Why does the upload method discards all the data returned by the upload?

[heprotecbuthealsoattac]

So the upload() method discards all the metadata which has to be re-queried later from an unreliable api that fails often. Could you add this metadata to eaither filelink object or keep it in memory somewhere so that it's easy to get?

[bkwi]

Hi @heprotecbuthealsoattac,

First of all, I'm sorry for such a late response.
Which metadata items do you need to re-query for most often? And how often would you say our api fails?
We are currently making some changes to our metadata API and I hope we'll be able to address reliability issues.

[heprotecbuthealsoattac]

@bkwi I need things like: 'url', 'container', 'filename', 'key', 'size'. All this data is availble to the python module after just uploading the file but is discarded later on which is a problem because I need to re-query the metadata api to get it back, however it did fail on me rather often. I was setting up a monitoring system that was uploading the files via python to 'emulate' user's interaction with filestack and had to resort to this kind of 'hack':

from filestack.uploads.multipart import multipart_upload

CLIENT = Client(FS_API_KEY)
metadata = multipart_upload(
        CLIENT.apikey,
        FILE_PATH,
        None,
        CLIENT.storage,
        params={
            'filename': filename,
            'path': path,
            'location': 's3',
            'access': 'private'
        },
        security=CLIENT.security
    )

    r = do_stuff_with_the_metadata(metadata'])

[bkwi]

we could try and add an additional attribute to Filelink class that would be populated only when object was created as a result of a file upload:

flink = client.upload(...)
flink.upload_response  # would contain all the data that was returned
# but
flink = Filelink(<HANDLE>)
flink.upload_response # would be None

[bkwi]

hi @heprotecbuthealsoattac

filestack-python==3.4.0 has just been released. It contains changes described in my previous comment. Sorry it took so long.