@x12i/content-fetcher 0.1.0

package/dist/storage.js

@@ -0,0 +1,109 @@
+import { ArtifactStoreError } from "./errors.js";
+import { sha256ContentHash } from "./util/hash.js";
+async function storeOperation(operation, run) {
+    try {
+        return await run();
+    }
+    catch (error) {
+        if (error instanceof ArtifactStoreError)
+            throw error;
+        throw new ArtifactStoreError(`Artifact store operation '${operation}' failed.`, operation);
+    }
+}
+/**
+ * Rights-gated, deduplicating persistence of raw artifacts:
+ * - raw bytes are stored only when the rights profile allows retention;
+ * - content already known for this source is never re-stored or overwritten;
+ * - every acquisition keeps its own record via putMetadata.
+ */
+export class ArtifactPersistence {
+    store;
+    constructor(store) {
+        this.store = store;
+    }
+    async persistArtifact(input) {
+        const contentHash = sha256ContentHash(input.bytes);
+        if (!input.rights.allowRawRetention) {
+            return {};
+        }
+        const existing = await storeOperation("findByContentHash", () => this.store.findByContentHash({ sourceId: input.sourceId, contentHash }));
+        if (existing !== null) {
+            return {
+                artifact: {
+                    artifactId: existing.artifactId,
+                    fileName: input.fileName,
+                    contentType: input.contentType,
+                    contentHash,
+                    byteLength: input.bytes.length,
+                },
+                duplicateOfArtifactId: existing.artifactId,
+            };
+        }
+        const stored = await storeOperation("putOriginal", () => this.store.putOriginal({
+            acquisitionId: input.acquisitionId,
+            fileName: input.fileName,
+            contentType: input.contentType,
+            bytes: input.bytes,
+            contentHash,
+            rights: input.rights,
+        }));
+        return {
+            artifact: {
+                artifactId: stored.artifactId,
+                fileName: input.fileName,
+                contentType: input.contentType,
+                contentHash,
+                byteLength: input.bytes.length,
+            },
+        };
+    }
+    /** Resolve a previous version's artifactId when the caller supplied its content hash. */
+    async resolvePreviousVersionId(request, policy) {
+        const hash = request.previousVersion?.contentHash;
+        if (hash === undefined)
+            return undefined;
+        const found = await storeOperation("findByContentHash", () => this.store.findByContentHash({ sourceId: policy.sourceId, contentHash: hash }));
+        return found?.artifactId;
+    }
+    /** Metadata + rights decision for one acquisition event. Never contains raw content or credentials. */
+    async writeAcquisitionMetadata(result, extras) {
+        await storeOperation("putMetadata", () => this.store.putMetadata({
+            acquisitionId: result.acquisitionId,
+            metadata: {
+                recordType: "acquisition",
+                ...extras,
+                acquisitionId: result.acquisitionId,
+                requestId: result.requestId,
+                sourceId: result.sourceId,
+                status: result.status,
+                channel: result.channel,
+                artifact: result.artifact,
+                version: result.version,
+                provenance: result.provenance,
+                rights: result.rights,
+                diagnostics: result.diagnostics,
+            },
+        }));
+    }
+    /** Compact audit record for every completed request. Sanitized by construction. */
+    async writeAuditRecord(result, requestedAt, completedAt) {
+        await storeOperation("putMetadata", () => this.store.putMetadata({
+            acquisitionId: result.acquisitionId,
+            metadata: {
+                recordType: "audit",
+                requestId: result.requestId,
+                acquisitionId: result.acquisitionId,
+                sourceId: result.sourceId,
+                status: result.status,
+                channel: result.channel,
+                requestedAt: requestedAt.toISOString(),
+                completedAt: completedAt.toISOString(),
+                ...(result.artifact?.contentHash !== undefined
+                    ? { contentHash: result.artifact.contentHash }
+                    : {}),
+                rightsProfileId: result.rights.profileId,
+            },
+        }));
+    }
+}
+//# sourceMappingURL=storage.js.map

package/dist/storage.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"storage.js","sourceRoot":"","sources":["../src/storage.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,kBAAkB,EAAE,MAAM,aAAa,CAAC;AACjD,OAAO,EAAE,iBAAiB,EAAE,MAAM,gBAAgB,CAAC;AA+BnD,KAAK,UAAU,cAAc,CAAI,SAAiB,EAAE,GAAqB;IACvE,IAAI,CAAC;QACH,OAAO,MAAM,GAAG,EAAE,CAAC;IACrB,CAAC;IAAC,OAAO,KAAK,EAAE,CAAC;QACf,IAAI,KAAK,YAAY,kBAAkB;YAAE,MAAM,KAAK,CAAC;QACrD,MAAM,IAAI,kBAAkB,CAC1B,6BAA6B,SAAS,WAAW,EACjD,SAAS,CACV,CAAC;IACJ,CAAC;AACH,CAAC;AAED;;;;;GAKG;AACH,MAAM,OAAO,mBAAmB;IACD;IAA7B,YAA6B,KAAoB;QAApB,UAAK,GAAL,KAAK,CAAe;IAAG,CAAC;IAErD,KAAK,CAAC,eAAe,CAAC,KAA2B;QAC/C,MAAM,WAAW,GAAG,iBAAiB,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC;QACnD,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,iBAAiB,EAAE,CAAC;YACpC,OAAO,EAAE,CAAC;QACZ,CAAC;QACD,MAAM,QAAQ,GAAG,MAAM,cAAc,CAAC,mBAAmB,EAAE,GAAG,EAAE,CAC9D,IAAI,CAAC,KAAK,CAAC,iBAAiB,CAAC,EAAE,QAAQ,EAAE,KAAK,CAAC,QAAQ,EAAE,WAAW,EAAE,CAAC,CACxE,CAAC;QACF,IAAI,QAAQ,KAAK,IAAI,EAAE,CAAC;YACtB,OAAO;gBACL,QAAQ,EAAE;oBACR,UAAU,EAAE,QAAQ,CAAC,UAAU;oBAC/B,QAAQ,EAAE,KAAK,CAAC,QAAQ;oBACxB,WAAW,EAAE,KAAK,CAAC,WAAW;oBAC9B,WAAW;oBACX,UAAU,EAAE,KAAK,CAAC,KAAK,CAAC,MAAM;iBAC/B;gBACD,qBAAqB,EAAE,QAAQ,CAAC,UAAU;aAC3C,CAAC;QACJ,CAAC;QACD,MAAM,MAAM,GAAG,MAAM,cAAc,CAAC,aAAa,EAAE,GAAG,EAAE,CACtD,IAAI,CAAC,KAAK,CAAC,WAAW,CAAC;YACrB,aAAa,EAAE,KAAK,CAAC,aAAa;YAClC,QAAQ,EAAE,KAAK,CAAC,QAAQ;YACxB,WAAW,EAAE,KAAK,CAAC,WAAW;YAC9B,KAAK,EAAE,KAAK,CAAC,KAAK;YAClB,WAAW;YACX,MAAM,EAAE,KAAK,CAAC,MAAM;SACrB,CAAC,CACH,CAAC;QACF,OAAO;YACL,QAAQ,EAAE;gBACR,UAAU,EAAE,MAAM,CAAC,UAAU;gBAC7B,QAAQ,EAAE,KAAK,CAAC,QAAQ;gBACxB,WAAW,EAAE,KAAK,CAAC,WAAW;gBAC9B,WAAW;gBACX,UAAU,EAAE,KAAK,CAAC,KAAK,CAAC,MAAM;aAC/B;SACF,CAAC;IACJ,CAAC;IAED,yFAAyF;IACzF,KAAK,CAAC,wBAAwB,CAC5B,OAA8C,EAC9C,MAAsC;QAEtC,MAAM,IAAI,GAAG,OAAO,CAAC,eAAe,EAAE,WAAW,CAAC;QAClD,IAAI,IAAI,KAAK,SAAS;YAAE,OAAO,SAAS,CAAC;QACzC,MAAM,KAAK,GAAG,MAAM,cAAc,CAAC,mBAAmB,EAAE,GAAG,EAAE,CAC3D,IAAI,CAAC,KAAK,CAAC,iBAAiB,CAAC,EAAE,QAAQ,EAAE,MAAM,CAAC,QAAQ,EAAE,WAAW,EAAE,IAAI,EAAE,CAAC,CAC/E,CAAC;QACF,OAAO,KAAK,EAAE,UAAU,CAAC;IAC3B,CAAC;IAED,uGAAuG;IACvG,KAAK,CAAC,wBAAwB,CAAC,MAAyB,EAAE,MAAc;QACtE,MAAM,cAAc,CAAC,aAAa,EAAE,GAAG,EAAE,CACvC,IAAI,CAAC,KAAK,CAAC,WAAW,CAAC;YACrB,aAAa,EAAE,MAAM,CAAC,aAAa;YACnC,QAAQ,EAAE;gBACR,UAAU,EAAE,aAAa;gBACzB,GAAG,MAAM;gBACT,aAAa,EAAE,MAAM,CAAC,aAAa;gBACnC,SAAS,EAAE,MAAM,CAAC,SAAS;gBAC3B,QAAQ,EAAE,MAAM,CAAC,QAAQ;gBACzB,MAAM,EAAE,MAAM,CAAC,MAAM;gBACrB,OAAO,EAAE,MAAM,CAAC,OAAO;gBACvB,QAAQ,EAAE,MAAM,CAAC,QAAQ;gBACzB,OAAO,EAAE,MAAM,CAAC,OAAO;gBACvB,UAAU,EAAE,MAAM,CAAC,UAAU;gBAC7B,MAAM,EAAE,MAAM,CAAC,MAAM;gBACrB,WAAW,EAAE,MAAM,CAAC,WAAW;aAChC;SACF,CAAC,CACH,CAAC;IACJ,CAAC;IAED,mFAAmF;IACnF,KAAK,CAAC,gBAAgB,CACpB,MAAyB,EACzB,WAAiB,EACjB,WAAiB;QAEjB,MAAM,cAAc,CAAC,aAAa,EAAE,GAAG,EAAE,CACvC,IAAI,CAAC,KAAK,CAAC,WAAW,CAAC;YACrB,aAAa,EAAE,MAAM,CAAC,aAAa;YACnC,QAAQ,EAAE;gBACR,UAAU,EAAE,OAAO;gBACnB,SAAS,EAAE,MAAM,CAAC,SAAS;gBAC3B,aAAa,EAAE,MAAM,CAAC,aAAa;gBACnC,QAAQ,EAAE,MAAM,CAAC,QAAQ;gBACzB,MAAM,EAAE,MAAM,CAAC,MAAM;gBACrB,OAAO,EAAE,MAAM,CAAC,OAAO;gBACvB,WAAW,EAAE,WAAW,CAAC,WAAW,EAAE;gBACtC,WAAW,EAAE,WAAW,CAAC,WAAW,EAAE;gBACtC,GAAG,CAAC,MAAM,CAAC,QAAQ,EAAE,WAAW,KAAK,SAAS;oBAC5C,CAAC,CAAC,EAAE,WAAW,EAAE,MAAM,CAAC,QAAQ,CAAC,WAAW,EAAE;oBAC9C,CAAC,CAAC,EAAE,CAAC;gBACP,eAAe,EAAE,MAAM,CAAC,MAAM,CAAC,SAAS;aACzC;SACF,CAAC,CACH,CAAC;IACJ,CAAC;CACF"}

package/dist/types.d.ts

@@ -0,0 +1,232 @@
+/**
+ * Public types for @x12i/content-fetcher.
+ *
+ * These shapes are part of the package contract and mirror the package
+ * requirement document exactly. Downstream packages (extractor, normalizer)
+ * consume {@link FetchedArtifactReference}; do not rename or widen fields.
+ */
+/** Content hash format used throughout the package. */
+export type Sha256ContentHash = `sha256:${string}`;
+/** Ways content may be acquired from an approved source. */
+export type AcquisitionChannel = "direct_http" | "official_api" | "rss_or_atom" | "permitted_browser" | "approved_archive" | "licensed_provider" | "manual_file";
+/** Terminal outcome of an acquisition request. There is no generic "failed". */
+export type AcquisitionStatus = "collected" | "not_modified" | "metadata_only" | "robots_disallowed" | "rate_limited" | "waf_or_bot_challenge" | "login_required" | "paywalled" | "geo_restricted" | "source_not_authorised" | "not_found" | "temporarily_unavailable" | "redirect_disallowed" | "content_too_large" | "unsupported_content_type" | "integrity_error" | "manual_review_required";
+/** Rights profile attached to every result and stored with every artifact. */
+export interface RightsProfile {
+    profileId: string;
+    allowRawRetention: boolean;
+    allowExtraction: boolean;
+    allowDisplay: boolean;
+    allowTraining: boolean;
+    retentionClass: string;
+}
+/**
+ * Policy controlling every acquisition for one source. A request is only
+ * executed when it satisfies the source's policy.
+ */
+export interface SourcePolicy {
+    sourceId: string;
+    publisher: string;
+    /** Hostnames permitted for this source. A hostname or any subdomain of an entry matches. */
+    allowedDomains: string[];
+    allowedProtocols: Array<"https" | "http">;
+    allowedChannels: AcquisitionChannel[];
+    /** Channel fallback order. Must be a subset of {@link allowedChannels}. */
+    channelOrder: AcquisitionChannel[];
+    robotsPolicy: "required" | "not_applicable";
+    browserAccess: "disabled" | "permitted";
+    authProfileId?: string;
+    archiveAccess: "disabled" | "permitted";
+    licensedProviderAccess: "disabled" | "permitted";
+    rateLimit: {
+        maxRequestsPerMinute: number;
+        maxConcurrentRequests: number;
+    };
+    limits: {
+        maxRedirects: number;
+        maxResponseBytes: number;
+    };
+    rights: RightsProfile;
+}
+/** A file supplied directly by an authorized user or system. */
+export interface ManualFileInput {
+    fileName: string;
+    contentType: string;
+    bytes: Uint8Array;
+}
+/** Exactly one field must be supplied. */
+export interface FetchRequestResource {
+    url?: string;
+    apiResource?: string;
+    feedEntryUrl?: string;
+    archiveReference?: string;
+    licensedResourceId?: string;
+    manualFile?: ManualFileInput;
+}
+/** Version identifiers from a previous acquisition of the same resource. */
+export interface PreviousVersion {
+    etag?: string;
+    lastModified?: string;
+    contentHash?: Sha256ContentHash;
+}
+/** The exact resource to acquire. The fetcher never locates related URLs. */
+export interface FetchRequest {
+    requestId: string;
+    sourceId: string;
+    resource: FetchRequestResource;
+    expectedContentTypes?: string[];
+    previousVersion?: PreviousVersion;
+}
+/** Public interface returned by {@link createContentFetcher}. */
+export interface ContentFetcher {
+    fetch(request: FetchRequest): Promise<AcquisitionResult>;
+}
+export interface AcquisitionArtifact {
+    artifactId: string;
+    fileName: string;
+    contentType: string;
+    contentHash: Sha256ContentHash;
+    byteLength: number;
+}
+export interface AcquisitionVersion {
+    etag?: string;
+    lastModified?: string;
+    previousVersionId?: string;
+    duplicateOfArtifactId?: string;
+}
+export interface AcquisitionProvenance {
+    requestedUrl?: string;
+    finalUrl?: string;
+    retrievedAt: string;
+    publisher: string;
+    archiveReference?: string;
+    feedEntryId?: string;
+    apiResource?: string;
+}
+export interface AcquisitionDiagnostics {
+    httpStatus?: number;
+    retryAfterSeconds?: number;
+    reasonCode?: string;
+    sanitizedMessage?: string;
+}
+/** Returned for every request that passes input validation. Never thrown for expected retrieval outcomes. */
+export interface AcquisitionResult {
+    requestId: string;
+    acquisitionId: string;
+    status: AcquisitionStatus;
+    sourceId: string;
+    channel?: AcquisitionChannel;
+    artifact?: AcquisitionArtifact;
+    version?: AcquisitionVersion;
+    provenance: AcquisitionProvenance;
+    rights: RightsProfile;
+    diagnostics: AcquisitionDiagnostics;
+}
+/** Handoff payload for the downstream extraction / normalization package. */
+export interface FetchedArtifactReference {
+    acquisitionId: string;
+    artifactId: string;
+    sourceId: string;
+    fileName: string;
+    contentType: string;
+    contentHash: Sha256ContentHash;
+    provenance: AcquisitionProvenance;
+    rights: RightsProfile;
+}
+/** Immutable raw-artifact storage owned by the host application. */
+export interface ArtifactStore {
+    putOriginal(input: {
+        acquisitionId: string;
+        fileName: string;
+        contentType: string;
+        bytes: Uint8Array;
+        contentHash: Sha256ContentHash;
+        rights: RightsProfile;
+    }): Promise<{
+        artifactId: string;
+    }>;
+    putMetadata(input: {
+        acquisitionId: string;
+        metadata: object;
+    }): Promise<void>;
+    findByContentHash(input: {
+        sourceId: string;
+        contentHash: Sha256ContentHash;
+    }): Promise<{
+        artifactId: string;
+    } | null>;
+}
+/** Minimal HTTP transport. Implementations must NOT follow redirects automatically. */
+export interface HttpClientRequestInit {
+    method?: "GET" | "HEAD";
+    headers?: Record<string, string>;
+    signal?: AbortSignal;
+}
+export interface HttpClientResponse {
+    status: number;
+    /** Header names in lowercase. */
+    headers: Record<string, string>;
+    body?: AsyncIterable<Uint8Array>;
+}
+export interface HttpClient {
+    request(url: string, init?: HttpClientRequestInit): Promise<HttpClientResponse>;
+}
+/** Result of a browser adapter retrieval. The adapter must not attempt to bypass challenges. */
+export interface BrowserFetchOutput {
+    /** HTTP-like status the browser interaction produced. */
+    status: number;
+    finalUrl?: string;
+    contentType?: string;
+    bytes?: Uint8Array;
+    headers?: Record<string, string>;
+    /** Set when the page was not retrievable for an access-restriction reason. */
+    blockedReason?: "captcha" | "waf" | "login" | "paywall" | "geo";
+}
+export interface BrowserAdapter {
+    fetchPage(input: {
+        url: string;
+        userAgent: string;
+    }): Promise<BrowserFetchOutput>;
+}
+/** Credentials supplied for a policy's authProfileId. Values are never persisted or logged. */
+export interface AuthCredentials {
+    type: "bearer" | "header";
+    /** Header name for {@link type}:"header". Defaults to x-api-key. */
+    headerName?: string;
+    value: string;
+}
+export interface AuthProvider {
+    getCredentials(input: {
+        authProfileId: string;
+        url: string;
+        channel: AcquisitionChannel;
+    }): Promise<AuthCredentials | null>;
+}
+/** Address resolution hook so SSRF validation can run after DNS resolution. */
+export interface DnsResolver {
+    lookup(hostname: string): Promise<{
+        addresses: string[];
+    }>;
+}
+/** Options for {@link createContentFetcher}. */
+export interface ContentFetcherOptions {
+    sourcePolicies: SourcePolicy[];
+    artifactStore: ArtifactStore;
+    /** Required (at creation time) when any policy enables a network channel. */
+    httpClient?: HttpClient;
+    /** Required (at fetch time) when a permitted_browser acquisition is attempted. */
+    browserAdapter?: BrowserAdapter;
+    /** Required (at fetch time) when a policy declaring an authProfileId is used. */
+    authProvider?: AuthProvider;
+    /** Default: system resolver. */
+    dnsResolver?: DnsResolver;
+    /** User agent for requests and robots.txt evaluation. Default: "x12i-content-fetcher/1.0". */
+    userAgent?: string;
+    /** Default: 10 minutes. */
+    robotsCacheTtlMs?: number;
+    /** Default: 30_000 ms. */
+    requestTimeoutMs?: number;
+    /** Clock injection for tests. Default: () => new Date(). */
+    now?: () => Date;
+}
+//# sourceMappingURL=types.d.ts.map

package/dist/types.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"types.d.ts","sourceRoot":"","sources":["../src/types.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAEH,uDAAuD;AACvD,MAAM,MAAM,iBAAiB,GAAG,UAAU,MAAM,EAAE,CAAC;AAEnD,4DAA4D;AAC5D,MAAM,MAAM,kBAAkB,GAC1B,aAAa,GACb,cAAc,GACd,aAAa,GACb,mBAAmB,GACnB,kBAAkB,GAClB,mBAAmB,GACnB,aAAa,CAAC;AAElB,gFAAgF;AAChF,MAAM,MAAM,iBAAiB,GACzB,WAAW,GACX,cAAc,GACd,eAAe,GACf,mBAAmB,GACnB,cAAc,GACd,sBAAsB,GACtB,gBAAgB,GAChB,WAAW,GACX,gBAAgB,GAChB,uBAAuB,GACvB,WAAW,GACX,yBAAyB,GACzB,qBAAqB,GACrB,mBAAmB,GACnB,0BAA0B,GAC1B,iBAAiB,GACjB,wBAAwB,CAAC;AAE7B,8EAA8E;AAC9E,MAAM,WAAW,aAAa;IAC5B,SAAS,EAAE,MAAM,CAAC;IAClB,iBAAiB,EAAE,OAAO,CAAC;IAC3B,eAAe,EAAE,OAAO,CAAC;IACzB,YAAY,EAAE,OAAO,CAAC;IACtB,aAAa,EAAE,OAAO,CAAC;IACvB,cAAc,EAAE,MAAM,CAAC;CACxB;AAED;;;GAGG;AACH,MAAM,WAAW,YAAY;IAC3B,QAAQ,EAAE,MAAM,CAAC;IACjB,SAAS,EAAE,MAAM,CAAC;IAElB,4FAA4F;IAC5F,cAAc,EAAE,MAAM,EAAE,CAAC;IACzB,gBAAgB,EAAE,KAAK,CAAC,OAAO,GAAG,MAAM,CAAC,CAAC;IAE1C,eAAe,EAAE,kBAAkB,EAAE,CAAC;IACtC,2EAA2E;IAC3E,YAAY,EAAE,kBAAkB,EAAE,CAAC;IAEnC,YAAY,EAAE,UAAU,GAAG,gBAAgB,CAAC;IAC5C,aAAa,EAAE,UAAU,GAAG,WAAW,CAAC;IACxC,aAAa,CAAC,EAAE,MAAM,CAAC;IAEvB,aAAa,EAAE,UAAU,GAAG,WAAW,CAAC;IACxC,sBAAsB,EAAE,UAAU,GAAG,WAAW,CAAC;IAEjD,SAAS,EAAE;QACT,oBAAoB,EAAE,MAAM,CAAC;QAC7B,qBAAqB,EAAE,MAAM,CAAC;KAC/B,CAAC;IAEF,MAAM,EAAE;QACN,YAAY,EAAE,MAAM,CAAC;QACrB,gBAAgB,EAAE,MAAM,CAAC;KAC1B,CAAC;IAEF,MAAM,EAAE,aAAa,CAAC;CACvB;AAED,gEAAgE;AAChE,MAAM,WAAW,eAAe;IAC9B,QAAQ,EAAE,MAAM,CAAC;IACjB,WAAW,EAAE,MAAM,CAAC;IACpB,KAAK,EAAE,UAAU,CAAC;CACnB;AAED,0CAA0C;AAC1C,MAAM,WAAW,oBAAoB;IACnC,GAAG,CAAC,EAAE,MAAM,CAAC;IACb,WAAW,CAAC,EAAE,MAAM,CAAC;IACrB,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,gBAAgB,CAAC,EAAE,MAAM,CAAC;IAC1B,kBAAkB,CAAC,EAAE,MAAM,CAAC;IAC5B,UAAU,CAAC,EAAE,eAAe,CAAC;CAC9B;AAED,4EAA4E;AAC5E,MAAM,WAAW,eAAe;IAC9B,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,WAAW,CAAC,EAAE,iBAAiB,CAAC;CACjC;AAED,6EAA6E;AAC7E,MAAM,WAAW,YAAY;IAC3B,SAAS,EAAE,MAAM,CAAC;IAClB,QAAQ,EAAE,MAAM,CAAC;IACjB,QAAQ,EAAE,oBAAoB,CAAC;IAC/B,oBAAoB,CAAC,EAAE,MAAM,EAAE,CAAC;IAChC,eAAe,CAAC,EAAE,eAAe,CAAC;CACnC;AAED,iEAAiE;AACjE,MAAM,WAAW,cAAc;IAC7B,KAAK,CAAC,OAAO,EAAE,YAAY,GAAG,OAAO,CAAC,iBAAiB,CAAC,CAAC;CAC1D;AAED,MAAM,WAAW,mBAAmB;IAClC,UAAU,EAAE,MAAM,CAAC;IACnB,QAAQ,EAAE,MAAM,CAAC;IACjB,WAAW,EAAE,MAAM,CAAC;IACpB,WAAW,EAAE,iBAAiB,CAAC;IAC/B,UAAU,EAAE,MAAM,CAAC;CACpB;AAED,MAAM,WAAW,kBAAkB;IACjC,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,iBAAiB,CAAC,EAAE,MAAM,CAAC;IAC3B,qBAAqB,CAAC,EAAE,MAAM,CAAC;CAChC;AAED,MAAM,WAAW,qBAAqB;IACpC,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,WAAW,EAAE,MAAM,CAAC;IACpB,SAAS,EAAE,MAAM,CAAC;IAClB,gBAAgB,CAAC,EAAE,MAAM,CAAC;IAC1B,WAAW,CAAC,EAAE,MAAM,CAAC;IACrB,WAAW,CAAC,EAAE,MAAM,CAAC;CACtB;AAED,MAAM,WAAW,sBAAsB;IACrC,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,iBAAiB,CAAC,EAAE,MAAM,CAAC;IAC3B,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,gBAAgB,CAAC,EAAE,MAAM,CAAC;CAC3B;AAED,6GAA6G;AAC7G,MAAM,WAAW,iBAAiB;IAChC,SAAS,EAAE,MAAM,CAAC;IAClB,aAAa,EAAE,MAAM,CAAC;IACtB,MAAM,EAAE,iBAAiB,CAAC;IAE1B,QAAQ,EAAE,MAAM,CAAC;IACjB,OAAO,CAAC,EAAE,kBAAkB,CAAC;IAE7B,QAAQ,CAAC,EAAE,mBAAmB,CAAC;IAE/B,OAAO,CAAC,EAAE,kBAAkB,CAAC;IAE7B,UAAU,EAAE,qBAAqB,CAAC;IAClC,MAAM,EAAE,aAAa,CAAC;IAEtB,WAAW,EAAE,sBAAsB,CAAC;CACrC;AAED,6EAA6E;AAC7E,MAAM,WAAW,wBAAwB;IACvC,aAAa,EAAE,MAAM,CAAC;IACtB,UAAU,EAAE,MAAM,CAAC;IACnB,QAAQ,EAAE,MAAM,CAAC;IACjB,QAAQ,EAAE,MAAM,CAAC;IACjB,WAAW,EAAE,MAAM,CAAC;IACpB,WAAW,EAAE,iBAAiB,CAAC;IAC/B,UAAU,EAAE,qBAAqB,CAAC;IAClC,MAAM,EAAE,aAAa,CAAC;CACvB;AAMD,oEAAoE;AACpE,MAAM,WAAW,aAAa;IAC5B,WAAW,CAAC,KAAK,EAAE;QACjB,aAAa,EAAE,MAAM,CAAC;QACtB,QAAQ,EAAE,MAAM,CAAC;QACjB,WAAW,EAAE,MAAM,CAAC;QACpB,KAAK,EAAE,UAAU,CAAC;QAClB,WAAW,EAAE,iBAAiB,CAAC;QAC/B,MAAM,EAAE,aAAa,CAAC;KACvB,GAAG,OAAO,CAAC;QAAE,UAAU,EAAE,MAAM,CAAA;KAAE,CAAC,CAAC;IAEpC,WAAW,CAAC,KAAK,EAAE;QACjB,aAAa,EAAE,MAAM,CAAC;QACtB,QAAQ,EAAE,MAAM,CAAC;KAClB,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC;IAElB,iBAAiB,CAAC,KAAK,EAAE;QACvB,QAAQ,EAAE,MAAM,CAAC;QACjB,WAAW,EAAE,iBAAiB,CAAC;KAChC,GAAG,OAAO,CAAC;QAAE,UAAU,EAAE,MAAM,CAAA;KAAE,GAAG,IAAI,CAAC,CAAC;CAC5C;AAED,uFAAuF;AACvF,MAAM,WAAW,qBAAqB;IACpC,MAAM,CAAC,EAAE,KAAK,GAAG,MAAM,CAAC;IACxB,OAAO,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;IACjC,MAAM,CAAC,EAAE,WAAW,CAAC;CACtB;AAED,MAAM,WAAW,kBAAkB;IACjC,MAAM,EAAE,MAAM,CAAC;IACf,iCAAiC;IACjC,OAAO,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;IAChC,IAAI,CAAC,EAAE,aAAa,CAAC,UAAU,CAAC,CAAC;CAClC;AAED,MAAM,WAAW,UAAU;IACzB,OAAO,CAAC,GAAG,EAAE,MAAM,EAAE,IAAI,CAAC,EAAE,qBAAqB,GAAG,OAAO,CAAC,kBAAkB,CAAC,CAAC;CACjF;AAED,gGAAgG;AAChG,MAAM,WAAW,kBAAkB;IACjC,yDAAyD;IACzD,MAAM,EAAE,MAAM,CAAC;IACf,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,WAAW,CAAC,EAAE,MAAM,CAAC;IACrB,KAAK,CAAC,EAAE,UAAU,CAAC;IACnB,OAAO,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;IACjC,8EAA8E;IAC9E,aAAa,CAAC,EAAE,SAAS,GAAG,KAAK,GAAG,OAAO,GAAG,SAAS,GAAG,KAAK,CAAC;CACjE;AAED,MAAM,WAAW,cAAc;IAC7B,SAAS,CAAC,KAAK,EAAE;QACf,GAAG,EAAE,MAAM,CAAC;QACZ,SAAS,EAAE,MAAM,CAAC;KACnB,GAAG,OAAO,CAAC,kBAAkB,CAAC,CAAC;CACjC;AAED,+FAA+F;AAC/F,MAAM,WAAW,eAAe;IAC9B,IAAI,EAAE,QAAQ,GAAG,QAAQ,CAAC;IAC1B,oEAAoE;IACpE,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,KAAK,EAAE,MAAM,CAAC;CACf;AAED,MAAM,WAAW,YAAY;IAC3B,cAAc,CAAC,KAAK,EAAE;QACpB,aAAa,EAAE,MAAM,CAAC;QACtB,GAAG,EAAE,MAAM,CAAC;QACZ,OAAO,EAAE,kBAAkB,CAAC;KAC7B,GAAG,OAAO,CAAC,eAAe,GAAG,IAAI,CAAC,CAAC;CACrC;AAED,+EAA+E;AAC/E,MAAM,WAAW,WAAW;IAC1B,MAAM,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC;QAAE,SAAS,EAAE,MAAM,EAAE,CAAA;KAAE,CAAC,CAAC;CAC5D;AAED,gDAAgD;AAChD,MAAM,WAAW,qBAAqB;IACpC,cAAc,EAAE,YAAY,EAAE,CAAC;IAC/B,aAAa,EAAE,aAAa,CAAC;IAE7B,6EAA6E;IAC7E,UAAU,CAAC,EAAE,UAAU,CAAC;IACxB,kFAAkF;IAClF,cAAc,CAAC,EAAE,cAAc,CAAC;IAChC,iFAAiF;IACjF,YAAY,CAAC,EAAE,YAAY,CAAC;IAC5B,gCAAgC;IAChC,WAAW,CAAC,EAAE,WAAW,CAAC;IAE1B,8FAA8F;IAC9F,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,2BAA2B;IAC3B,gBAAgB,CAAC,EAAE,MAAM,CAAC;IAC1B,0BAA0B;IAC1B,gBAAgB,CAAC,EAAE,MAAM,CAAC;IAC1B,4DAA4D;IAC5D,GAAG,CAAC,EAAE,MAAM,IAAI,CAAC;CAClB"}

package/dist/types.js

@@ -0,0 +1,9 @@
+/**
+ * Public types for @x12i/content-fetcher.
+ *
+ * These shapes are part of the package contract and mirror the package
+ * requirement document exactly. Downstream packages (extractor, normalizer)
+ * consume {@link FetchedArtifactReference}; do not rename or widen fields.
+ */
+export {};
+//# sourceMappingURL=types.js.map

package/dist/types.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"types.js","sourceRoot":"","sources":["../src/types.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG"}

package/dist/util/bytes.d.ts

@@ -0,0 +1,21 @@
+/** Concatenate byte chunks. */
+export declare function concatBytes(chunks: Uint8Array[]): Uint8Array;
+export declare function utf8Bytes(text: string): Uint8Array;
+export interface BodyReadResult {
+    bytes: Uint8Array;
+    /** True when the stream produced more than `maxBytes` bytes. */
+    exceeded: boolean;
+}
+/**
+ * Read an async byte stream under a byte budget.
+ *
+ * - `maxBytes`: hard cap. Reading past it sets `exceeded` and stops iteration.
+ * - `prefixOnly` + `prefixBytes`: stop as soon as the prefix budget is met
+ *   (used to sniff classification markers without downloading a full body).
+ */
+export declare function readBody(body: AsyncIterable<Uint8Array> | undefined, options: {
+    maxBytes: number;
+    prefixBytes?: number;
+    prefixOnly?: boolean;
+}): Promise<BodyReadResult>;
+//# sourceMappingURL=bytes.d.ts.map

package/dist/util/bytes.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"bytes.d.ts","sourceRoot":"","sources":["../../src/util/bytes.ts"],"names":[],"mappings":"AAAA,+BAA+B;AAC/B,wBAAgB,WAAW,CAAC,MAAM,EAAE,UAAU,EAAE,GAAG,UAAU,CAU5D;AAID,wBAAgB,SAAS,CAAC,IAAI,EAAE,MAAM,GAAG,UAAU,CAElD;AAED,MAAM,WAAW,cAAc;IAC7B,KAAK,EAAE,UAAU,CAAC;IAClB,gEAAgE;IAChE,QAAQ,EAAE,OAAO,CAAC;CACnB;AAED;;;;;;GAMG;AACH,wBAAsB,QAAQ,CAC5B,IAAI,EAAE,aAAa,CAAC,UAAU,CAAC,GAAG,SAAS,EAC3C,OAAO,EAAE;IAAE,QAAQ,EAAE,MAAM,CAAC;IAAC,WAAW,CAAC,EAAE,MAAM,CAAC;IAAC,UAAU,CAAC,EAAE,OAAO,CAAA;CAAE,GACxE,OAAO,CAAC,cAAc,CAAC,CAoBzB"}

package/dist/util/bytes.js

@@ -0,0 +1,44 @@
+/** Concatenate byte chunks. */
+export function concatBytes(chunks) {
+    let total = 0;
+    for (const chunk of chunks)
+        total += chunk.length;
+    const out = new Uint8Array(total);
+    let offset = 0;
+    for (const chunk of chunks) {
+        out.set(chunk, offset);
+        offset += chunk.length;
+    }
+    return out;
+}
+const encoder = new TextEncoder();
+export function utf8Bytes(text) {
+    return encoder.encode(text);
+}
+/**
+ * Read an async byte stream under a byte budget.
+ *
+ * - `maxBytes`: hard cap. Reading past it sets `exceeded` and stops iteration.
+ * - `prefixOnly` + `prefixBytes`: stop as soon as the prefix budget is met
+ *   (used to sniff classification markers without downloading a full body).
+ */
+export async function readBody(body, options) {
+    const chunks = [];
+    let total = 0;
+    if (body) {
+        for await (const chunk of body) {
+            chunks.push(chunk);
+            total += chunk.length;
+            if (total > options.maxBytes) {
+                return { bytes: concatBytes(chunks), exceeded: true };
+            }
+            if (options.prefixOnly &&
+                options.prefixBytes !== undefined &&
+                total >= options.prefixBytes) {
+                break;
+            }
+        }
+    }
+    return { bytes: concatBytes(chunks), exceeded: false };
+}
+//# sourceMappingURL=bytes.js.map

package/dist/util/bytes.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"bytes.js","sourceRoot":"","sources":["../../src/util/bytes.ts"],"names":[],"mappings":"AAAA,+BAA+B;AAC/B,MAAM,UAAU,WAAW,CAAC,MAAoB;IAC9C,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,KAAK,MAAM,KAAK,IAAI,MAAM;QAAE,KAAK,IAAI,KAAK,CAAC,MAAM,CAAC;IAClD,MAAM,GAAG,GAAG,IAAI,UAAU,CAAC,KAAK,CAAC,CAAC;IAClC,IAAI,MAAM,GAAG,CAAC,CAAC;IACf,KAAK,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;QAC3B,GAAG,CAAC,GAAG,CAAC,KAAK,EAAE,MAAM,CAAC,CAAC;QACvB,MAAM,IAAI,KAAK,CAAC,MAAM,CAAC;IACzB,CAAC;IACD,OAAO,GAAG,CAAC;AACb,CAAC;AAED,MAAM,OAAO,GAAG,IAAI,WAAW,EAAE,CAAC;AAElC,MAAM,UAAU,SAAS,CAAC,IAAY;IACpC,OAAO,OAAO,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;AAC9B,CAAC;AAQD;;;;;;GAMG;AACH,MAAM,CAAC,KAAK,UAAU,QAAQ,CAC5B,IAA2C,EAC3C,OAAyE;IAEzE,MAAM,MAAM,GAAiB,EAAE,CAAC;IAChC,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,IAAI,IAAI,EAAE,CAAC;QACT,IAAI,KAAK,EAAE,MAAM,KAAK,IAAI,IAAI,EAAE,CAAC;YAC/B,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YACnB,KAAK,IAAI,KAAK,CAAC,MAAM,CAAC;YACtB,IAAI,KAAK,GAAG,OAAO,CAAC,QAAQ,EAAE,CAAC;gBAC7B,OAAO,EAAE,KAAK,EAAE,WAAW,CAAC,MAAM,CAAC,EAAE,QAAQ,EAAE,IAAI,EAAE,CAAC;YACxD,CAAC;YACD,IACE,OAAO,CAAC,UAAU;gBAClB,OAAO,CAAC,WAAW,KAAK,SAAS;gBACjC,KAAK,IAAI,OAAO,CAAC,WAAW,EAC5B,CAAC;gBACD,MAAM;YACR,CAAC;QACH,CAAC;IACH,CAAC;IACD,OAAO,EAAE,KAAK,EAAE,WAAW,CAAC,MAAM,CAAC,EAAE,QAAQ,EAAE,KAAK,EAAE,CAAC;AACzD,CAAC"}

package/dist/util/hash.d.ts

@@ -0,0 +1,5 @@
+import type { Sha256ContentHash } from "../types.js";
+export declare function sha256Hex(bytes: Uint8Array): string;
+export declare function sha256ContentHash(bytes: Uint8Array): Sha256ContentHash;
+export declare function isSha256ContentHash(value: unknown): value is Sha256ContentHash;
+//# sourceMappingURL=hash.d.ts.map

package/dist/util/hash.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"hash.d.ts","sourceRoot":"","sources":["../../src/util/hash.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EAAE,iBAAiB,EAAE,MAAM,aAAa,CAAC;AAErD,wBAAgB,SAAS,CAAC,KAAK,EAAE,UAAU,GAAG,MAAM,CAEnD;AAED,wBAAgB,iBAAiB,CAAC,KAAK,EAAE,UAAU,GAAG,iBAAiB,CAEtE;AAED,wBAAgB,mBAAmB,CAAC,KAAK,EAAE,OAAO,GAAG,KAAK,IAAI,iBAAiB,CAE9E"}

package/dist/util/hash.js

@@ -0,0 +1,11 @@
+import { createHash } from "node:crypto";
+export function sha256Hex(bytes) {
+    return createHash("sha256").update(bytes).digest("hex");
+}
+export function sha256ContentHash(bytes) {
+    return `sha256:${sha256Hex(bytes)}`;
+}
+export function isSha256ContentHash(value) {
+    return typeof value === "string" && /^sha256:[0-9a-f]{64}$/.test(value);
+}
+//# sourceMappingURL=hash.js.map

package/dist/util/hash.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"hash.js","sourceRoot":"","sources":["../../src/util/hash.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,UAAU,EAAE,MAAM,aAAa,CAAC;AAGzC,MAAM,UAAU,SAAS,CAAC,KAAiB;IACzC,OAAO,UAAU,CAAC,QAAQ,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;AAC1D,CAAC;AAED,MAAM,UAAU,iBAAiB,CAAC,KAAiB;IACjD,OAAO,UAAU,SAAS,CAAC,KAAK,CAAC,EAAE,CAAC;AACtC,CAAC;AAED,MAAM,UAAU,mBAAmB,CAAC,KAAc;IAChD,OAAO,OAAO,KAAK,KAAK,QAAQ,IAAI,uBAAuB,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;AAC1E,CAAC"}

package/dist/util/ids.d.ts

@@ -0,0 +1,3 @@
+export declare function newAcquisitionId(): string;
+export declare function newVersionId(): string;
+//# sourceMappingURL=ids.d.ts.map

package/dist/util/ids.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"ids.d.ts","sourceRoot":"","sources":["../../src/util/ids.ts"],"names":[],"mappings":"AAEA,wBAAgB,gBAAgB,IAAI,MAAM,CAEzC;AAED,wBAAgB,YAAY,IAAI,MAAM,CAErC"}

package/dist/util/ids.js

@@ -0,0 +1,8 @@
+import { randomUUID } from "node:crypto";
+export function newAcquisitionId() {
+    return `acq_${randomUUID()}`;
+}
+export function newVersionId() {
+    return `ver_${randomUUID()}`;
+}
+//# sourceMappingURL=ids.js.map

package/dist/util/ids.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"ids.js","sourceRoot":"","sources":["../../src/util/ids.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,UAAU,EAAE,MAAM,aAAa,CAAC;AAEzC,MAAM,UAAU,gBAAgB;IAC9B,OAAO,OAAO,UAAU,EAAE,EAAE,CAAC;AAC/B,CAAC;AAED,MAAM,UAAU,YAAY;IAC1B,OAAO,OAAO,UAAU,EAAE,EAAE,CAAC;AAC/B,CAAC"}

package/dist/validate-request.d.ts

@@ -0,0 +1,9 @@
+import type { FetchRequest, FetchRequestResource } from "./types.js";
+export type ResourceKind = "url" | "api_resource" | "feed_entry" | "archive_reference" | "licensed_resource" | "manual_file";
+/**
+ * Runtime validation of a fetch request against the request contract.
+ * Throws InvalidFetchRequestError; returns a normalized deep copy.
+ */
+export declare function validateFetchRequest(input: unknown): FetchRequest;
+export declare function resourceKindOf(resource: FetchRequestResource): ResourceKind;
+//# sourceMappingURL=validate-request.d.ts.map

package/dist/validate-request.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"validate-request.d.ts","sourceRoot":"","sources":["../src/validate-request.ts"],"names":[],"mappings":"AASA,OAAO,KAAK,EAAE,YAAY,EAAE,oBAAoB,EAAE,MAAM,YAAY,CAAC;AAWrE,MAAM,MAAM,YAAY,GACpB,KAAK,GACL,cAAc,GACd,YAAY,GACZ,mBAAmB,GACnB,mBAAmB,GACnB,aAAa,CAAC;AAelB;;;GAGG;AACH,wBAAgB,oBAAoB,CAAC,KAAK,EAAE,OAAO,GAAG,YAAY,CAwGjE;AAED,wBAAgB,cAAc,CAAC,QAAQ,EAAE,oBAAoB,GAAG,YAAY,CAO3E"}

package/dist/validate-request.js

@@ -0,0 +1,137 @@
+import { InvalidFetchRequestError } from "./errors.js";
+import { isSha256ContentHash } from "./util/hash.js";
+import { isMimeTypeLike, isNonEmptyString, isRecord, isStringArray, isUint8Array, } from "./runtime-utils.js";
+const RESOURCE_KEYS = [
+    "url",
+    "apiResource",
+    "feedEntryUrl",
+    "archiveReference",
+    "licensedResourceId",
+    "manualFile",
+];
+function fail(code, message) {
+    throw new InvalidFetchRequestError(message, code);
+}
+function isAbsoluteUrl(raw) {
+    try {
+        const url = new URL(raw);
+        return url.hostname !== "";
+    }
+    catch {
+        return false;
+    }
+}
+/**
+ * Runtime validation of a fetch request against the request contract.
+ * Throws InvalidFetchRequestError; returns a normalized deep copy.
+ */
+export function validateFetchRequest(input) {
+    if (!isRecord(input))
+        fail("request_not_object", "FetchRequest must be an object.");
+    if (!isNonEmptyString(input.requestId)) {
+        fail("missing_request_id", "requestId must be a non-empty string.");
+    }
+    if (!isNonEmptyString(input.sourceId)) {
+        fail("missing_source_id", "sourceId must be a non-empty string.");
+    }
+    if (!isRecord(input.resource)) {
+        fail("missing_resource", "resource must be an object.");
+    }
+    const resourceInput = input.resource;
+    const supplied = RESOURCE_KEYS.filter((key) => resourceInput[key] !== undefined);
+    if (supplied.length === 0) {
+        fail("resource_missing", "resource must supply exactly one resource reference.");
+    }
+    if (supplied.length > 1) {
+        fail("resource_ambiguous", `resource must supply exactly one resource reference, got: ${supplied.join(", ")}.`);
+    }
+    const kind = supplied[0];
+    const resource = {};
+    if (kind === "manualFile") {
+        const manual = input.resource.manualFile;
+        if (!isRecord(manual) ||
+            !isNonEmptyString(manual.fileName) ||
+            !isMimeTypeLike(manual.contentType) ||
+            !isUint8Array(manual.bytes) ||
+            manual.bytes.length === 0) {
+            fail("manual_file_invalid", "manualFile must supply a non-empty fileName, a MIME contentType, and at least one byte.");
+        }
+        resource.manualFile = {
+            fileName: manual.fileName,
+            contentType: manual.contentType.trim(),
+            bytes: manual.bytes,
+        };
+    }
+    else {
+        const value = input.resource[kind];
+        if (!isNonEmptyString(value)) {
+            fail(`${kind}_invalid`, `resource.${kind} must be a non-empty string.`);
+        }
+        if (!isAbsoluteUrl(value)) {
+            fail(`${kind}_not_absolute`, `resource.${kind} must be an absolute URL with a host.`);
+        }
+        resource[kind] = value;
+    }
+    let expectedContentTypes;
+    if (input.expectedContentTypes !== undefined) {
+        if (!isStringArray(input.expectedContentTypes) || input.expectedContentTypes.length === 0) {
+            fail("expected_content_types_invalid", "expectedContentTypes must be a non-empty string array when provided.");
+        }
+        for (const entry of input.expectedContentTypes) {
+            if (!isMimeTypeLike(entry)) {
+                fail("expected_content_types_invalid", `expectedContentTypes contains an invalid MIME pattern: ${entry}.`);
+            }
+        }
+        expectedContentTypes = input.expectedContentTypes.map((e) => e.trim().toLowerCase());
+    }
+    let previousVersion;
+    if (input.previousVersion !== undefined) {
+        const pv = input.previousVersion;
+        if (!isRecord(pv))
+            fail("previous_version_invalid", "previousVersion must be an object when provided.");
+        if (pv.etag !== undefined && !isNonEmptyString(pv.etag)) {
+            fail("previous_version_invalid", "previousVersion.etag must be a non-empty string.");
+        }
+        if (pv.lastModified !== undefined && !isNonEmptyString(pv.lastModified)) {
+            fail("previous_version_invalid", "previousVersion.lastModified must be a non-empty string.");
+        }
+        if (pv.contentHash !== undefined) {
+            const normalized = String(pv.contentHash).toLowerCase();
+            if (!isSha256ContentHash(normalized)) {
+                fail("previous_version_invalid", "previousVersion.contentHash must match sha256:<64 lowercase hex>.");
+            }
+            previousVersion = {
+                ...(pv.etag !== undefined ? { etag: pv.etag } : {}),
+                ...(pv.lastModified !== undefined ? { lastModified: pv.lastModified } : {}),
+                ...(pv.contentHash !== undefined ? { contentHash: normalized } : {}),
+            };
+        }
+        else {
+            previousVersion = {
+                ...(pv.etag !== undefined ? { etag: pv.etag } : {}),
+                ...(pv.lastModified !== undefined ? { lastModified: pv.lastModified } : {}),
+            };
+        }
+    }
+    return {
+        requestId: input.requestId,
+        sourceId: input.sourceId,
+        resource,
+        ...(expectedContentTypes !== undefined ? { expectedContentTypes } : {}),
+        ...(previousVersion !== undefined ? { previousVersion } : {}),
+    };
+}
+export function resourceKindOf(resource) {
+    if (resource.url !== undefined)
+        return "url";
+    if (resource.apiResource !== undefined)
+        return "api_resource";
+    if (resource.feedEntryUrl !== undefined)
+        return "feed_entry";
+    if (resource.archiveReference !== undefined)
+        return "archive_reference";
+    if (resource.licensedResourceId !== undefined)
+        return "licensed_resource";
+    return "manual_file";
+}
+//# sourceMappingURL=validate-request.js.map