add turndown for m->d formatting

2023-08-26 01:43:17 +12:00 · 2023-08-26 01:43:17 +12:00 · 8c4e16e255
commit 8c4e16e255
parent 27b8c547e3
5 changed files with 297 additions and 16 deletions
--- a/m2d/converters/event-to-message.js
+++ b/m2d/converters/event-to-message.js
@ -2,19 +2,41 @@

 const Ty = require("../../types")
 const DiscordTypes = require("discord-api-types/v10")
-const markdown = require("discord-markdown")
+const chunk = require("chunk-text")
+const TurndownService = require("turndown")

 const passthrough = require("../../passthrough")
 const { sync, db, discord } = passthrough
 /** @type {import("../../matrix/file")} */
 const file = sync.require("../../matrix/file")

+// https://github.com/mixmark-io/turndown/blob/97e4535ca76bb2e70d9caa2aa4d4686956b06d44/src/utilities.js#L26C28-L33C2
+const BLOCK_ELEMENTS = [
+	"ADDRESS", "ARTICLE", "ASIDE", "AUDIO", "BLOCKQUOTE", "BODY", "CANVAS",
+	"CENTER", "DD", "DETAILS", "DIR", "DIV", "DL", "DT", "FIELDSET", "FIGCAPTION", "FIGURE",
+	"FOOTER", "FORM", "FRAMESET", "H1", "H2", "H3", "H4", "H5", "H6", "HEADER",
+	"HGROUP", "HR", "HTML", "ISINDEX", "LI", "MAIN", "MENU", "NAV", "NOFRAMES",
+	"NOSCRIPT", "OL", "OUTPUT", "P", "PRE", "SECTION", "SUMMARY", "TABLE", "TBODY", "TD",
+	"TFOOT", "TH", "THEAD", "TR", "UL"
+]
+
+const turndownService = new TurndownService({
+	hr: "----"
+})
+
+turndownService.addRule("strikethrough", {
+	filter: ["del", "s", "strike"],
+	replacement: function (content) {
+		return "~~" + content + "~~"
+	}
+})
+
 /**
 * @param {Ty.Event.Outer<Ty.Event.M_Room_Message>} event
 */
 function eventToMessage(event) {
 	/** @type {(DiscordTypes.RESTPostAPIWebhookWithTokenJSONBody & {files?: {name: string, file: Buffer}[]})[]} */
-	const messages = []
+	let messages = []

 	let displayName = event.sender
 	let avatarURL = undefined
@ -24,20 +46,51 @@ function eventToMessage(event) {
 		// TODO: get the media repo domain and the avatar url from the matrix member event
 	}

-	if (event.content.msgtype === "m.text") {
-		messages.push({
-			content: event.content.body,
-			username: displayName,
-			avatar_url: avatarURL
-		})
-	} else if (event.content.msgtype === "m.emote") {
-		messages.push({
-			content: `\* _${displayName} ${event.content.body}_`,
-			username: displayName,
-			avatar_url: avatarURL
+	// Convert content depending on what the message is
+	let content = event.content.body // ultimate fallback
+	if (event.content.format === "org.matrix.custom.html" && event.content.formatted_body) {
+		let input = event.content.formatted_body
+		if (event.content.msgtype === "m.emote") {
+			input = `* ${displayName} ${input}`
+		}
+
+		// Note: Element's renderers on Web and Android currently collapse whitespace, like the browser does. Turndown also collapses whitespace which is good for me.
+		// If later I'm using a client that doesn't collapse whitespace and I want turndown to follow suit, uncomment the following line of code, and it Just Works:
+		// input = input.replace(/ /g, "&nbsp;")
+		// There is also a corresponding test to uncomment, named "event2message: whitespace is retained"
+
+		// The matrix spec hasn't decided whether \n counts as a newline or not, but I'm going to count it, because if it's in the data it's there for a reason.
+		// But I should not count it if it's between block elements.
+		input = input.replace(/(<\/?([^ >]+)[^>]*>)?\n(<\/?([^ >]+)[^>]*>)?/g, (whole, beforeContext, beforeTag, afterContext, afterTag) => {
+			if (typeof beforeTag !== "string" && typeof afterTag !== "string") {
+				return "<br>"
+			}
+			beforeContext = beforeContext || ""
+			beforeTag = beforeTag || ""
+			afterContext = afterContext || ""
+			afterTag = afterTag || ""
+			if (!BLOCK_ELEMENTS.includes(beforeTag.toUpperCase()) && !BLOCK_ELEMENTS.includes(afterTag.toUpperCase())) {
+				return beforeContext + "<br>" + afterContext
+			} else {
+				return whole
+			}
 		})
+
+		// @ts-ignore
+		content = turndownService.turndown(input)
+
+		// It's optimised for commonmark, we need to replace the space-space-newline with just newline
+		content = content.replace(/  \n/g, "\n")
 	}

+	// Split into 2000 character chunks
+	const chunks = chunk(content, 2000)
+	messages = messages.concat(chunks.map(content => ({
+		content,
+		username: displayName,
+		avatar_url: avatarURL
+	})))
+
 	return messages
 }

--- a/m2d/converters/event-to-message.test.js
+++ b/m2d/converters/event-to-message.test.js
@ -4,6 +4,12 @@ const {test} = require("supertape")
 const {eventToMessage} = require("./event-to-message")
 const data = require("../../test/data")

+function sameFirstContentAndWhitespace(t, a, b) {
+	const a2 = JSON.stringify(a[0].content)
+	const b2 = JSON.stringify(b[0].content)
+	t.equal(a2, b2)
+}
+
 test("event2message: janky test", t => {
 	t.deepEqual(
 		eventToMessage({
@ -28,6 +34,165 @@ test("event2message: janky test", t => {
 	)
 })

+test("event2message: basic html is converted to markdown", t => {
+	t.deepEqual(
+		eventToMessage({
+			content: {
+				msgtype: "m.text",
+				body: "wrong body",
+				format: "org.matrix.custom.html",
+				formatted_body: "this <strong>is</strong> a <strong><em>test</em></strong> of <del>formatting</del>"
+			},
+			event_id: "$g07oYSZFWBkxohNEfywldwgcWj1hbhDzQ1sBAKvqOOU",
+			origin_server_ts: 1688301929913,
+			room_id: "!kLRqKKUQXcibIMtOpl:cadence.moe",
+			sender: "@cadence:cadence.moe",
+			type: "m.room.message",
+			unsigned: {
+				age: 405299
+			}
+		}),
+		[{
+			username: "cadence",
+			content: "this **is** a **_test_** of ~~formatting~~",
+			avatar_url: undefined
+		}]
+	)
+})
+
+test("event2message: markdown syntax is escaped", t => {
+	t.deepEqual(
+		eventToMessage({
+			content: {
+				msgtype: "m.text",
+				body: "wrong body",
+				format: "org.matrix.custom.html",
+				formatted_body: "this **is** an <strong><em>extreme</em></strong> \\*test\\* of"
+			},
+			event_id: "$g07oYSZFWBkxohNEfywldwgcWj1hbhDzQ1sBAKvqOOU",
+			origin_server_ts: 1688301929913,
+			room_id: "!kLRqKKUQXcibIMtOpl:cadence.moe",
+			sender: "@cadence:cadence.moe",
+			type: "m.room.message",
+			unsigned: {
+				age: 405299
+			}
+		}),
+		[{
+			username: "cadence",
+			content: "this \\*\\*is\\*\\* an **_extreme_** \\\\\\*test\\\\\\* of",
+			avatar_url: undefined
+		}]
+	)
+})
+
+test("event2message: html lines are bridged correctly", t => {
+	t.deepEqual(
+		eventToMessage({
+			content: {
+				msgtype: "m.text",
+				body: "wrong body",
+				format: "org.matrix.custom.html",
+				formatted_body: "<p>paragraph one<br>line <em>two</em><br>line three<br><br>paragraph two\nline <em>two</em>\nline three\n\nparagraph three</p><p>paragraph four\nline two<br>line three\nline four</p>paragraph five"
+			},
+			event_id: "$g07oYSZFWBkxohNEfywldwgcWj1hbhDzQ1sBAKvqOOU",
+			origin_server_ts: 1688301929913,
+			room_id: "!kLRqKKUQXcibIMtOpl:cadence.moe",
+			sender: "@cadence:cadence.moe",
+			type: "m.room.message",
+			unsigned: {
+				age: 405299
+			}
+		}),
+		[{
+			username: "cadence",
+			content: "paragraph one\nline _two_\nline three\n\nparagraph two\nline _two_\nline three\n\nparagraph three\n\nparagraph four\nline two\nline three\nline four\n\nparagraph five",
+			avatar_url: undefined
+		}]
+	)
+})
+
+/*test("event2message: whitespace is retained", t => {
+	t.deepEqual(
+		eventToMessage({
+			content: {
+				msgtype: "m.text",
+				body: "wrong body",
+				format: "org.matrix.custom.html",
+				formatted_body: "line one: test       test<br>line two: <strong>test</strong>       <strong>test</strong><br>line three: <strong>test       test</strong><br>line four: test<strong>       </strong>test<br>       line five"
+			},
+			event_id: "$g07oYSZFWBkxohNEfywldwgcWj1hbhDzQ1sBAKvqOOU",
+			origin_server_ts: 1688301929913,
+			room_id: "!kLRqKKUQXcibIMtOpl:cadence.moe",
+			sender: "@cadence:cadence.moe",
+			type: "m.room.message",
+			unsigned: {
+				age: 405299
+			}
+		}),
+		[{
+			username: "cadence",
+			content: "line one: test       test\nline two: **test**       **test**\nline three: **test       test**\nline four: test       test\n       line five",
+			avatar_url: undefined
+		}]
+	)
+})*/
+
+test("event2message: whitespace is collapsed", t => {
+	sameFirstContentAndWhitespace(
+		t,
+		eventToMessage({
+			content: {
+				msgtype: "m.text",
+				body: "wrong body",
+				format: "org.matrix.custom.html",
+				formatted_body: "line one: test       test<br>line two: <strong>test</strong>       <strong>test</strong><br>line three: <strong>test       test</strong><br>line four: test<strong>       </strong>test<br>       line five"
+			},
+			event_id: "$g07oYSZFWBkxohNEfywldwgcWj1hbhDzQ1sBAKvqOOU",
+			origin_server_ts: 1688301929913,
+			room_id: "!kLRqKKUQXcibIMtOpl:cadence.moe",
+			sender: "@cadence:cadence.moe",
+			type: "m.room.message",
+			unsigned: {
+				age: 405299
+			}
+		}),
+		[{
+			username: "cadence",
+			content: "line one: test test\nline two: **test** **test**\nline three: **test test**\nline four: test test\nline five",
+			avatar_url: undefined
+		}]
+	)
+})
+
+test("event2message: lists are bridged correctly", t => {
+	sameFirstContentAndWhitespace(
+		t,
+		eventToMessage({
+			"type": "m.room.message",
+			"sender": "@cadence:cadence.moe",
+			"content": {
+				"msgtype": "m.text",
+				"body": "* line one\n* line two\n* line three\n  * nested one\n  * nested two\n* line four",
+				"format": "org.matrix.custom.html",
+				"formatted_body": "<ul>\n<li>line one</li>\n<li>line two</li>\n<li>line three\n<ul>\n<li>nested one</li>\n<li>nested two</li>\n</ul>\n</li>\n<li>line four</li>\n</ul>\n"
+			},
+			"origin_server_ts": 1692967314062,
+			"unsigned": {
+				"age": 112,
+				"transaction_id": "m1692967313951.441"
+			},
+			"event_id": "$l-xQPY5vNJo3SNxU9d8aOWNVD1glMslMyrp4M_JEF70",
+			"room_id": "!BpMdOUkWWhFxmTrENV:cadence.moe"
+		}),
+		[{
+			username: "cadence",
+			content: "*   line one\n*   line two\n*   line three\n    *   nested one\n    *   nested two\n*   line four",
+			avatar_url: undefined
+		}]
+	)
+})
+
 test("event2message: long messages are split", t => {
 	t.deepEqual(
 		eventToMessage({
@ -55,3 +220,29 @@ test("event2message: long messages are split", t => {
 		}]
 	)
 })
+
+test("event2message: m.emote markdown syntax is escaped", t => {
+	t.deepEqual(
+		eventToMessage({
+			content: {
+				msgtype: "m.emote",
+				body: "wrong body",
+				format: "org.matrix.custom.html",
+				formatted_body: "shows you **her** <strong><em>extreme</em></strong> \\*test\\* of"
+			},
+			event_id: "$g07oYSZFWBkxohNEfywldwgcWj1hbhDzQ1sBAKvqOOU",
+			origin_server_ts: 1688301929913,
+			room_id: "!kLRqKKUQXcibIMtOpl:cadence.moe",
+			sender: "@cadence:cadence.moe",
+			type: "m.room.message",
+			unsigned: {
+				age: 405299
+			}
+		}),
+		[{
+			username: "cadence",
+			content: "\\* cadence shows you \\*\\*her\\*\\* **_extreme_** \\\\\\*test\\\\\\* of",
+			avatar_url: undefined
+		}]
+	)
+})
--- a/package-lock.json
+++ b/package-lock.json
@ -10,6 +10,7 @@
      "license": "MIT",
      "dependencies": {
        "better-sqlite3": "^8.3.0",
+        "chunk-text": "^2.0.1",
        "cloudstorm": "^0.8.0",
        "discord-markdown": "git+https://git.sr.ht/~cadence/nodejs-discord-markdown#440130ef343c8183a81c7c09809731484aa3a182",
        "heatsync": "^2.4.1",
@ -20,7 +21,8 @@
        "node-fetch": "^2.6.7",
        "prettier-bytes": "^1.0.4",
        "snowtransfer": "^0.8.0",
-        "try-to-catch": "^3.0.1"
+        "try-to-catch": "^3.0.1",
+        "turndown": "^7.1.2"
      },
      "devDependencies": {
        "@types/node": "^18.16.0",
@ -732,6 +734,18 @@
      "resolved": "https://registry.npmjs.org/chownr/-/chownr-1.1.4.tgz",
      "integrity": "sha512-jJ0bqzaylmJtVnNgzTeSOs8DPavpbYgEr/b0YL8/2GO3xJEhInFmhKMUnEJQjZumK7KXGFhUy89PrsJWlakBVg=="
    },
+    "node_modules/chunk-text": {
+      "version": "2.0.1",
+      "resolved": "https://registry.npmjs.org/chunk-text/-/chunk-text-2.0.1.tgz",
+      "integrity": "sha512-ER6TSpe2DT4wjOVOKJ3FFAYv7wE77HA/Ztz88Peiv3lq/2oVMsItYJJsVVI0xNZM8cdImOOTNqlw+LQz7gYdJg==",
+      "dependencies": {
+        "runes": "^0.4.3"
+      },
+      "bin": {
+        "chunk": "bin/server.js",
+        "chunk-text": "bin/server.js"
+      }
+    },
    "node_modules/ci-info": {
      "version": "3.8.0",
      "resolved": "https://registry.npmjs.org/ci-info/-/ci-info-3.8.0.tgz",
@ -1057,6 +1071,11 @@
        "simple-markdown": "^0.7.2"
      }
    },
+    "node_modules/domino": {
+      "version": "2.1.6",
+      "resolved": "https://registry.npmjs.org/domino/-/domino-2.1.6.tgz",
+      "integrity": "sha512-3VdM/SXBZX2omc9JF9nOPCtDaYQ67BGp5CoLpIQlO2KCAPETs8TcDHacF26jXadGbvUteZzRTeos2fhID5+ucQ=="
+    },
    "node_modules/ee-first": {
      "version": "1.1.1",
      "resolved": "https://registry.npmjs.org/ee-first/-/ee-first-1.1.1.tgz",
@ -2646,6 +2665,14 @@
        "node": "*"
      }
    },
+    "node_modules/runes": {
+      "version": "0.4.3",
+      "resolved": "https://registry.npmjs.org/runes/-/runes-0.4.3.tgz",
+      "integrity": "sha512-K6p9y4ZyL9wPzA+PMDloNQPfoDGTiFYDvdlXznyGKgD10BJpcAosvATKrExRKOrNLgD8E7Um7WGW0lxsnOuNLg==",
+      "engines": {
+        "node": ">=4.0.0"
+      }
+    },
    "node_modules/safe-buffer": {
      "version": "5.2.1",
      "resolved": "https://registry.npmjs.org/safe-buffer/-/safe-buffer-5.2.1.tgz",
@ -3216,6 +3243,14 @@
        "node": "*"
      }
    },
+    "node_modules/turndown": {
+      "version": "7.1.2",
+      "resolved": "https://registry.npmjs.org/turndown/-/turndown-7.1.2.tgz",
+      "integrity": "sha512-ntI9R7fcUKjqBP6QU8rBK2Ehyt8LAzt3UBT9JR9tgo6GtuKvyUzpayWmeMKJw1DPdXzktvtIT8m2mVXz+bL/Qg==",
+      "dependencies": {
+        "domino": "^2.1.6"
+      }
+    },
    "node_modules/type-is": {
      "version": "1.6.18",
      "resolved": "https://registry.npmjs.org/type-is/-/type-is-1.6.18.tgz",
--- a/package.json
+++ b/package.json
@ -16,6 +16,7 @@
  "license": "MIT",
  "dependencies": {
    "better-sqlite3": "^8.3.0",
+    "chunk-text": "^2.0.1",
    "cloudstorm": "^0.8.0",
    "discord-markdown": "git+https://git.sr.ht/~cadence/nodejs-discord-markdown#440130ef343c8183a81c7c09809731484aa3a182",
    "heatsync": "^2.4.1",
@ -26,7 +27,8 @@
    "node-fetch": "^2.6.7",
    "prettier-bytes": "^1.0.4",
    "snowtransfer": "^0.8.0",
-    "try-to-catch": "^3.0.1"
+    "try-to-catch": "^3.0.1",
+    "turndown": "^7.1.2"
  },
  "devDependencies": {
    "@types/node": "^18.16.0",
--- a/types.d.ts
+++ b/types.d.ts
@ -67,7 +67,7 @@ export namespace Event {
 	}

 	export type M_Room_Message = {
-		msgtype: "m.text"
+		msgtype: "m.text" | "m.emote"
 		body: string
 		format?: "org.matrix.custom.html"
 		formatted_body?: string