visgl · ibgreen · Oct 18, 2024 · Oct 18, 2024
diff --git a/docs/arrowjs/api-reference/builder.md b/docs/arrowjs/api-reference/builder.md
@@ -77,3 +77,19 @@ type BuilderOptions<T extends DataType = any, TNull = any> {
 
 `makeBuilder()` returns `Builder` which is a base class for the various that Arrow JS builder subclasses that
 construct Arrow Vectors from JavaScript values.
+
+### append()
+
+### flush()
+
+```ts
+builder.flush(): Data
+```
+
+### finish()
+
+```ts
+builder.finish();
+```
+
+When the builder is no longer needed, the application should call `builder.finish()`.
diff --git a/docs/arrowjs/developer-guide/builders.md b/docs/arrowjs/developer-guide/builders.md
@@ -23,3 +23,29 @@ const utf8Vector = utf8Builder.finish().toVector();
 console.log(utf8Vector.toJSON());
 // > ["hello", null, "world", null]
 ```
+
+One way to build a table with multiple columns is to create an arrow `Struct` field type using the fields in the table's schema,
+and then create a `Data` object using that `Field` object and the data
+
+```ts
+function buildTable(arrowSchema: arrow.Schema, const data: any[][]) {
+  const arrowBuilders = this.arrowSchema.fields.map((field) => arrow.makeBuilder({type: field.type, [null]));
+
+  // Application data
+  const row = [column0value, column1Value, ...];
+
+  for (let i = 0; i < this.arrowBuilders.length; i++) {
+    arrowBuilders[i].append(row[i]);
+  }
+
+  const arrowDatas = arrowBuilders.map((builder) => builder.flush());
+  const structField = new arrow.Struct(arrowSchema.fields);
+  const arrowStructData = new arrow.Data(structField, 0, length, 0, undefined, arrowDatas);
+  const arrowRecordBatch = new arrow.RecordBatch(arrowSchema, arrowStructData);
+  const arrowTable = new arrow.Table([arrowRecordBatch])
+
+  arrowBuilders.forEach((builder) => builder.finish());
+
+  return arrowTable;
+}
+```
diff --git a/modules/bson/src/bson-format.ts b/modules/bson/src/bson-format.ts
@@ -0,0 +1,15 @@
+// loaders.gl
+// SPDX-License-Identifier: MIT
+// Copyright (c) vis.gl contributors
+
+import type {Format} from '@loaders.gl/loader-utils';
+
+export const BSONFormat = {
+  name: 'BSON',
+  id: 'bson',
+  module: 'bson',
+  extensions: ['bson'],
+  mimeTypes: ['application/bson'],
+  category: 'json',
+  binary: true
+} as const satisfies Format;
diff --git a/modules/bson/src/bson-loader.ts b/modules/bson/src/bson-loader.ts
@@ -1,9 +1,11 @@
 // loaders.gl
 // SPDX-License-Identifier: MIT
 // Copyright (c) vis.gl contributors
+
 import type {LoaderWithParser, LoaderOptions} from '@loaders.gl/loader-utils';
 import type {ParseBSONOptions} from './lib/parsers/parse-bson';
 import {parseBSONSync} from './lib/parsers/parse-bson';
+import {BSONFormat} from './bson-format';
 
 // __VERSION__ is injected by babel-plugin-version-inline
 // @ts-ignore TS2304: Cannot find name '__VERSION__'.
@@ -18,6 +20,7 @@ export type BSONLoaderOptions = LoaderOptions & {
 };
 
 export const BSONLoader = {
+  ...BSONFormat,
   dataType: null as unknown as Record<string, unknown>,
   batchType: null as never,
   name: 'BSON',

diff --git a/modules/bson/src/bson-writer.ts b/modules/bson/src/bson-writer.ts
@@ -5,6 +5,7 @@
 import type {WriterWithEncoder, WriterOptions} from '@loaders.gl/loader-utils';
 import type {EncodeBSONOptions} from './lib/encoders/encode-bson';
 import {encodeBSONSync} from './lib/encoders/encode-bson';
+import {BSONFormat} from './bson-format';
 
 // __VERSION__ is injected by babel-plugin-version-inline
 // @ts-ignore TS2304: Cannot find name '__VERSION__'.
@@ -15,6 +16,7 @@ export type BSONWriterOptions = WriterOptions & {
 }
 
 export const BSONWriter = {
+  ...BSONFormat,
   name: 'BSON',
   id: 'bson',
   module: 'bson',

diff --git a/modules/draco/src/draco-writer.ts b/modules/draco/src/draco-writer.ts
@@ -35,6 +35,7 @@ export const DracoWriter = {
   module: 'draco',
   version: VERSION,
   extensions: ['drc'],
+  mimeTypes: ['application/octet-stream'],
   options: {
     draco: DEFAULT_DRACO_WRITER_OPTIONS
   },

diff --git a/modules/gis/src/index.ts b/modules/gis/src/index.ts
@@ -73,7 +73,7 @@ export {parseWKTCRS} from './lib//wkt-crs/parse-wkt-crs';
 export type {EncodeWKTCRSOptions} from './lib//wkt-crs/encode-wkt-crs';
 export {encodeWKTCRS} from './lib//wkt-crs/encode-wkt-crs';
 
-// Arrow Geometries
+// GEOARROW
 export type {
   BinaryDataFromGeoArrow,
   BinaryGeometriesFromArrowOptions
@@ -87,7 +87,6 @@ export {
   getMeanCentersFromBinaryGeometries
 } from './lib/feature-collection-converters/convert-geoarrow-to-binary-features';
 
-// GEOARROW
 export {convertGeoArrowGeometryToGeoJSON} from './lib/geometry-converters/convert-geoarrow-to-geojson';
 export {getGeometryColumnsFromSchema} from './lib/geoarrow/geoarrow-metadata';
 export {updateBoundsFromGeoArrowSamples} from './lib/geoarrow/get-arrow-bounds';

diff --git a/modules/images/src/image-writer.ts b/modules/images/src/image-writer.ts
@@ -21,6 +21,7 @@ export const ImageWriter = {
   module: 'images',
   version: VERSION,
   extensions: ['jpeg'],
+  mimeTypes: [],
   options: {
     image: {
       mimeType: 'image/png',

diff --git a/modules/loader-utils/src/format-types.ts b/modules/loader-utils/src/format-types.ts
@@ -0,0 +1,24 @@
+/**
+ * A worker loader definition that can be used with `@loaders.gl/core` functions
+ */
+export type Format = {
+  /** Human readable name */
+  name: string;
+  /** Unique lower-case id string for this format. Used for e.g. LoaderOptions */
+  id: string;
+  /** loaders.gl module that contains the implementation of this format */
+  module: string;
+  /** Which category does this loader belong to */
+  category?: string;
+  /** File extensions that are potential matches with this loader. */
+  extensions: string[];
+  /** MIMETypes that indicate a match with this loader. @note Some MIMETypes are generic and supported by many loaders */
+  mimeTypes: string[];
+  /** Is this a binary format */
+  binary?: boolean;
+  /** Is this a text format */
+  text?: boolean;
+
+  /** Test some initial bytes of content to see if this loader might be a match */
+  tests?: (((ArrayBuffer: ArrayBuffer) => boolean) | ArrayBuffer | string)[];
+};
diff --git a/modules/loader-utils/src/index.ts b/modules/loader-utils/src/index.ts
@@ -20,6 +20,10 @@ export type {
   FetchLike
 } from './types';
 
+// formats
+
+export type {Format} from './format-types';
+
 // loaders
 
 export type {

diff --git a/modules/loader-utils/src/loader-types.ts b/modules/loader-utils/src/loader-types.ts
@@ -2,10 +2,8 @@
 // SPDX-License-Identifier: MIT
 // Copyright (c) vis.gl contributors
 
-import {
-  FetchLike,
-  TransformBatches /* , DataType, SyncDataType, BatchableDataType */
-} from './types';
+import type {Format} from './format-types';
+import {FetchLike, TransformBatches} from './types';
 import {ReadableFile} from './lib/files/file';
 
 // LOADERS
@@ -113,7 +111,7 @@ type PreloadOptions = {
 /**
  * A worker loader definition that can be used with `@loaders.gl/core` functions
  */
-export type Loader<DataT = any, BatchT = any, LoaderOptionsT = LoaderOptions> = {
+export type Loader<DataT = any, BatchT = any, LoaderOptionsT = LoaderOptions> = Format & {
   /** The result type of this loader  */
   dataType?: DataT;
   /** The batched result type of this loader  */
@@ -123,26 +121,24 @@ export type Loader<DataT = any, BatchT = any, LoaderOptionsT = LoaderOptions> =
   options: LoaderOptionsT;
   /** Deprecated Options */
   deprecatedOptions?: Record<string, string | Record<string, string>>;
+  /** Version should be injected by build tools */
+  version: string;
+  /** A boolean, or a URL */
+  worker?: string | boolean;
+  // end Worker
 
   /** Human readable name */
   name: string;
   /** id should be the same as the field used in LoaderOptions */
   id: string;
   /** module is used to generate worker threads, need to be the module directory name */
   module: string;
-  /** Version should be injected by build tools */
-  version: string;
-  /** A boolean, or a URL */
-  worker?: string | boolean;
-  // end Worker
-
   /** Which category does this loader belong to */
   category?: string;
   /** File extensions that are potential matches with this loader. */
   extensions: string[];
   /** MIMETypes that indicate a match with this loader. @note Some MIMETypes are generic and supported by many loaders */
   mimeTypes: string[];
-
   /** Is the input of this loader binary */
   binary?: boolean;
   /** Is the input of this loader text */

diff --git a/modules/loader-utils/src/source-types.ts b/modules/loader-utils/src/source-types.ts
@@ -2,6 +2,7 @@
 // SPDX-License-Identifier: MIT
 // Copyright (c) vis.gl contributors
 
+import type {Format} from './format-types';
 import type {DataSource, DataSourceOptions} from './lib/sources/data-source';
 
 /**
@@ -15,7 +16,7 @@ export interface Source<
     unknown,
     DataSourceOptions
   >
-> {
+> extends Format {
   /** Type of source created by this service */
   dataSource?: DataSourceT;
   /** Type of options used when creating sources */

diff --git a/modules/loader-utils/src/writer-types.ts b/modules/loader-utils/src/writer-types.ts
@@ -2,6 +2,8 @@
 // SPDX-License-Identifier: MIT
 // Copyright (c) vis.gl contributors
 
+import type {Format} from './format-types';
+
 // WRITERS
 
 /** Options for writers */
@@ -24,39 +26,35 @@ export type WriterOptions = {
  * A writer definition that can be used with `@loaders.gl/core` functions
  */
 // eslint-disable-next-line @typescript-eslint/no-unused-vars
-export type Writer<DataT = unknown, BatchT = unknown, WriterOptionsT = WriterOptions> = {
+export type Writer<DataT = unknown, BatchT = unknown, WriterOptionsT = WriterOptions> = Format & {
   /** The result type of this loader  */
   dataType?: DataT;
   /** The batched result type of this loader  */
   batchType?: BatchT;
+  /** Version should be injected by build tools */
+  version: string;
+  /** A boolean, or a URL */
+  worker?: string | boolean;
+  // end Worker
+  options: WriterOptionsT;
+  deprecatedOptions?: Record<string, string>;
 
   /** Human readable name */
   name: string;
   /** id should be the same as the field used in LoaderOptions */
   id: string;
   /** module is used to generate worker threads, need to be the module directory name */
   module: string;
-  /** Version should be injected by build tools */
-  version: string;
-  /** A boolean, or a URL */
-  worker?: string | boolean;
-  // end Worker
-
   /** Which category does this loader belong to */
   category?: string;
   /** File extensions that are potential matches with this loader. */
   extensions: string[];
   /** MIMETypes that indicate a match with this loader. @note Some MIMETypes are generic and supported by many loaders */
   mimeTypes?: string[];
-
   /** Is the input of this loader binary */
   binary?: boolean;
   /** Is the input of this loader text */
   text?: boolean;
-
-  /** Default options for this writer */
-  options: WriterOptionsT;
-  deprecatedOptions?: Record<string, string>;
 };
 
 /**

diff --git a/modules/schema/src/categories/category-table.ts b/modules/schema/src/categories/category-table.ts
@@ -79,7 +79,7 @@ export type TableBatch =
   | ArrowTableBatch;
 
 /** Batch for a table organized as an array of rows, each row is an array of values */
-export type ArrayRowTableBatch = Batch & {
+export type ArrayRowTableBatch<MetadataT = unknown> = Batch<MetadataT> & {
   shape: 'array-row-table';
   schema?: Schema;
   schemaType?: 'explicit' | 'deduced';
@@ -88,7 +88,7 @@ export type ArrayRowTableBatch = Batch & {
 };
 
 /** Batch for a table organized as an array of rows, each row is an object mapping columns to values */
-export type ObjectRowTableBatch = Batch & {
+export type ObjectRowTableBatch<MetadataT = unknown> = Batch<MetadataT> & {
   shape: 'object-row-table';
   schema?: Schema;
   schemaType?: 'explicit' | 'deduced';
@@ -97,7 +97,7 @@ export type ObjectRowTableBatch = Batch & {
 };
 
 /** Batch for a table organized as an array of rows, each row is an array of values */
-export type GeoJSONTableBatch = Batch & {
+export type GeoJSONTableBatch<MetadataT = unknown> = Batch<MetadataT> & {
   shape: 'geojson-table';
   schema?: Schema;
   schemaType?: 'explicit' | 'deduced';
@@ -107,7 +107,7 @@ export type GeoJSONTableBatch = Batch & {
 };
 
 /** Batch for a table organized as a map of columns, each column is an array of value */
-export type ColumnarTableBatch = Batch & {
+export type ColumnarTableBatch<MetadataT = unknown> = Batch<MetadataT> & {
   shape: 'columnar-table';
   schemaType?: 'explicit' | 'deduced';
   schema?: Schema;
@@ -116,7 +116,7 @@ export type ColumnarTableBatch = Batch & {
 };
 
 /** Batch that wraps an Apache Arrow RecordBatch */
-export type ArrowTableBatch = Batch & {
+export type ArrowTableBatch<MetadataT = unknown> = Batch<MetadataT> & {
   shape: 'arrow-table';
   schemaType?: 'explicit' | 'deduced';
   schema?: Schema;

diff --git a/modules/schema/src/types/batch.ts b/modules/schema/src/types/batch.ts
@@ -12,9 +12,11 @@ type ApacheRecordBatch = unknown;
  * @see parseInBatches()
  * @see loadInBatches()
  */
-export type Batch = {
+export type Batch<MetadataT = unknown> = {
   /** A batch can contain metadata, data, or in case of unstructured data (JSON) */
   batchType: 'data' | 'metadata' | 'partial-result' | 'final-result';
+  /** Metadata for this batch if batchType === 'metadata' */
+  metadata?: MetadataT;
   /** A string identifying the shape of data in this batch (table, etc) */
   shape: string;
   /** Schema of the data in this batch */

diff --git a/modules/textures/src/compressed-texture-writer.ts b/modules/textures/src/compressed-texture-writer.ts
@@ -31,6 +31,7 @@ export const CompressedTextureWriter = {
   version: VERSION,
 
   extensions: ['dds'],
+  mimeTypes: ['image/vnd-ms.dds', 'image/x-dds', 'application/octet-stream'],
 
   options: {
     texture: {

diff --git a/modules/textures/src/ktx2-basis-writer.ts b/modules/textures/src/ktx2-basis-writer.ts
@@ -28,6 +28,7 @@ export const KTX2BasisWriter = {
   version: VERSION,
 
   extensions: ['ktx2'],
+  mimeTypes: ['image/ktx2'],
   options: {
     ['ktx2-basis-writer']: {
       useSRGB: false,

diff --git a/modules/wkt/src/twkb-writer.ts b/modules/wkt/src/twkb-writer.ts
@@ -23,6 +23,7 @@ export const TWKBWriter = {
   module: 'wkt',
   version: VERSION,
   extensions: ['twkb'],
+  mimeTypes: ['application/text'],
   encode: async (geometry: Geometry, options?: TWKBWriterOptions) =>
     convertGeometryToTWKB(geometry, options?.twkb),
   encodeSync: (geometry: Geometry, options?: TWKBWriterOptions) =>

diff --git a/modules/wkt/src/wkb-writer.ts b/modules/wkt/src/wkb-writer.ts
@@ -29,6 +29,7 @@ export const WKBWriter = {
   module: 'wkt',
   version: VERSION,
   extensions: ['wkb'],
+  mimeTypes: ['application/wkb', 'application/octet-stream'],
   options: {
     wkb: {
       hasZ: false,

diff --git a/modules/wkt/src/wkt-writer.ts b/modules/wkt/src/wkt-writer.ts
@@ -20,6 +20,7 @@ export const WKTWriter = {
   module: 'wkt',
   version: VERSION,
   extensions: ['wkt'],
+  mimeTypes: ['application/wkt', 'text/plain'],
   text: true,
   encode: async (geometry: Geometry) => convertGeometryToWKTSync(geometry),
   encodeSync: convertGeometryToWKTSync,