jdwyah

SemanticLogger.sync!
SemanticLogger.default_level = :trace # Prefab will take over the filtering
SemanticLogger.clear_appenders!
SemanticLogger.add_appender(
  io: $stdout,
  formatter: Rails.env.development? ? :color : :json,
  filter: Prefab.log_filter
)

jdwyah

module EzConfig
  class CallInterceptor < GRPC::ClientInterceptor
    def cancel!
      @call.instance_variable_get("@wrapped").instance_variable_get("@call").cancel
    end

    def server_streamer(request:, call:, method:, metadata:, &block)
      @call = call
      yield
    end

jdwyah

name: landings
image: ezcater-landings
slack: j-growth

deployables:
  web:
    type: web
    port: 3000
    command: "bundle exec rails server -p 3000 -b 0.0.0.0"
    replicas: 2

jdwyah

create or replace function
 public.significance(control_size integer,
              control_conversion integer,
              experiment_size integer,
              experiment_conversion integer)
 returns float
 stable as $$
   from scipy.stats import norm

   def standard_error(sample_size, successes):

jdwyah

     daily AS (

        SELECT
          experiment_arm.day,
          experiment_arm.experiment_test,
          experiment_arm.cumulative_exposures :: INT      experiment_cumulative_exposures,
          experiment_arm.cumulative_order_placed_1 :: INT experiment_cumulative_conversions,

          control_arm.cumulative_exposures :: INT         control_cumulative_exposures,
          control_arm.cumulative_order_placed_1 :: INT    control_cumulative_conversions

jdwyah

      cumulative AS (
        SELECT
          DAY,
          experiment_test,
          experiment_result,
          exposures,
          sum(exposures)
          OVER (
            PARTITION BY experiment_test, experiment_result
            ORDER BY DAY

jdwyah

      results AS (
        SELECT
          event,
          experiment_test,
          experiment_result,
          day,
          count(DISTINCT first_exposures.tracking_id)            exposures,
          count(DISTINCT conversions_order_placed_1.tracking_id) order_placed_1,
          count(DISTINCT conversions_order_placed_1.tracking_id) /
          count(DISTINCT first_exposures.tracking_id) :: FLOAT   conversion_rate

jdwyah

 WITH first_exposures AS (
      SELECT
        tracking_id,
        created_at,
        event,
        experiment_test,
        experiment_result,
        date_trunc('day', created_at) AS day
      FROM warehouse.events
      WHERE category = 'experiment'

jdwyah

CREATE TABLE events (
  event_guid           CHARACTER VARYING(128),
  tracking_id          CHARACTER VARYING(128),
  category             CHARACTER VARYING(128) encode TEXT255,
  sub_category         CHARACTER VARYING(128),
  event                CHARACTER VARYING(128) encode TEXT255,
  event_name           CHARACTER VARYING(128) encode TEXT255,  
  value_decimal        NUMERIC(38, 6),
  value_string         CHARACTER VARYING(255),
  created_at           TIMESTAMP WITHOUT TIME ZONE,

jdwyah

WITH monthly_usage AS (
    SELECT
      user_id,
      date_part('month', age(created_at, '1970-01-01')) +
      12 * date_part('year', age(created_at, '1970-01-01')) AS time_period
    FROM orders
    WHERE order_state = 'completed'
    GROUP BY 1, 2
    ORDER BY 1, 2)