NXPY-178: Use a uniq ID for the S3 direct upload key

[BoboTiG]

Before the patch, the key was computed using the blob name.
And it would fail if the name contains disallowed characters like parenthesis:

HTTPError(400), error: 'java.lang.IllegalArgumentException: Invalid key: file (1).bin'

So the new code now uses a UUID.

[nuxeojenkins]

View issues in JIRA:

• NXPY-178: Use a uniq ID for the S3 direct upload key
• NXPY-177: Prevent AttributeError when fetching the server version and the response is bad

[sourcery-ai]

Sourcery Code Quality Report (beta)

❌  Merging this PR will decrease code quality in the affected files by 0.01 out of 10.

Quality metrics	Before	After	Change
Complexity	1.01	1.02	0.01 🔴
Method Length	34.29	34.62	0.33 🔴
Quality	9.10	9.09	-0.01 🔴

Other metrics	Before	After	Change
Lines	1901	1923	22

Changed files	Quality Before	Quality After	Quality Change
nuxeo/client.py	8.48	8.48	0.00
nuxeo/models.py	9.53	9.53	0.00
nuxeo/uploads.py	8.57	8.54	-0.03 🔴
nuxeo/handlers/s3.py	8.30	8.29	-0.01 🔴
tests/test_upload_s3.py	8.86	8.84	-0.02 🔴

Here are some functions in these files that still need a tune-up:

File	Function	Complexity	Length	Overall	Recommendation
nuxeo/client.py	NuxeoClient.request	13	254.67	4.46	Split out functionality
nuxeo/handlers/s3.py	ChunkUploaderS3.iter_upload	7	165.92	5.90	Split out functionality
tests/test_upload_s3.py	test_upload_chunked_resume	3	191.74	5.98	Split out functionality
nuxeo/client.py	NuxeoClient.__init__	2	144.79	6.70	Split out functionality

Please see our documentation here for details on how these metrics are calculated.

We are actively working on this report - lots more documentation and extra metrics to come!

Let us know what you think of it by mentioning @sourcery-ai in a comment.