Add Support For File-Like Object

[SwordSam]

When I read from S3 bucket which is actually a .dlis file:

s3_client = boto3.client('s3') s3_response_object = s3_client.get_object(Bucket=x['bucket'], Key=x['s3_key']) object_content = s3_response_object['Body'].read()

I am not able to pass it to dlisio.load() as it only accepts a path to a file while my object_content is a BytesIO object.
Is there any workaround or fix for it?
Thanks,

[achaikou]

Hi,

Thank you for reporting this issue.

Unfortunately as of now dlisio does not have cloud support [ #163 ]
We consider supporting it in the future and will take it into consideration that we have users who would appreciate the feature.

I am not familiar with File-Like Objects, so the only workaround I can think of is to convert BytesIO into a temporary file first (which seems to be possible) and then read it as normal.

[SwordSam]

Hi,

Thanks for the prompt reply.
Correct temporary files are what I am using as workaround for now.

[jokva]

I am not familiar with File-Like Objects, so the only workaround I can think of is to convert BytesIO into a temporary file first (which seems to be possible) and then read it as normal.

This is pretty much the use case we design lfp for, so "all" we need is a Python file-like backend for it. It's a reasonably easy thing to do, but of course takes a while with testing and whatnot.