AWS Lambda error No module named 'StringIO' or No module named 'StringIO' - amazon-web-services

Question

I try to use AWS Lambda for mass email sending, the code we use as the link below:
https://aws.amazon.com/cn/premiumsupport/knowledge-center/mass-email-ses-lambda/
from __future__ import print_function
import StringIO
import csv
import json
import os
import urllib
import zlib
from time import strftime, gmtime
from email.mime.multipart import MIMEMultipart
from email.mime.text import MIMEText
import boto3
import botocore
import concurrent.futures
__author__ = 'Said Ali Samed'
__date__ = '10/04/2016'
__version__ = '1.0'
# Get Lambda environment variables
region = os.environ['us-east-1']
max_threads = os.environ['10']
text_message_file = os.environ['email_body.txt']
html_message_file = os.environ['email_body.html']
# Initialize clients
s3 = boto3.client('s3', region_name=region)
ses = boto3.client('ses', region_name=region)
send_errors = []
mime_message_text = ''
mime_message_html = ''
def current_time():
return strftime("%Y-%m-%d %H:%M:%S UTC", gmtime())
def mime_email(subject, from_address, to_address, text_message=None, html_message=None):
msg = MIMEMultipart('alternative')
msg['Subject'] = subject
msg['From'] = from_address
msg['To'] = to_address
if text_message:
msg.attach(MIMEText(text_message, 'plain'))
if html_message:
msg.attach(MIMEText(html_message, 'html'))
return msg.as_string()
def send_mail(from_address, to_address, message):
global send_errors
try:
response = ses.send_raw_email(
Source=from_address,
Destinations=[
to_address,
],
RawMessage={
'Data': message
}
)
if not isinstance(response, dict): # log failed requests only
send_errors.append('%s, %s, %s' % (current_time(), to_address, response))
except botocore.exceptions.ClientError as e:
send_errors.append('%s, %s, %s, %s' %
(current_time(),
to_address,
', '.join("%s=%r" % (k, v) for (k, v) in e.response['ResponseMetadata'].iteritems()),
e.message))
def lambda_handler(event, context):
global send_errors
global mime_message_text
global mime_message_html
try:
# Read the uploaded csv file from the bucket into python dictionary list
bucket = event['Records'][0]['s3']['bucket']['name']
key = urllib.unquote_plus(event['Records'][0]['s3']['object']['key']).decode('utf8')
response = s3.get_object(Bucket=bucket, Key=key)
body = zlib.decompress(response['Body'].read(), 16+zlib.MAX_WBITS)
reader = csv.DictReader(StringIO.StringIO(body),
fieldnames=['from_address', 'to_address', 'subject', 'message'])
# Read the message files
try:
response = s3.get_object(Bucket=bucket, Key=text_message_file)
mime_message_text = response['Body'].read()
except:
mime_message_text = None
print('Failed to read text message file. Did you upload %s?' % text_message_file)
try:
response = s3.get_object(Bucket=bucket, Key=html_message_file)
mime_message_html = response['Body'].read()
except:
mime_message_html = None
print('Failed to read html message file. Did you upload %s?' % html_message_file)
if not mime_message_text and not mime_message_html:
raise ValueError('Cannot continue without a text or html message file.')
# Send in parallel using several threads
e = concurrent.futures.ThreadPoolExecutor(max_workers=max_threads)
for row in reader:
from_address = row['from_address'].strip()
to_address = row['to_address'].strip()
subject = row['subject'].strip()
message = mime_email(subject, from_address, to_address, mime_message_text, mime_message_html)
e.submit(send_mail, from_address, to_address, message)
e.shutdown()
except Exception as e:
print(e.message + ' Aborting...')
raise e
print('Send email complete.')
# Remove the uploaded csv file
try:
response = s3.delete_object(Bucket=bucket, Key=key)
if 'ResponseMetadata' in response.keys() and response['ResponseMetadata']['HTTPStatusCode'] == 204:
print('Removed s3://%s/%s' % (bucket, key))
except Exception as e:
print(e)
# Upload errors if any to S3
if len(send_errors) > 0:
try:
result_data = '\n'.join(send_errors)
logfile_key = key.replace('.csv.gz', '') + '_error.log'
response = s3.put_object(Bucket=bucket, Key=logfile_key, Body=result_data)
if 'ResponseMetadata' in response.keys() and response['ResponseMetadata']['HTTPStatusCode'] == 200:
print('Send email errors saved in s3://%s/%s' % (bucket, logfile_key))
except Exception as e:
print(e)
raise e
# Reset publish error log
send_errors = []
if __name__ == "__main__":
json_content = json.loads(open('event.json', 'r').read())
lambda_handler(json_content, None)
but it has problem when i choose python 2.7.the error is
module initialization error 'us-east-1'
when i choose python 3.6 the error is
Unable to import module 'lambda_function': No module named 'StringIO'
anyone can tell me what is the problem it is ?

Answer 1

From Python v3, the StringIO module has gone. Instead, import the io module and use io.StringIO.
The problem with the v27 version is presumably that the following statement is failing:
region = os.environ['us-east-1']
This will result in a KeyError if us-east-1 is not an available environment variable. Instead use AWS_REGION or AWS_DEFAULT_REGION. See the full list of Lambda environment variables.

Answer 2

Please set the environment variables as described in step 4 of the article:
"Configure Lambda environment variables appropriate to your usage scenario. For example, the following variables would be valid for a given use case:
REGION=us-east-1, MAX_THREADS=10, TEXT_MESSAGE_FILE=email_body.txt, HTML_MESSAGE_FILE=email_body.html."
What was done (as per the code provided in the question) is replacing names of environment variables with their values, which means that python is looking for e.g. 'us-east-1' environment variable which isn't there...
This is the original code
# Get Lambda environment variables
region = os.environ['REGION']
max_threads = os.environ['MAX_THREADS']
text_message_file = os.environ['TEXT_MESSAGE_FILE']
html_message_file = os.environ['HTML_MESSAGE_FILE']
You can also hard-code the values, like below:
# Get Lambda environment variables
region = 'us-east-1'
max_threads = '10'
text_message_file = 'email_body.txt'
html_message_file = 'email_body.html'
but I'd suggest to set the environment variables instead (and use the version of script provided by the article author). When it comes to setting environment variables in Lambda, see this article :)