twenty-app-intake 0.4.0 → 0.5.0

package/src/logic-functions/replay-bulk.mjs

@@ -521,7 +521,7 @@ var FlattenStage = class {
 
 // src/constants/field-map.ts
 function normaliseKey(key) {
-  return key.replace(/([A-Z])/g, "_$1").toLowerCase().replace(/[-\s.]/g, "_").replace(/_{2,}/g, "_");
+  return key.replace(/([a-z0-9])([A-Z])/g, "$1_$2").replace(/([A-Z]+)([A-Z][a-z])/g, "$1_$2").toLowerCase().replace(/[-\s.]/g, "_").replace(/_{2,}/g, "_");
 }
 var FIELD_MAP = {
   // ── Phone ─────────────────────────────────────────────────────────────────
@@ -828,15 +828,56 @@ function parseTargetObject(raw) {
       return "auto";
   }
 }
-function normalizeFields(flat, rules = []) {
-  const sorted = [...rules].sort((a, b) => b.priority - a.priority);
+var PROTECTED_CANONICAL_NAMES = /* @__PURE__ */ new Set([
+  "email",
+  "phone",
+  "firstName",
+  "lastName",
+  "fullName"
+]);
+function compileRule(rule) {
+  try {
+    return { ...rule, matcher: new RegExp(rule.inputPattern, "i") };
+  } catch {
+    return { ...rule, matcher: null };
+  }
+}
+function ruleMatches(rule, key, lookupKey) {
+  if (rule.inputPattern === key || rule.inputPattern === lookupKey) return true;
+  if (!rule.matcher) return false;
+  return rule.matcher.test(key) || rule.matcher.test(lookupKey);
+}
+function normalizeFields(flat, rules = [], warnings = []) {
+  const compiled = [...rules].sort((a, b) => b.priority - a.priority).map(compileRule);
   return Object.entries(flat).map(
-    ([key, value]) => normalizeField(key, value, sorted)
+    ([key, value]) => normalizeField(key, value, compiled, warnings)
   );
 }
-function normalizeField(key, value, rules) {
+function normalizeField(key, value, rules, warnings) {
   const lookupKey = normaliseKey(key);
   const builtIn = FIELD_MAP[lookupKey];
+  for (const rule of rules) {
+    if (!ruleMatches(rule, key, lookupKey)) continue;
+    if (builtIn && PROTECTED_CANONICAL_NAMES.has(builtIn.canonicalName)) {
+      const displaces = rule.canonicalName !== builtIn.canonicalName || actionForRule(rule) !== "field";
+      if (displaces) {
+        warnings.push(
+          `Rule "${rule.inputPattern}" matched "${key}", which the built-in map holds as the contact's ${builtIn.canonicalName}. The built-in mapping was kept: a contact with no ${builtIn.canonicalName} cannot be found again, so every submission would create a duplicate. Point the rule at a key the built-in map does not already own.`
+        );
+        continue;
+      }
+    }
+    return {
+      originalKey: key,
+      canonicalName: rule.canonicalName,
+      twentyType: typeForRule(rule),
+      action: actionForRule(rule),
+      value,
+      source: "rule",
+      targetObject: parseTargetObject(rule.targetObject),
+      mergeStrategy: parseFieldMergeStrategy(rule.mergeStrategy)
+    };
+  }
   if (builtIn) {
     return {
       originalKey: key,
@@ -848,26 +889,6 @@ function normalizeField(key, value, rules) {
       targetObject: "auto"
     };
   }
-  for (const rule of rules) {
-    let matches = false;
-    try {
-      matches = rule.inputPattern === key || rule.inputPattern === lookupKey || new RegExp(rule.inputPattern, "i").test(key);
-    } catch {
-      matches = rule.inputPattern === key;
-    }
-    if (matches) {
-      return {
-        originalKey: key,
-        canonicalName: rule.canonicalName,
-        twentyType: typeForRule(rule),
-        action: actionForRule(rule),
-        value,
-        source: "rule",
-        targetObject: parseTargetObject(rule.targetObject),
-        mergeStrategy: parseFieldMergeStrategy(rule.mergeStrategy)
-      };
-    }
-  }
   const detectedType = detectType(value);
   const classification = classify(key, value);
   const prefix = OPPORTUNITY_KEY_PREFIXES.find((p) => lookupKey.startsWith(p));
@@ -982,7 +1003,7 @@ var NormalizeStage = class {
       return err(ErrorCode.INTERNAL_ERROR, "FlattenStage must run before NormalizeStage");
     }
     const rules = await this.loadRules(ctx.source?.id);
-    const normalized = normalizeFields(ctx.flat, rules);
+    const normalized = normalizeFields(ctx.flat, rules, ctx.warnings);
     const { crmFields, noteFields } = partition(normalized);
     return ok({
       ...ctx,
@@ -1446,6 +1467,183 @@ function formatValue(value) {
   return String(value);
 }
 
+// src/constants/free-email-domains.ts
+var FREE_EMAIL_DOMAINS = /* @__PURE__ */ new Set([
+  // Google
+  "gmail.com",
+  "googlemail.com",
+  // Microsoft
+  "outlook.com",
+  "outlook.co.uk",
+  "outlook.fr",
+  "outlook.de",
+  "outlook.es",
+  "hotmail.com",
+  "hotmail.co.uk",
+  "hotmail.fr",
+  "hotmail.de",
+  "hotmail.it",
+  "hotmail.es",
+  "live.com",
+  "live.co.uk",
+  "live.nl",
+  "msn.com",
+  "passport.com",
+  // Yahoo
+  "yahoo.com",
+  "yahoo.co.uk",
+  "yahoo.co.jp",
+  "yahoo.ca",
+  "yahoo.com.au",
+  "yahoo.de",
+  "yahoo.fr",
+  "yahoo.es",
+  "yahoo.it",
+  "yahoo.in",
+  "ymail.com",
+  "rocketmail.com",
+  // Apple
+  "icloud.com",
+  "me.com",
+  "mac.com",
+  // AOL / Verizon / legacy US ISPs — common on older consumer accounts
+  "aol.com",
+  "aim.com",
+  "verizon.net",
+  "att.net",
+  "sbcglobal.net",
+  "bellsouth.net",
+  "comcast.net",
+  "cox.net",
+  "charter.net",
+  "earthlink.net",
+  "juno.com",
+  "netzero.net",
+  "roadrunner.com",
+  "rr.com",
+  "optonline.net",
+  "frontier.com",
+  "windstream.net",
+  // Privacy-oriented providers
+  "proton.me",
+  "protonmail.com",
+  "protonmail.ch",
+  "pm.me",
+  "tutanota.com",
+  "tutanota.de",
+  "tuta.io",
+  "tuta.com",
+  "hushmail.com",
+  "fastmail.com",
+  "fastmail.fm",
+  "posteo.de",
+  "mailbox.org",
+  // Other large free providers
+  "gmx.com",
+  "gmx.net",
+  "gmx.de",
+  "gmx.at",
+  "gmx.ch",
+  "mail.com",
+  "email.com",
+  "usa.com",
+  "inbox.com",
+  "zoho.com",
+  "zohomail.com",
+  "yandex.com",
+  "yandex.ru",
+  "ya.ru",
+  "mail.ru",
+  "bk.ru",
+  "list.ru",
+  "inbox.ru",
+  "rambler.ru",
+  "qq.com",
+  "foxmail.com",
+  "163.com",
+  "126.com",
+  "sina.com",
+  "sina.cn",
+  "sohu.com",
+  "naver.com",
+  "hanmail.net",
+  "daum.net",
+  "nate.com",
+  "seznam.cz",
+  "wp.pl",
+  "o2.pl",
+  "onet.pl",
+  "interia.pl",
+  "libero.it",
+  "virgilio.it",
+  "alice.it",
+  "tiscali.it",
+  "orange.fr",
+  "wanadoo.fr",
+  "free.fr",
+  "laposte.net",
+  "sfr.fr",
+  "bbox.fr",
+  "neuf.fr",
+  "web.de",
+  "t-online.de",
+  "freenet.de",
+  "arcor.de",
+  "btinternet.com",
+  "sky.com",
+  "talktalk.net",
+  "virginmedia.com",
+  "ntlworld.com",
+  "bigpond.com",
+  "bigpond.net.au",
+  "optusnet.com.au",
+  "iinet.net.au",
+  "xtra.co.nz",
+  "terra.com.br",
+  "uol.com.br",
+  "bol.com.br",
+  "ig.com.br",
+  "rediffmail.com",
+  "sify.com",
+  "indiatimes.com",
+  // Throwaway / disposable services — never a real employer
+  "mailinator.com",
+  "guerrillamail.com",
+  "guerrillamail.net",
+  "sharklasers.com",
+  "10minutemail.com",
+  "temp-mail.org",
+  "tempmail.com",
+  "throwawaymail.com",
+  "yopmail.com",
+  "trashmail.com",
+  "getnada.com",
+  "dispostable.com",
+  "maildrop.cc",
+  "mintemail.com",
+  "fakeinbox.com",
+  "spamgourmet.com",
+  "mailnesia.com",
+  "moakt.com",
+  "emailondeck.com",
+  "burnermail.io",
+  "anonaddy.com",
+  "simplelogin.io",
+  "duck.com",
+  "relay.firefox.com",
+  // Placeholder domains that show up in test and pasted payloads
+  "example.com",
+  "example.org",
+  "example.net",
+  "test.com",
+  "localhost"
+]);
+function isFreeEmailDomain(host) {
+  const normalised = host.trim().toLowerCase().replace(/^www\./, "").replace(/\.$/, "");
+  if (!normalised) return false;
+  return FREE_EMAIL_DOMAINS.has(normalised);
+}
+
 // src/lib/lookup.ts
 async function firstMatch(collection, query) {
   try {
@@ -1463,11 +1661,16 @@ async function findCompanyByDomain(domain) {
   if (!domain) return null;
   const host = extractHost(domain);
   if (!host) return null;
+  if (isFreeEmailDomain(host)) return null;
   return firstMatch(
     "companies",
     `filter=domainName.primaryLinkUrl[like]:%25${encodeURIComponent(host)}%25`
   );
 }
+function isCompanyDomain(domain) {
+  const host = extractHost(domain);
+  return host.length > 0 && !isFreeEmailDomain(host);
+}
 function extractHost(url) {
   try {
     return new URL(url).hostname.replace(/^www\./, "");
@@ -1588,6 +1791,17 @@ var UpsertRecordStage = class {
     if (companyName && !standard["name"]) standard["name"] = companyName;
     if (!standard["domainName"] && assembled["domainName"]) standard["domainName"] = assembled["domainName"];
     if (!standard["address"] && assembled["address"]) standard["address"] = assembled["address"];
+    const proposedDomain = typeof standard["domainName"] === "string" ? standard["domainName"] : standard["domainName"]?.["primaryLinkUrl"];
+    if (typeof proposedDomain === "string" && proposedDomain && !isCompanyDomain(proposedDomain)) {
+      delete standard["domainName"];
+      ctx.warnings.push(
+        `"${extractHost(proposedDomain)}" is an email provider, not a company domain \u2014 it was left off the company record.`
+      );
+    }
+    if (!standard["name"] && !standard["domainName"]) {
+      ctx.warnings.push("No company name or domain in this payload \u2014 no company was created.");
+      return void 0;
+    }
     const domain = standard["domainName"]?.["primaryLinkUrl"];
     let companyId;
     if (domain) {