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
feat: copy as markdown in the dataframe table (#7507)
Myles Scolnick1f17c1e6290d
diff --git a/frontend/src/components/data-table/download-actions.tsx b/frontend/src/components/data-table/download-actions.tsxindex 8145d50698..501b99fb76 100644--- a/frontend/src/components/data-table/download-actions.tsx+++ b/frontend/src/components/data-table/download-actions.tsx@@ -5,6 +5,7 @@ import {BrickWallIcon,ChevronDownIcon,ClipboardListIcon,+ FileTextIcon,TableIcon,} from "lucide-react";import React from "react";@@ -13,7 +14,11 @@ import { logNever } from "@/utils/assertNever";import { copyToClipboard } from "@/utils/copy";import { downloadByURL } from "@/utils/download";import { prettyError } from "@/utils/errors";-import { jsonParseWithSpecialChar, jsonToTSV } from "@/utils/json/json-parser";+import {+ jsonParseWithSpecialChar,+ jsonToMarkdown,+ jsonToTSV,+} from "@/utils/json/json-parser";import { Button } from "../ui/button";import {DropdownMenu,@@ -70,6 +75,12 @@ const clipboardOptions = [description: "Comma-separated values",icon: TableIcon,},+ {+ label: "Markdown",+ format: "markdown",+ description: "Preserves hyperlinks and formatting",+ icon: FileTextIcon,+ },] as const;export const DownloadAs: React.FC<DownloadActionProps> = (props) => {@@ -115,6 +126,12 @@ export const DownloadAs: React.FC<DownloadActionProps> = (props) => {text = csv;break;}+ case "markdown": {+ const downloadUrl = await getDownloadUrl("json");+ const json = await fetchJson(downloadUrl);+ text = jsonToMarkdown(json);+ break;+ }default:logNever(format);return;diff --git a/frontend/src/utils/__tests__/json-parser.test.ts b/frontend/src/utils/__tests__/json-parser.test.tsindex 5b98162230..e0777bc23e 100644--- a/frontend/src/utils/__tests__/json-parser.test.ts+++ b/frontend/src/utils/__tests__/json-parser.test.ts@@ -1,6 +1,10 @@/* Copyright 2024 Marimo. All rights reserved. */import { expect, it } from "vitest";-import { jsonParseWithSpecialChar, jsonToTSV } from "../json/json-parser";+import {+ jsonParseWithSpecialChar,+ jsonToMarkdown,+ jsonToTSV,+} from "../json/json-parser";it("can jsonParseWithSpecialChar happy path", () => {expect(jsonParseWithSpecialChar('"hello"')).toEqual("hello");@@ -131,3 +135,194 @@ it("handles NaN values in TSV", () => {expect(jsonToTSV([{ a: Number.NaN, b: 1 }], locale)).toEqual("a\tb\nNaN\t1");});++it("can convert json to markdown - basic table", () => {+ expect(jsonToMarkdown([])).toMatchInlineSnapshot(`""`);++ expect(jsonToMarkdown([{ a: 1, b: 2 }])).toMatchInlineSnapshot(`+ "| a | b |+ |---|---|+ | 1 | 2 |"+ `);++ expect(+ jsonToMarkdown([+ { a: 1, b: 2 },+ { a: 3, b: 4 },+ ]),+ ).toMatchInlineSnapshot(`+ "| a | b |+ |---|---|+ | 1 | 2 |+ | 3 | 4 |"+ `);++ expect(+ jsonToMarkdown([+ { name: "Alice", age: 30 },+ { name: "Bob", age: 25 },+ ]),+ ).toMatchInlineSnapshot(`+ "| name | age |+ |---|---|+ | Alice | 30 |+ | Bob | 25 |"+ `);+});++it("can convert json to markdown - with URLs", () => {+ expect(+ jsonToMarkdown([+ { name: "Google", url: "https://google.com" },+ { name: "GitHub", url: "https://github.com" },+ ]),+ ).toMatchInlineSnapshot(`+ "| name | url |+ |---|---|+ | Google | [https://google.com](https://google.com) |+ | GitHub | [https://github.com](https://github.com) |"+ `);++ // Mixed content - URL and text in same cell+ expect(+ jsonToMarkdown([{ info: "Visit https://example.com for more" }]),+ ).toMatchInlineSnapshot(`+ "| info |+ |---|+ | Visit [https://example.com](https://example.com) for more |"+ `);++ // Multiple URLs in one cell+ expect(+ jsonToMarkdown([{ links: "https://google.com and https://github.com" }]),+ ).toMatchInlineSnapshot(`+ "| links |+ |---|+ | [https://google.com](https://google.com) and [https://github.com](https://github.com) |"+ `);+});++it("can convert json to markdown - handles nulls and undefined", () => {+ expect(+ jsonToMarkdown([{ a: null, b: undefined, c: 1 }]),+ ).toMatchInlineSnapshot(`+ "| a | b | c |+ |---|---|---|+ | | | 1 |"+ `);++ expect(+ jsonToMarkdown([+ { a: 1, b: 2 },+ { a: null, b: 3 },+ ]),+ ).toMatchInlineSnapshot(`+ "| a | b |+ |---|---|+ | 1 | 2 |+ | | 3 |"+ `);+});++it("can convert json to markdown - handles special characters", () => {+ // Pipes need to be escaped since they're markdown table delimiters+ expect(+ jsonToMarkdown([{ a: "value|with|pipes", b: "normal" }]),+ ).toMatchInlineSnapshot(`+ "| a | b |+ |---|---|+ | value\\|with\\|pipes | normal |"+ `);++ // Newlines should be replaced with spaces+ expect(+ jsonToMarkdown([{ a: "line1\nline2", b: "normal" }]),+ ).toMatchInlineSnapshot(`+ "| a | b |+ |---|---|+ | line1 line2 | normal |"+ `);++ // Backslashes should be escaped+ expect(jsonToMarkdown([{ path: "C:\\Users\\Name" }])).toMatchInlineSnapshot(`+ "| path |+ |---|+ | C:\\\\Users\\\\Name |"+ `);+});++it("can convert json to markdown - handles different data types", () => {+ expect(+ jsonToMarkdown([{ str: "text", num: 42, bool: true, nil: null }]),+ ).toMatchInlineSnapshot(`+ "| str | num | bool | nil |+ |---|---|---|---|+ | text | 42 | true | |"+ `);++ // Numbers (including NaN and Infinity)+ expect(+ jsonToMarkdown([+ {+ a: Number.NaN,+ b: Number.POSITIVE_INFINITY,+ c: Number.NEGATIVE_INFINITY,+ },+ ]),+ ).toMatchInlineSnapshot(`+ "| a | b | c |+ |---|---|---|+ | NaN | Infinity | -Infinity |"+ `);++ // Arrays and objects should be stringified+ expect(jsonToMarkdown([{ data: [1, 2, 3] }])).toMatchInlineSnapshot(`+ "| data |+ |---|+ | [1,2,3] |"+ `);++ expect(+ jsonToMarkdown([{ data: { nested: "value" } }]),+ ).toMatchInlineSnapshot(`+ "| data |+ |---|+ | {"nested":"value"} |"+ `);+});++it("can convert json to markdown - handles existing markdown links", () => {+ // When input already contains a markdown link, it should be preserved as-is+ expect(+ jsonToMarkdown([{ link: "[Google](https://google.com)" }]),+ ).toMatchInlineSnapshot(`+ "| link |+ |---|+ | [Google](https://google.com) |"+ `);++ // Multiple existing markdown links+ expect(+ jsonToMarkdown([+ { text: "[Google](https://google.com) and [GitHub](https://github.com)" },+ ]),+ ).toMatchInlineSnapshot(`+ "| text |+ |---|+ | [Google](https://google.com) and [GitHub](https://github.com) |"+ `);++ // Mix of existing markdown links and plain URLs+ expect(+ jsonToMarkdown([+ {+ mixed:+ "[Google](https://google.com) and plain https://example.com and [GitHub](https://github.com)",+ },+ ]),+ ).toMatchInlineSnapshot(`+ "| mixed |+ |---|+ | [Google](https://google.com) and plain [https://example.com](https://example.com) and [GitHub](https://github.com) |"+ `);+});diff --git a/frontend/src/utils/json/json-parser.ts b/frontend/src/utils/json/json-parser.tsindex 87667e15c5..7224a214fb 100644--- a/frontend/src/utils/json/json-parser.ts+++ b/frontend/src/utils/json/json-parser.ts@@ -96,6 +96,95 @@ export function jsonToTSV(json: Record<string, unknown>[], locale: string) {return `${keys.join("\t")}\n${values.join("\n")}`;}+/**+ * Converts JSON data to a Markdown table format+ * Detects URLs and converts them to markdown links [url](url)+ */+export function jsonToMarkdown(json: Record<string, unknown>[]): string {+ if (json.length === 0) {+ return "";+ }++ const keys = Object.keys(json[0]);++ const header = `| ${keys.join(" | ")} |`;+ const separator = `|${keys.map(() => "---").join("|")}|`;++ const rows = json.map((row) => {+ const cells = keys.map((key) => formatValueForMarkdown(row[key]));+ return `| ${cells.join(" | ")} |`;+ });++ return [header, separator, ...rows].join("\n");+}++/**+ * Formats a value for markdown export+ * - Detects URLs and converts them to markdown links+ * - Skips URLs that are already in markdown link format+ * - Escapes special markdown characters+ * - Handles nulls, arrays, and objects+ */+function formatValueForMarkdown(value: unknown): string {+ // Handle nulls and undefined+ if (value === null || value === undefined) {+ return "";+ }++ // Handle arrays and objects by stringifying+ if (Array.isArray(value)) {+ return escapeMarkdownTableChars(JSON.stringify(value));+ }++ if (typeof value === "object") {+ return escapeMarkdownTableChars(JSON.stringify(value));+ }++ // Convert to string+ let stringValue = String(value);++ // Regex to match existing markdown links: [text](url)+ const markdownLinkRegex = /\[([^\]]+)\]\((https?:\/\/[^)]+)\)/g;++ // Temporarily replace existing markdown links with placeholders+ const placeholders = new Map<string, string>();+ let placeholderIndex = 0;++ stringValue = stringValue.replace(markdownLinkRegex, (match) => {+ const placeholder = `__MDLINK_${placeholderIndex++}__`;+ placeholders.set(placeholder, match);+ return placeholder;+ });++ // Convert plain URLs to markdown links+ const urlRegex = /(https?:\/\/\S+)/g;+ stringValue = stringValue.replace(urlRegex, "[$1]($1)");++ // Restore markdown links from placeholders+ for (const [placeholder, original] of placeholders) {+ stringValue = stringValue.replace(placeholder, original);+ }++ // Escape special markdown table characters+ return escapeMarkdownTableChars(stringValue);+}++/**+ * Escapes special characters in markdown tables+ */+function escapeMarkdownTableChars(value: string): string {+ return (+ value+ // Escape backslashes first (must be first to avoid double-escaping)+ .replaceAll("\\", "\\\\")+ // Escape pipes (markdown table delimiter)+ .replaceAll("|", "\\|")+ // Replace newlines with spaces+ .replaceAll("\n", " ")+ .replaceAll("\r", "")+ );+}+/** Adapted from https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/BigInt#use_within_json */function sanitizeBigInt(value: unknown): unknown {if (