@gscdump/engine 0.11.1 → 0.11.3

package/dist/_chunks/dispatch.mjs

@@ -1,4 +1,6 @@
 var AnalyzerCapabilityError = class extends Error {
+	tool;
+	missing;
 	constructor(tool, missing) {
 		super(`analyzer "${tool}" requires capabilities [${missing.join(", ")}] not provided by source`);
 		this.tool = tool;

package/dist/entities.mjs

@@ -26,6 +26,88 @@ function hashUrl(url) {
 	}
 	return (hi >>> 0).toString(16).padStart(8, "0") + (lo >>> 0).toString(16).padStart(8, "0");
 }
+const INSPECTION_PARQUET_COLUMNS = [
+	{
+		name: "urlHash",
+		type: "VARCHAR",
+		nullable: false
+	},
+	{
+		name: "url",
+		type: "VARCHAR",
+		nullable: false
+	},
+	{
+		name: "inspectedAt",
+		type: "VARCHAR",
+		nullable: false
+	},
+	{
+		name: "indexStatus",
+		type: "VARCHAR",
+		nullable: true
+	},
+	{
+		name: "lastCrawlTime",
+		type: "VARCHAR",
+		nullable: true
+	},
+	{
+		name: "googleCanonical",
+		type: "VARCHAR",
+		nullable: true
+	},
+	{
+		name: "userCanonical",
+		type: "VARCHAR",
+		nullable: true
+	},
+	{
+		name: "coverageState",
+		type: "VARCHAR",
+		nullable: true
+	},
+	{
+		name: "robotsTxtState",
+		type: "VARCHAR",
+		nullable: true
+	},
+	{
+		name: "indexingState",
+		type: "VARCHAR",
+		nullable: true
+	},
+	{
+		name: "pageFetchState",
+		type: "VARCHAR",
+		nullable: true
+	},
+	{
+		name: "mobileUsabilityVerdict",
+		type: "VARCHAR",
+		nullable: true
+	},
+	{
+		name: "richResultsVerdict",
+		type: "VARCHAR",
+		nullable: true
+	},
+	{
+		name: "scheduleNextAt",
+		type: "BIGINT",
+		nullable: true
+	},
+	{
+		name: "scheduleConsecutiveUnchanged",
+		type: "INTEGER",
+		nullable: true
+	},
+	{
+		name: "schedulePolicyVersion",
+		type: "INTEGER",
+		nullable: true
+	}
+];
 function createInspectionStore(opts) {
 	const hash = opts.hash ?? hashUrl;
 	const ds = opts.dataSource;
@@ -117,88 +199,6 @@ function createInspectionStore(opts) {
 		}
 	};
 }
-const INSPECTION_PARQUET_COLUMNS = [
-	{
-		name: "urlHash",
-		type: "VARCHAR",
-		nullable: false
-	},
-	{
-		name: "url",
-		type: "VARCHAR",
-		nullable: false
-	},
-	{
-		name: "inspectedAt",
-		type: "VARCHAR",
-		nullable: false
-	},
-	{
-		name: "indexStatus",
-		type: "VARCHAR",
-		nullable: true
-	},
-	{
-		name: "lastCrawlTime",
-		type: "VARCHAR",
-		nullable: true
-	},
-	{
-		name: "googleCanonical",
-		type: "VARCHAR",
-		nullable: true
-	},
-	{
-		name: "userCanonical",
-		type: "VARCHAR",
-		nullable: true
-	},
-	{
-		name: "coverageState",
-		type: "VARCHAR",
-		nullable: true
-	},
-	{
-		name: "robotsTxtState",
-		type: "VARCHAR",
-		nullable: true
-	},
-	{
-		name: "indexingState",
-		type: "VARCHAR",
-		nullable: true
-	},
-	{
-		name: "pageFetchState",
-		type: "VARCHAR",
-		nullable: true
-	},
-	{
-		name: "mobileUsabilityVerdict",
-		type: "VARCHAR",
-		nullable: true
-	},
-	{
-		name: "richResultsVerdict",
-		type: "VARCHAR",
-		nullable: true
-	},
-	{
-		name: "scheduleNextAt",
-		type: "BIGINT",
-		nullable: true
-	},
-	{
-		name: "scheduleConsecutiveUnchanged",
-		type: "INTEGER",
-		nullable: true
-	},
-	{
-		name: "schedulePolicyVersion",
-		type: "INTEGER",
-		nullable: true
-	}
-];
 function sitemapIndexKey(ctx) {
 	return ctx.siteId ? `u_${ctx.userId}/${ctx.siteId}/entities/sitemaps/index.json` : `u_${ctx.userId}/entities/sitemaps/index.json`;
 }

package/dist/rollups.d.mts

@@ -1,4 +1,4 @@
-import { a as DataSource } from "./_chunks/storage.mjs";
+import { a as DataSource, c as FileSetRef } from "./_chunks/storage.mjs";
 import { t as ColumnDef } from "./_chunks/schema.mjs";
 import { TenantCtx } from "gscdump/contracts";
 import * as _$_gscdump_engine_contracts0 from "@gscdump/engine/contracts";
@@ -13,10 +13,7 @@ interface RollupCtx extends TenantCtx {
 interface RollupEngine {
   runSQL: (opts: {
     ctx: TenantCtx;
-    fileSets: Record<string, {
-      table: _$_gscdump_engine_contracts0.TableName;
-      partitions?: string[];
-    }>;
+    fileSets: Record<string, FileSetRef>;
     table?: _$_gscdump_engine_contracts0.TableName;
     sql: string;
     params?: unknown[];

package/dist/rollups.mjs

@@ -343,17 +343,17 @@ const indexingHealthRollup = {
 		if (!await dataSource.head?.(key)) return { days: [] };
 		const sql = `
       SELECT
-        substr(inspectedAt, 1, 10) AS date,
+        substr(CAST(inspectedAt AS VARCHAR), 1, 10) AS date,
         COUNT(*)::BIGINT AS total_urls,
-        SUM(CASE WHEN indexStatus = 'PASS' THEN 1 ELSE 0 END)::BIGINT AS indexed_count,
-        SUM(CASE WHEN pageFetchState = 'SOFT_404' THEN 1 ELSE 0 END)::BIGINT AS soft_404,
-        SUM(CASE WHEN pageFetchState = 'REDIRECT_ERROR' THEN 1 ELSE 0 END)::BIGINT AS redirect,
-        SUM(CASE WHEN pageFetchState = 'NOT_FOUND' THEN 1 ELSE 0 END)::BIGINT AS not_found,
-        SUM(CASE WHEN mobileUsabilityVerdict = 'PASS' THEN 1 ELSE 0 END)::BIGINT AS mobile_passes,
-        SUM(CASE WHEN richResultsVerdict = 'PASS' THEN 1 ELSE 0 END)::BIGINT AS rich_results_passes,
-        SUM(CASE WHEN userCanonical IS NOT NULL AND googleCanonical IS NOT NULL AND userCanonical <> googleCanonical THEN 1 ELSE 0 END)::BIGINT AS canonical_mismatches
+        SUM(CASE WHEN CAST(indexStatus AS VARCHAR) = 'PASS' THEN 1 ELSE 0 END)::BIGINT AS indexed_count,
+        SUM(CASE WHEN CAST(pageFetchState AS VARCHAR) = 'SOFT_404' THEN 1 ELSE 0 END)::BIGINT AS soft_404,
+        SUM(CASE WHEN CAST(pageFetchState AS VARCHAR) = 'REDIRECT_ERROR' THEN 1 ELSE 0 END)::BIGINT AS redirect,
+        SUM(CASE WHEN CAST(pageFetchState AS VARCHAR) = 'NOT_FOUND' THEN 1 ELSE 0 END)::BIGINT AS not_found,
+        SUM(CASE WHEN CAST(mobileUsabilityVerdict AS VARCHAR) = 'PASS' THEN 1 ELSE 0 END)::BIGINT AS mobile_passes,
+        SUM(CASE WHEN CAST(richResultsVerdict AS VARCHAR) = 'PASS' THEN 1 ELSE 0 END)::BIGINT AS rich_results_passes,
+        SUM(CASE WHEN userCanonical IS NOT NULL AND googleCanonical IS NOT NULL AND CAST(userCanonical AS VARCHAR) <> CAST(googleCanonical AS VARCHAR) THEN 1 ELSE 0 END)::BIGINT AS canonical_mismatches
       FROM read_parquet({{INSPECTIONS}}, union_by_name = true)
-      WHERE substr(inspectedAt, 1, 10) >= '${utcDateMinusDays(builtAt, 90)}'
+      WHERE substr(CAST(inspectedAt AS VARCHAR), 1, 10) >= '${utcDateMinusDays(builtAt, 90)}'
       GROUP BY 1
       ORDER BY 1
     `;

package/dist/source/index.mjs

@@ -2,6 +2,7 @@ import { h as resolveToSQL, n as pgResolverAdapter, s as assertDimensionsSupport
 import { i as getFilterDimensions } from "../_chunks/resolver.mjs";
 import { n as runAnalyzerFromSource } from "../_chunks/dispatch.mjs";
 var AttachedTableMissingError = class extends Error {
+	missing;
 	constructor(missing) {
 		super(`attached-table source: required table(s) not attached: ${missing.join(", ")}`);
 		this.missing = missing;

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@gscdump/engine",
   "type": "module",
-  "version": "0.11.1",
+  "version": "0.11.3",
   "description": "Append-only Parquet/DuckDB storage engine + planner + adapters for the gscdump pipeline. Node + edge runtimes; opt-in heavy peers.",
   "author": {
     "name": "Harlan Wilton",
@@ -152,8 +152,8 @@
   },
   "peerDependencies": {
     "@duckdb/duckdb-wasm": "^1.32.0",
-    "hyparquet": "^1.25.6",
-    "hyparquet-writer": "^0.14.0"
+    "hyparquet": "^1.25.8",
+    "hyparquet-writer": "^0.15.1"
   },
   "peerDependenciesMeta": {
     "@duckdb/duckdb-wasm": {
@@ -169,14 +169,14 @@
   "dependencies": {
     "drizzle-orm": "^0.45.2",
     "proper-lockfile": "^4.1.2",
-    "gscdump": "0.11.1"
+    "gscdump": "0.11.3"
   },
   "devDependencies": {
     "@duckdb/duckdb-wasm": "^1.32.0",
     "@types/proper-lockfile": "^4.1.4",
     "aws4fetch": "^1.0.20",
-    "hyparquet": "^1.25.6",
-    "hyparquet-writer": "^0.14.0",
+    "hyparquet": "^1.25.8",
+    "hyparquet-writer": "^0.15.1",
     "tsx": "^4.21.0",
     "vitest": "^4.1.5"
   },