Problem description

Be able to read public GCS files without providing credentials.

Steps/code to reproduce the problem

path = "gs://tensorflow-nightly/prod/tensorflow/release/ubuntu_16/gpu_py37_full/nightly_release/18/20190813-010608/github/tensorflow/pip_pkg/tf_nightly_gpu-1.15.0.dev20190813-cp37-cp37m-linux_x86_64.whl"

import smart_open
try:
    f = smart_open.s

As anyone who reads this might, there's classification and regression. There's a special case of regression that some people call "ordinal regression". This involves predicting a number (regression, duh), but on an ordinal discrete scale. For instance, in the IMdB dataset, the goal is to predict a rating that is in {0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10}. In other words, ordinal regression is a mix of

Implement progressive versions of hopping and tumbling windows:

• Both window macro methods should get added versions that take an additional parameter
• The parameter should represent the time interval that should be used to produce intermediate results of aggregations
• The parameter should be a clean divisor of the tumble size for tumbling windows and the hop size for hopping windows

Hello, I have a CSV file that has 9 features and 9 expected targets, and I want to test 2 regression models on this data (that should be generated as a stream).

When I test the MultiTargetRegressionHoeffdingTree and RegressorChain on this data I get a bad R2-score, but when I tried normalizing my data with scikit-learn I get a pretty good R2-score. The problem is that I should not use sci

It is currently hard for users to track which versions of dependencies they are getting and which versions they should use when adding extra dependencies to their projects. This results in code like this in our own example projects:

libraryDependencies ++= Seq(
        "com.lightbend.akka"     %% "akka-stream-alpakka-file"  % "1.1.2",
        "com.typesafe.akka"      %% "akka-http-spray-js

Is your feature request related to a problem? Please describe.
Today the user needs to deploy udf jars and reference data csvs manually to the blob location

Describe the solution you'd like
Enable the user to choose a file on a local disk which the web portal will then upload to the right location

目前要kbar必須一次獲取整日kbar資料。是否能在Streaming新增訂閱Kbar onopen onclose events，在開新bar或收bar時收到資料就好，這樣速度比較快，server loading也比較低，謝謝。

I totally forgot we have machinery for this in the "muti program" tests. We can likely reuse this on the "external ctrl-c" tests as well!