dorfteich/packages/shared/src/editor-schema/schema.ts

import { Schema } from 'prosemirror-model';
import { tableNodes } from 'prosemirror-tables';

/**
 * The one ProseMirror schema Dorfteich documents are written in (ADR 0004).
 * Editor (TipTap, #25), server-side derivation (this package), and plugin
 * validation (ADR 0008) all import this schema instead of defining their
 * own, so "valid document" means the same thing everywhere.
 *
 * Node names `wikilink` and `plugin_block` are reserved for later stories
 * (wikilinks, plugin-defined block types) — do not repurpose them.
 */
export const editorSchema = new Schema({
  nodes: {
    doc: { content: 'block+' },

    paragraph: {
      group: 'block',
      content: 'inline*',
      parseDOM: [{ tag: 'p' }],
      toDOM: () => ['p', 0],
    },

    heading: {
      group: 'block',
      content: 'inline*',
      defining: true,
      attrs: { level: { default: 1, validate: 'number' } },
      parseDOM: [1, 2, 3, 4].map((level) => ({ tag: `h${level}`, attrs: { level } })),
      toDOM: (node) => [`h${node.attrs.level as number}`, 0],
    },

    blockquote: {
      group: 'block',
      content: 'block+',
      parseDOM: [{ tag: 'blockquote' }],
      toDOM: () => ['blockquote', 0],
    },

    code_block: {
      group: 'block',
      content: 'text*',
      marks: '',
      code: true,
      defining: true,
      whitespace: 'pre',
      parseDOM: [{ tag: 'pre', preserveWhitespace: 'full' }],
      toDOM: () => ['pre', ['code', 0]],
    },

    horizontal_rule: {
      group: 'block',
      parseDOM: [{ tag: 'hr' }],
      toDOM: () => ['hr'],
    },

    bullet_list: {
      group: 'block',
      content: 'list_item+',
      parseDOM: [{ tag: 'ul' }],
      toDOM: () => ['ul', 0],
    },

    ordered_list: {
      group: 'block',
      content: 'list_item+',
      attrs: { order: { default: 1, validate: 'number' } },
      parseDOM: [{ tag: 'ol' }],
      toDOM: (node) =>
        node.attrs.order === 1 ? ['ol', 0] : ['ol', { start: node.attrs.order }, 0],
    },

    list_item: {
      content: 'paragraph block*',
      parseDOM: [{ tag: 'li' }],
      toDOM: () => ['li', 0],
    },

    task_list: {
      group: 'block',
      content: 'task_item+',
      parseDOM: [{ tag: 'ul[data-type="task_list"]' }],
      toDOM: () => ['ul', { 'data-type': 'task_list' }, 0],
    },

    task_item: {
      content: 'paragraph block*',
      attrs: { checked: { default: false, validate: 'boolean' } },
      parseDOM: [{ tag: 'li[data-type="task_item"]' }],
      toDOM: (node) => [
        'li',
        { 'data-type': 'task_item', 'data-checked': String(node.attrs.checked) },
        0,
      ],
    },

    text: { group: 'inline' },

    hard_break: {
      group: 'inline',
      inline: true,
      selectable: false,
      parseDOM: [{ tag: 'br' }],
      toDOM: () => ['br'],
    },

    image: {
      group: 'inline',
      inline: true,
      atom: true,
      attrs: {
        fileId: { validate: 'string' },
        alt: { default: '', validate: 'string' },
        width: { default: null },
      },
      parseDOM: [{ tag: 'img[data-file-id]' }],
      toDOM: (node) => [
        'img',
        {
          'data-file-id': node.attrs.fileId as string,
          alt: node.attrs.alt as string,
          width: node.attrs.width as number | null,
        },
      ],
    },

    ...tableNodes({ tableGroup: 'block', cellContent: 'block+', cellAttributes: {} }),
  },

  marks: {
    bold: {
      parseDOM: [{ tag: 'strong' }, { tag: 'b' }],
      toDOM: () => ['strong', 0],
    },

    italic: {
      parseDOM: [{ tag: 'em' }, { tag: 'i' }],
      toDOM: () => ['em', 0],
    },

    code: {
      parseDOM: [{ tag: 'code' }],
      toDOM: () => ['code', 0],
    },

    strikethrough: {
      parseDOM: [{ tag: 's' }, { tag: 'del' }],
      toDOM: () => ['s', 0],
    },

    link: {
      inclusive: false,
      attrs: { href: { validate: 'string' } },
      parseDOM: [{ tag: 'a[href]' }],
      toDOM: (mark) => ['a', { href: mark.attrs.href as string }, 0],
    },
  },
});

/** Link protocols allowed in editor content (security.md §Content). */
export const ALLOWED_LINK_PROTOCOLS = ['http:', 'https:', 'mailto:'] as const;

/** Rejects `javascript:`/`data:`/etc. hrefs; also rejects unparseable input. */
export function isAllowedLinkHref(href: string): boolean {
  try {
    // A base is only needed to resolve protocol-relative/relative inputs;
    // those never carry an executable scheme, so any base works here.
    const url = new URL(href, 'https://dorfteich.invalid');
    return (ALLOWED_LINK_PROTOCOLS as readonly string[]).includes(url.protocol);
  } catch {
    return false;
  }
}