Airbyte Topiary is an open-source configuration and deployment management tool for Airbyte. As this tool is early in development, I highly recommend reading below before using the tool to avoid irreversible (potentially unexpected) changes to your Airbyte deployment.
No releases yet.
git clone github.com/garden-of-delete/airbyte-topiarypython3 -m venv .venv
Activate the environment with . .venv/bin/activatepip3 install requirements.txtpython topiary.py will display help and usage.Airbyte Topiary allows configuration for an airbyte deployment to be moved to and from yaml files through interaction with the Airbyte API. Provided .yml configuration is first validated, but care should be taken to ensure all the details are correct. Check the examples/ directory to see some example configurations.
Sources require the following:
name: a name given to the source. should be unique across the whole Airbyte deployment
sourceName: the name associated with the Airbyte connector. e.g. GitHub, Slack. Used to choose the right connector type when creating a new source.
connectionConfiguration: specific to each source. Check the documentation for that source to get a list. Will include things like:
access_token / api_token / some_other_secret
repository
start_date (provided as a standard timestamp YYYY-MM-DDThr:mm:ssZ)
Optionally, a sourceId (uuid) can be provided to bypass using name to check if the source already exists in the Airbyte deployment during a sync operation.
Same as Sources, but will probably have more destination specific details in the connectionConfiguration section. For example, the BigQuery destination requires something like:
big_query_client_buffer_size_mb: 15
credentials_json: '**********'
dataset_id: somedataset
dataset_location: US
project_id: some-projectOptionally, a destinationId (uuid) can be provided to bypass using name to check if the source already exists in the Airbyte deployment during a sync operation.
Connections require the following:
sourceName or sourceId: used to identify the source. Id will be tried first, then name.
destinationName or destinationId: used to identify the destination. Id will be tried first, then name.
connectionName or connectionId: used to provide a name for a new connection (not visible in Airbyte's GUI), or to target an existing connection for changes
prefix: prefixes the tables produced by the connection. For example github_superset_
namespaceDefinition: tells the connection to use the namespace configuration (schema / dataset information, other details) of the source, destination, or custom. I personally leave the namespace configuration up to the destination (destination).
schedule:
units: number of units of time as an integer
timeUnit: units of time used (hours, days, etc)
status: active or inactive. Note: an "active" connector with a schedule will start a sync attempt in Airbyte immediately upon creation.
Optionally, a syncCatalog can also be specified. This monstrosity is specific to each source and contains the configuration for each of the streams in the connection. Since the syncCatalog as expected by the Airbyte API is not particularly human readable, Topiary provides some options here:
syncCatalog is not provided, Topiary will retrieve the default sync catalog from the source and use that. Note, the default syncCatalog has all available streams selected with the default sync mode (usually "Full Refresh - Overwrite/Append")syncCatalog for an existing connection, I would recommend first syncing the connection to yaml before making changes and applying back to Airbyte. Read the "The Sync Workflow" section below to see how to do this.A connection connecting a GitHub source to a BigQuery destination might look something like this (no SyncCatalog provided, so defaults will be used):
- sourceName: apache/superset
destinationName: community-data-bq
prefix: 'github_superset_'
namespaceDefinition: destination
schedule:
units: 24
timeUnit: hours
status: 'inactive'Airbyte Topiary supports a number of workflows designed to make managing Airbyte deployments at scale easier. These are:
All the sync workflows described below are accessed through the sync master mode like so:
python topiary.py sync ...
In all cases, a configuration origin, which follows the sync command, and a --target are required.
Topiary will use the .yaml or .yml file extensions following the source and --target arguments to choose the right sync workflow.
During setup, Airbyte creates a default workspace called 'default'. Topiary allows the user to specify an alternative existing workspace by name using the optional --workspace argument, followed by the name of the workspace.
The yaml to deployment workflow takes a .yaml file as the origin and applies the configuration contained within to a destination Airbyte deployment.
These optional arguments can be used in combination to define what to apply the sync operation to:
--sources--destinations--connections--all (same as --sources --destinations --connections)Note: if none of the four optional arguments above are given, no changes will be made to the --target.
Basic usage could be something like:
python topiary.py sync config.yml --target http://123.456.789.0:8081 --all
Almost all sources and destinations will have associated secrets. topiary ignores any secrets specified in the source config.yml. Secrets are specified separately using the --secrets argument, followed by a .yaml file. For example:
python topiary.py sync config.yml --target http://123.456.789.0:8081 --secrets secrets.yml --all
There are a number of additional optional parameters that modify how a sync operation is carried out:
--wipe removes all sources, destinations, and connectors before applying config.yml--backup followed by a filename. Dumps the full configuration of airbyte to the specified file before applying --wipe and config.yml--validate validates the sources, destinations, and connections on the destination Airbyte deployment after applying changes.Used together, a realistic invocation of topiary might look something like:
python topiary.py sync config.yml --target http://123.456.789.0:8081 --secrets secrets.yml --all --validate --backup backup_config.yml
If the yaml file specifies connectors with valid sourceId, destinationId, or connectionId matching matching Airbyte deployment, or failing that, valid names, then topiary will attempt to modify the existing source/destination/connection instead of creating a new one.
An existing Airbyte deployment can be written to a .yaml by following the --target argument with a filename having the .yaml or .yml extension. For example:
python topiary.py sync http://123.456.789.0:8081 --target my_deployment.yml
will write the configuration of all sources, destinations, and connections to my_deployment.yml.
Note in this case, no --secrets file is specified, since it has no meaning in this workflow. Secrets can't be extracted from the Airbyte API.
The wipe mode deletes sources, destinations, connections or any combination in an existing Airbyte deployment.
python topiary.py wipe http://123.456.789.0:8081 --all
As with syncing a .yaml file to a deployment, these optional arguments can be used in combination to define what to apply the wipe operation to:
--sources--destinations--connections--all (same as --sources --destinations --connections)Note: the --wipe argument when used in the sync workflow will wipe ALL sources/destinations/connections, not just those specified.
The validate mode validates sources, destinations, connections or any combination in an existing Airbyte deployment.
python topiary.py validate http://123.456.789.0:8081 --all
As with wipe mode, these optional arguments can be used in combination to define what to apply the validate operation to:
--sources--destinations--connections--all (same as --sources --destinations --connections)This is a small project I've been building in my free time, so there isn't much structure needed around contributing (for now). Check the issue list, open an issue for your change if needed, fork the project, modify it, then open a PR :)
Thanks to Abhi and the Airbyte team for being responsive to questions and feedback during the development process. Also big thanks to the team at Preset.io for supporting the concept and my use of Airbyte while employed there.
Copyright 2021 Robert Stolz
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.