Use full url to download files.

[crisely09]

Related to https://github.com/BlueBrain/nexus-forge/issues/282 The idea is to check if the contentUrl is already expanded and use it as given. When the file_id is passed, then do the standard url construction, i.e. endpoint/org/proj/file_id .

[codecov-commenter]

Codecov Report

Merging #283 (c992e76) into master (e4edeca) will decrease coverage by 0.10%. The diff coverage is 8.69%.

@@            Coverage Diff             @@
##           master     #283      +/-   ##
==========================================
- Coverage   72.85%   72.75%   -0.10%     
==========================================
  Files          88       88              
  Lines        5574     5583       +9     
==========================================
+ Hits         4061     4062       +1     
- Misses       1513     1521       +8     

Flag	Coverage Δ
unittests	72.75% <8.69%> (-0.10%)	:arrow_down:

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
kgforge/specializations/resources/datasets.py	62.88% <0.00%> (ø)
kgforge/specializations/stores/bluebrain_nexus.py	22.43% <7.69%> (+0.09%)	:arrow_up:
kgforge/core/archetypes/store.py	58.87% <12.50%> (-2.70%)	:arrow_down:

[MFSY]

Looks good. Fixed the following errors (introduced in previous merged commits) occurring when downloading:

• Before, _download_many were failing if at least one resource did not have a requested content_type even if other resources in the list had. Now, only resources with the provided content_type are considered when downloading
• Before, Dataset.download() was raising a ValueError when 'follow' was provided because "source" argument has a default value and providing 'source' and 'follow' arguments was not allowed. Now, warnings are raised (asking to choose one of the two and stating that 'source' will be deprecated) and 'follow' is always used when provided.

[crisely09]

I see that you added the solution to the cross_bucket problem here, that is good, two birds in one shot. Maybe we could add an example for downloading using cross_bucket=True if it doesn't already exist.

[MFSY]

I see that you added the solution to the cross_bucket problem here, that is good, two birds in one shot. Maybe we could add an example for downloading using cross_bucket=True if it doesn't already exist.

It is mentioned both in the Querying and Dataset notebooks.