isabella232 / pipelinewise

Data Pipeline Framework using the singer.io spec

PipelineWise

PipelineWise is a Data Pipeline Framework using the Singer.io specification to ingest and replicate data from various sources to various destinations. Documentation is available at https://transferwise.github.io/pipelinewise/

Features

• Built with ELT in mind: PipelineWise fits into the ELT landscape and is not a traditional ETL tool. PipelineWise aims to reproduce the data from the source to an Analytics-Data-Store in as close to the original format as possible. Some minor load time transformations are supported but complex mapping and joins have to be done in the Analytics-Data-Store to extract meaning.

• Managed Schema Changes: When source data changes, PipelineWise detects the change and alters the schema in your Analytics-Data-Store automatically

• Load time transformations: Ideal place to obfuscate, mask or filter sensitive data that should never be replicated in the Data Warehouse

• YAML based configuration: Data pipelines are defined as YAML files, ensuring that the entire configuration is kept under version control

• Lightweight: No daemons or database setup are required

• Extensible: PipelineWise is using Singer.io compatible taps and target connectors. New connectors can be added to PipelineWise with relatively small effort

Table of Contents

• Connectors
  • Running from docker
  • Building from source
• Developing with Docker
• Links
• License

Connectors

Tap extracts data from any source and write it to a standard stream in a JSON-based format, and target consumes data from taps and do something with it, like load it into a file, API or database

Type	Name	Latest Version	Description
Tap	Postgres		Extracts data from PostgreSQL databases. Supporting Log-Based Inremental, Key-Based Incremental and Full Table replications
Tap	MySQL		Extracts data from MySQL databases. Supporting Log-Based Inremental, Key-Based Incremental and Full Table replications
Tap	Oracle		Extracts data from Oracle databases. Supporting Log-Based Inremental, Key-Based Incremental and Full Table replications
Tap	Kafka		Extracts data from Kafka topics
Tap	AdWords		Extracts data Google Ads API (former Google Adwords) using OAuth and support incremental loading based on input state
Tap	S3 CSV		Extracts data from S3 csv files (currently a fork of tap-s3-csv because we wanted to use our own auth method)
Tap	Zendesk		Extracts data from Zendesk using OAuth and Key-Based incremental replications
Tap	Snowflake		Extracts data from Snowflake databases. Supporting Key-Based Incremental and Full Table replications
Tap	Salesforce		Extracts data from Salesforce database using BULK and REST extraction API with Key-Based incremental replications
Tap	Jira		Extracts data from Atlassian Jira using Base auth or OAuth credentials
Target	Postgres		Loads data from any tap into PostgreSQL database
Target	Redshift		Loads data from any tap into Amazon Redshift Data Warehouse
Target	Snowflake		Loads data from any tap into Snowflake Data Warehouse
Target	S3 CSV		Uploads data from any tap to S3 in CSV format
Transform	Field		Transforms fields from any tap and sends the results to any target. Recommended for data masking/ obfuscation

Running from docker

If you have Docker installed then using docker is the easiest and recommended method of start using PipelineWise.

1. Build an executable docker images that has every required dependency and it's isolated from your host system:

   $ docker build -t pipelinewise:latest .

2. Once the image is ready, create an alias to the docker wrapper script:

   $ alias pipelinewise="$(PWD)/bin/pipelinewise-docker"

3. Check if the installation was successfully by running the pipelinewise status command:

   $ pipelinewise status
   
   Tap ID    Tap Type      Target ID     Target Type      Enabled    Status    Last Sync    Last Sync Result
   --------  ------------  ------------  ---------------  ---------  --------  -----------  ------------------
   0 pipeline(s)

You can run any pipelinewise command at this point. Tutorials to create and run pipelines is at https://transferwise.github.io/pipelinewise/installation_guide/creating_pipelines.html .

PS:

For the tests to work, run the commands inside a PipelineWise container. To create, start and get a bash shell in the container:

$ docker-compose up -d pipelinewise
$ docker exec -it pipelinewise_dev bash

Building from source

1. Make sure that every dependencies installed on your system:

   • Python 3.x
   • python3-dev
   • python3-venv
   • postgresql

2. Run the install script that installs the PipelineWise CLI and every supported singer connectors into separated virtual environments:

   $ ./install.sh --connectors=all

   Press Y to accept the license agreement of the required singer components. To automate the installation and accept every license agreement run ./install --acceptlicenses Use the optional --connectors=...,... argument to install only a specific list of singer connectors.

3. To start CLI you need to activate the CLI virtual environment and has to set PIPELINEWISE_HOME environment variable:

   $ source {ACTUAL_ABSOLUTE_PATH}/.virtualenvs/pipelinewise/bin/activate
   $ export PIPELINEWISE_HOME={ACTUAL_ABSOLUTE_PATH}

   (The ACTUAL_ABSOLUTE_PATH differs on every system, the install script prints you the correct command that fits to your system once the installation completed)

4. Check if the installation was successfully by running the pipelinewise status command:

   $ pipelinewise status
   
   Tap ID    Tap Type      Target ID     Target Type      Enabled    Status    Last Sync    Last Sync Result
   --------  ------------  ------------  ---------------  ---------  --------  -----------  ------------------
   0 pipeline(s)

You can run any pipelinewise command at this point. Tutorials to create and run pipelines is at https://transferwise.github.io/pipelinewise/installation_guide/creating_pipelines.html .

To run unit tests:

$ pytest --ignore tests/end_to_end

Note: End-to-end tests are ignored because it requires specific environment with source and target databases. You need to use the docker development environment to run end to end tests. Read more details at Developing with Docker.

To run unit tests and generate code coverage:

$ coverage run -m pytest --ignore tests/end_to_end && coverage report

To generate HTML coverage report.

$ coverage run -m pytest --ignore tests/end_to_end && coverage html -d coverage_html

Note: The HTML report will be generated in coverage_html/index.html

Developing with Docker

If you have Docker and Docker Compose installed, you can create a local development environment that includes not only the PipelineWise executables but a pre-configured development project as well with some databases as source and targets for a more convenient development experience.

For further instructions about setting up local development environment go to Test Project for Docker Development Environment.

Contribution

To add new taps and targets follow the instructions on Contribution Page

Links

• PipelineWise documentation
• Singer ETL specification
• Singer.io community slack channel

License

Apache License Version 2.0

See LICENSE to see the full text.