feat: render richer blocks and inline nodes (task 0004)

Extend the core parser and the CLI renderer to cover the everyday Markdown
body beyond headings and paragraphs: lists, blockquotes, fenced code,
thematic breaks, and inline emphasis, strong, code spans, and links.

Add remark-gfm to the pipeline so tables and strikethrough parse as their
own nodes. Any block or inline node core does not model falls through to
RawBlock / RawInline, whose verbatim value is position-sliced from the
original input so unmodelled constructs round-trip byte-for-byte.
This commit was merged in pull request #3.
This commit is contained in:
2026-07-29 07:34:56 -04:00
parent ab309d3226
commit 22befaaa71
9 changed files with 681 additions and 26 deletions

View File

@@ -1,16 +1,17 @@
import type { PhrasingContent, Root, RootContent } from "mdast";
import type { ListItem, Node, PhrasingContent, Root, RootContent } from "mdast";
import remarkFrontmatter from "remark-frontmatter";
import remarkGfm from "remark-gfm";
import remarkParse from "remark-parse";
import { unified } from "unified";
import { parse as parseYaml } from "yaml";
import type { Block, DocumentAST, Frontmatter, InlineNode } from "./types.ts";
import type { Block, DocumentAST, Frontmatter, InlineNode, ListItemBlock } from "./types.ts";
const processor = unified().use(remarkParse).use(remarkFrontmatter);
const processor = unified().use(remarkParse).use(remarkFrontmatter).use(remarkGfm);
export function parse(input: string): DocumentAST {
const tree = processor.parse(input);
const frontmatter = extractFrontmatter(tree);
const blocks = tree.children.flatMap(translateBlock);
const blocks = tree.children.flatMap((node) => translateBlock(node, input));
return { frontmatter, blocks, diagnostics: [] };
}
@@ -26,21 +27,74 @@ function extractFrontmatter(tree: Root): Frontmatter {
return {};
}
function translateBlock(node: RootContent): Block[] {
if (node.type === "heading") {
return [{ type: "heading", level: node.depth, children: translateInline(node.children) }];
function translateBlock(node: RootContent, input: string): Block[] {
switch (node.type) {
case "heading":
return [
{ type: "heading", level: node.depth, children: translateInline(node.children, input) },
];
case "paragraph":
return [{ type: "paragraph", children: translateInline(node.children, input) }];
case "list":
return [
{
type: "list",
ordered: node.ordered ?? false,
items: node.children.map((item) => translateListItem(item, input)),
},
];
case "blockquote":
return [
{
type: "blockquote",
children: node.children.flatMap((child) => translateBlock(child, input)),
},
];
case "code":
return [{ type: "code", ...(node.lang ? { lang: node.lang } : {}), value: node.value }];
case "thematicBreak":
return [{ type: "thematicBreak" }];
// Frontmatter is captured separately and must not double as a block.
case "yaml":
return [];
default:
return [{ type: "raw", value: slice(node, input) }];
}
if (node.type === "paragraph") {
return [{ type: "paragraph", children: translateInline(node.children) }];
}
return [];
}
function translateInline(nodes: PhrasingContent[]): InlineNode[] {
return nodes.flatMap((node) => {
if (node.type === "text") {
return [{ type: "text", value: node.value }];
function translateListItem(item: ListItem, input: string): ListItemBlock {
return {
type: "listItem",
children: item.children.flatMap((child) => translateBlock(child, input)),
};
}
function translateInline(nodes: PhrasingContent[], input: string): InlineNode[] {
return nodes.flatMap((node): InlineNode[] => {
switch (node.type) {
case "text":
return [{ type: "text", value: node.value }];
case "emphasis":
return [{ type: "emphasis", children: translateInline(node.children, input) }];
case "strong":
return [{ type: "strong", children: translateInline(node.children, input) }];
case "inlineCode":
return [{ type: "codeSpan", value: node.value }];
case "link":
return [{ type: "link", href: node.url, children: translateInline(node.children, input) }];
default:
return [{ type: "rawInline", value: slice(node, input) }];
}
return [];
});
}
// Verbatim source for an unmodelled node, taken by position so it round-trips
// byte-for-byte rather than being re-stringified through remark.
function slice(node: Node, input: string): string {
const start = node.position?.start.offset;
const end = node.position?.end.offset;
if (start === undefined || end === undefined) {
return "";
}
return input.slice(start, end);
}