@bluefields/db 0.2.0

package/dist/schema/scrape-cache.js

@@ -0,0 +1,23 @@
+// SPDX-License-Identifier: Apache-2.0
+import { integer, pgTable, primaryKey, text, timestamp, uuid } from 'drizzle-orm/pg-core';
+import { customers } from './customers.js';
+/**
+ * scrape_cache — per-customer URL-level cache for /scrape maxAge (Phase A.5).
+ *
+ * Composite PK on (customer_id, url_canonical). One row per URL per
+ * customer — upserted on every fresh scrape so we always reflect the
+ * latest fetch. Cross-customer sharing happens deeper at the extractor
+ * cache layer (content_hash + schema_hash).
+ */
+export const scrapeCache = pgTable('scrape_cache', {
+    customerId: uuid('customer_id')
+        .notNull()
+        .references(() => customers.id, { onDelete: 'cascade' }),
+    urlCanonical: text('url_canonical').notNull(),
+    extractionId: uuid('extraction_id'),
+    contentHash: text('content_hash').notNull(),
+    finalUrl: text('final_url').notNull(),
+    responseStatus: integer('response_status').notNull(),
+    fetchedAt: timestamp('fetched_at', { withTimezone: true }).notNull().defaultNow(),
+}, (table) => [primaryKey({ columns: [table.customerId, table.urlCanonical] })]);
+//# sourceMappingURL=scrape-cache.js.map

package/dist/schema/scrape-cache.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"scrape-cache.js","sourceRoot":"","sources":["../../src/schema/scrape-cache.ts"],"names":[],"mappings":"AAAA,sCAAsC;AACtC,OAAO,EAAE,OAAO,EAAE,OAAO,EAAE,UAAU,EAAE,IAAI,EAAE,SAAS,EAAE,IAAI,EAAE,MAAM,qBAAqB,CAAC;AAE1F,OAAO,EAAE,SAAS,EAAE,MAAM,gBAAgB,CAAC;AAE3C;;;;;;;GAOG;AACH,MAAM,CAAC,MAAM,WAAW,GAAG,OAAO,CAChC,cAAc,EACd;IACE,UAAU,EAAE,IAAI,CAAC,aAAa,CAAC;SAC5B,OAAO,EAAE;SACT,UAAU,CAAC,GAAG,EAAE,CAAC,SAAS,CAAC,EAAE,EAAE,EAAE,QAAQ,EAAE,SAAS,EAAE,CAAC;IAC1D,YAAY,EAAE,IAAI,CAAC,eAAe,CAAC,CAAC,OAAO,EAAE;IAC7C,YAAY,EAAE,IAAI,CAAC,eAAe,CAAC;IACnC,WAAW,EAAE,IAAI,CAAC,cAAc,CAAC,CAAC,OAAO,EAAE;IAC3C,QAAQ,EAAE,IAAI,CAAC,WAAW,CAAC,CAAC,OAAO,EAAE;IACrC,cAAc,EAAE,OAAO,CAAC,iBAAiB,CAAC,CAAC,OAAO,EAAE;IACpD,SAAS,EAAE,SAAS,CAAC,YAAY,EAAE,EAAE,YAAY,EAAE,IAAI,EAAE,CAAC,CAAC,OAAO,EAAE,CAAC,UAAU,EAAE;CAClF,EACD,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,UAAU,CAAC,EAAE,OAAO,EAAE,CAAC,KAAK,CAAC,UAAU,EAAE,KAAK,CAAC,YAAY,CAAC,EAAE,CAAC,CAAC,CAC7E,CAAC"}

package/dist/schema/sessions.d.ts

@@ -0,0 +1,153 @@
+/**
+ * sessions — named, per-customer cookie/header bundles for reuse
+ * across /scrape, /crawl, and /batch calls (Phase G).
+ *
+ * Cookies + headers are AES-256-GCM ciphertext, stored as a versioned
+ * JSON envelope by the helpers in @bluefields/control/sessions.ts.
+ * Unique on (customer_id, name) so save/use lookups are O(1).
+ */
+export declare const sessions: import("drizzle-orm/pg-core").PgTableWithColumns<{
+    name: "sessions";
+    schema: undefined;
+    columns: {
+        id: import("drizzle-orm/pg-core").PgColumn<{
+            name: "id";
+            tableName: "sessions";
+            dataType: "string";
+            columnType: "PgUUID";
+            data: string;
+            driverParam: string;
+            notNull: true;
+            hasDefault: true;
+            isPrimaryKey: true;
+            isAutoincrement: false;
+            hasRuntimeDefault: false;
+            enumValues: undefined;
+            baseColumn: never;
+            identity: undefined;
+            generated: undefined;
+        }, {}, {}>;
+        customerId: import("drizzle-orm/pg-core").PgColumn<{
+            name: "customer_id";
+            tableName: "sessions";
+            dataType: "string";
+            columnType: "PgUUID";
+            data: string;
+            driverParam: string;
+            notNull: true;
+            hasDefault: false;
+            isPrimaryKey: false;
+            isAutoincrement: false;
+            hasRuntimeDefault: false;
+            enumValues: undefined;
+            baseColumn: never;
+            identity: undefined;
+            generated: undefined;
+        }, {}, {}>;
+        name: import("drizzle-orm/pg-core").PgColumn<{
+            name: "name";
+            tableName: "sessions";
+            dataType: "string";
+            columnType: "PgText";
+            data: string;
+            driverParam: string;
+            notNull: true;
+            hasDefault: false;
+            isPrimaryKey: false;
+            isAutoincrement: false;
+            hasRuntimeDefault: false;
+            enumValues: [string, ...string[]];
+            baseColumn: never;
+            identity: undefined;
+            generated: undefined;
+        }, {}, {}>;
+        cookiesCiphertext: import("drizzle-orm/pg-core").PgColumn<{
+            name: "cookies_ciphertext";
+            tableName: "sessions";
+            dataType: "string";
+            columnType: "PgText";
+            data: string;
+            driverParam: string;
+            notNull: true;
+            hasDefault: false;
+            isPrimaryKey: false;
+            isAutoincrement: false;
+            hasRuntimeDefault: false;
+            enumValues: [string, ...string[]];
+            baseColumn: never;
+            identity: undefined;
+            generated: undefined;
+        }, {}, {}>;
+        headersCiphertext: import("drizzle-orm/pg-core").PgColumn<{
+            name: "headers_ciphertext";
+            tableName: "sessions";
+            dataType: "string";
+            columnType: "PgText";
+            data: string;
+            driverParam: string;
+            notNull: false;
+            hasDefault: false;
+            isPrimaryKey: false;
+            isAutoincrement: false;
+            hasRuntimeDefault: false;
+            enumValues: [string, ...string[]];
+            baseColumn: never;
+            identity: undefined;
+            generated: undefined;
+        }, {}, {}>;
+        expiresAt: import("drizzle-orm/pg-core").PgColumn<{
+            name: "expires_at";
+            tableName: "sessions";
+            dataType: "date";
+            columnType: "PgTimestamp";
+            data: Date;
+            driverParam: string;
+            notNull: false;
+            hasDefault: false;
+            isPrimaryKey: false;
+            isAutoincrement: false;
+            hasRuntimeDefault: false;
+            enumValues: undefined;
+            baseColumn: never;
+            identity: undefined;
+            generated: undefined;
+        }, {}, {}>;
+        createdAt: import("drizzle-orm/pg-core").PgColumn<{
+            name: "created_at";
+            tableName: "sessions";
+            dataType: "date";
+            columnType: "PgTimestamp";
+            data: Date;
+            driverParam: string;
+            notNull: true;
+            hasDefault: true;
+            isPrimaryKey: false;
+            isAutoincrement: false;
+            hasRuntimeDefault: false;
+            enumValues: undefined;
+            baseColumn: never;
+            identity: undefined;
+            generated: undefined;
+        }, {}, {}>;
+        updatedAt: import("drizzle-orm/pg-core").PgColumn<{
+            name: "updated_at";
+            tableName: "sessions";
+            dataType: "date";
+            columnType: "PgTimestamp";
+            data: Date;
+            driverParam: string;
+            notNull: true;
+            hasDefault: true;
+            isPrimaryKey: false;
+            isAutoincrement: false;
+            hasRuntimeDefault: false;
+            enumValues: undefined;
+            baseColumn: never;
+            identity: undefined;
+            generated: undefined;
+        }, {}, {}>;
+    };
+    dialect: "pg";
+}>;
+export type Session = typeof sessions.$inferSelect;
+export type NewSession = typeof sessions.$inferInsert;

package/dist/schema/sessions.js

@@ -0,0 +1,27 @@
+// SPDX-License-Identifier: Apache-2.0
+import { index, pgTable, text, timestamp, unique, uuid } from 'drizzle-orm/pg-core';
+import { customers } from './customers.js';
+/**
+ * sessions — named, per-customer cookie/header bundles for reuse
+ * across /scrape, /crawl, and /batch calls (Phase G).
+ *
+ * Cookies + headers are AES-256-GCM ciphertext, stored as a versioned
+ * JSON envelope by the helpers in @bluefields/control/sessions.ts.
+ * Unique on (customer_id, name) so save/use lookups are O(1).
+ */
+export const sessions = pgTable('sessions', {
+    id: uuid('id').primaryKey().defaultRandom(),
+    customerId: uuid('customer_id')
+        .notNull()
+        .references(() => customers.id, { onDelete: 'cascade' }),
+    name: text('name').notNull(),
+    cookiesCiphertext: text('cookies_ciphertext').notNull(),
+    headersCiphertext: text('headers_ciphertext'),
+    expiresAt: timestamp('expires_at', { withTimezone: true }),
+    createdAt: timestamp('created_at', { withTimezone: true }).notNull().defaultNow(),
+    updatedAt: timestamp('updated_at', { withTimezone: true }).notNull().defaultNow(),
+}, (table) => [
+    unique('sessions_customer_name_key').on(table.customerId, table.name),
+    index('idx_sessions_customer_name').on(table.customerId, table.name),
+]);
+//# sourceMappingURL=sessions.js.map

package/dist/schema/sessions.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"sessions.js","sourceRoot":"","sources":["../../src/schema/sessions.ts"],"names":[],"mappings":"AAAA,sCAAsC;AACtC,OAAO,EAAE,KAAK,EAAE,OAAO,EAAE,IAAI,EAAE,SAAS,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,qBAAqB,CAAC;AAEpF,OAAO,EAAE,SAAS,EAAE,MAAM,gBAAgB,CAAC;AAE3C;;;;;;;GAOG;AACH,MAAM,CAAC,MAAM,QAAQ,GAAG,OAAO,CAC7B,UAAU,EACV;IACE,EAAE,EAAE,IAAI,CAAC,IAAI,CAAC,CAAC,UAAU,EAAE,CAAC,aAAa,EAAE;IAC3C,UAAU,EAAE,IAAI,CAAC,aAAa,CAAC;SAC5B,OAAO,EAAE;SACT,UAAU,CAAC,GAAG,EAAE,CAAC,SAAS,CAAC,EAAE,EAAE,EAAE,QAAQ,EAAE,SAAS,EAAE,CAAC;IAC1D,IAAI,EAAE,IAAI,CAAC,MAAM,CAAC,CAAC,OAAO,EAAE;IAC5B,iBAAiB,EAAE,IAAI,CAAC,oBAAoB,CAAC,CAAC,OAAO,EAAE;IACvD,iBAAiB,EAAE,IAAI,CAAC,oBAAoB,CAAC;IAC7C,SAAS,EAAE,SAAS,CAAC,YAAY,EAAE,EAAE,YAAY,EAAE,IAAI,EAAE,CAAC;IAC1D,SAAS,EAAE,SAAS,CAAC,YAAY,EAAE,EAAE,YAAY,EAAE,IAAI,EAAE,CAAC,CAAC,OAAO,EAAE,CAAC,UAAU,EAAE;IACjF,SAAS,EAAE,SAAS,CAAC,YAAY,EAAE,EAAE,YAAY,EAAE,IAAI,EAAE,CAAC,CAAC,OAAO,EAAE,CAAC,UAAU,EAAE;CAClF,EACD,CAAC,KAAK,EAAE,EAAE,CAAC;IACT,MAAM,CAAC,4BAA4B,CAAC,CAAC,EAAE,CAAC,KAAK,CAAC,UAAU,EAAE,KAAK,CAAC,IAAI,CAAC;IACrE,KAAK,CAAC,4BAA4B,CAAC,CAAC,EAAE,CAAC,KAAK,CAAC,UAAU,EAAE,KAAK,CAAC,IAAI,CAAC;CACrE,CACF,CAAC"}

package/dist/schema/snapshots.d.ts

@@ -0,0 +1,382 @@
+/**
+ * Snapshots — the immutable provenance table.
+ *
+ * SPEC DEVIATION: The data-model spec writes `id UUID PRIMARY KEY` with
+ * `PARTITION BY RANGE (fetched_at)`. Postgres requires the partition column
+ * to appear in every unique/primary-key constraint on a partitioned table,
+ * so a sole-`id` PK is invalid SQL. We use composite primary key
+ * `(id, fetched_at)` to make the partition design actually work, matching
+ * the pattern already used in `cost_records` and `audit_log`. The chunk's
+ * acceptance criterion "migration runs cleanly" forces this fix.
+ *
+ * Consequence: FKs that reference `snapshots(id)` from non-partitioned
+ * tables (e.g. `extractions.snapshot_id`) cannot be expressed at the DB
+ * level — referential integrity is enforced by application code only.
+ */
+export declare const snapshots: import("drizzle-orm/pg-core").PgTableWithColumns<{
+    name: "snapshots";
+    schema: undefined;
+    columns: {
+        id: import("drizzle-orm/pg-core").PgColumn<{
+            name: "id";
+            tableName: "snapshots";
+            dataType: "string";
+            columnType: "PgUUID";
+            data: string;
+            driverParam: string;
+            notNull: true;
+            hasDefault: true;
+            isPrimaryKey: false;
+            isAutoincrement: false;
+            hasRuntimeDefault: false;
+            enumValues: undefined;
+            baseColumn: never;
+            identity: undefined;
+            generated: undefined;
+        }, {}, {}>;
+        urlCanonical: import("drizzle-orm/pg-core").PgColumn<{
+            name: "url_canonical";
+            tableName: "snapshots";
+            dataType: "string";
+            columnType: "PgText";
+            data: string;
+            driverParam: string;
+            notNull: true;
+            hasDefault: false;
+            isPrimaryKey: false;
+            isAutoincrement: false;
+            hasRuntimeDefault: false;
+            enumValues: [string, ...string[]];
+            baseColumn: never;
+            identity: undefined;
+            generated: undefined;
+        }, {}, {}>;
+        urlSurt: import("drizzle-orm/pg-core").PgColumn<{
+            name: "url_surt";
+            tableName: "snapshots";
+            dataType: "string";
+            columnType: "PgText";
+            data: string;
+            driverParam: string;
+            notNull: true;
+            hasDefault: false;
+            isPrimaryKey: false;
+            isAutoincrement: false;
+            hasRuntimeDefault: false;
+            enumValues: [string, ...string[]];
+            baseColumn: never;
+            identity: undefined;
+            generated: undefined;
+        }, {}, {}>;
+        canonVersion: import("drizzle-orm/pg-core").PgColumn<{
+            name: "canon_version";
+            tableName: "snapshots";
+            dataType: "number";
+            columnType: "PgSmallInt";
+            data: number;
+            driverParam: string | number;
+            notNull: true;
+            hasDefault: false;
+            isPrimaryKey: false;
+            isAutoincrement: false;
+            hasRuntimeDefault: false;
+            enumValues: undefined;
+            baseColumn: never;
+            identity: undefined;
+            generated: undefined;
+        }, {}, {}>;
+        contentHash: import("drizzle-orm/pg-core").PgColumn<{
+            name: "content_hash";
+            tableName: "snapshots";
+            dataType: "string";
+            columnType: "PgText";
+            data: string;
+            driverParam: string;
+            notNull: true;
+            hasDefault: false;
+            isPrimaryKey: false;
+            isAutoincrement: false;
+            hasRuntimeDefault: false;
+            enumValues: [string, ...string[]];
+            baseColumn: never;
+            identity: undefined;
+            generated: undefined;
+        }, {}, {}>;
+        fetchedAt: import("drizzle-orm/pg-core").PgColumn<{
+            name: "fetched_at";
+            tableName: "snapshots";
+            dataType: "date";
+            columnType: "PgTimestamp";
+            data: Date;
+            driverParam: string;
+            notNull: true;
+            hasDefault: false;
+            isPrimaryKey: false;
+            isAutoincrement: false;
+            hasRuntimeDefault: false;
+            enumValues: undefined;
+            baseColumn: never;
+            identity: undefined;
+            generated: undefined;
+        }, {}, {}>;
+        fetcherId: import("drizzle-orm/pg-core").PgColumn<{
+            name: "fetcher_id";
+            tableName: "snapshots";
+            dataType: "string";
+            columnType: "PgText";
+            data: string;
+            driverParam: string;
+            notNull: true;
+            hasDefault: false;
+            isPrimaryKey: false;
+            isAutoincrement: false;
+            hasRuntimeDefault: false;
+            enumValues: [string, ...string[]];
+            baseColumn: never;
+            identity: undefined;
+            generated: undefined;
+        }, {}, {}>;
+        proxyEgressIp: import("drizzle-orm/pg-core").PgColumn<{
+            name: "proxy_egress_ip";
+            tableName: "snapshots";
+            dataType: "string";
+            columnType: "PgInet";
+            data: string;
+            driverParam: string;
+            notNull: true;
+            hasDefault: false;
+            isPrimaryKey: false;
+            isAutoincrement: false;
+            hasRuntimeDefault: false;
+            enumValues: undefined;
+            baseColumn: never;
+            identity: undefined;
+            generated: undefined;
+        }, {}, {}>;
+        userAgent: import("drizzle-orm/pg-core").PgColumn<{
+            name: "user_agent";
+            tableName: "snapshots";
+            dataType: "string";
+            columnType: "PgText";
+            data: string;
+            driverParam: string;
+            notNull: true;
+            hasDefault: false;
+            isPrimaryKey: false;
+            isAutoincrement: false;
+            hasRuntimeDefault: false;
+            enumValues: [string, ...string[]];
+            baseColumn: never;
+            identity: undefined;
+            generated: undefined;
+        }, {}, {}>;
+        responseStatus: import("drizzle-orm/pg-core").PgColumn<{
+            name: "response_status";
+            tableName: "snapshots";
+            dataType: "number";
+            columnType: "PgInteger";
+            data: number;
+            driverParam: string | number;
+            notNull: true;
+            hasDefault: false;
+            isPrimaryKey: false;
+            isAutoincrement: false;
+            hasRuntimeDefault: false;
+            enumValues: undefined;
+            baseColumn: never;
+            identity: undefined;
+            generated: undefined;
+        }, {}, {}>;
+        responseHeaders: import("drizzle-orm/pg-core").PgColumn<{
+            name: "response_headers";
+            tableName: "snapshots";
+            dataType: "json";
+            columnType: "PgJsonb";
+            data: unknown;
+            driverParam: unknown;
+            notNull: true;
+            hasDefault: false;
+            isPrimaryKey: false;
+            isAutoincrement: false;
+            hasRuntimeDefault: false;
+            enumValues: undefined;
+            baseColumn: never;
+            identity: undefined;
+            generated: undefined;
+        }, {}, {}>;
+        rawHtmlR2Key: import("drizzle-orm/pg-core").PgColumn<{
+            name: "raw_html_r2_key";
+            tableName: "snapshots";
+            dataType: "string";
+            columnType: "PgText";
+            data: string;
+            driverParam: string;
+            notNull: true;
+            hasDefault: false;
+            isPrimaryKey: false;
+            isAutoincrement: false;
+            hasRuntimeDefault: false;
+            enumValues: [string, ...string[]];
+            baseColumn: never;
+            identity: undefined;
+            generated: undefined;
+        }, {}, {}>;
+        rawHtmlSizeBytes: import("drizzle-orm/pg-core").PgColumn<{
+            name: "raw_html_size_bytes";
+            tableName: "snapshots";
+            dataType: "number";
+            columnType: "PgBigInt53";
+            data: number;
+            driverParam: string | number;
+            notNull: true;
+            hasDefault: false;
+            isPrimaryKey: false;
+            isAutoincrement: false;
+            hasRuntimeDefault: false;
+            enumValues: undefined;
+            baseColumn: never;
+            identity: undefined;
+            generated: undefined;
+        }, {}, {}>;
+        fetchDurationMs: import("drizzle-orm/pg-core").PgColumn<{
+            name: "fetch_duration_ms";
+            tableName: "snapshots";
+            dataType: "number";
+            columnType: "PgInteger";
+            data: number;
+            driverParam: string | number;
+            notNull: true;
+            hasDefault: false;
+            isPrimaryKey: false;
+            isAutoincrement: false;
+            hasRuntimeDefault: false;
+            enumValues: undefined;
+            baseColumn: never;
+            identity: undefined;
+            generated: undefined;
+        }, {}, {}>;
+        attestationSignature: import("drizzle-orm/pg-core").PgColumn<{
+            name: "attestation_signature";
+            tableName: "snapshots";
+            dataType: "string";
+            columnType: "PgText";
+            data: string;
+            driverParam: string;
+            notNull: true;
+            hasDefault: false;
+            isPrimaryKey: false;
+            isAutoincrement: false;
+            hasRuntimeDefault: false;
+            enumValues: [string, ...string[]];
+            baseColumn: never;
+            identity: undefined;
+            generated: undefined;
+        }, {}, {}>;
+        attestationKeyId: import("drizzle-orm/pg-core").PgColumn<{
+            name: "attestation_key_id";
+            tableName: "snapshots";
+            dataType: "string";
+            columnType: "PgText";
+            data: string;
+            driverParam: string;
+            notNull: true;
+            hasDefault: false;
+            isPrimaryKey: false;
+            isAutoincrement: false;
+            hasRuntimeDefault: false;
+            enumValues: [string, ...string[]];
+            baseColumn: never;
+            identity: undefined;
+            generated: undefined;
+        }, {}, {}>;
+        attestationAlgorithm: import("drizzle-orm/pg-core").PgColumn<{
+            name: "attestation_algorithm";
+            tableName: "snapshots";
+            dataType: "string";
+            columnType: "PgText";
+            data: string;
+            driverParam: string;
+            notNull: true;
+            hasDefault: true;
+            isPrimaryKey: false;
+            isAutoincrement: false;
+            hasRuntimeDefault: false;
+            enumValues: [string, ...string[]];
+            baseColumn: never;
+            identity: undefined;
+            generated: undefined;
+        }, {}, {}>;
+        manifestVersion: import("drizzle-orm/pg-core").PgColumn<{
+            name: "manifest_version";
+            tableName: "snapshots";
+            dataType: "number";
+            columnType: "PgSmallInt";
+            data: number;
+            driverParam: string | number;
+            notNull: true;
+            hasDefault: true;
+            isPrimaryKey: false;
+            isAutoincrement: false;
+            hasRuntimeDefault: false;
+            enumValues: undefined;
+            baseColumn: never;
+            identity: undefined;
+            generated: undefined;
+        }, {}, {}>;
+        encryptionKeyId: import("drizzle-orm/pg-core").PgColumn<{
+            name: "encryption_key_id";
+            tableName: "snapshots";
+            dataType: "string";
+            columnType: "PgText";
+            data: string;
+            driverParam: string;
+            notNull: false;
+            hasDefault: false;
+            isPrimaryKey: false;
+            isAutoincrement: false;
+            hasRuntimeDefault: false;
+            enumValues: [string, ...string[]];
+            baseColumn: never;
+            identity: undefined;
+            generated: undefined;
+        }, {}, {}>;
+        costMicroUsd: import("drizzle-orm/pg-core").PgColumn<{
+            name: "cost_micro_usd";
+            tableName: "snapshots";
+            dataType: "number";
+            columnType: "PgInteger";
+            data: number;
+            driverParam: string | number;
+            notNull: true;
+            hasDefault: false;
+            isPrimaryKey: false;
+            isAutoincrement: false;
+            hasRuntimeDefault: false;
+            enumValues: undefined;
+            baseColumn: never;
+            identity: undefined;
+            generated: undefined;
+        }, {}, {}>;
+        createdAt: import("drizzle-orm/pg-core").PgColumn<{
+            name: "created_at";
+            tableName: "snapshots";
+            dataType: "date";
+            columnType: "PgTimestamp";
+            data: Date;
+            driverParam: string;
+            notNull: true;
+            hasDefault: true;
+            isPrimaryKey: false;
+            isAutoincrement: false;
+            hasRuntimeDefault: false;
+            enumValues: undefined;
+            baseColumn: never;
+            identity: undefined;
+            generated: undefined;
+        }, {}, {}>;
+    };
+    dialect: "pg";
+}>;
+export type Snapshot = typeof snapshots.$inferSelect;
+export type NewSnapshot = typeof snapshots.$inferInsert;
+export declare const SNAPSHOTS_PARTITION_COLUMN = "fetched_at";

package/dist/schema/snapshots.js

@@ -0,0 +1,60 @@
+import { bigint, index, inet, integer, jsonb, pgTable, primaryKey, smallint, text, timestamp, uuid, } from 'drizzle-orm/pg-core';
+/**
+ * Snapshots — the immutable provenance table.
+ *
+ * SPEC DEVIATION: The data-model spec writes `id UUID PRIMARY KEY` with
+ * `PARTITION BY RANGE (fetched_at)`. Postgres requires the partition column
+ * to appear in every unique/primary-key constraint on a partitioned table,
+ * so a sole-`id` PK is invalid SQL. We use composite primary key
+ * `(id, fetched_at)` to make the partition design actually work, matching
+ * the pattern already used in `cost_records` and `audit_log`. The chunk's
+ * acceptance criterion "migration runs cleanly" forces this fix.
+ *
+ * Consequence: FKs that reference `snapshots(id)` from non-partitioned
+ * tables (e.g. `extractions.snapshot_id`) cannot be expressed at the DB
+ * level — referential integrity is enforced by application code only.
+ */
+export const snapshots = pgTable('snapshots', {
+    id: uuid('id').notNull().defaultRandom(),
+    urlCanonical: text('url_canonical').notNull(),
+    // SURT form for host-level clustering and ops (Review #2)
+    urlSurt: text('url_surt').notNull(),
+    // which rules produced url_canonical (Review #2)
+    canonVersion: smallint('canon_version').notNull(),
+    // SHA-256 hex of the raw response body
+    contentHash: text('content_hash').notNull(),
+    fetchedAt: timestamp('fetched_at', { withTimezone: true }).notNull(),
+    // which worker
+    fetcherId: text('fetcher_id').notNull(),
+    // the actual IP the fetch came from
+    proxyEgressIp: inet('proxy_egress_ip').notNull(),
+    userAgent: text('user_agent').notNull(),
+    responseStatus: integer('response_status').notNull(),
+    responseHeaders: jsonb('response_headers').notNull(),
+    // content-addressed R2 key: 'content/<first2-of-sha>/<sha256>' (Review #2)
+    rawHtmlR2Key: text('raw_html_r2_key').notNull(),
+    rawHtmlSizeBytes: bigint('raw_html_size_bytes', { mode: 'number' }).notNull(),
+    fetchDurationMs: integer('fetch_duration_ms').notNull(),
+    attestationSignature: text('attestation_signature').notNull(),
+    attestationKeyId: text('attestation_key_id').notNull(),
+    // IANA JOSE algorithm name (RFC 8037); supports future hybrid PQC (Review #3)
+    attestationAlgorithm: text('attestation_algorithm').notNull().default('EdDSA'),
+    // which manifest schema produced the signed bytes (Review #3)
+    manifestVersion: smallint('manifest_version').notNull().default(1),
+    // NULL = plaintext blob in R2; non-NULL = encrypted at rest (v2; Review #2)
+    encryptionKeyId: text('encryption_key_id'),
+    // track cost per fetch (Review #11)
+    costMicroUsd: integer('cost_micro_usd').notNull(),
+    createdAt: timestamp('created_at', { withTimezone: true }).notNull().defaultNow(),
+}, (table) => [
+    // Composite PK required for partitioned tables (see spec-deviation note above).
+    primaryKey({ columns: [table.id, table.fetchedAt] }),
+    index('idx_snap_url_time').on(table.urlCanonical, table.fetchedAt.desc()),
+    index('idx_snap_surt_time').on(table.urlSurt, table.fetchedAt.desc()),
+    index('idx_snap_hash').on(table.contentHash),
+]);
+// SQL-level metadata: partitioning + partman bootstrap. Drizzle ORM has no
+// native partitioning support; the partition clause and partman.create_parent
+// call live in the raw migration file (0001_initial.sql).
+export const SNAPSHOTS_PARTITION_COLUMN = 'fetched_at';
+//# sourceMappingURL=snapshots.js.map

package/dist/schema/snapshots.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"snapshots.js","sourceRoot":"","sources":["../../src/schema/snapshots.ts"],"names":[],"mappings":"AAEA,OAAO,EACL,MAAM,EACN,KAAK,EACL,IAAI,EACJ,OAAO,EACP,KAAK,EACL,OAAO,EACP,UAAU,EACV,QAAQ,EACR,IAAI,EACJ,SAAS,EACT,IAAI,GACL,MAAM,qBAAqB,CAAC;AAE7B;;;;;;;;;;;;;;GAcG;AACH,MAAM,CAAC,MAAM,SAAS,GAAG,OAAO,CAC9B,WAAW,EACX;IACE,EAAE,EAAE,IAAI,CAAC,IAAI,CAAC,CAAC,OAAO,EAAE,CAAC,aAAa,EAAE;IACxC,YAAY,EAAE,IAAI,CAAC,eAAe,CAAC,CAAC,OAAO,EAAE;IAC7C,0DAA0D;IAC1D,OAAO,EAAE,IAAI,CAAC,UAAU,CAAC,CAAC,OAAO,EAAE;IACnC,iDAAiD;IACjD,YAAY,EAAE,QAAQ,CAAC,eAAe,CAAC,CAAC,OAAO,EAAE;IACjD,uCAAuC;IACvC,WAAW,EAAE,IAAI,CAAC,cAAc,CAAC,CAAC,OAAO,EAAE;IAC3C,SAAS,EAAE,SAAS,CAAC,YAAY,EAAE,EAAE,YAAY,EAAE,IAAI,EAAE,CAAC,CAAC,OAAO,EAAE;IACpE,eAAe;IACf,SAAS,EAAE,IAAI,CAAC,YAAY,CAAC,CAAC,OAAO,EAAE;IACvC,oCAAoC;IACpC,aAAa,EAAE,IAAI,CAAC,iBAAiB,CAAC,CAAC,OAAO,EAAE;IAChD,SAAS,EAAE,IAAI,CAAC,YAAY,CAAC,CAAC,OAAO,EAAE;IACvC,cAAc,EAAE,OAAO,CAAC,iBAAiB,CAAC,CAAC,OAAO,EAAE;IACpD,eAAe,EAAE,KAAK,CAAC,kBAAkB,CAAC,CAAC,OAAO,EAAE;IACpD,2EAA2E;IAC3E,YAAY,EAAE,IAAI,CAAC,iBAAiB,CAAC,CAAC,OAAO,EAAE;IAC/C,gBAAgB,EAAE,MAAM,CAAC,qBAAqB,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE,CAAC,CAAC,OAAO,EAAE;IAC7E,eAAe,EAAE,OAAO,CAAC,mBAAmB,CAAC,CAAC,OAAO,EAAE;IACvD,oBAAoB,EAAE,IAAI,CAAC,uBAAuB,CAAC,CAAC,OAAO,EAAE;IAC7D,gBAAgB,EAAE,IAAI,CAAC,oBAAoB,CAAC,CAAC,OAAO,EAAE;IACtD,8EAA8E;IAC9E,oBAAoB,EAAE,IAAI,CAAC,uBAAuB,CAAC,CAAC,OAAO,EAAE,CAAC,OAAO,CAAC,OAAO,CAAC;IAC9E,8DAA8D;IAC9D,eAAe,EAAE,QAAQ,CAAC,kBAAkB,CAAC,CAAC,OAAO,EAAE,CAAC,OAAO,CAAC,CAAC,CAAC;IAClE,4EAA4E;IAC5E,eAAe,EAAE,IAAI,CAAC,mBAAmB,CAAC;IAC1C,oCAAoC;IACpC,YAAY,EAAE,OAAO,CAAC,gBAAgB,CAAC,CAAC,OAAO,EAAE;IACjD,SAAS,EAAE,SAAS,CAAC,YAAY,EAAE,EAAE,YAAY,EAAE,IAAI,EAAE,CAAC,CAAC,OAAO,EAAE,CAAC,UAAU,EAAE;CAClF,EACD,CAAC,KAAK,EAAE,EAAE,CAAC;IACT,gFAAgF;IAChF,UAAU,CAAC,EAAE,OAAO,EAAE,CAAC,KAAK,CAAC,EAAE,EAAE,KAAK,CAAC,SAAS,CAAC,EAAE,CAAC;IACpD,KAAK,CAAC,mBAAmB,CAAC,CAAC,EAAE,CAAC,KAAK,CAAC,YAAY,EAAE,KAAK,CAAC,SAAS,CAAC,IAAI,EAAE,CAAC;IACzE,KAAK,CAAC,oBAAoB,CAAC,CAAC,EAAE,CAAC,KAAK,CAAC,OAAO,EAAE,KAAK,CAAC,SAAS,CAAC,IAAI,EAAE,CAAC;IACrE,KAAK,CAAC,eAAe,CAAC,CAAC,EAAE,CAAC,KAAK,CAAC,WAAW,CAAC;CAC7C,CACF,CAAC;AAKF,2EAA2E;AAC3E,8EAA8E;AAC9E,0DAA0D;AAC1D,MAAM,CAAC,MAAM,0BAA0B,GAAG,YAAY,CAAC"}