mdream 0.13.3 → 0.15.0

package/README.md

@@ -311,6 +311,117 @@ htmlToMarkdown(html, { plugins: [plugin] })
 
 The extraction plugin provides memory-efficient element extraction with full text content and attributes, perfect for SEO analysis, content discovery, and data mining.
 
+## Markdown Splitting
+
+Split HTML into chunks during conversion for LLM context windows, vector databases, or document processing.
+
+### Basic Chunking
+
+```ts
+import { TAG_H2 } from 'mdream'
+import { htmlToMarkdownSplitChunks } from 'mdream/splitter'
+
+const html = `
+  <h1>Documentation</h1>
+  <h2>Installation</h2>
+  <p>Install via npm...</p>
+  <h2>Usage</h2>
+  <p>Use it like this...</p>
+`
+
+const chunks = htmlToMarkdownSplitChunks(html, {
+  headersToSplitOn: [TAG_H2], // Split on h2 headers
+  chunkSize: 1000, // Max chars per chunk
+  chunkOverlap: 200, // Overlap for context
+  stripHeaders: true // Remove headers from content
+})
+
+// Each chunk includes content and metadata
+chunks.forEach((chunk) => {
+  console.log(chunk.content)
+  console.log(chunk.metadata.headers) // { h1: "Documentation", h2: "Installation" }
+  console.log(chunk.metadata.code) // Language if chunk contains code
+  console.log(chunk.metadata.loc) // Line numbers
+})
+```
+
+### Streaming Chunks (Memory Efficient)
+
+For large documents, use the generator version to process chunks one at a time:
+
+```ts
+import { htmlToMarkdownSplitChunksStream } from 'mdream/splitter'
+
+// Process chunks incrementally - lower memory usage
+for (const chunk of htmlToMarkdownSplitChunksStream(html, options)) {
+  await processChunk(chunk) // Handle each chunk as it's generated
+
+  // Can break early if you found what you need
+  if (foundTarget)
+    break
+}
+```
+
+**Benefits of streaming:**
+- Lower memory usage - chunks aren't stored in an array
+- Early termination - stop processing when you find what you need
+- Better for large documents
+
+### Splitting Options
+
+```ts
+interface SplitterOptions {
+  // Structural splitting
+  headersToSplitOn?: number[] // TAG_H1, TAG_H2, etc. Default: [TAG_H2-TAG_H6]
+
+  // Size-based splitting
+  chunkSize?: number // Max chunk size. Default: 1000
+  chunkOverlap?: number // Overlap between chunks. Default: 200
+  lengthFunction?: (text: string) => number // Custom length (e.g., token count)
+
+  // Output formatting
+  stripHeaders?: boolean // Remove headers from content. Default: true
+  returnEachLine?: boolean // Split into individual lines. Default: false
+
+  // Standard options
+  origin?: string // Base URL for links/images
+  plugins?: Plugin[] // Apply plugins during conversion
+}
+```
+
+### Chunk Metadata
+
+Each chunk includes rich metadata for context:
+
+```ts
+interface MarkdownChunk {
+  content: string
+  metadata: {
+    headers?: Record<string, string> // Header hierarchy: { h1: "Title", h2: "Section" }
+    code?: string // Code block language if present
+    loc?: { // Line number range
+      lines: { from: number, to: number }
+    }
+  }
+}
+```
+
+### Use with Presets
+
+Combine splitting with presets for optimized output:
+
+```ts
+import { TAG_H2 } from 'mdream'
+import { withMinimalPreset } from 'mdream/preset/minimal'
+import { htmlToMarkdownSplitChunks } from 'mdream/splitter'
+
+const chunks = htmlToMarkdownSplitChunks(html, withMinimalPreset({
+  headersToSplitOn: [TAG_H2],
+  chunkSize: 500,
+  origin: 'https://example.com'
+}))
+```
+
 ## Credits
 
 - [ultrahtml](https://github.com/natemoo-re/ultrahtml): HTML parsing inspiration

package/dist/_chunks/{const-BOAJ1T5c.mjs → const-Bf_XN9U9.mjs}

@@ -29,10 +29,7 @@ function collectNodeContent(node, content, state) {
 */
 function assembleBufferedContent(state) {
 	const fragments = [];
-	for (const [regionId, content] of Array.from(state.regionContentBuffers.entries())) {
-		const include = state.regionToggles.get(regionId);
-		if (include) fragments.push(...content);
-	}
+	for (const [regionId, content] of Array.from(state.regionContentBuffers.entries())) if (state.regionToggles.get(regionId)) fragments.push(...content);
 	state.regionToggles.clear();
 	state.regionContentBuffers.clear();
 	return fragments.join("").trimStart();
@@ -285,4 +282,4 @@ const LIST_ITEM_SPACING = [1, 0];
 const TABLE_ROW_SPACING = [0, 1];
 
 //#endregion
-export { BLOCKQUOTE_SPACING, DEFAULT_BLOCK_SPACING, ELEMENT_NODE, HTML_ENTITIES, LIST_ITEM_SPACING, MARKDOWN_CODE_BLOCK, MARKDOWN_EMPHASIS, MARKDOWN_HORIZONTAL_RULE, MARKDOWN_INLINE_CODE, MARKDOWN_STRIKETHROUGH, MARKDOWN_STRONG, MAX_TAG_ID, NO_SPACING, NodeEventEnter, NodeEventExit, TABLE_ROW_SPACING, TAG_A, TAG_ABBR, TAG_ADDRESS, TAG_AREA, TAG_ARTICLE, TAG_ASIDE, TAG_AUDIO, TAG_B, TAG_BASE, TAG_BDO, TAG_BLOCKQUOTE, TAG_BODY, TAG_BR, TAG_BUTTON, TAG_CANVAS, TAG_CAPTION, TAG_CENTER, TAG_CITE, TAG_CODE, TAG_COL, TAG_DD, TAG_DEL, TAG_DETAILS, TAG_DFN, TAG_DIALOG, TAG_DIV, TAG_DL, TAG_DT, TAG_EM, TAG_EMBED, TAG_FIELDSET, TAG_FIGCAPTION, TAG_FIGURE, TAG_FOOTER, TAG_FORM, TAG_H1, TAG_H2, TAG_H3, TAG_H4, TAG_H5, TAG_H6, TAG_HEAD, TAG_HEADER, TAG_HR, TAG_HTML, TAG_I, TAG_IFRAME, TAG_IMG, TAG_INPUT, TAG_INS, TAG_KBD, TAG_KEYGEN, TAG_LABEL, TAG_LEGEND, TAG_LI, TAG_LINK, TAG_MAIN, TAG_MAP, TAG_MARK, TAG_META, TAG_METER, TAG_NAV, TAG_NOFRAMES, TAG_NOSCRIPT, TAG_OBJECT, TAG_OL, TAG_OPTION, TAG_P, TAG_PARAM, TAG_PLAINTEXT, TAG_PRE, TAG_PROGRESS, TAG_Q, TAG_RP, TAG_RT, TAG_RUBY, TAG_SAMP, TAG_SCRIPT, TAG_SECTION, TAG_SELECT, TAG_SMALL, TAG_SOURCE, TAG_SPAN, TAG_STRONG, TAG_STYLE, TAG_SUB, TAG_SUMMARY, TAG_SUP, TAG_SVG, TAG_TABLE, TAG_TBODY, TAG_TD, TAG_TEMPLATE, TAG_TEXTAREA, TAG_TFOOT, TAG_TH, TAG_THEAD, TAG_TIME, TAG_TITLE, TAG_TR, TAG_TRACK, TAG_U, TAG_UL, TAG_VAR, TAG_VIDEO, TAG_WBR, TAG_XMP, TEXT_NODE, TagIdMap, assembleBufferedContent, collectNodeContent, createBufferRegion };
+export { TAG_H2 as $, TAG_TBODY as $t, TAG_BUTTON as A, TAG_P as At, TAG_DFN as B, TAG_SCRIPT as Bt, TAG_AUDIO as C, TAG_METER as Ct, TAG_BLOCKQUOTE as D, TAG_OBJECT as Dt, TAG_BDO as E, TAG_NOSCRIPT as Et, TAG_CODE as F, TAG_Q as Ft, TAG_EM as G, TAG_SPAN as Gt, TAG_DIV as H, TAG_SELECT as Ht, TAG_COL as I, TAG_RP as It, TAG_FIGCAPTION as J, TAG_SUB as Jt, TAG_EMBED as K, TAG_STRONG as Kt, TAG_DD as L, TAG_RT as Lt, TAG_CAPTION as M, TAG_PLAINTEXT as Mt, TAG_CENTER as N, TAG_PRE as Nt, TAG_BODY as O, TAG_OL as Ot, TAG_CITE as P, TAG_PROGRESS as Pt, TAG_H1 as Q, TAG_TABLE as Qt, TAG_DEL as R, TAG_RUBY as Rt, TAG_ASIDE as S, TAG_META as St, TAG_BASE as T, TAG_NOFRAMES as Tt, TAG_DL as U, TAG_SMALL as Ut, TAG_DIALOG as V, TAG_SECTION as Vt, TAG_DT as W, TAG_SOURCE as Wt, TAG_FOOTER as X, TAG_SUP as Xt, TAG_FIGURE as Y, TAG_SUMMARY as Yt, TAG_FORM as Z, TAG_SVG as Zt, TAG_A as _, TagIdMap as _n, TAG_LI as _t, LIST_ITEM_SPACING as a, TAG_THEAD as an, TAG_HEADER as at, TAG_AREA as b, createBufferRegion as bn, TAG_MAP as bt, MARKDOWN_HORIZONTAL_RULE as c, TAG_TR as cn, TAG_I as ct, MARKDOWN_STRONG as d, TAG_UL as dn, TAG_INPUT as dt, TAG_TD as en, TAG_H3 as et, MAX_TAG_ID as f, TAG_VAR as fn, TAG_INS as ft, TABLE_ROW_SPACING as g, TEXT_NODE as gn, TAG_LEGEND as gt, NodeEventExit as h, TAG_XMP as hn, TAG_LABEL as ht, HTML_ENTITIES as i, TAG_TH as in, TAG_HEAD as it, TAG_CANVAS as j, TAG_PARAM as jt, TAG_BR as k, TAG_OPTION as kt, MARKDOWN_INLINE_CODE as l, TAG_TRACK as ln, TAG_IFRAME as lt, NodeEventEnter as m, TAG_WBR as mn, TAG_KEYGEN as mt, DEFAULT_BLOCK_SPACING as n, TAG_TEXTAREA as nn, TAG_H5 as nt, MARKDOWN_CODE_BLOCK as o, TAG_TIME as on, TAG_HR as ot, NO_SPACING as p, TAG_VIDEO as pn, TAG_KBD as pt, TAG_FIELDSET as q, TAG_STYLE as qt, ELEMENT_NODE as r, TAG_TFOOT as rn, TAG_H6 as rt, MARKDOWN_EMPHASIS as s, TAG_TITLE as sn, TAG_HTML as st, BLOCKQUOTE_SPACING as t, TAG_TEMPLATE as tn, TAG_H4 as tt, MARKDOWN_STRIKETHROUGH as u, TAG_U as un, TAG_IMG as ut, TAG_ABBR as v, assembleBufferedContent as vn, TAG_LINK as vt, TAG_B as w, TAG_NAV as wt, TAG_ARTICLE as x, TAG_MARK as xt, TAG_ADDRESS as y, collectNodeContent as yn, TAG_MAIN as yt, TAG_DETAILS as z, TAG_SAMP as zt };

package/dist/_chunks/{extraction-BPaDGYvv.mjs → extraction-BA9MDtq3.mjs}

@@ -1,4 +1,4 @@
-import { createPlugin } from "./plugin-DrovQriD.mjs";
+import { t as createPlugin } from "./plugin-CjWWQTuL.mjs";
 
 //#region src/libs/query-selector.ts
 /**
@@ -28,8 +28,7 @@ function createClassSelector(selector) {
 	return {
 		matches: (element) => {
 			if (!element.attributes?.class) return false;
-			const classes = element.attributes.class.trim().split(" ").filter(Boolean);
-			return classes.includes(className);
+			return element.attributes.class.trim().split(" ").filter(Boolean).includes(className);
 		},
 		toString: () => `.${className}`
 	};
@@ -109,7 +108,7 @@ function extractionPlugin(selectors) {
 		matcher: parseSelector(selector),
 		callback
 	}));
-	const trackedElements = new Map();
+	const trackedElements = /* @__PURE__ */ new Map();
 	return createPlugin({
 		onNodeEnter(element) {
 			matcherCallbacks.forEach(({ matcher, callback }) => {
@@ -126,7 +125,6 @@ function extractionPlugin(selectors) {
 				if (tracked) tracked.textContent += textNode.value;
 				currentParent = currentParent.parent;
 			}
-			return void 0;
 		},
 		onNodeExit(element, state) {
 			const tracked = trackedElements.get(element);
@@ -143,4 +141,4 @@ function extractionPlugin(selectors) {
 }
 
 //#endregion
-export { extractionPlugin, parseSelector };
+export { parseSelector as n, extractionPlugin as t };

package/dist/_chunks/{llms-txt-DC12yO2l.mjs → llms-txt-T79S7X24.mjs}

@@ -1,7 +1,7 @@
-import { htmlToMarkdown } from "./src-C3QpB75q.mjs";
-import { extractionPlugin } from "./extraction-BPaDGYvv.mjs";
-import { readFile } from "node:fs/promises";
-import { basename, dirname, relative, sep } from "pathe";
+import { t as htmlToMarkdown } from "./src-BJpipdul.mjs";
+import { t as extractionPlugin } from "./extraction-BA9MDtq3.mjs";
+import { mkdir, open, readFile } from "node:fs/promises";
+import { basename, dirname, join, relative, sep } from "pathe";
 import { glob } from "tinyglobby";
 
 //#region src/llms-txt.ts
@@ -13,28 +13,27 @@ function extractMetadata(html, url) {
 	let description = "";
 	let keywords = "";
 	let author = "";
-	const extractionPluginInstance = extractionPlugin({
-		"title": (element) => {
-			if (!title && element.textContent) title = element.textContent.trim();
-		},
-		"meta[name=\"description\"]": (element) => {
-			if (!description && element.attributes?.content) description = element.attributes.content.trim();
-		},
-		"meta[property=\"og:description\"]": (element) => {
-			if (!description && element.attributes?.content) description = element.attributes.content.trim();
-		},
-		"meta[name=\"keywords\"]": (element) => {
-			if (!keywords && element.attributes?.content) keywords = element.attributes.content.trim();
-		},
-		"meta[name=\"author\"]": (element) => {
-			if (!author && element.attributes?.content) author = element.attributes.content.trim();
-		},
-		"meta[property=\"og:title\"]": (element) => {
-			if (!title && element.attributes?.content) title = element.attributes.content.trim();
-		}
-	});
 	htmlToMarkdown(html, {
-		plugins: [extractionPluginInstance],
+		plugins: [extractionPlugin({
+			"title": (element) => {
+				if (!title && element.textContent) title = element.textContent.trim();
+			},
+			"meta[name=\"description\"]": (element) => {
+				if (!description && element.attributes?.content) description = element.attributes.content.trim();
+			},
+			"meta[property=\"og:description\"]": (element) => {
+				if (!description && element.attributes?.content) description = element.attributes.content.trim();
+			},
+			"meta[name=\"keywords\"]": (element) => {
+				if (!keywords && element.attributes?.content) keywords = element.attributes.content.trim();
+			},
+			"meta[name=\"author\"]": (element) => {
+				if (!author && element.attributes?.content) author = element.attributes.content.trim();
+			},
+			"meta[property=\"og:title\"]": (element) => {
+				if (!title && element.attributes?.content) title = element.attributes.content.trim();
+			}
+		})],
 		origin: url
 	});
 	return {
@@ -113,8 +112,7 @@ function generateLlmsTxtContent(files, options) {
 * Parse frontmatter from markdown content
 */
 function parseFrontmatter(content) {
-	const frontmatterRegex = /^---\n([\s\S]*?)\n---\n([\s\S]*)$/;
-	const match = content.match(frontmatterRegex);
+	const match = content.match(/^---\n([\s\S]*?)\n---\n([\s\S]*)$/);
 	if (!match) return {
 		frontmatter: null,
 		body: content
@@ -127,8 +125,7 @@ function parseFrontmatter(content) {
 		const colonIndex = line.indexOf(":");
 		if (colonIndex > 0) {
 			const key = line.substring(0, colonIndex).trim();
-			const value = line.substring(colonIndex + 1).trim();
-			frontmatter[key] = value;
+			frontmatter[key] = line.substring(colonIndex + 1).trim();
 		}
 	}
 	return {
@@ -172,11 +169,10 @@ function generateLlmsFullTxtContent(files, options) {
 				if (file.metadata.keywords) metadata.keywords = file.metadata.keywords;
 				if (file.metadata.author) metadata.author = file.metadata.author;
 			}
-			const mergedFrontmatter = frontmatter ? {
+			const frontmatterString = serializeFrontmatter(frontmatter ? {
 				...frontmatter,
 				...metadata
-			} : metadata;
-			const frontmatterString = serializeFrontmatter(mergedFrontmatter);
+			} : metadata);
 			let contentBody = frontmatter ? body : file.content;
 			const titleLine = contentBody.trim().split("\n")[0];
 			if (titleLine === file.title || titleLine === `# ${file.title}`) contentBody = contentBody.trim().split("\n").slice(1).join("\n").trimStart();
@@ -191,8 +187,7 @@ function generateLlmsFullTxtContent(files, options) {
 function generateMarkdownFilesContent(files) {
 	const markdownFiles = [];
 	for (const file of files) {
-		const urlPath = file.url === "/" ? "index" : file.url.replace(/^\//, "").replace(/\/$/, "");
-		const mdPath = `md/${urlPath}.md`;
+		const mdPath = `md/${file.url === "/" ? "index" : file.url.replace(/^\//, "").replace(/\/$/, "")}.md`;
 		markdownFiles.push({
 			path: mdPath,
 			content: file.content
@@ -220,6 +215,99 @@ async function generateLlmsTxtArtifacts(options) {
 		processedFiles: files
 	};
 }
+/**
+* Create a WritableStream that generates llms.txt artifacts by streaming pages to disk
+*
+* Writes llms.txt (and optionally llms-full.txt) incrementally as pages are written,
+* never keeping full content in memory. Creates outputDir recursively if needed.
+*
+* @example
+* ```typescript
+* const stream = createLlmsTxtStream({
+*   siteName: 'My Docs',
+*   description: 'Documentation site',
+*   origin: 'https://example.com',
+*   generateFull: true,
+*   outputDir: './dist',
+* })
+*
+* const writer = stream.getWriter()
+* await writer.write({
+*   title: 'Home',
+*   content: '# Welcome\n\nHome page content.',
+*   url: '/',
+* })
+* await writer.close()
+* ```
+*
+* @param options - Configuration options
+* @returns WritableStream that accepts ProcessedFile objects
+*/
+function createLlmsTxtStream(options = {}) {
+	const { siteName = "Site", description, origin = "", generateFull, outputDir = process.cwd() } = options;
+	let llmsTxtHandle;
+	let llmsFullTxtHandle;
+	return new WritableStream({
+		async start() {
+			await mkdir(outputDir, { recursive: true });
+			llmsTxtHandle = await open(join(outputDir, "llms.txt"), "w");
+			let header = `# ${siteName}\n\n`;
+			if (description) header += `> ${description}\n\n`;
+			header += `## Pages\n\n`;
+			await llmsTxtHandle.write(header);
+			if (generateFull) {
+				llmsFullTxtHandle = await open(join(outputDir, "llms-full.txt"), "w");
+				let fullHeader = `# ${siteName}\n\n`;
+				if (description) fullHeader += `> ${description}\n\n`;
+				await llmsFullTxtHandle.write(fullHeader);
+			}
+		},
+		async write(file) {
+			const desc = file.metadata?.description;
+			const descText = desc ? `: ${desc.substring(0, 100)}${desc.length > 100 ? "..." : ""}` : "";
+			let chunk = "";
+			if (file.filePath && file.filePath.endsWith(".md")) {
+				const relativePath = relative(outputDir, file.filePath);
+				chunk = `- [${file.title}](${relativePath})${descText}\n`;
+			} else {
+				const url = file.url.startsWith("http://") || file.url.startsWith("https://") ? file.url : origin + file.url;
+				chunk = `- [${file.title}](${url})${descText}\n`;
+			}
+			await llmsTxtHandle?.write(chunk);
+			if (generateFull && llmsFullTxtHandle) {
+				const url = file.url.startsWith("http://") || file.url.startsWith("https://") ? file.url : origin ? origin + file.url : file.url;
+				const { frontmatter, body } = parseFrontmatter(file.content);
+				const metadata = {
+					title: file.title,
+					url
+				};
+				if (file.filePath) metadata.file = relative(outputDir, file.filePath);
+				if (file.metadata) {
+					if (file.metadata.description) metadata.description = file.metadata.description;
+					if (file.metadata.keywords) metadata.keywords = file.metadata.keywords;
+					if (file.metadata.author) metadata.author = file.metadata.author;
+				}
+				const frontmatterString = serializeFrontmatter(frontmatter ? {
+					...frontmatter,
+					...metadata
+				} : metadata);
+				let contentBody = frontmatter ? body : file.content;
+				const titleLine = contentBody.trim().split("\n")[0];
+				if (titleLine === file.title || titleLine === `# ${file.title}`) contentBody = contentBody.trim().split("\n").slice(1).join("\n").trimStart();
+				const fullChunk = `---\n${frontmatterString}\n---\n\n${contentBody}\n\n---\n\n`;
+				await llmsFullTxtHandle.write(fullChunk);
+			}
+		},
+		async close() {
+			await llmsTxtHandle?.close();
+			await llmsFullTxtHandle?.close();
+		},
+		async abort(reason) {
+			await llmsTxtHandle?.close();
+			await llmsFullTxtHandle?.close();
+		}
+	});
+}
 
 //#endregion
-export { generateLlmsTxtArtifacts };
+export { generateLlmsTxtArtifacts as n, createLlmsTxtStream as t };

package/dist/_chunks/{markdown-processor-f7XT0--8.mjs → markdown-processor-D26Uo5td.mjs}

@@ -1,4 +1,4 @@
-import { BLOCKQUOTE_SPACING, DEFAULT_BLOCK_SPACING, ELEMENT_NODE, HTML_ENTITIES, LIST_ITEM_SPACING, MARKDOWN_CODE_BLOCK, MARKDOWN_EMPHASIS, MARKDOWN_HORIZONTAL_RULE, MARKDOWN_INLINE_CODE, MARKDOWN_STRIKETHROUGH, MARKDOWN_STRONG, MAX_TAG_ID, NO_SPACING, NodeEventEnter, NodeEventExit, TABLE_ROW_SPACING, TAG_A, TAG_ABBR, TAG_ADDRESS, TAG_AREA, TAG_ASIDE, TAG_AUDIO, TAG_B, TAG_BASE, TAG_BDO, TAG_BLOCKQUOTE, TAG_BODY, TAG_BR, TAG_BUTTON, TAG_CANVAS, TAG_CENTER, TAG_CITE, TAG_CODE, TAG_COL, TAG_DD, TAG_DEL, TAG_DETAILS, TAG_DFN, TAG_DIALOG, TAG_DIV, TAG_DL, TAG_DT, TAG_EM, TAG_EMBED, TAG_FIELDSET, TAG_FOOTER, TAG_FORM, TAG_H1, TAG_H2, TAG_H3, TAG_H4, TAG_H5, TAG_H6, TAG_HEAD, TAG_HR, TAG_I, TAG_IFRAME, TAG_IMG, TAG_INPUT, TAG_INS, TAG_KBD, TAG_KEYGEN, TAG_LABEL, TAG_LEGEND, TAG_LI, TAG_LINK, TAG_MAP, TAG_MARK, TAG_META, TAG_METER, TAG_NAV, TAG_NOFRAMES, TAG_NOSCRIPT, TAG_OL, TAG_OPTION, TAG_P, TAG_PARAM, TAG_PLAINTEXT, TAG_PRE, TAG_PROGRESS, TAG_Q, TAG_RP, TAG_RT, TAG_RUBY, TAG_SAMP, TAG_SCRIPT, TAG_SELECT, TAG_SMALL, TAG_SOURCE, TAG_SPAN, TAG_STRONG, TAG_STYLE, TAG_SUB, TAG_SUMMARY, TAG_SUP, TAG_SVG, TAG_TABLE, TAG_TBODY, TAG_TD, TAG_TEMPLATE, TAG_TEXTAREA, TAG_TFOOT, TAG_TH, TAG_THEAD, TAG_TIME, TAG_TITLE, TAG_TR, TAG_TRACK, TAG_U, TAG_UL, TAG_VAR, TAG_VIDEO, TAG_WBR, TAG_XMP, TEXT_NODE, TagIdMap, assembleBufferedContent, collectNodeContent } from "./const-BOAJ1T5c.mjs";
+import { $ as TAG_H2, $t as TAG_TBODY, A as TAG_BUTTON, At as TAG_P, B as TAG_DFN, Bt as TAG_SCRIPT, C as TAG_AUDIO, Ct as TAG_METER, D as TAG_BLOCKQUOTE, E as TAG_BDO, Et as TAG_NOSCRIPT, F as TAG_CODE, Ft as TAG_Q, G as TAG_EM, Gt as TAG_SPAN, H as TAG_DIV, Ht as TAG_SELECT, I as TAG_COL, It as TAG_RP, Jt as TAG_SUB, K as TAG_EMBED, Kt as TAG_STRONG, L as TAG_DD, Lt as TAG_RT, Mt as TAG_PLAINTEXT, N as TAG_CENTER, Nt as TAG_PRE, O as TAG_BODY, Ot as TAG_OL, P as TAG_CITE, Pt as TAG_PROGRESS, Q as TAG_H1, Qt as TAG_TABLE, R as TAG_DEL, Rt as TAG_RUBY, S as TAG_ASIDE, St as TAG_META, T as TAG_BASE, Tt as TAG_NOFRAMES, U as TAG_DL, Ut as TAG_SMALL, V as TAG_DIALOG, W as TAG_DT, Wt as TAG_SOURCE, X as TAG_FOOTER, Xt as TAG_SUP, Yt as TAG_SUMMARY, Z as TAG_FORM, Zt as TAG_SVG, _ as TAG_A, _n as TagIdMap, _t as TAG_LI, a as LIST_ITEM_SPACING, an as TAG_THEAD, b as TAG_AREA, bt as TAG_MAP, c as MARKDOWN_HORIZONTAL_RULE, cn as TAG_TR, ct as TAG_I, d as MARKDOWN_STRONG, dn as TAG_UL, dt as TAG_INPUT, en as TAG_TD, et as TAG_H3, f as MAX_TAG_ID, fn as TAG_VAR, ft as TAG_INS, g as TABLE_ROW_SPACING, gn as TEXT_NODE, gt as TAG_LEGEND, h as NodeEventExit, hn as TAG_XMP, ht as TAG_LABEL, i as HTML_ENTITIES, in as TAG_TH, it as TAG_HEAD, j as TAG_CANVAS, jt as TAG_PARAM, k as TAG_BR, kt as TAG_OPTION, l as MARKDOWN_INLINE_CODE, ln as TAG_TRACK, lt as TAG_IFRAME, m as NodeEventEnter, mn as TAG_WBR, mt as TAG_KEYGEN, n as DEFAULT_BLOCK_SPACING, nn as TAG_TEXTAREA, nt as TAG_H5, o as MARKDOWN_CODE_BLOCK, on as TAG_TIME, ot as TAG_HR, p as NO_SPACING, pn as TAG_VIDEO, pt as TAG_KBD, q as TAG_FIELDSET, qt as TAG_STYLE, r as ELEMENT_NODE, rn as TAG_TFOOT, rt as TAG_H6, s as MARKDOWN_EMPHASIS, sn as TAG_TITLE, t as BLOCKQUOTE_SPACING, tn as TAG_TEMPLATE, tt as TAG_H4, u as MARKDOWN_STRIKETHROUGH, un as TAG_U, ut as TAG_IMG, v as TAG_ABBR, vn as assembleBufferedContent, vt as TAG_LINK, w as TAG_B, wt as TAG_NAV, xt as TAG_MARK, y as TAG_ADDRESS, yn as collectNodeContent, z as TAG_DETAILS, zt as TAG_SAMP } from "./const-Bf_XN9U9.mjs";
 
 //#region src/tags.ts
 function resolveUrl(url, origin) {
@@ -6,15 +6,9 @@ function resolveUrl(url, origin) {
 	if (url.startsWith("//")) return `https:${url}`;
 	if (url.startsWith("#")) return url;
 	if (origin) {
-		if (url.startsWith("/") && origin) {
-			const cleanOrigin = origin.endsWith("/") ? origin.slice(0, -1) : origin;
-			return `${cleanOrigin}${url}`;
-		}
+		if (url.startsWith("/") && origin) return `${origin.endsWith("/") ? origin.slice(0, -1) : origin}${url}`;
 		if (url.startsWith("./")) return `${origin}/${url.slice(2)}`;
-		if (!url.startsWith("http")) {
-			const cleanUrl = url.startsWith("/") ? url.slice(1) : url;
-			return `${origin}/${cleanUrl}`;
-		}
+		if (!url.startsWith("http")) return `${origin}/${url.startsWith("/") ? url.slice(1) : url}`;
 	}
 	return url;
 }
@@ -157,10 +151,7 @@ const tagHandlers = {
 	},
 	[TAG_CODE]: {
 		enter: ({ node }) => {
-			if ((node.depthMap[TAG_PRE] || 0) > 0) {
-				const language = getLanguageFromClass(node.attributes?.class);
-				return `${MARKDOWN_CODE_BLOCK}${language}\n`;
-			}
+			if ((node.depthMap[TAG_PRE] || 0) > 0) return `${MARKDOWN_CODE_BLOCK}${getLanguageFromClass(node.attributes?.class)}\n`;
 			return MARKDOWN_INLINE_CODE;
 		},
 		exit: ({ node }) => {
@@ -179,9 +170,7 @@ const tagHandlers = {
 			if (isInsideTableCell(node)) return "<li>";
 			const depth = (node.depthMap[TAG_UL] || 0) + (node.depthMap[TAG_OL] || 0) - 1;
 			const isOrdered = node.parent?.tagId === TAG_OL;
-			const indent = "  ".repeat(Math.max(0, depth));
-			const marker = isOrdered ? `${node.index + 1}. ` : "- ";
-			return `${indent}${marker}`;
+			return `${"  ".repeat(Math.max(0, depth))}${isOrdered ? `${node.index + 1}. ` : "- "}`;
 		},
 		exit: ({ node }) => isInsideTableCell(node) ? "</li>" : void 0,
 		spacing: LIST_ITEM_SPACING
@@ -194,8 +183,7 @@ const tagHandlers = {
 			if (!node.attributes?.href) return "";
 			const href = resolveUrl(node.attributes?.href || "", state.options?.origin);
 			let title = node.attributes?.title;
-			const lastContent = state.lastContentCache;
-			if (lastContent === title) title = "";
+			if (state.lastContentCache === title) title = "";
 			return title ? `](${href} "${title}")` : `](${href})`;
 		},
 		collapsesInnerWhiteSpace: true,
@@ -204,9 +192,7 @@ const tagHandlers = {
 	},
 	[TAG_IMG]: {
 		enter: ({ node, state }) => {
-			const alt = node.attributes?.alt || "";
-			const src = resolveUrl(node.attributes?.src || "", state.options?.origin);
-			return `![${alt}](${src})`;
+			return `![${node.attributes?.alt || ""}](${resolveUrl(node.attributes?.src || "", state.options?.origin)})`;
 		},
 		collapsesInnerWhiteSpace: true,
 		isSelfClosing: true,
@@ -241,15 +227,14 @@ const tagHandlers = {
 				state.tableRenderedTable = true;
 				const alignments = state.tableColumnAlignments;
 				while (alignments.length < state.tableCurrentRowCells) alignments.push("");
-				const alignmentMarkers = alignments.map((align) => {
+				return ` |\n| ${alignments.map((align) => {
 					switch (align) {
 						case "left": return ":---";
 						case "center": return ":---:";
 						case "right": return "---:";
 						default: return "---";
 					}
-				});
-				return ` |\n| ${alignmentMarkers.join(" | ")} |`;
+				}).join(" | ")} |`;
 			}
 			return " |";
 		},
@@ -637,17 +622,15 @@ function isWhitespace(charCode) {
 */
 function parseHtml(html, options = {}) {
 	const events = [];
-	const state = {
-		depthMap: new Uint8Array(MAX_TAG_ID),
-		depth: 0,
-		plugins: options.plugins || []
-	};
-	const remainingHtml = parseHtmlInternal(html, state, (event) => {
-		events.push(event);
-	});
 	return {
 		events,
-		remainingHtml
+		remainingHtml: parseHtmlInternal(html, {
+			depthMap: new Uint8Array(MAX_TAG_ID),
+			depth: 0,
+			plugins: options.plugins || []
+		}, (event) => {
+			events.push(event);
+		})
 	};
 }
 /**
@@ -840,8 +823,7 @@ function processClosingTag(htmlChunk, position, state, handleEvent) {
 	const chunkLength = htmlChunk.length;
 	let foundClose = false;
 	while (i < chunkLength) {
-		const charCode = htmlChunk.charCodeAt(i);
-		if (charCode === GT_CHAR) {
+		if (htmlChunk.charCodeAt(i) === GT_CHAR) {
 			foundClose = true;
 			break;
 		}
@@ -852,8 +834,7 @@ function processClosingTag(htmlChunk, position, state, handleEvent) {
 		newPosition: position,
 		remainingText: htmlChunk.substring(position)
 	};
-	const tagName = htmlChunk.substring(tagNameStart, i).toLowerCase();
-	const tagId = TagIdMap[tagName] ?? -1;
+	const tagId = TagIdMap[htmlChunk.substring(tagNameStart, i).toLowerCase()] ?? -1;
 	if (state.currentNode?.tagHandler?.isNonNesting && tagId !== state.currentNode.tagId) return {
 		complete: false,
 		newPosition: position,
@@ -885,16 +866,15 @@ function closeNode(node, state, handleEvent) {
 		const prefix = node.attributes?.title || node.attributes?.["aria-label"] || "";
 		if (prefix) {
 			node.childTextNodeIndex = 1;
-			const textNode = {
-				type: TEXT_NODE,
-				value: prefix,
-				parent: node,
-				index: 0,
-				depth: node.depth + 1
-			};
 			handleEvent({
 				type: NodeEventEnter,
-				node: textNode
+				node: {
+					type: TEXT_NODE,
+					value: prefix,
+					parent: node,
+					index: 0,
+					depth: node.depth + 1
+				}
 			});
 			for (const parent of traverseUpToFirstBlockNode(node)) parent.childTextNodeIndex = (parent.childTextNodeIndex || 0) + 1;
 		}
@@ -1262,8 +1242,8 @@ function calculateNewLineConfig(node) {
 function createMarkdownProcessor(options = {}) {
 	const state = {
 		options,
-		regionToggles: new Map(),
-		regionContentBuffers: new Map(),
+		regionToggles: /* @__PURE__ */ new Map(),
+		regionContentBuffers: /* @__PURE__ */ new Map(),
 		depthMap: new Uint8Array(MAX_TAG_ID)
 	};
 	state.regionToggles.set(0, true);
@@ -1315,8 +1295,7 @@ function createMarkdownProcessor(options = {}) {
 			const res = handler[eventFn](context);
 			if (res) output.push(res);
 		}
-		const newLineConfig = calculateNewLineConfig(node);
-		const configuredNewLines = newLineConfig[eventType] || 0;
+		const configuredNewLines = calculateNewLineConfig(node)[eventType] || 0;
 		const newLines = Math.max(0, configuredNewLines - lastNewLines);
 		if (newLines > 0) {
 			if (!buff.length) {
@@ -1332,13 +1311,10 @@ function createMarkdownProcessor(options = {}) {
 				const isInlineElement = node.tagHandler?.isInline;
 				const collapsesWhiteSpace = node.tagHandler?.collapsesInnerWhiteSpace;
 				const hasSpacing = node.tagHandler?.spacing && Array.isArray(node.tagHandler.spacing);
-				const isBlockElement = !isInlineElement && !collapsesWhiteSpace && configuredNewLines > 0;
-				const shouldTrim = (!isInlineElement || eventType === NodeEventExit) && !isBlockElement && !(collapsesWhiteSpace && eventType === NodeEventEnter) && !(hasSpacing && eventType === NodeEventEnter);
-				if (shouldTrim) {
+				if ((!isInlineElement || eventType === NodeEventExit) && !(!isInlineElement && !collapsesWhiteSpace && configuredNewLines > 0) && !(collapsesWhiteSpace && eventType === NodeEventEnter) && !(hasSpacing && eventType === NodeEventEnter)) {
 					const originalLength = lastFragment.length;
 					const trimmed = lastFragment.trimEnd();
-					const trimmedChars = originalLength - trimmed.length;
-					if (trimmedChars > 0) {
+					if (originalLength - trimmed.length > 0) {
 						if (buff?.length && buff[buff.length - 1] === lastFragment) buff[buff.length - 1] = trimmed;
 					}
 				}
@@ -1352,12 +1328,11 @@ function createMarkdownProcessor(options = {}) {
 	* Process HTML string and generate events
 	*/
 	function processHtml(html) {
-		const parseState = {
+		parseHtmlStream(html, {
 			depthMap: state.depthMap,
 			depth: 0,
 			plugins: state.options?.plugins || []
-		};
-		parseHtmlStream(html, parseState, (event) => {
+		}, (event) => {
 			processPluginsForEvent(event, state.options?.plugins, state, processEvent);
 		});
 	}
@@ -1365,18 +1340,14 @@ function createMarkdownProcessor(options = {}) {
 	* Get the final markdown output
 	*/
 	function getMarkdown() {
-		const assembledContent = assembleBufferedContent(state);
-		return assembledContent.trimEnd();
+		return assembleBufferedContent(state).trimEnd();
 	}
 	/**
 	* Get new markdown content since the last call (for streaming)
 	*/
 	function getMarkdownChunk() {
 		const fragments = [];
-		for (const [regionId, content] of Array.from(state.regionContentBuffers.entries())) {
-			const include = state.regionToggles.get(regionId);
-			if (include) fragments.push(...content);
-		}
+		for (const [regionId, content] of Array.from(state.regionContentBuffers.entries())) if (state.regionToggles.get(regionId)) fragments.push(...content);
 		const currentContent = fragments.join("").trimStart();
 		const newContent = currentContent.slice(lastYieldedLength);
 		lastYieldedLength = currentContent.length;
@@ -1393,4 +1364,4 @@ function createMarkdownProcessor(options = {}) {
 const MarkdownProcessor = createMarkdownProcessor;
 
 //#endregion
-export { MarkdownProcessor, createMarkdownProcessor, parseHtml, parseHtmlStream, processPluginsForEvent };
+export { parseHtmlStream as a, parseHtml as i, createMarkdownProcessor as n, processPluginsForEvent as r, MarkdownProcessor as t };

package/dist/_chunks/{minimal-co1tIZYm.mjs → minimal-BiDhcwif.mjs}

@@ -1,5 +1,5 @@
-import { TAG_ASIDE, TAG_BUTTON, TAG_EMBED, TAG_FIELDSET, TAG_FIGURE, TAG_FOOTER, TAG_FORM, TAG_IFRAME, TAG_INPUT, TAG_NAV, TAG_OBJECT, TAG_SELECT, TAG_TEXTAREA } from "./const-BOAJ1T5c.mjs";
-import { filterPlugin, frontmatterPlugin, isolateMainPlugin, tailwindPlugin } from "./plugins-C5_irVJs.mjs";
+import { A as TAG_BUTTON, Dt as TAG_OBJECT, Ht as TAG_SELECT, K as TAG_EMBED, S as TAG_ASIDE, X as TAG_FOOTER, Y as TAG_FIGURE, Z as TAG_FORM, dt as TAG_INPUT, lt as TAG_IFRAME, nn as TAG_TEXTAREA, q as TAG_FIELDSET, wt as TAG_NAV } from "./const-Bf_XN9U9.mjs";
+import { a as filterPlugin, i as frontmatterPlugin, r as isolateMainPlugin, t as tailwindPlugin } from "./plugins-DJnqR2fA.mjs";
 
 //#region src/preset/minimal.ts
 /**
@@ -37,4 +37,4 @@ function withMinimalPreset(options = {}) {
 }
 
 //#endregion
-export { withMinimalPreset };
+export { withMinimalPreset as t };

package/dist/_chunks/{plugin-DrovQriD.mjs → plugin-CjWWQTuL.mjs}

@@ -9,4 +9,4 @@ function createPlugin(plugin) {
 }
 
 //#endregion
-export { createPlugin };
+export { createPlugin as t };

package/dist/_chunks/{plugin-CgnpSqtP.d.mts → plugin-D5soyEXm.d.mts}

@@ -1,4 +1,4 @@
-import { Plugin } from "./types-DqiI86yW.mjs";
+import { d as Plugin } from "./types-CT4ZxeOH.mjs";
 
 //#region src/pluggable/plugin.d.ts
 
@@ -9,4 +9,4 @@ import { Plugin } from "./types-DqiI86yW.mjs";
  */
 declare function createPlugin<T extends Partial<Plugin>>(plugin: T): Plugin;
 //#endregion
-export { createPlugin as createPlugin$1 };
+export { createPlugin as t };