Formatted output for extensibility

[ykakarap]

What would you like to be added:

I would like to add support for formatted outputs like yaml and json. As mentioned in #21 and #27 Zeitgeist can be used by other systems that rely on the output of Zeitgeist to perform some actions. To allow smoother integrations with other systems, eg: build system, dependency bots, etc, we need a machine readable output.

Example:
For the following dependency.yaml :

dependencies:
- name: terraform 
  version: 0.12.3
  upstream:
    flavour: github
    url: hashicorp/terraform
- name: awesome-cli
  version: 79b5b78183f2585689780fbdef1395b06d05ed25
  scheme: random
  upstream:
    flavour: github
    url: project/awesome-cli
    branch: master

The output (in yaml format) can be :

- name: terraform  # example of a component where a new version is detected
  version: 0.12.3
  nextVersion: 0.14.7 # new version available that satisfies the given constraints
- name: awesome-cli  # example of a component where no new version is detected
  version: 79b5b78183f2585689780fbdef1395b06d05ed25
  nextVersion: 79b5b78183f2585689780fbdef1395b06d05ed25

The output format can be specified by --output-format and the accepted values will be json and yaml.
Would also be better to add a -o flag to support writing the output to a target file. Would default to writing the output to dependency_output.(yaml|json)

Why is this needed:

As mentioned in #21 and #27 Zeitgeist can be used by other systems that rely on the output of Zeitgeist to perform some actions. To allow smoother integrations with other systems, eg: build system, dependency bots, etc, we need a machine readable output.

Additional note:
An issue along the same lines is recorded at #21. This issue will be used to track the efforts on this feature as it is going to cover a broader list of output formats, namely yaml and json.

[ykakarap]

/assign

[chlunde]

Cool! What do you think the next step is? I think there are two alternatives:

• Adding subcommands/flags to this binary to perform the replacement
• Adding position output to allow other tools to perform the replacement without reimplementing parts of zeitgeist

I think the schema needs to be extended with offsets and file names for each match/version number zeitgeist finds if option 2 is the path forward. But this would also be a very low level API so I'm not sure if there would be more than one (bug free) consumer of such an API :)

Either way, I think this is a perfect first step for both options 👍

[ykakarap]

The issue is addressed in this #31

/close

[k8s-ci-robot]

@ykakarap: Closing this issue.

In response to this:

The issue is addressed in this #31

/close

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.