Support MVTec anomaly data and add UnionMerge

[wonjuleee]

Summary

How to test

Checklist

• I submit my changes into the develop branch
• I have added description of my changes into CHANGELOG
• I have updated the documentation accordingly
• I have added tests to cover my changes
• I have linked related issues

License

• I submit my code changes under the same MIT License that covers the project.
  Feel free to contact the maintainers if that's a concern.
• I have updated the license header for each file (see an example below)

# Copyright (C) 2021 Intel Corporation
#
# SPDX-License-Identifier: MIT

[codecov-commenter]

Codecov Report

Base: 77.60% // Head: 77.72% // Increases project coverage by +0.12% 🎉

Coverage data is based on head (38f665d) compared to base (3cc2e1c).
Patch coverage: 87.17% of modified lines in pull request are covered.

Additional details and impacted files

@@             Coverage Diff             @@
##           develop     #810      +/-   ##
===========================================
+ Coverage    77.60%   77.72%   +0.12%     
===========================================
  Files          179      184       +5     
  Lines        23238    23570     +332     
  Branches      5023     5082      +59     
===========================================
+ Hits         18034    18320     +286     
- Misses        4141     4169      +28     
- Partials      1063     1081      +18     

Flag	Coverage Δ
macos-11_Python-3.10	77.65% <87.17%> (+0.12%)	⬆️
macos-11_Python-3.7	77.23% <86.84%> (+0.13%)	⬆️
macos-11_Python-3.8	77.60% <87.17%> (+0.13%)	⬆️
macos-11_Python-3.9	77.61% <87.17%> (+0.13%)	⬆️
ubuntu-20.04_Python-3.10	?
ubuntu-20.04_Python-3.7	?
ubuntu-20.04_Python-3.8	?
ubuntu-20.04_Python-3.9	?
windows-2019_Python-3.10	77.60% <87.15%> (+0.13%)	⬆️
windows-2019_Python-3.7	77.17% <86.81%> (+0.13%)	⬆️
windows-2019_Python-3.8	77.54% <87.15%> (+0.13%)	⬆️
windows-2019_Python-3.9	77.54% <87.15%> (+0.12%)	⬆️

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

Impacted Files	Coverage Δ
datumaro/cli/commands/merge.py	20.20% <0.00%> (ø)
datumaro/cli/contexts/model.py	21.64% <0.00%> (-0.17%)	⬇️
datumaro/components/project.py	79.18% <0.00%> (-0.06%)	⬇️
datumaro/plugins/data_formats/mvtec/exporter.py	72.72% <72.72%> (ø)
datumaro/plugins/transforms.py	76.72% <75.00%> (+0.03%)	⬆️
datumaro/components/merger.py	76.74% <76.74%> (ø)
datumaro/components/visualizer.py	93.39% <87.50%> (+0.25%)	⬆️
datumaro/components/operations.py	83.82% <87.73%> (-0.19%)	⬇️
datumaro/components/dataset.py	82.30% <90.90%> (-0.04%)	⬇️
datumaro/components/annotation.py	91.50% <91.80%> (+<0.01%)	⬆️
... and 14 more

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

☔ View full report at Codecov.
📢 Do you have feedback about the report comment? Let us know in this issue.

[vinnamkim]

Descriptions are not updated.

[wonjuleee]

I simply updated the description. Do we need to add more comments here?
For instance, "when categories information between subsets, e.g., train and validation, are different, we need to set the merging policy as union. Please refer to the MVTec anomaly dataset."
If so, how about IntersectMerge case?

[vinnamkim]

I don't know much about it except IntersectMerge do more than this to merge annotations. So let's leave it as it is for now. It would be good to enhance those descriptions afterwards. I think that it can be done with a future work we found for refactoring merger codes to resolve their circular dependency problems.

[wonjuleee]

agree, let me prepare the next PR.

[vinnamkim]

LGTM.