A reactive notebook for Python — run reproducible experiments, query with SQL, execute as a script, deploy as an app, and version with git. Stored as pure Python. All in a modern, AI-native editor.
Python artificial-intelligencedagdata-sciencedata-visualizationdataflowdeveloper-toolsmachine-learningnotebookspipelinepythonreactivesqlweb-app
/* Copyright 2026 Marimo. All rights reserved. */
// @ts-expect-error - no types
import * as vl from "vega-loader";
import type { DataFormat } from "./types";
import type { DataType } from "./vega-loader";
// Re-export the vega-loader functions to add TypeScript types
export function read<T = object>(
data: string | Record<string, unknown> | Record<string, unknown>[],
format:
| DataFormat
| {
type: DataFormat["type"];
parse: "auto";
}
| {
type: DataFormat["type"];
parse: Record<string, DataType>;
}
| undefined,
): T[] {
return vl.read(data, format);
}
export interface Loader {
load(
uri: string,
options?: unknown,
): Promise<string | Record<string, unknown> | Record<string, unknown>[]>;
sanitize(uri: string, options?: unknown): Promise<{ href: string }>;
http(uri: string, options?: unknown): Promise<string>;
file(filename: string): Promise<string>;
}
export function createLoader(): Loader {
return vl.loader();
}
export type VegaDataType =
| "boolean"
| "integer"
| "number"
| "date"
| "string"
| "unknown";
export type FieldTypes = Record<string, VegaDataType>;
export const typeParsers: Record<VegaDataType, (value: string) => unknown> =
vl.typeParsers;
export type { DataType } from "@/core/kernel/messages";