normalize emails on bulk invite, normalize/lowercase emails on invite… (#2184)

* normalize emails on bulk invite, normalize/lowercase emails on invite matching

* fix validate_email import

backend/danswer/auth/users.py

@@ -8,6 +8,8 @@ from email.mime.text import MIMEText
 from typing import Optional
 from typing import Tuple
 
+from email_validator import EmailNotValidError
+from email_validator import validate_email
 from fastapi import APIRouter
 from fastapi import Depends
 from fastapi import HTTPException
@@ -62,7 +64,6 @@ from danswer.utils.telemetry import optional_telemetry
 from danswer.utils.telemetry import RecordType
 from danswer.utils.variable_functionality import fetch_versioned_implementation
 
-
 logger = setup_logger()
 
 
@@ -105,7 +106,27 @@ def user_needs_to_be_verified() -> bool:
 
 def verify_email_is_invited(email: str) -> None:
     whitelist = get_invited_users()
-    if (whitelist and email not in whitelist) or not email:
+    if not whitelist:
+        return
+
+    if not email:
+        raise PermissionError("Email must be specified")
+
+    email_info = validate_email(email)  # can raise EmailNotValidError
+
+    for email_whitelist in whitelist:
+        try:
+            # normalized emails are now being inserted into the db
+            # we can remove this normalization on read after some time has passed
+            email_info_whitelist = validate_email(email_whitelist)
+        except EmailNotValidError:
+            continue
+
+        # oddly, normalization does not include lowercasing the user part of the
+        # email address ... which we want to allow
+        if email_info.normalized.lower() == email_info_whitelist.normalized.lower():
+            return
+
     raise PermissionError("User not on allowed user whitelist")
 
 

backend/danswer/server/manage/users.py

@@ -2,6 +2,7 @@ import re
 from datetime import datetime
 from datetime import timezone
 
+from email_validator import validate_email
 from fastapi import APIRouter
 from fastapi import Body
 from fastapi import Depends
@@ -159,12 +160,18 @@ def bulk_invite_users(
     emails: list[str] = Body(..., embed=True),
     current_user: User | None = Depends(current_admin_user),
 ) -> int:
+    """emails are string validated. If any email fails validation, no emails are
+    invited and an exception is raised."""
     if current_user is None:
         raise HTTPException(
             status_code=400, detail="Auth is disabled, cannot invite users"
         )
 
-    all_emails = list(set(emails) | set(get_invited_users()))
+    normalized_emails = []
+    for email in emails:
+        email_info = validate_email(email)  # can raise EmailNotValidError
+        normalized_emails.append(email_info.normalized)  # type: ignore
+    all_emails = list(set(normalized_emails) | set(get_invited_users()))
     return write_invited_users(all_emails)