Merge branch 'main' into add_beam_serverless_ingest

ai_ta_backend/emails.py

@@ -4,16 +4,25 @@
 from email.mime.multipart import MIMEMultipart
 
 
-def send_email(subject, body_text, sender, receipients):
+def send_email(subject: str, body_text: str, sender: str, receipients: list, bcc_receipients: list):
+    """
+    Send an email using the AWS SES service
+    :param subject: The subject of the email
+    :param body_text: The body of the email
+    :param sender: The email address of the sender
+    :param receipients: A list of email addresses to send the email to
+    :param bcc_receipients: A list of email addresses to send the email to as BCC
+    :return: A string indicating the result of the email send operation
+
+    """
     # Create message content
     message = MIMEMultipart("alternative")
     message["Subject"] = subject
     message["From"] = sender
+    message["To"] = ", ".join(receipients)
 
-    if len(receipients) == 1:
-        message["To"] = receipients[0]
-    else:
-        message["To"] = ", ".join(receipients)
+    if len(bcc_receipients) > 0:
+        message["Bcc"] = ", ".join(bcc_receipients)
 
     # Add plain text part
     part1 = MIMEText(body_text, "plain")
@@ -24,6 +33,6 @@ def send_email(subject, body_text, sender, receipients):
     # Connect to SMTP server
     with smtplib.SMTP_SSL(os.getenv('SES_HOST'), os.getenv('SES_PORT')) as server: # type: ignore
         server.login(os.getenv('USERNAME_SMTP'), os.getenv('PASSWORD_SMTP'))    # type: ignore
-        server.sendmail(sender, receipients, message.as_string())
+        server.sendmail(sender, receipients + bcc_receipients, message.as_string())
 
     return "Email sent successfully!"

ai_ta_backend/export_data.py

@@ -19,7 +19,9 @@
 
 def export_documents_json(course_name: str, from_date='', to_date=''):
   """
-  This function exports the documents to a csv file.
+  This function exports the documents to a json file.
+  1. If the number of documents is greater than 1000, it calls a background task to upload the documents to S3.
+  2. If the number of documents is less than 1000, it fetches the documents and zips them.
   Args:
       course_name (str): The name of the course.
       from_date (str, optional): The start date for the data export. Defaults to ''.
@@ -198,14 +200,27 @@ def export_data_in_bg(response, download_type, course_name, s3_path):
     course_metadata = response.json()
     course_metadata = json.loads(course_metadata['result'])
     admin_emails = course_metadata['course_admins']
+    bcc_emails = []
+
+    # check for Kastan's email and move to bcc
+    if '[email protected]' in admin_emails:
+      admin_emails.remove('[email protected]')
+      bcc_emails.append('[email protected]')
+
+    # add course owner email to admin_emails
     admin_emails.append(course_metadata['course_owner'])
     admin_emails = list(set(admin_emails))
     print("admin_emails: ", admin_emails)
+    print("bcc_emails: ", bcc_emails)
+
+    # add a check for emails, don't send email if no admin emails
+    if len(admin_emails) == 0:
+      return "No admin emails found. Email not sent."
 
     # send email to admins
     subject = "UIUC.chat Data Export Complete for " + course_name
     body_text = "The data export for " + course_name + " is complete.\n\nYou can download the file from the following link: \n\n" + s3_url + "\n\nThis link will expire in 48 hours."
-    email_status = send_email(subject, body_text, os.getenv('EMAIL_SENDER'), admin_emails)
+    email_status = send_email(subject, body_text, os.getenv('EMAIL_SENDER'), admin_emails, bcc_emails)
     print("email_status: ", email_status)
 
     return "File uploaded to S3. Email sent to admins."