lite-email-parser 1.0.0 → 1.0.1

package/dist/index.d.ts

@@ -1,2 +1,2 @@
-export {};
+export * from './parser.js';
 //# sourceMappingURL=index.d.ts.map

package/dist/index.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":""}
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA,cAAc,aAAa,CAAA"}

package/dist/index.js

@@ -1,2 +1 @@
-console.log("hi");
-export {};
+export * from './parser.js';

package/dist/parser.d.ts

@@ -0,0 +1,30 @@
+import { ParsedMail } from "mailparser";
+export interface IFile {
+    name?: string;
+    url: string;
+    type?: string;
+    size?: number;
+    tags?: Record<string, string>;
+}
+export interface UploadFileOptions {
+    buffer: Buffer | string;
+    contentType: string;
+    filename: string;
+}
+export type UploadFunction = (file: UploadFileOptions) => Promise<string>;
+export declare function processInlineImages(html: string, cidMap: Map<string, string>, uploader: UploadFunction): Promise<string>;
+export declare function processEmailAttachments(parsedEmail: ParsedMail, uploader: UploadFunction): Promise<{
+    attachments: IFile[];
+    inlineMap: Map<string, string>;
+}>;
+export declare function extractFirstMessage(html: string): string;
+export interface ParseEmailOptions {
+    buffer: Buffer | string;
+    uploader: UploadFunction;
+}
+export interface ParseEmailResult {
+    attachments: IFile[];
+    response: string;
+}
+export declare function parseEmail(options: ParseEmailOptions): Promise<ParseEmailResult>;
+//# sourceMappingURL=parser.d.ts.map

package/dist/parser.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"parser.d.ts","sourceRoot":"","sources":["../src/parser.ts"],"names":[],"mappings":"AAAA,OAAO,EAAgB,UAAU,EAAE,MAAM,YAAY,CAAC;AAItD,MAAM,WAAW,KAAK;IAClB,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,GAAG,EAAE,MAAM,CAAC;IACZ,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,IAAI,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;CACjC;AAED,MAAM,WAAW,iBAAiB;IAC9B,MAAM,EAAE,MAAM,GAAG,MAAM,CAAC;IACxB,WAAW,EAAE,MAAM,CAAC;IACpB,QAAQ,EAAE,MAAM,CAAC;CACpB;AAED,MAAM,MAAM,cAAc,GAAG,CAAC,IAAI,EAAE,iBAAiB,KAAK,OAAO,CAAC,MAAM,CAAC,CAAC;AAE1E,wBAAsB,mBAAmB,CACrC,IAAI,EAAE,MAAM,EACZ,MAAM,EAAE,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,EAC3B,QAAQ,EAAE,cAAc,GACzB,OAAO,CAAC,MAAM,CAAC,CA8CjB;AAED,wBAAsB,uBAAuB,CACzC,WAAW,EAAE,UAAU,EACvB,QAAQ,EAAE,cAAc,GACzB,OAAO,CAAC;IAAE,WAAW,EAAE,KAAK,EAAE,CAAC;IAAC,SAAS,EAAE,GAAG,CAAC,MAAM,EAAE,MAAM,CAAC,CAAA;CAAE,CAAC,CA6CnE;AAED,wBAAgB,mBAAmB,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM,CA0FxD;AAED,MAAM,WAAW,iBAAiB;IAC9B,MAAM,EAAE,MAAM,GAAG,MAAM,CAAC;IACxB,QAAQ,EAAE,cAAc,CAAC;CAC5B;AAED,MAAM,WAAW,gBAAgB;IAC7B,WAAW,EAAE,KAAK,EAAE,CAAC;IACrB,QAAQ,EAAE,MAAM,CAAC;CACpB;AAED,wBAAsB,UAAU,CAAC,OAAO,EAAE,iBAAiB,GAAG,OAAO,CAAC,gBAAgB,CAAC,CAmBtF"}

package/dist/parser.js

@@ -0,0 +1,178 @@
+import { simpleParser } from "mailparser";
+import * as cheerio from "cheerio";
+import { v4 as uuidv4 } from "uuid";
+export async function processInlineImages(html, cidMap, uploader) {
+    const $ = cheerio.load(html);
+    // remove width and height attributes from <img> tags
+    $('img').each((_, el) => {
+        $(el).removeAttr('width');
+        $(el).removeAttr('height');
+    });
+    // replace inline attachments
+    $('img[src^="cid:"]').each((_, el) => {
+        const src = $(el).attr('src');
+        const cid = src?.replace('cid:', '').trim();
+        if (cid && cidMap.has(cid)) {
+            $(el).attr('src', cidMap.get(cid));
+        }
+    });
+    // handle base64 images
+    const base64Images = $('img[src^="data:"]').toArray();
+    for (const el of base64Images) {
+        const src = $(el).attr('src');
+        if (!src)
+            continue;
+        const matches = src.match(/^data:(image\/(\w+));base64,(.*)$/);
+        if (matches) {
+            const contentType = matches[1];
+            const base64Data = matches[3];
+            const filename = `inline-${uuidv4()}`;
+            const buffer = Buffer.from(base64Data, 'base64');
+            try {
+                const url = await uploader({
+                    buffer,
+                    contentType,
+                    filename
+                });
+                $(el).attr('src', url);
+            }
+            catch (err) {
+                console.error(`Failed to upload base64 inline image: ${err}`);
+            }
+        }
+    }
+    return $.html();
+}
+export async function processEmailAttachments(parsedEmail, uploader) {
+    const attachments = [];
+    const inlineMap = new Map();
+    if (parsedEmail.attachments && parsedEmail.attachments.length > 0) {
+        for (const att of parsedEmail.attachments) {
+            const filename = att.filename || uuidv4();
+            const contentType = att.contentType || 'application/octet-stream';
+            try {
+                const publicUrl = await uploader({
+                    buffer: att.content,
+                    contentType,
+                    filename
+                });
+                const attachment = {
+                    name: att.filename,
+                    url: publicUrl,
+                    type: att.contentType,
+                    size: att.size,
+                };
+                attachment.tags = att.related
+                    ? { inlineAttachment: 'true' }
+                    : { attachment: 'true' };
+                attachments.push(attachment);
+                // Map Content-ID to URL (mailparser wraps CID in < >, strip them)
+                if (att.contentId) {
+                    inlineMap.set(att.contentId.replace(/[<>]/g, ''), publicUrl);
+                }
+                // Also map by filename so we can match email-provider proxy URLs
+                if (att.filename) {
+                    inlineMap.set(att.filename, publicUrl);
+                }
+            }
+            catch (error) {
+                console.error(`Failed to upload attachment ${filename}: ${error}`);
+            }
+        }
+    }
+    return { attachments, inlineMap };
+}
+export function extractFirstMessage(html) {
+    const $ = cheerio.load(html);
+    // Gmail signature
+    $('span.gmail_signature_prefix').remove();
+    $('div.gmail_signature').remove();
+    // Gmail quoted replies
+    $('.gmail_chip').remove();
+    $('.gmail_quote').nextAll().remove();
+    $('.gmail_quote').remove();
+    // Outlook signature / forwarded messages
+    $('#divRplyFwdMsg').nextAll().remove();
+    $('#divRplyFwdMsg').remove();
+    // Outlook signature marker: <div id="Signature">
+    $('div#Signature').remove();
+    // Apple Mail signature
+    // Apple Mail wraps signatures in <div dir="ltr" class="AppleMailSignature">
+    $('div.AppleMailSignature').remove();
+    // Apple Mail quoted replies
+    // Apple Mail wraps quoted content in <blockquote type="cite">
+    $('blockquote[type="cite"]').nextAll().remove();
+    $('blockquote[type="cite"]').remove();
+    // Fallback for other email providers
+    // Matches any div whose id or class contains "signature" (case-insensitive)
+    $('div')
+        .filter((_, el) => {
+        const id = ($(el).attr('id') || '').toLowerCase();
+        const cls = ($(el).attr('class') || '').toLowerCase();
+        return id.includes('signature') || cls.includes('signature');
+    })
+        .remove();
+    // Fallback for quoted replies from other email providers
+    // Matches any div whose id or class contains "quote" or "reply" (case-insensitive)
+    $('div')
+        .filter((_, el) => {
+        const id = ($(el).attr('id') || '').toLowerCase();
+        const cls = ($(el).attr('class') || '').toLowerCase();
+        return (id.includes('quote') ||
+            id.includes('reply') ||
+            cls.includes('quote') ||
+            cls.includes('reply'));
+    })
+        .each((_, el) => {
+        $(el).nextAll().remove();
+        $(el).remove();
+    });
+    // Remove confidentiality disclaimers
+    $('body')
+        .find('div, p, font, span, td')
+        .filter(function () {
+        // Skip if this element has child elements that aren't just <br>
+        if ($(this).children().not('br').length > 0) {
+            return false;
+        }
+        const elementText = $(this).text().toLowerCase();
+        return [
+            'this e-mail and any attachments may contain confidential',
+            'sent from'
+        ].some((text) => elementText.includes(text));
+    })
+        .remove();
+    // Clean empty divs
+    $('div')
+        .filter((_, el) => {
+        // Avoid removing divs that are direct children of list items, as this can break bullet points
+        if ($(el).closest('li').length > 0) {
+            return false;
+        }
+        // Avoid removing divs that contain images (inline images, etc.)
+        if ($(el).find('img').length > 0) {
+            return false;
+        }
+        return ($(el).text().trim().length === 0 &&
+            $(el).children().not('br').length === 0);
+    })
+        .remove();
+    return $('body').html() || '';
+}
+export async function parseEmail(options) {
+    const { buffer, uploader } = options;
+    const parsedEmail = await simpleParser(buffer);
+    const { attachments, inlineMap } = await processEmailAttachments(parsedEmail, uploader);
+    let response = parsedEmail.html;
+    if (response) {
+        response = await processInlineImages(response, inlineMap, uploader);
+        response = extractFirstMessage(response);
+    }
+    else {
+        response = parsedEmail.text || '';
+    }
+    return {
+        attachments,
+        response,
+    };
+}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "lite-email-parser",
-  "version": "1.0.0",
+  "version": "1.0.1",
   "description": "Parse email to get signature, replies, attachments and inline images",
   "main": "dist/index.js",
   "types": "dist/index.d.ts",
@@ -9,8 +9,19 @@
   ],
   "scripts": {
     "build": "npx tsc",
-    "start": "node ./dist/src/index.js"
+    "start": "node ./dist/src/index.js",
+    "test": "npx -y tsx sample/sample.ts"
   },
   "author": "philipedc",
-  "type": "module"
+  "type": "module",
+  "dependencies": {
+    "cheerio": "^1.2.0",
+    "mailparser": "^3.9.14",
+    "uuid": "^14.0.1"
+  },
+  "devDependencies": {
+    "@types/mailparser": "^3.4.6",
+    "@types/uuid": "^10.0.0",
+    "typescript": "^7.0.2"
+  }
 }