@gscdump/engine 0.4.0 → 0.6.2

package/dist/schema.mjs

@@ -1,139 +1,2 @@
-import { date, doublePrecision, getTableConfig, integer, pgTable, varchar } from "drizzle-orm/pg-core";
-function metricCols() {
-	return {
-		clicks: integer("clicks").notNull(),
-		impressions: integer("impressions").notNull(),
-		sum_position: doublePrecision("sum_position").notNull()
-	};
-}
-const dateCol = () => date("date").notNull();
-const pages = pgTable("pages", {
-	url: varchar("url").notNull(),
-	date: dateCol(),
-	...metricCols()
-});
-const keywords = pgTable("keywords", {
-	query: varchar("query").notNull(),
-	query_canonical: varchar("query_canonical"),
-	date: dateCol(),
-	...metricCols()
-});
-const countries = pgTable("countries", {
-	country: varchar("country").notNull(),
-	date: dateCol(),
-	...metricCols()
-});
-const devices = pgTable("devices", {
-	device: varchar("device").notNull(),
-	date: dateCol(),
-	...metricCols()
-});
-const page_keywords = pgTable("page_keywords", {
-	url: varchar("url").notNull(),
-	query: varchar("query").notNull(),
-	query_canonical: varchar("query_canonical"),
-	date: dateCol(),
-	...metricCols()
-});
-const search_appearance = pgTable("search_appearance", {
-	searchAppearance: varchar("searchAppearance").notNull(),
-	date: dateCol(),
-	...metricCols()
-});
-const drizzleSchema = {
-	pages,
-	keywords,
-	countries,
-	devices,
-	page_keywords,
-	search_appearance
-};
-const TABLE_METADATA = {
-	pages: {
-		sortKey: ["date", "url"],
-		version: 1
-	},
-	keywords: {
-		sortKey: ["date", "query"],
-		version: 2
-	},
-	countries: {
-		sortKey: ["date", "country"],
-		version: 1
-	},
-	devices: {
-		sortKey: ["date", "device"],
-		version: 1
-	},
-	page_keywords: {
-		sortKey: [
-			"date",
-			"url",
-			"query"
-		],
-		version: 2
-	},
-	search_appearance: {
-		sortKey: ["date", "searchAppearance"],
-		version: 1
-	}
-};
-function pgSqlTypeToColumnType(sqlType) {
-	const t = sqlType.toLowerCase();
-	if (t.startsWith("varchar") || t === "text" || t.startsWith("char")) return "VARCHAR";
-	if (t === "date" || t.startsWith("timestamp")) return "DATE";
-	if (t.startsWith("double") || t === "real" || t.startsWith("numeric") || t.startsWith("decimal")) return "DOUBLE";
-	if (t === "bigint" || t === "int8") return "BIGINT";
-	if (t === "integer" || t === "int" || t === "int4" || t === "smallint" || t === "int2") return "INTEGER";
-	throw new Error(`unmapped pg type '${sqlType}' — extend pgSqlTypeToColumnType in @gscdump/engine/schema`);
-}
-function tableSchemaFrom(tableName) {
-	const columns = getTableConfig(drizzleSchema[tableName]).columns.map((col) => ({
-		name: col.name,
-		type: pgSqlTypeToColumnType(col.getSQLType()),
-		nullable: !col.notNull
-	}));
-	const meta = TABLE_METADATA[tableName];
-	return {
-		name: tableName,
-		columns,
-		sortKey: meta.sortKey,
-		version: meta.version
-	};
-}
-const METRIC_TABLES = [
-	"pages",
-	"keywords",
-	"countries",
-	"devices",
-	"page_keywords",
-	"search_appearance"
-];
-const SCHEMAS = Object.fromEntries(METRIC_TABLES.map((t) => [t, tableSchemaFrom(t)]));
-function currentSchemaVersion(table) {
-	return SCHEMAS[table].version;
-}
-function schemaFor(table) {
-	return SCHEMAS[table];
-}
-function allTables() {
-	return METRIC_TABLES;
-}
-function inferTable(dimensions) {
-	const dims = new Set(dimensions);
-	const hasPage = dims.has("page");
-	const hasQuery = dims.has("query");
-	if (hasPage && hasQuery) return "page_keywords";
-	if (hasQuery) return "keywords";
-	if (hasPage) return "pages";
-	if (dims.has("country")) return "countries";
-	if (dims.has("device")) return "devices";
-	if (dims.has("searchAppearance")) return "search_appearance";
-	return "keywords";
-}
-function dimensionToColumn(dim, _table) {
-	if (dim === "page") return "url";
-	if (dim === "queryCanonical") return "query_canonical";
-	return dim;
-}
+import { a as inferTable, c as countries, d as keywords, f as page_keywords, i as dimensionToColumn, l as devices, m as search_appearance, n as allTables, o as schemaFor, p as pages, r as currentSchemaVersion, s as TABLE_METADATA, t as SCHEMAS, u as drizzleSchema } from "./_chunks/schema.mjs";
 export { SCHEMAS, TABLE_METADATA, allTables, countries, currentSchemaVersion, devices, dimensionToColumn, drizzleSchema, inferTable, keywords, page_keywords, pages, schemaFor, search_appearance };

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@gscdump/engine",
   "type": "module",
-  "version": "0.4.0",
+  "version": "0.6.2",
   "description": "Append-only Parquet/DuckDB storage engine + planner + adapters for the gscdump pipeline. Node + edge runtimes; opt-in heavy peers.",
   "author": {
     "name": "Harlan Wilton",
@@ -121,7 +121,7 @@
   "peerDependencies": {
     "@duckdb/duckdb-wasm": "^1.32.0",
     "better-sqlite3": "^12.9.0",
-    "hyparquet": "^1.25.1",
+    "hyparquet": "^1.25.6",
     "hyparquet-writer": "^0.13.0",
     "wa-sqlite": "^1.0.0"
   },
@@ -145,16 +145,16 @@
   "dependencies": {
     "drizzle-orm": "^0.45.2",
     "proper-lockfile": "^4.1.2",
-    "gscdump": "0.4.0"
+    "gscdump": "0.6.2"
   },
   "devDependencies": {
     "@duckdb/duckdb-wasm": "^1.32.0",
     "@types/proper-lockfile": "^4.1.4",
     "aws4fetch": "^1.0.20",
     "better-sqlite3": "^12.9.0",
-    "hyparquet": "^1.25.1",
+    "hyparquet": "^1.25.6",
     "hyparquet-writer": "^0.13.0",
-    "tsx": "^4.19.2",
+    "tsx": "^4.21.0",
     "vitest": "^4.1.5"
   },
   "scripts": {