ml-auto-build-segments

Introduction

This Python Library allows you to automatically add attributes to Parent Segment in Treasure Data and Auto-build Folders and Segments in Audience Studio.

Inputs

• config/auto_build.yml : This is the .yaml config file that defines the params for the Python functions that add new tables to Parent Segment via API, and use the column_name where main_attr: yes param flags which attribute will be used for the segment creation rules. Example config/auto_build.yml file with notes on each param below:

#####################################################################
########################## GLOBAL PARAMS ############################
#####################################################################
sink_database: ml_db                        #database where all model output tables will be saved
unique_user_id: td_canonical_id             #the main join key that will be used to join behaviors tables to customers table      
api_endpoint: 'https://api.treasuredata.com'
ps_name: 'ENTER PS NAME'                 #name of Parent Segment where ML Folders & Audiences will be built
v5_flag: 1                                  #indicate if you want to build Segments in V5 or V4
folder_depth: 10                            #determines how deeply into nested folders in Audience Studio the code will scan to find the Segments you want 
ps_stats_table: 'ps_stats'                  #stores list of all PS, Folders, and Segments in given 

######### Params for new attributes to add to PS #####
src_table:                                  #params for ML Model output table to be added to PS and used for Audience Creation
  name: next_best_product_final
  join_key: td_canonical_id
  cols:                                     #list of columns from src_table to be added to PS as attributes
    - name: nbp_helmets
      type: number
      build_segments: yes                   #if = 'yes', this column is used as the attribute value for building auto segments               
    - name: next_best_product
      type: string
      build_segments: yes                   #if = 'yes', this column is used as the attribute value for building auto segments

########## Auto Build Segment Params ############
rerun_ps: yes                             #if = 'yes', ps will be re-run after the ML model table is added to PS
ml_folder: '[ML] Base Audiences'          #name of Base Folder where all ML Model Audiences will be created
sub_folder: 'Auto-Segmentation'           #name of Sub-Folder where specific Audiences to current WF ML Model will be added
attr_group: 'ML Models'                  #name of Attribute Group where attributes added to Parent Segment will be grouped in Audience Studio

Copyright © 2022 Treasure Data, Inc. (or its affiliates). All rights reserved