parser.models.S3Object should allow size of >= 0

[byron70]

Attempting to parse a S3 PUT event in an AWS Lambda where the S3 object size is 0 bytes, yields a pydantic.error_wrappers.ValidationError indicating size must be greater than 0.

Expected Behavior

Since 0 byte objects in S3 are allowed, it is expected these types of S3 events would be validated without issue when parsing the Lambda event.

Current Behavior

Receiving the following error attempting to parse a 0 byte S3 PUT event.

parse(event=event, model=models.S3Model, envelope=envelopes.SnsSqsEnvelope)

pydantic.error_wrappers.ValidationError: 1 validation error for S3Model
Records -> 0 -> s3 -> object -> size
  ensure this value is greater than 0 (type=value_error.number.not_gt; limit_value=0)

Possible Solution

Change S3Object to the following:

class S3Object(BaseModel):
    key: str
    size: NonNegativeInt
    eTag: str
    sequencer: str
    versionId: Optional[str]

Steps to Reproduce (for bugs)

from aws_lambda_powertools.utilities.parser import parse, models, envelopes

event = {
    "Records": [
        {
            "eventVersion": "2.1",
            "eventSource": "aws:s3",
            "awsRegion": "us-east-2",
            "eventTime": "2021-03-25T22:41:47.873Z",
            "eventName": "ObjectCreated:Put",
            "userIdentity": {"principalId": "AWS:ARODJFUCGRB5ZBMY74FY:foo.bar"},
            "requestParameters": {"sourceIPAddress": "127.0.0.1"},
            "responseElements": {
                "x-amz-request-id": "foo",
                "x-amz-id-2": "foo",
            },
            "s3": {
                "s3SchemaVersion": "1.0",
                "configurationId": "tf-s3-topic-cur",
                "bucket": {
                    "name": "bucket",
                    "ownerIdentity": {"principalId": "ME"},
                    "arn": "arn:aws:s3:::bucket",
                },
                "object": {
                    "key": "reports/test/test.json",
                    "size": 0,
                    "eTag": "foo",
                    "versionId": "foo",
                    "sequencer": "foo",
                },
            },
        }
    ]
}
parse(event=event, model=models.S3Model)

Environment

aws-lambda-powertools==1.13.0
python 3.8

[michaelbrewer]

@byron70 Otherwise, is there a feature missing in the data for S3 Object?

• https://awslabs.github.io/aws-lambda-powertools-python/utilities/data_classes/#s3
• https://awslabs.github.io/aws-lambda-powertools-python/api/utilities/data_classes/s3_event.html

[byron70]

@byron70 Otherwise, is there a feature missing in the data for S3 Object?

Best I can tell everything else seems to be in order.

[heitorlessa]

Fix is in - @byron70 could you double check in case I've missed anything obvious? Had to upgrade Pydantic to 1.8 as this new type was introduced too.

#375

Once merged, this will make to the next release hopefully mid next week

[heitorlessa]

hey @byron70 this is now out as part of the 1.14.0 release