Update README.md (chimpler#14)

* Update README.md
* Support object lambda

Author: mjgp2

README.md

@@ -6,7 +6,7 @@ In order to support development either on RDS or locally, we implemented our own
 the one provided in RDS. It was implemented in Python using the boto3 library.
 
 ## Installation
-Make sure boto3 is installed using the default Python 2 installed on your computer.
+Make sure boto3 is installed using the default Python 3 installed on your computer.
 On MacOS, this can be done as follows:
 
     sudo /usr/bin/easy_install boto3
@@ -35,7 +35,7 @@ Then install `postgres-aws-s3`:
 
 Finally in Postgres:
 ```postgresql
-psql> CREATE EXTENSION plpythonu;
+psql> CREATE EXTENSION plpython3u;
 psql> CREATE EXTENSION aws_s3;
 ``` 
 
@@ -395,7 +395,7 @@ psql> CREATE EXTENSION aws_s3;
 Set the endpoint url and the aws keys to use s3 (in localstack you can set the aws creds to any non-empty string):
 ```
 psql> SET aws_s3.endpoint_url TO 'http://localstack:4566';
-psql> SET aws_s3.s3.aws_access_key_id TO 'dummy';
+psql> SET aws_s3.aws_access_key_id TO 'dummy';
 psql> SET aws_s3.secret_access_key TO 'dummy';
 ```
 

aws_s3--0.0.1.sql

@@ -82,24 +82,25 @@ AS $$
         'endpoint_url': endpoint_url if endpoint_url else default_aws_settings.get('endpoint_url')
     }
 
-    s3 = boto3.client(
+    s3 = boto3.resource(
         's3',
         region_name=region,
         **aws_settings
     )
 
-    response = s3.head_object(Bucket=bucket, Key=file_path)
+    obj = s3.Object(bucket, file_path)
+    response = obj.get()
     content_encoding = response.get('ContentEncoding')
+    body = response['Body']
 
     with tempfile.NamedTemporaryFile() as fd:
         if content_encoding and content_encoding.lower() == 'gzip':
-            with tempfile.NamedTemporaryFile() as gzfd:
-                s3.download_fileobj(bucket, file_path, gzfd)
-                gzfd.flush()
-                gzfd.seek(0)
-                shutil.copyfileobj(gzip.GzipFile(fileobj=gzfd, mode='rb'), fd)
+            with gzip.GzipFile(fileobj=body) as gzipfile:
+                while fd.write(gzipfile.read(204800)):
+                    pass
         else:
-                s3.download_fileobj(bucket, file_path, fd)
+            while fd.write(body.read(204800)):
+                pass
         fd.flush()
         formatted_column_list = "({column_list})".format(column_list=column_list) if column_list else ''
         res = plpy.execute("COPY {table_name} {formatted_column_list} FROM {filename} {options};".format(