SegHouse

Send Segment event from S3 files to Data Warehouses.

Features

• Stores Segment events from S3 files to Data Warehouse

• Supported warehouses

  • ClickHouse

• Fixes data type issues

Installation

This tool depends on awscli.

• pip3 install awscli
• aws configure

Install SegHouse.

• pip install seghouse

Python version >= 3.7 is required. If installed python version is less than 3.7 then first install python3.7.

• sudo apt install python3.7
• sudo apt-get install build-essential libssl-dev libffi-dev python3.7-dev
• Then install seghouse using python3.7 -m pip install seghouse command.

• Remember to run all SegHouse commands using python3.7. Example

  python3.7 -m seghouse send ...

Commands Overview

• Send s3 segment files to warehouse.

  • Command : seghouse send --config-file ~/example-seghouse-config.yml --s3-dir "s3://company/clickstream/example_app/android" --namespace example_app_android
  • The command expects to find json.gz files in the S3 path. All these files will be parsed according to Segment Spec and events will be stored in destination warehouses.
  • The configuration file looks like this.

# Configure the warehouses credentials here.
# The user should have create database, create table and add column permissions.
warehouses:
  - type: clickhouse
    host: clickhouse_host_ip
    port: 9000
    user: clickhouse_user
    password: clickhouse_password

# Specify fields that should be skipped
skip_fields:
  - 'field_to_be_skipped_1'
  - 'field_to_be_skipped_2'

# Specify additional timestamp fields
# A new field will be created by converting timestamp to given timezone
extra_timestamps:
  timestamp_ist: Asia/Kolkata