tplm-lang 0.3.2 → 0.3.4

package/README.md

@@ -84,6 +84,16 @@ const tpl = fromBigQueryTable({
   credentialsPath: "./credentials.json",
 });
 
+// Pre-built connection (any Malloy Connection instance)
+import { fromConnection } from "tplm-lang";
+import { BigQueryConnection } from "@malloydata/db-bigquery";
+const conn = new BigQueryConnection("bigquery", {}, { credentials });
+const tpl = fromConnection({
+  connection: conn,
+  table: "project.dataset.sales",
+  dialect: "bigquery",
+});
+
 // Then query the same way
 const { html } = await tpl.query("TABLE ROWS region * revenue.sum;");
 ```
@@ -184,18 +194,29 @@ COLS education * (revenue.sum ACROSS COLS | revenue.mean)
 ### Easy Connectors (Recommended)
 
 ```typescript
-import { fromCSV, fromDuckDBTable, fromBigQueryTable } from "tplm-lang";
+import {
+  fromCSV, fromDuckDBTable, fromBigQueryTable, fromConnection
+} from "tplm-lang";
 
 // CSV or Parquet files
 const tpl = fromCSV("data/sales.csv");
 const tpl = fromDuckDBTable("data/sales.parquet");
 
-// BigQuery
+// BigQuery (file-based credentials)
 const tpl = fromBigQueryTable({
   table: "project.dataset.sales",
   credentialsPath: "./credentials.json",
 });
 
+// Pre-built connection (inject any Malloy Connection instance)
+import { BigQueryConnection } from "@malloydata/db-bigquery";
+const conn = new BigQueryConnection("bigquery", {}, { credentials });
+const tpl = fromConnection({
+  connection: conn,
+  table: "project.dataset.sales",
+  dialect: "bigquery", // or "duckdb"
+});
+
 // Add computed dimensions with .extend()
 const tplWithDims = tpl.extend(`
   dimension: region is pick 'North' when region_code = 1 else 'South'

package/dist/compiler/grid-spec-builder.js

@@ -16,10 +16,17 @@ let currentOrderingProvider;
 // DATE NORMALIZATION
 // ---
 /**
- * Normalize a raw data value: converts Date objects to formatted strings
+ * Normalize a raw data value: converts date-like objects to formatted strings
  * so they display cleanly as headers and work correctly with Set-based
  * deduplication and value-based cell lookup.
  *
+ * Handles two cases:
+ * 1. JS Date objects (from DuckDB connector)
+ * 2. BigQuery SDK wrapper objects (BigQueryDate, BigQueryTimestamp,
+ *    BigQueryDatetime) which are plain objects with a `.value` string
+ *    property and no custom toString() — producing "[object Object]"
+ *    if left unconverted.
+ *
  * Dates at midnight UTC (date-only) become "YYYY-MM-DD".
  * Timestamps with a time component become "YYYY-MM-DD HH:MM:SS".
  */
@@ -30,19 +37,38 @@ function normalizeDimValue(value) {
             ? iso.slice(0, 10)
             : iso.slice(0, 19).replace("T", " ");
     }
+    // Handle BigQuery SDK date types (BigQueryDate, BigQueryTimestamp, etc.)
+    // These are plain objects with a .value string property.
+    if (value !== null &&
+        typeof value === "object" &&
+        !Array.isArray(value) &&
+        "value" in value &&
+        typeof value.value === "string") {
+        const str = value.value;
+        // ISO timestamp at midnight → date-only
+        if (/^\d{4}-\d{2}-\d{2}T00:00:00(\.0+)?Z?$/.test(str)) {
+            return str.slice(0, 10);
+        }
+        // ISO timestamp with time → "YYYY-MM-DD HH:MM:SS"
+        if (/^\d{4}-\d{2}-\d{2}T/.test(str)) {
+            return str.slice(0, 19).replace("T", " ");
+        }
+        // Already formatted (BigQueryDate "2025-08-01", BigQueryDatetime "2025-08-01 00:00:00")
+        return str;
+    }
     return value;
 }
-/**
- * Recursively normalize all Date values in query result rows.
- * This ensures Date objects are converted to formatted strings before
- * any downstream processing (header extraction, cell indexing, etc.).
- */
+/** Check if an object is a date-like wrapper (BigQuery SDK types) vs a nested row */
+function isDateLikeWrapper(obj) {
+    return "value" in obj && typeof obj.value === "string";
+}
 function normalizeResultRow(row) {
     const normalized = {};
     for (const key of Object.keys(row)) {
         const val = row[key];
         if (Array.isArray(val)) {
-            normalized[key] = val.map((item) => item && typeof item === "object" && !Array.isArray(item) && !(item instanceof Date)
+            normalized[key] = val.map((item) => item && typeof item === "object" && !Array.isArray(item)
+                && !(item instanceof Date) && !isDateLikeWrapper(item)
                 ? normalizeResultRow(item)
                 : normalizeDimValue(item));
         }

package/dist/executor/index.d.ts

@@ -50,6 +50,12 @@ export declare function createLocalConnection(): Promise<Connection>;
  */
 export declare function getConnection(): Promise<Connection>;
 export declare function getConnectionType(): ConnectionType | null;
+/**
+ * Inject a pre-built Malloy connection directly.
+ * Use this when you already have a configured Connection instance
+ * (e.g., a BigQueryConnection with in-memory credentials).
+ */
+export declare function setConnection(conn: Connection, type: ConnectionType): void;
 export declare function listDatasets(): Promise<string[]>;
 export declare function listTables(dataset: string): Promise<{
     name: string;
@@ -80,4 +86,4 @@ export declare function executeMalloy(malloySource: string, options?: ExecuteOpt
  */
 export declare function executeSQL(sql: string): Promise<any[]>;
 export { Runtime };
-export type { DuckDBConnection, BigQueryConnection };
+export type { Connection, DuckDBConnection, BigQueryConnection };

package/dist/executor/index.js

@@ -134,6 +134,16 @@ export function getConnectionType() {
     }
     return currentConnectionType;
 }
+/**
+ * Inject a pre-built Malloy connection directly.
+ * Use this when you already have a configured Connection instance
+ * (e.g., a BigQueryConnection with in-memory credentials).
+ */
+export function setConnection(conn, type) {
+    connectionInstance = conn;
+    currentConnectionType = type;
+    pendingOptions = null;
+}
 // ---
 // SCHEMA EXPLORATION (BigQuery only)
 // ---