At this moment relu_layer op doesn't allow threshold configuration, and legacy RELU op allows that.
We should add configuration option to relu_layer.

Describe the bug
If I pass 'Next 7 days' into the dateRange in query, the response is not a range over the next 7 days but only the day IN 7 days.

To Reproduce
Steps to reproduce the behavior:

1. Build a query like this:

const query: Query = {
    measures: [
      measure
    ],
    timeDimensions: [
      {
        dimension: 'Your.Dimension,
        dateRange: 'Next 7

• Delta Lake 1.0.0
• Spark 3.1.2
• Scala 2.12
• AdoptOpenJDK-11.0.11+9 (build 11.0.11+9)

The following code gives a NullPointerException. This is for a directory-based delta table that does not exist and uses a generated column.

import io.delta.tables.DeltaTable
DeltaTable.create
  .addColumn(
    DeltaTable.columnBuilder("value")
      .generatedAlwaysAs("true")
      .nullab

Travis CI tests are very flaky lately and need some additional check/debugging.

Some of the tests, which are failing:

[info] - should start job, return JobStarted (async) and invoke callback *** FAILED *** (161 milliseconds)
[info]   0 was not equal to 1 (JobManagerActorSpec.scala:487)
[info]   org.scalatest.exceptions.TestFailedException:
[info]   at org.scalatest.MatchersHelper$.

I have a simple regression task (using a LightGBMRegressor) where I want to penalize negative predictions more than positive ones. Is there a way to achieve this with the default regression LightGBM objectives (see https://lightgbm.readthedocs.io/en/latest/Parameters.html)? If not, is it somehow possible to define (many example for default LightGBM model) and pass a custom regression objective?