBA,MAAM,UAAU,2BAA2B,CAGzC,SAAuB,EACvB,MAAM,GAA6B,EAAE;IAErC,MAAM,SAAS,GAAG,SAAS,CAAC,MAAM,CAAC,mBAAmB,CAAC,CAAC;IACxD,MAAM,SAAS,GAAG,SAAS,CAAC,MAAM,CAAC,oBAAoB,CAAC,CAAC;IACzD,MAAM,MAAM,GAAG,SAAS,CAAC,MAAM,CAAC,iBAAiB,CAAC,CAAC;IACnD,MAAM,qBAAqB,GAAG,MAAM,CAAC,mBAAmB,KAAK,SAAS,CAAC;IACvE,MAAM,GAAG,GAAG,MAAM,CAAC,GAAG,IAAI,IAAI,IAAI,EAAE,CAAC;IACrC,MAAM,MAAM,GAAG,IAAI,IAAI,CAAC,GAAG,CAAC,CAAC;IAC7B,MAAM,CAAC,WAAW,CAChB,MAAM,CAAC,WAAW,EAAE,GAAG,eAAe,CAAC,MAAM,CAAC,kBAAkB,EAAE,EAAE,CAAC,CACtE,CAAC;IACF,MAAM,QAAQ,GAAQ,EAAE,CAAC;IACzB,MAAM,QAAQ,GAAQ,EAAE,CAAC;IACzB,MAAM,OAAO,GAA8C,EAAE,CAAC;IAC9D,KAAK,MAAM,IAAI,IAAI,SAAS,EAAE,CAAC;QAC7B,IAAI,MAAqE,CAAC;QAC1E,IAAI,MAAM,CAAC,UAAU;YAAE,MAAM,GAAG,KAAK,CAAC;aACjC,IAAI,SAAS,CAAC,GAAG,CAAC,IAAI,CAAC,UAAU,CAAC;YAAE,MAAM,GAAG,WAAW,CAAC;aACzD,IACH,SAAS,CAAC,GAAG,CAAC,IAAI,CAAC,UAAU,CAAC;YAC9B,IAAI,CAAC,QAAQ,EAAE,SAAS,KAAK,IAAI;YAEjC,MAAM,GAAG,WAAW,CAAC;aAClB,IAAI,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,UAAU,CAAC,IAAI,IAAI,CAAC,QAAQ,EAAE,MAAM,KAAK,IAAI;YACpE,MAAM,GAAG,QAAQ,CAAC;aACf,IAAI,CAAC,qBAAqB,EAAE,CAAC;YAChC,MAAM,UAAU,GAAG,UAAU,CAAC,IAAI,CAAC,gBAAgB,CAAC,CAAC;YACrD,IAAI,CAAC,UAAU;gBAAE,MAAM,GAAG,SAAS,CAAC;iBAC/B,IAAI,UAAU,IAAI,MAAM;gBAAE,MAAM,GAAG,QAAQ,CAAC;QACnD,CAAC;QACD,IAAI,MAAM,EAAE,CAAC;YACX,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YACpB,OAAO,CAAC,IAAI,CAAC,UAAU,CAAC,GAAG,MAAM,CAAC;QACpC,CAAC;aAAM,CAAC;YACN,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QACtB,CAAC;IACH,CAAC;IACD,OAAO,EAAE,QAAQ,EAAE,QAAQ,EAAE,OAAO,EAAE,CAAC;AACzC,CAAC;AAED,SAAS,SAAS,CAAC,MAAqC;IACtD,OAAO,IAAI,GAAG,CAAC,CAAC,MAAM,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC;AAC9E,CAAC;AAED,SAAS,UAAU,CAAC,KAAyB;IAC3C,IAAI,CAAC,KAAK;QAAE,OAAO,IAAI,CAAC;IACxB,MAAM,IAAI,GAAG,IAAI,IAAI,CAAC,KAAK,CAAC,CAAC;IAC7B,OAAO,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,OAAO,EAAE,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC;AACpD,CAAC;AAED,SAAS,eAAe,CAAC,KAAyB,EAAE,QAAgB;IAClE,OAAO,MAAM,CAAC,SAAS,CAAC,KAAK,CAAC,IAAI,KAAM,GAAG,CAAC,CAAC,CAAC,CAAC,KAAM,CAAC,CAAC,CAAC,QAAQ,CAAC;AACnE,CAAC","sourcesContent":["export interface IngestionInventoryCandidate {\n  externalId: string;\n  sourceModifiedAt?: string;\n  metadata?: Record<string, unknown>;\n}\n\nexport interface IngestionHydrationPolicy {\n  selectedExternalIds?: readonly string[];\n  canonicalExternalIds?: readonly string[];\n  pinnedExternalIds?: readonly string[];\n  recentWindowMonths?: number;\n  hydrateAll?: boolean;\n  now?: Date;\n}\n\nexport interface IngestionHydrationSelection<T> {\n  selected: T[];\n  deferred: T[];\n  reasons: Record<\n    string,\n    \"confirmed\" | \"canonical\" | \"pinned\" | \"recent\" | \"undated\" | \"all\"\n  >;\n}\n\nexport function selectInventoryForHydration<\n  T extends IngestionInventoryCandidate,\n>(\n  inventory: readonly T[],\n  policy: IngestionHydrationPolicy = {},\n): IngestionHydrationSelection<T> {\n  const confirmed = stringSet(policy.selectedExternalIds);\n  const canonical = stringSet(policy.canonicalExternalIds);\n  const pinned = stringSet(policy.pinnedExternalIds);\n  const hasConfirmedSelection = policy.selectedExternalIds !== undefined;\n  const now = policy.now ?? new Date();\n  const cutoff = new Date(now);\n  cutoff.setUTCMonth(\n    cutoff.getUTCMonth() - positiveInteger(policy.recentWindowMonths, 12),\n  );\n  const selected: T[] = [];\n  const deferred: T[] = [];\n  const reasons: IngestionHydrationSelection<T>[\"reasons\"] = {};\n  for (const item of inventory) {\n    let reason: IngestionHydrationSelection<T>[\"reasons\"][string] | undefined;\n    if (policy.hydrateAll) reason = \"all\";\n    else if (confirmed.has(item.externalId)) reason = \"confirmed\";\n    else if (\n      canonical.has(item.externalId) ||\n      item.metadata?.canonical === true\n    )\n      reason = \"canonical\";\n    else if (pinned.has(item.externalId) || item.metadata?.pinned === true)\n      reason = \"pinned\";\n    else if (!hasConfirmedSelection) {\n      const modifiedAt = parsedDate(item.sourceModifiedAt);\n      if (!modifiedAt) reason = \"undated\";\n      else if (modifiedAt >= cutoff) reason = \"recent\";\n    }\n    if (reason) {\n      selected.push(item);\n      reasons[item.externalId] = reason;\n    } else {\n      deferred.push(item);\n    }\n  }\n  return { selected, deferred, reasons };\n}\n\nfunction stringSet(values: readonly string[] | undefined): Set<string> {\n  return new Set((values ?? []).map((value) => value.trim()).filter(Boolean));\n}\n\nfunction parsedDate(value: string | undefined): Date | null {\n  if (!value) return null;\n  const date = new Date(value);\n  return Number.isNaN(date.getTime()) ? null : date;\n}\n\nfunction positiveInteger(value: number | undefined, fallback: number): number {\n  return Number.isInteger(value) && value! > 0 ? value! : fallback;\n}\n"]}

package/dist/ingestion/website.d.ts

@@ -0,0 +1,27 @@
+export interface WebsiteAsset {
+    url: string;
+    kind: "image" | "video" | "audio" | "font" | "stylesheet" | "script";
+    role?: "logo" | "open-graph";
+}
+export interface WebsiteDesignTokens {
+    colors: string[];
+    typography: Array<{
+        family: string;
+        size: string;
+        weight: string;
+        lineHeight: string;
+        letterSpacing: string;
+    }>;
+    spacing: string[];
+    radii: string[];
+    cssVariables: Record<string, string>;
+}
+export interface WebsiteExtraction {
+    title: string;
+    text: string;
+    assets: WebsiteAsset[];
+    internalLinks: string[];
+    designTokens: WebsiteDesignTokens;
+}
+export declare function extractStaticWebsiteContext(html: string, baseUrl: string): WebsiteExtraction;
+//# sourceMappingURL=website.d.ts.map

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

@@ -0,0 +1 @@
+{"version":3,"file":"website.d.ts","sourceRoot":"","sources":["../../src/ingestion/website.ts"],"names":[],"mappings":"AAEA,MAAM,WAAW,YAAY;IAC3B,GAAG,EAAE,MAAM,CAAC;IACZ,IAAI,EAAE,OAAO,GAAG,OAAO,GAAG,OAAO,GAAG,MAAM,GAAG,YAAY,GAAG,QAAQ,CAAC;IACrE,IAAI,CAAC,EAAE,MAAM,GAAG,YAAY,CAAC;CAC9B;AAED,MAAM,WAAW,mBAAmB;IAClC,MAAM,EAAE,MAAM,EAAE,CAAC;IACjB,UAAU,EAAE,KAAK,CAAC;QAChB,MAAM,EAAE,MAAM,CAAC;QACf,IAAI,EAAE,MAAM,CAAC;QACb,MAAM,EAAE,MAAM,CAAC;QACf,UAAU,EAAE,MAAM,CAAC;QACnB,aAAa,EAAE,MAAM,CAAC;KACvB,CAAC,CAAC;IACH,OAAO,EAAE,MAAM,EAAE,CAAC;IAClB,KAAK,EAAE,MAAM,EAAE,CAAC;IAChB,YAAY,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;CACtC;AAED,MAAM,WAAW,iBAAiB;IAChC,KAAK,EAAE,MAAM,CAAC;IACd,IAAI,EAAE,MAAM,CAAC;IACb,MAAM,EAAE,YAAY,EAAE,CAAC;IACvB,aAAa,EAAE,MAAM,EAAE,CAAC;IACxB,YAAY,EAAE,mBAAmB,CAAC;CACnC;AAED,wBAAgB,2BAA2B,CACzC,IAAI,EAAE,MAAM,EACZ,OAAO,EAAE,MAAM,GACd,iBAAiB,CA2EnB"}