mailpop 1.0.7 → 1.0.9

package/dist/index.js

@@ -9,6 +9,7 @@ import { readFileSync } from 'fs';
 import path from 'path';
 import { fileURLToPath } from 'url';
 import { normalizeDomain, findWebsiteInRow } from './utils/normalize.js';
+import { verifyEmailFallback } from './utils/validators.js';
 const __dirname = path.dirname(fileURLToPath(import.meta.url));
 const pkg = JSON.parse(readFileSync(path.resolve(__dirname, '../package.json'), 'utf-8'));
 const version = pkg.version || 'unknown';
@@ -135,8 +136,7 @@ Options:
         await Logger.error('app-initialize-fail', undefined, undefined, `Failed to read input CSV headers: ${errorMsg}`);
         process.exit(1);
     }
-    // Construct combined output headers, preserving original columns and adding new ones
-    const outputHeaders = [...inputHeaders];
+    // Construct combined output headers, placing new email columns next to the website column
     const newColumns = [
         'email',
         'email_source',
@@ -144,11 +144,20 @@ Options:
         'confidence_score',
         'discovery_method',
     ];
-    for (const col of newColumns) {
-        if (!outputHeaders.includes(col)) {
-            outputHeaders.push(col);
+    // Filter out any existing occurrences of these columns to avoid duplicates
+    const cleanInputHeaders = inputHeaders.filter((h) => !newColumns.includes(h));
+    // Find standard website column names
+    const websiteKey = cleanInputHeaders.find((h) => ['website', 'websiteurl', 'website_url', 'url', 'site', 'web'].includes(h.toLowerCase().trim()));
+    const outputHeaders = [];
+    for (const h of cleanInputHeaders) {
+        outputHeaders.push(h);
+        if (h === websiteKey) {
+            outputHeaders.push(...newColumns);
         }
     }
+    if (!websiteKey) {
+        outputHeaders.push(...newColumns);
+    }
     // 2. Initialize crawler and browser
     crawlerInstance = new Crawler();
     await crawlerInstance.initialize(config.headless);
@@ -216,15 +225,33 @@ Options:
                     return;
                 }
                 // Map crawling result, retaining all original row keys
+                let selectedEmail = result.selectedEmail ? result.selectedEmail.email : '';
+                let emailSource = result.selectedEmail ? result.selectedEmail.emailSource : '';
+                let emailType = result.selectedEmail ? result.selectedEmail.emailType : '';
+                let confidenceScore = result.selectedEmail
+                    ? String(result.selectedEmail.confidenceScore)
+                    : '';
+                let discoveryMethod = result.selectedEmail ? result.selectedEmail.discoveryMethod : '';
+                // If no email detected, try to fall back to hello@domain
+                if (!selectedEmail) {
+                    const fallbackDomain = normalizeDomain(target.domain);
+                    const fallbackEmail = `hello@${fallbackDomain}`;
+                    const isFallbackValid = await verifyEmailFallback(fallbackEmail);
+                    if (isFallbackValid) {
+                        selectedEmail = fallbackEmail;
+                        emailSource = target.website;
+                        emailType = 'role';
+                        confidenceScore = '50';
+                        discoveryMethod = 'fallback-hello';
+                    }
+                }
                 const outputRow = {
                     ...row,
-                    email: result.selectedEmail ? result.selectedEmail.email : '',
-                    email_source: result.selectedEmail ? result.selectedEmail.emailSource : '',
-                    email_type: result.selectedEmail ? result.selectedEmail.emailType : '',
-                    confidence_score: result.selectedEmail
-                        ? String(result.selectedEmail.confidenceScore)
-                        : '',
-                    discovery_method: result.selectedEmail ? result.selectedEmail.discoveryMethod : '',
+                    email: selectedEmail,
+                    email_source: emailSource,
+                    email_type: emailType,
+                    confidence_score: confidenceScore,
+                    discovery_method: discoveryMethod,
                 };
                 // Append output row incrementally matching the dynamic headers list
                 await appendCsvRow(outputPath, outputRow, outputHeaders, false);

package/dist/utils/validators.js

@@ -1,6 +1,9 @@
 import { normalizeDomain } from './normalize.js';
 import { config } from '../config.js';
+import dns from 'dns/promises';
 const EMAIL_REGEX = /^[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\.[a-zA-Z]{2,}$/;
+const MAX_EMAIL_LENGTH = 254;
+const MAX_LOCAL_PART_LENGTH = 64;
 const REJECTED_PREFIXES = [
     'noreply',
     'no-reply',
@@ -33,6 +36,9 @@ const REJECTED_DOMAINS = [
  * @param email - The email to check.
  */
 export function isValidEmail(email) {
+    if (email.length > MAX_EMAIL_LENGTH) {
+        return false;
+    }
     if (!EMAIL_REGEX.test(email)) {
         return false;
     }
@@ -42,6 +48,9 @@ export function isValidEmail(email) {
     }
     const localPart = parts[0].toLowerCase().trim();
     const domainPart = parts[1].toLowerCase().trim();
+    if (localPart.length > MAX_LOCAL_PART_LENGTH) {
+        return false;
+    }
     // Reject blacklisted prefixes
     if (REJECTED_PREFIXES.includes(localPart)) {
         return false;
@@ -59,6 +68,10 @@ export function isValidEmail(email) {
     if (REJECTED_DOMAINS.includes(domainPart)) {
         return false;
     }
+    // Reject Sentry ingest reporting domains
+    if (domainPart.includes('sentry.io')) {
+        return false;
+    }
     // Simple heuristics for temporary or obviously fake emails
     if (localPart.startsWith('noreply') ||
         localPart.startsWith('no-reply') ||
@@ -93,3 +106,41 @@ export function isDomainMatch(email, targetDomainOrUrl) {
     const targetDomain = normalizeDomain(targetDomainOrUrl);
     return emailDomain === targetDomain || emailDomain.endsWith('.' + targetDomain);
 }
+/**
+ * Verifies if a fallback email is valid using Disify API with local DNS MX lookup fallback.
+ * @param email - The fallback email to verify.
+ */
+export async function verifyEmailFallback(email) {
+    if (!isValidEmail(email)) {
+        return false;
+    }
+    const domain = email.split('@')[1];
+    if (!domain) {
+        return false;
+    }
+    // 1. Try Disify email verification API
+    try {
+        const controller = new AbortController();
+        const timeoutId = setTimeout(() => controller.abort(), 3000);
+        const res = await fetch(`https://api.disify.com/v1/email/${email}`, {
+            signal: controller.signal,
+        });
+        clearTimeout(timeoutId);
+        if (res.ok) {
+            const data = (await res.json());
+            // Email is valid if format matches, not disposable, and DNS MX records exist
+            return data.format && !data.disposable && data.dns;
+        }
+    }
+    catch (_e) {
+        // If API fails, fall back to direct DNS check
+    }
+    // 2. DNS MX records lookup fallback
+    try {
+        const mx = await dns.resolveMx(domain);
+        return mx && mx.length > 0;
+    }
+    catch (_e) {
+        return false;
+    }
+}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "mailpop",
-  "version": "1.0.7",
+  "version": "1.0.9",
   "description": "Production-ready public contact email discovery tool from company websites.",
   "type": "module",
   "main": "dist/index.js",