Key given by lambda S3 event cannot be used when containing non-ASCII characters

Question

I have a Python lambda script that shrinks images as they are uploaded to S3. When the uploaded filename contains non-ASCII characters (Hebrew in my case), I cannot get the object (Forbidden as if the file doesn't exist).

Here's (some of) my code:

s3_client = boto3.client('s3')
def handler(event, context):
    for record in event['Records']:
        bucket = record['s3']['bucket']['name']
        key = record['s3']['object']['key']
        s3_client.download_file(bucket, key, "/tmp/somefile")

This raises An error occurred (403) when calling the HeadObject operation: Forbidden: ClientError. I also see in the log that the key contains characters like %D7%92.

Following the web I also tried to unquote the key according to some sources (http://blog.rackspace.com/the-devnull-s3-bucket-hacking-with-aws-lambda-and-python/) like so, with no luck:

key = urllib.unquote_plus(record['s3']['object']['key'])

Same error, although this time the log states that I'm trying to retrieve a key with characters like this: פ×קס×.

Note that this script is verified to work on English keys, and the tests were done on keys with no spaces.

Answer 1

#This worked for me
import urllib.parse
encodedStr = 'My+name+is+Tarak'
urllib.parse.unquote_plus(encodedStr)
"My name is Tarak"

Answer 2

I had a similar problem. I solved it adding an encode before doing the unquote:

key = urllib.unquote_plus(event['Records'][0]['s3']['object']['key'].encode("utf8"))