@postalsys/email-text-tools 1.1.0 → 1.2.0

package/.ncurc.js

@@ -0,0 +1,8 @@
+module.exports = {
+    upgrade: true,
+    reject: [
+        // Block package upgrades that moved to ESM
+        'node-fetch',
+        'nanoid'
+    ]
+};

package/examples/reply.js

@@ -0,0 +1,23 @@
+'use strict';
+
+const fs = require('fs').promises;
+const extractReply = require('../lib/extract-reply');
+const simpleParser = require('mailparser').simpleParser;
+
+async function main() {
+    const eml = await fs.readFile(process.argv[2]);
+
+    const parsed = await simpleParser(eml);
+
+    const replyText = await extractReply(
+        {
+            html: parsed.html,
+            text: parsed.text
+        },
+        process.env.OPENAI_API_KEY
+    );
+
+    console.log(replyText);
+}
+
+main();

package/lib/extract-reply.js

@@ -0,0 +1,89 @@
+'use strict';
+
+const packageData = require('../package.json');
+const htmlToText = require('../lib/html-to-text');
+const nodeFetch = require('node-fetch');
+const fetchCmd = global.fetch || nodeFetch;
+
+const MAX_TEXT_LENGTH = 6 * 1024;
+const OPENAI_API_URL = 'https://api.openai.com/v1/chat/completions';
+
+async function extractReply(message, apiToken) {
+    let text = message.text;
+    if (!text && message.html) {
+        text = htmlToText(message.text);
+    }
+
+    text = text.trim();
+    if (text.length > MAX_TEXT_LENGTH) {
+        text = text.substring(0, MAX_TEXT_LENGTH);
+    }
+
+    let headers = {
+        'User-Agent': `${packageData.name}/${packageData.version}`,
+        Authorization: `Bearer ${apiToken}`,
+        'Content-Type': 'application/json'
+    };
+
+    let payload = {
+        model: 'gpt-3.5-turbo',
+        messages: [
+            {
+                role: 'user',
+                content: `Extract the reply text without the final courtesy closing from the following email message:
+
+${text}`
+            }
+        ]
+    };
+
+    let res = await fetchCmd(OPENAI_API_URL, {
+        method: 'post',
+        headers,
+        body: JSON.stringify(payload)
+    });
+
+    let data = await res.json();
+
+    if (!res.ok) {
+        if (data && data.error) {
+            let error = new Error(data.error.message || data.error);
+            if (data.error.code) {
+                error.code = data.error.code;
+            }
+
+            error.statusCode = res.status;
+            throw error;
+        }
+
+        let error = new Error('Failed to run API request');
+        error.statusCode = res.status;
+        throw error;
+    }
+
+    if (!data) {
+        throw new Error(`Failed to POST API request`);
+    }
+
+    let output =
+        data &&
+        data.choices &&
+        data.choices
+            .filter(msg => msg && msg.message && msg.message.role === 'assistant' && msg.message.content)
+            .sort((a, b) => ((a && a.index) || 0) - ((b && b.index) || 0))
+            .map(msg => msg.message.content)
+            .join('\n')
+            .trim()
+            .replace(/^\s*"|"\s*$/g, '')
+            .trim();
+
+    const response = {
+        id: data && data.id,
+        tokens: data && data.usage && data.usage.total_tokens,
+        replyText: output
+    };
+
+    return response;
+}
+
+module.exports = extractReply;

package/lib/mime-html.js

@@ -7,6 +7,30 @@ const he = require('he');
 
 const textToHtml = require('./text-to-html');
 
+const FORBID_TAGS = [
+    'title',
+    'link',
+    'meta',
+    'base',
+    'basefont',
+    'frame',
+    'iframe',
+    'frameset',
+    'template',
+    'script',
+    'noframes',
+    'noscript',
+    'object',
+    'embed',
+    'dialog',
+    'canvas',
+    'audio',
+    'video',
+    'applet'
+];
+// keep the style at first
+const FORBID_TAGS_ALL = ['style'].concat(FORBID_TAGS);
+
 // define styles for plaintext emails
 const INLINE_STYLE_BLOCK = `<style>
 
@@ -63,7 +87,7 @@ function mimeHtml(message) {
     const DOMPurify = createDOMPurify(window);
 
     // first pass, outputs a fixed HTML page
-    html = DOMPurify.sanitize(html, { WHOLE_DOCUMENT: true, FORCE_BODY: false, FORBID_TAGS: ['title', 'link', 'meta'] });
+    html = DOMPurify.sanitize(html, { WHOLE_DOCUMENT: true, FORCE_BODY: false, FORBID_TAGS });
 
     try {
         //inline styles
@@ -74,7 +98,12 @@ function mimeHtml(message) {
     }
 
     // second pass, outputs a BODY DOM element
-    const dom = DOMPurify.sanitize(html, { RETURN_DOM: true, WHOLE_DOCUMENT: false, FORCE_BODY: true, FORBID_TAGS: ['style', 'title', 'link', 'meta'] });
+    const dom = DOMPurify.sanitize(html, {
+        RETURN_DOM: true,
+        WHOLE_DOCUMENT: false,
+        FORCE_BODY: true,
+        FORBID_TAGS: FORBID_TAGS_ALL
+    });
 
     // mark all links to be opened in a new window
     for (let elm of dom.querySelectorAll('a')) {

package/lib/text-to-html.js

@@ -134,14 +134,14 @@ function textLinesToHtml(lines) {
             }
         } else if (i < lines.length - 1 && lines[i + 1].trim()) {
             let escaped = linkifyAndHtmlEscape(line);
-            if (/^[ \t]*[-_]+[ \t]*$/.test(escaped)) {
+            if (/^[ \t]*[-_]{2,}[ \t]*$/.test(escaped)) {
                 textparts.push(`<hr />`);
             } else {
                 textparts.push(`${escaped}<br />`);
             }
         } else {
             let escaped = linkifyAndHtmlEscape(line);
-            if (/^[ \t]*[-_]+[ \t]*$/.test(escaped)) {
+            if (/^[ \t]*[-_]{2,}[ \t]*$/.test(escaped)) {
                 escaped = `<hr />`;
             }
             textparts.push(escaped);

package/package.json

@@ -1,6 +1,6 @@
 {
     "name": "@postalsys/email-text-tools",
-    "version": "1.1.0",
+    "version": "1.2.0",
     "description": "Email text tools",
     "main": "index.js",
     "scripts": {
@@ -19,19 +19,21 @@
     "homepage": "https://github.com/postalsys/email-text-tools#readme",
     "devDependencies": {
         "eslint-config-nodemailer": "1.2.0",
-        "eslint-config-prettier": "8.5.0"
+        "eslint-config-prettier": "8.7.0",
+        "mailparser": "3.6.4"
     },
     "dependencies": {
         "dompurify": "3.0.1",
         "handlebars": "4.7.7",
         "he": "1.2.0",
-        "html-to-text": "8.2.1",
+        "html-to-text": "9.0.4",
         "js-beautify": "1.14.7",
         "jsdom": "21.1.0",
         "juice": "8.1.0",
         "linkify-it": "4.0.1",
-        "moment-timezone": "0.5.39",
-        "node-html-parser": "6.1.4",
-        "tlds": "1.233.0"
+        "moment-timezone": "0.5.41",
+        "node-fetch": "2.6.7",
+        "node-html-parser": "6.1.5",
+        "tlds": "1.236.0"
     }
 }