feature-engineering

Discussed in microsoft/nni#4070

[2021-08-14 10:13:41] INFO (NNIDataStore) Datastore initialization done
[2021-08-14 10:13:41] INFO (RestServer) RestServer start
[2021-08-14 10:13:41] INFO (RestServer) RestServer base port is 8080
[2021-08-14 10:13:41] I

• A better name date of birth column in our mock dataset would be birthday.

• src/CMakeLists.txt
• docs/cn/deploy.md
• python/test/sql_smoke_test.py
• src/codec/codec_bench_test.cc

Expected Behavior

When an entity is removed from the feature repo, it should be removed from the feature registry.

Current Behavior

Entities are only added, never removed from the Registry. This is a storage leak of sorts.

Steps to reproduce

• feast init
• In the new feature repo, feast apply
• Add a new entity in the feature repo. feast apply.
• Remove the new en

Problem
Some of our transformers & estimators are not thoroughly tested or not tested at all.

Solution
Use OpTransformerSpec and OpEstimatorSpec base test specs to provide tests for all existing transformers & estimators.

Details in discussion mljar/mljar-supervised#421

Right now, component_graph.get_component expects a string which is the unique name used to find a component in the graph (ex: "My Label Encoder", and not "Label Encoder"). This makes it difficult to find components of a specific class, ex LabelEncoder instances. We should add a method to help with this!

API should take into consideration what happens if we have multiple of the same component