thirteen37

id: example_router
routes:
  - id: tensorflow_model
    type: PROXY
    endpoint: "http://api.golabs.io/v2/5e4caccc310000e2cad8c071"
  - id: control
    type: PROXY
    endpoint: "http://api.golabs.io/v2/5e4cacd4310000e1cad8c073"
strategy:
  type: fiber.DefaultTuringRoutingStrategy

thirteen37

id: example_router
routes:
  - id: tensorflow_model
    type: PROXY
    endpoint: "http://api.golabs.io/v2/5e4caccc310000e2cad8c071"
  - id: control
    type: PROXY
    endpoint: "http://api.golabs.io/v2/5e4cacd4310000e1cad8c073"
strategy:
  type: fiber.DefaultTuringRoutingStrategy

thirteen37

# Batch retrieval from warehouse store at training
features = feast_client.get_batch_features(
    entity_rows = df, # dataframe containing entity values to join features to
    feature_refs = [ "humidity", "rainfall", "temperature" ] 
)
# Online retrieval from low-latency store at serving
features = feast_client.get_online_features(
    entity_rows = entity_rows, # list of entity values to join features to
    feature_refs = [ "humidity", "rainfall", "temperature" ] 
)

thirteen37

apiVersion: v2
scheduledPipelines:
- # reference a remote uri that contains the pipeline specification
  pipelineRef: https://source.golabs.io/ds/example/blob/master/cw.yaml
  # specifies when the scheduler should start running this pipeline
  startDate: 2019-02-28
  # the repeat interval for this job, using cron expression
  schedule: "*/5 * * * *"
  # additional no of tries to run the task upon failure
  retries: 1

thirteen37

# name of the pipeline
name: mypipeline
# default Docker image for each task
image: alpine
# default resources required for each task
resources:
  cpu: 2
  memory: 4GB
  disk: 80GB
# default working directory for each task, can reference a Git repo

thirteen37

Yu-Xi's user guide

Principles

• Default to "yes": I trust you know what to do, so you rarely need to ask me for affirmation. Exceptions: you aren't clear what to do, or your choices may cause irreversible harm.
• Default to "open": Ask Me Anything. My intention is to be open and honest. I prefer to over-communicate than under-communicate. Likewise, I greatly value feedback so please be open and honest with me too.

thirteen37

class ApacheFlink < Formula
  desc "Scalable batch and stream data processing"
  homepage "https://flink.apache.org/"
  url "http://www.apache.org/dyn/closer.lua?path=/flink/flink-1.4.0/flink-1.4.0-bin-scala_2.11.tgz"
  version "1.4.0"
  sha256 "e334dd99bb20f14f4be789d8a38253815ee22bcc07defabf388c87d20c7935b4"
  head "https://github.com/apache/flink.git"

  bottle :unneeded

thirteen37

Install Java 8 on C.H.I.P.

C.H.I.P. or PocketC.H.I.P. is a Debian Jessie-based system. We can use the Ubuntu Trusty repos:

sudo sh -c "echo \"deb http://ppa.launchpad.net/webupd8team/java/ubuntu trusty main\" > /etc/apt/sources.list.d/java8.list"
sudo sh -c "echo \"deb-src http://ppa.launchpad.net/webupd8team/java/ubuntu trusty main\" >> /etc/apt/sources.list.d/java8.list"
sudo apt-key adv --keyserver keyserver.ubuntu.com --recv-keys EEA14886
sudo apt-get update
sudo apt-get install -y oracle-java8-installer oracle-java8-set-default

thirteen37

Keybase proof

I hereby claim:

• I am thirteen37 on github.
• I am limyuxi (https://keybase.io/limyuxi) on keybase.
• I have a public key ASBuMYdfi5rEMZ4EQrHbFiidiuwTtueOjYVLNAei-jPXJAo

To claim this, I am signing this object:

thirteen37

The main trick here is the new Multiarch on Ubuntu Saucy and the need to enable it separately for Crouton.

1. Install Crouton

2. Install saucy

   sudo sh -e ~/Downloads/crouton -t xfce -r saucy
   

3. Start the chroot