@query-farm/vgi 0.23.0 → 0.24.1

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@query-farm/vgi",
-  "version": "0.23.0",
+  "version": "0.24.1",
   "description": "TypeScript framework for extending DuckDB with user-defined functions via the Vector Gateway Interface (VGI) over Arrow IPC.",
   "license": "SEE LICENSE IN LICENSE",
   "author": "Query Farm LLC (https://query.farm)",
@@ -69,15 +69,18 @@
   "files": [
     "dist",
     "src",
+    "scripts/postinstall.cjs",
     "!src/**/__tests__/**",
     "!src/**/*.test.ts",
     "!src/**/*.spec.ts"
   ],
+  "//imports": "`browser` maps to arrow-js DELIBERATELY, not flechette. Measured on the client surface (VgiClient + Arguments + batch helpers), a minified browser bundle is 84 KB gzip via arrow-js vs 116 KB via flechette — arrow-js tree-shakes to the subset the client touches, while flechette links as a unit. Flechette still wins for `worker-cf`, which exercises far more of the Arrow API and cannot take a peer dep. Keep `browser` AFTER `workerd`/`worker`: conditions match in key order, so a workerd build (which sets both) must hit flechette first.",
   "imports": {
     "#arrow-impl": {
       "flechette": "./src/arrow/impl-flechette/index.ts",
       "workerd": "./src/arrow/impl-flechette/index.ts",
       "worker": "./src/arrow/impl-flechette/index.ts",
+      "browser": "./src/arrow/impl-arrowjs/index.ts",
       "default": "./src/arrow/impl-arrowjs/index.ts"
     }
   },
@@ -87,12 +90,12 @@
   },
   "peerDependencies": {
     "@query-farm/apache-arrow": "^21.1.1",
-    "@query-farm/vgi-rpc": "^0.15.4"
+    "@query-farm/vgi-rpc": ">=0.18.1 <0.19.0"
   },
   "devDependencies": {
     "@duckdb/node-api": "^1.5.2-r.1",
     "@query-farm/apache-arrow": "^21.1.1",
-    "@query-farm/vgi-rpc": "^0.15.4",
+    "@query-farm/vgi-rpc": "^0.18.1",
     "@types/bun": "latest",
     "typescript": "^6.0.3"
   },

package/scripts/postinstall.cjs

@@ -0,0 +1,47 @@
+#!/usr/bin/env node
+// Copyright 2025, 2026 Query Farm LLC - https://query.farm
+// Postinstall: deduplicate @query-farm/apache-arrow across @query-farm/vgi-rpc
+// to avoid dual-package instanceof issues.
+//
+// This is a LOCAL DEVELOPMENT-only fixup for the vgi-typescript monorepo, where
+// @query-farm/vgi-rpc is a `file:` link with its own nested copy of apache-arrow.
+// It must NOT run when this package is installed as a dependency of someone
+// else's project — so it no-ops unless the sibling vgi-rpc-typescript checkout
+// is present next to this repo. (Published consumers get a single hoisted
+// apache-arrow from the registry, so no dedup is needed there.)
+const fs = require("fs");
+const path = require("path");
+
+const PKG = "@query-farm/apache-arrow";
+
+// Resolve everything relative to the repo root (this file lives in <repo>/scripts),
+// not the process cwd, so behavior doesn't depend on where the installer runs us.
+const repoRoot = path.resolve(__dirname, "..");
+
+// Dev-monorepo signal: the sibling vgi-rpc source checkout. When this package is
+// installed as a dependency, that path resolves inside the consumer's node_modules
+// and won't exist — so we bail and leave their tree untouched.
+const devSibling = path.resolve(repoRoot, "..", "vgi-rpc-typescript", "package.json");
+if (!fs.existsSync(devSibling)) {
+  process.exit(0);
+}
+
+// Deduplicate: replace vgi-rpc's nested copy with a symlink to ours.
+//   Without this, two separate DataType classes exist at runtime and
+//   instanceof checks fail across package boundaries.
+//   Bun uses per-file symlinks for file: deps, so module resolution follows
+//   the real path. We must target the REAL node_modules of vgi-rpc-typescript.
+const ourArrow = path.join(repoRoot, "node_modules", PKG);
+const rpcPkgLink = path.join(repoRoot, "node_modules", "@query-farm", "vgi-rpc", "package.json");
+try {
+  const realPkgPath = fs.realpathSync(rpcPkgLink);
+  const rpcRoot = path.dirname(realPkgPath);
+  const rpcArrow = path.join(rpcRoot, "node_modules", PKG);
+  const st = fs.lstatSync(rpcArrow);
+  if (!st.isSymbolicLink()) {
+    fs.rmSync(rpcArrow, { recursive: true });
+    fs.symlinkSync(ourArrow, rpcArrow);
+  }
+} catch (e) {
+  if (e.code !== "ENOENT") throw e;
+}

package/src/catalog/descriptors.ts

@@ -173,4 +173,22 @@ export interface CatalogDescriptor {
    * advertises a null `source_url`.
    */
   sourceUrl?: string;
+  /**
+   * Functions this catalog asks the client to *additionally* publish into its
+   * global (DuckDB `system.main`) function namespace, on top of the normal
+   * schema-qualified registration. Surfaced as IPC-serialized `FunctionInfo`
+   * records on `catalog_attach.global_functions`.
+   *
+   * Every entry must also be declared in exactly one of this catalog's
+   * schemas — the `FunctionInfo.schema_name` it carries is the bind-dispatch
+   * key, not a sentinel. Mirrors vgi-python's `Catalog.global_functions`.
+   */
+  globalFunctions?: VgiFunction[];
+  /**
+   * Prefix applied client-side to every `globalFunctions` entry's name when it
+   * is published globally (e.g. prefix `vgi_example` + `global_scalar` →
+   * `vgi_example_global_scalar`). The advertised `FunctionInfo.name` stays
+   * unprefixed. Surfaced as `catalog_attach.global_function_prefix`.
+   */
+  globalFunctionPrefix?: string;
 }

package/src/catalog/read-only.ts

@@ -17,6 +17,7 @@ import {
   type IndexConstraintType,
   type CopyFromFormatInfo,
   encodeAttachCatalogInfo,
+  encodeFunctionInfo,
 } from "./interface.js";
 import type { CatalogDescriptor, SchemaDescriptor, TableDescriptor, ViewDescriptor, MacroDescriptor, SettingDescriptor, SecretTypeDescriptor, ForeignKeyDef, DefaultValue } from "./descriptors.js";
 import { serializeColumnStatistics } from "../util/statistics.js";
@@ -119,6 +120,11 @@ export class ReadOnlyCatalogInterface extends CatalogInterface {
       // Advertise true so DuckDB routes catalog_table_column_statistics_get
       // for tables whose TableInfo.supports_column_statistics is also true.
       supports_column_statistics: true,
+      // Functions the client should ALSO publish into its global (system.main)
+      // namespace, under `global_function_prefix` + "_" + name. Empty unless
+      // the descriptor opts in via `globalFunctions`.
+      global_functions: this._globalFunctionInfos().map(encodeFunctionInfo),
+      global_function_prefix: this._descriptor.globalFunctionPrefix ?? "",
       resolved_data_version: null,
       resolved_implementation_version: null,
     };
@@ -374,7 +380,18 @@ export class ReadOnlyCatalogInterface extends CatalogInterface {
         if (t === "AGGREGATE_FUNCTION") return (f.kind as string) === "aggregate";
         return true;
       })
-      .map((f) => {
+      .map((f) => this._functionToInfo(f, name));
+  }
+
+  /**
+   * Build the wire `FunctionInfo` for one declared function.
+   *
+   * `schemaName` is the schema the function actually lives in — it is the
+   * bind-dispatch key, so a function advertised through more than one route
+   * (schema contents *and* `global_functions`) must carry the same value in
+   * both. Shared by `schemaContentsFunctions` and `_globalFunctionInfos`.
+   */
+  private _functionToInfo(f: VgiFunction, name: string): FunctionInfo {
         const meta = resolveMetadata(f);
         const argSchema = argumentSpecsToSchema(f.argumentSpecs);
         const argBytes = serializeSchema(argSchema);
@@ -474,7 +491,39 @@ export class ReadOnlyCatalogInterface extends CatalogInterface {
             secret_name: s.secret_name,
           })),
         };
-      });
+  }
+
+  /**
+   * Build `FunctionInfo` records for `CatalogDescriptor.globalFunctions`.
+   *
+   * Each entry's `schema_name` is the schema it actually lives in — the
+   * globally visible name is derived client-side from
+   * `global_function_prefix`, so `name` stays unprefixed here. An entry not
+   * declared in any schema is a descriptor bug, so it throws rather than
+   * advertising a function no bind can resolve. Mirrors vgi-python's
+   * `_global_function_infos`.
+   */
+  private _globalFunctionInfos(): FunctionInfo[] {
+    const globals = this._descriptor.globalFunctions ?? [];
+    if (globals.length === 0) return [];
+
+    const schemaOf = new Map<VgiFunction, string>();
+    for (const sch of this._descriptor.schemas) {
+      for (const fn of sch.functions ?? []) {
+        if (!schemaOf.has(fn)) schemaOf.set(fn, sch.name);
+      }
+    }
+
+    return globals.map((fn) => {
+      const schemaName = schemaOf.get(fn);
+      if (schemaName === undefined) {
+        throw new Error(
+          `Catalog '${this._descriptor.name}': '${fn.meta.name}' is listed in ` +
+            `globalFunctions but is not declared in any schema.`,
+        );
+      }
+      return this._functionToInfo(fn, schemaName);
+    });
   }
 
   override schemaContentsMacros(

package/src/client/client.ts

@@ -628,6 +628,10 @@ export class VgiClient {
       comment: inner.comment ?? null,
       tags: deserializeTags(inner.tags),
       supports_column_statistics: inner.supports_column_statistics ?? false,
+      global_functions: inner.global_functions
+        ? (Array.isArray(inner.global_functions) ? inner.global_functions : [...inner.global_functions]).map(toUint8Array)
+        : [],
+      global_function_prefix: inner.global_function_prefix ?? "",
       resolved_data_version: inner.resolved_data_version ?? null,
       resolved_implementation_version: inner.resolved_implementation_version ?? null,
     };

package/src/generated/vgi-client.ts

@@ -2,7 +2,7 @@
 // GENERATED by vgi.codegen.ts_client. DO NOT EDIT BY HAND.
 //
 // Generator: vgi-gen-ts-client v1
-// Content hash: 80ec56060e88
+// Content hash: e0862da99312
 //
 // To regenerate:
 //   uv run --project ~/Development/vgi-python vgi-gen-ts-client \
@@ -384,6 +384,8 @@ export interface CatalogAttachResult {
   comment?: string | null;
   tags: Record<string, string>;
   supports_column_statistics: boolean;
+  global_functions: Uint8Array[];
+  global_function_prefix: string;
   resolved_data_version?: string | null;
   resolved_implementation_version?: string | null;
 }

package/src/generated/vgi-protocol-schemas.ts

@@ -2,7 +2,7 @@
 // GENERATED by vgi.codegen.ts_schemas. DO NOT EDIT BY HAND.
 //
 // Generator: vgi-gen-ts-schemas v2
-// Content hash: 214d149e8f9c
+// Content hash: ae533ce2196c
 //
 // To regenerate:
 //   uv run --project ~/Development/vgi-python vgi-gen-ts-schemas \
@@ -265,6 +265,8 @@ export const CatalogAttachResultSchema = schema([
   field("comment", utf8(), true),
   field("tags", map(field("key", utf8(), false), field("value", utf8(), true), false), false),
   field("supports_column_statistics", bool(), false),
+  field("global_functions", list(field("item", binary(), true)), false),
+  field("global_function_prefix", utf8(), false),
   field("resolved_data_version", utf8(), true),
   field("resolved_implementation_version", utf8(), true),
 ]);

package/src/protocol/handlers/catalog/admin.ts

@@ -108,6 +108,12 @@ export function registerCatalogAdminMethods(protocol: Protocol, getCatalog: GetC
         // is also true. Catalogs that never serve column stats can override
         // this in attach() via CatalogAttachResult.supports_column_statistics.
         supports_column_statistics: result.supports_column_statistics ?? true,
+        // Protocol 1.3.0: functions the worker asks the client to publish into
+        // its global namespace. Empty unless the catalog advertises them —
+        // but the columns must still be present, since the extension pins the
+        // full CatalogAttachResult field list on the response schema.
+        global_functions: result.global_functions ?? [],
+        global_function_prefix: result.global_function_prefix ?? "",
         resolved_data_version: result.resolved_data_version ?? null,
         resolved_implementation_version: result.resolved_implementation_version ?? null,
       }, CatalogAttachResultSchema);