add a list of dataset libraries (#743)

see

huggingface-internal/moon-landing#9352 (comment)

Author: severo

packages/tasks/src/dataset-libraries.ts

@@ -0,0 +1,59 @@
+/**
+ * Elements configurable by a dataset library.
+ */
+export interface DatasetLibraryUiElement {
+	/**
+	 * Pretty name of the library.
+	 * displayed (in tags?, and) on the main
+	 * call-to-action button on the dataset page.
+	 */
+	prettyLabel: string;
+	/**
+	 * Repo name of the library's (usually on GitHub) code repo
+	 */
+	repoName: string;
+	/**
+	 * URL to library's (usually on GitHub) code repo
+	 */
+	repoUrl: string;
+	/**
+	 * URL to library's docs
+	 */
+	docsUrl?: string;
+}
+
+export const DATASET_LIBRARIES_UI_ELEMENTS = {
+	mlcroissant: {
+		prettyLabel: "mlcroissant",
+		repoName: "croissant",
+		repoUrl: "https://github.com/mlcommons/croissant/tree/main/python/mlcroissant",
+		docsUrl: "https://github.com/mlcommons/croissant/blob/main/python/mlcroissant/README.md",
+	},
+	webdataset: {
+		prettyLabel: "WebDataset",
+		repoName: "webdataset",
+		repoUrl: "https://github.com/webdataset/webdataset",
+		docsUrl: "https://huggingface.co/docs/hub/datasets-webdataset",
+	},
+	datasets: {
+		prettyLabel: "Datasets",
+		repoName: "datasets",
+		repoUrl: "https://github.com/huggingface/datasets",
+		docsUrl: "https://huggingface.co/docs/hub/datasets-usage",
+	},
+	pandas: {
+		prettyLabel: "pandas",
+		repoName: "pandas",
+		repoUrl: "https://github.com/pandas-dev/pandas",
+		docsUrl: "https://huggingface.co/docs/hub/datasets-pandas",
+	},
+	dask: {
+		prettyLabel: "Dask",
+		repoName: "dask",
+		repoUrl: "https://github.com/dask/dask",
+		docsUrl: "https://huggingface.co/docs/hub/datasets-dask",
+	},
+} satisfies Record<string, DatasetLibraryUiElement>;
+
+/// List of the dataset libraries supported by the Hub
+export type DatasetLibraryKey = keyof typeof DATASET_LIBRARIES_UI_ELEMENTS;

packages/tasks/src/index.ts

@@ -49,3 +49,6 @@ export { SKUS, DEFAULT_MEMORY_OPTIONS } from "./hardware";
 export type { HardwareSpec, SkuType } from "./hardware";
 export { LOCAL_APPS } from "./local-apps";
 export type { LocalApp, LocalAppKey } from "./local-apps";
+
+export { DATASET_LIBRARIES_UI_ELEMENTS } from "./dataset-libraries";
+export type { DatasetLibraryUiElement, DatasetLibraryKey } from "./dataset-libraries";