Skip to content

coreycaverly/gordian

 
 

Repository files navigation

Gordian

Latest Release Latest PyPi Version codecov Python Build Status Docker Build Status

Gordian applies transformations to files in github repositories and create PRs for the owners of the repositories to review and merge them.

This project grew from a need to keep various kubernetes services consistent and roll out changes at scale. The main use case for this tool is to make changes to configuration files across multiple repositories simultaneously.

Usage

Search and Replace

You can use the docker image to search and replace various strings across repositories. It supports simple and complex search/replace strings operations.

docker run --rm -it argoprojlabs/gordian:latest -h
usage: gordian [-h] [-c CONFIG_FILE] [-g GITHUB_API] --pr PR_MESSAGE [-v] [-d]
               [-b BRANCH] [-t TARGET_BRANCH] [-l PR_LABELS [PR_LABELS ...]]
               [-M | -m | -p]
               [--description DESCRIPTION | --description-file DESCRIPTION_FILE]
               [--force-changelog FORCE_CHANGELOG] -s SEARCH -r REPLACE

optional arguments:
  -h, --help            show this help message and exit
  -c CONFIG_FILE, --config CONFIG_FILE
                        Config file path. (default: config.yaml)
  -g GITHUB_API, --github-api GITHUB_API
                        Github API URL (default: None)
  --pr PR_MESSAGE       Pull request name. (default: None)
  -v, --verbose
  -d, --dry-run         Enable dry run mode (default: False)
  -b BRANCH, --branch BRANCH
                        Branch name to use (default: None)
  -t TARGET_BRANCH, --target-branch TARGET_BRANCH
                        Target branch (default: master)
  -l PR_LABELS [PR_LABELS ...], --labels PR_LABELS [PR_LABELS ...]
                        List of space separated label names you wish to add to
                        your pull request(s) (default: [])
  -F FILE, --file FILE
                        File to change, currently only supported with
                        PlainTextUpdater. (default: None)
  -M, --major           Bump the major version. (default: None)
  -m, --minor           Bump the minor version. (default: None)
  -p, --patch           Bump the patch version. (default: None)
  --description DESCRIPTION
                        Description to be passed to the PR. (default: )
  --description-file DESCRIPTION_FILE
                        Local file path for the description to be passed to
                        the PR. (default: None)
  --force-changelog FORCE_CHANGELOG
                        Fail if changelog does not exist or cannot be parsed
                        (default: None)
  -s SEARCH, --search SEARCH
                        The string to search for in config files. (default:
                        None)
  -r REPLACE, --replace REPLACE
                        The string that will replace instances of the searched
                        string. (default: None)

Simple transformations

You can use the command line interface to make simple changes across various JSON and YAML files, as shown in this example that modifies a kubernetes API Version.

The following command will update the repositories listed in the default config file config.yaml by making a new pull request to the target master branch using a new origin branch update_k8s_apiversion by updating all files that contains apiVersion: apps/v1beta2 to replace it with apiVersion: apps/v1. The new PR will have the labels k8s.1.16 and support assigned to it and the minor version will be bumped.

docker run --rm -it argoprojlabs/gordian:latest -b "update_k8s_apiversion" --pr "update_k8s_apiversion" -s "apiVersion: apps/v1beta2" -r "apiVersion: apps/v1" -l k8s.1.16 -v -m

Complex transformations

You can use the python script interface to make complex changes across various JSON and YAML files, as shown in this example that modifies a kubernetes resource. You can see more examples in the examples directory.

import sys
from gordian.gordian import get_basic_parser, apply_transformations
from gordian.transformations import Transformation

class PreScale(Transformation):

    def __init__(self, args, repo):
        super().__init__(args, repo)
        self.environments = args.environments

    def run(self):
        for env in self.environments:
            objects = self.repo.get_objects(f'overlays/{env}/envconfig-values.yaml')

            min_replicas = None
            for obj in objects:
                if obj['kind'] != 'HorizontalPodAutoscaler':
                    continue

                if obj['spec']['minReplicas'] != obj['spec']['maxReplicas']:
                    min_replicas = obj['spec']['minReplicas']
                    obj['spec']['maxReplicas'] = min_replicas

            if min_replicas is not None:
                objects.save(f'Setting maxRelicas = minReplicas = {min_replicas}', self.dry_run)

        self.repo.changelog.added('Set max replicas equal to min replicas', 'TICKET-1234')
        self.repo.changelog.save('Update changelog', self.dry_run)

if __name__ == '__main__':
    parser = get_basic_parser()
    parser.add_argument(
        '-e', '--environments',
        required=False,
        dest='environments',
        default=['prd'],
        action='append',
        help='Environments to update.'
    )
    args = parser.parse_args(sys.argv[1:])
    apply_transformations(args, [PreScale])

Dependencies

  • config.yaml (required) - list of repositories you wish to modify
  • GIT_USERNAME (optional) - your Github username
  • GIT_PASSWORD (optional) - your Github password or Personal Access Token
  • GIT_TOKEN (optional) - Github Personal Access Token that grants write access to the specified repositories

Authentication

Two methods of authentication are available:

  • Using a Personal Access Token
  • Using a Github Username & Password

A Github Personal Access Token, Github Username and Github Password can also be passed in via the token=, username= and password= named parameters. The passed value will always take precedence over any environment variable. (Added in 3.5.0)

Authentication - Personal Access Token

A Personal Access Token can be used in two ways:

  • Setting the GIT_TOKEN environment variable
  • Passing the token= named parameter

The Personal Access Token must have write access to any specified repositories you wish to submit changes to.

Authentication - Github Username & Password

A Github Username and Password combination can be used in two ways:

  • Setting the GIT_USERNAME and GIT_PASSWORD environment variables
  • Passing the username= and password= parameters

The user must have write access to any specified repositories you wish to submit changes to.

The GIT_PASSWORD or password= may also contain a Personal Access Token instead of the account password.

Development

The simplest way to hit the ground running if you want to contribute with code is using docker.

Launch a python container

localhost$ docker run --rm -it  -v $(pwd):$(pwd) -w $(pwd) python:3.7-stretch bash

Install the project and test dependencies in developer mode

container# pip install -e .[test]

Run the tests

container# pytest
=========================================== test session starts ============================================
platform linux -- Python 3.7.1, pytest-4.5.0, py-1.8.0, pluggy-0.11.0
rootdir: /Users/user/git/argoproj-labs
plugins: requests-mock-1.6.0, cov-2.7.1
collected 33 items

....
================================== 33 passed, 2 warnings in 1.73 seconds ===================================

Support

Creators

Contributors

About

No description, website, or topics provided.

Resources

License

Code of conduct

Stars

Watchers

Forks

Packages

No packages published

Languages

  • Python 98.2%
  • Dockerfile 1.2%
  • Shell 0.6%