Replies: 1 comment 2 replies
-
Thanks, @skshetry! What about this:
I'd consider machine-readable and greppable (pipeable?) separately, since json for example may not actually be easy to parse with grep/awk/etc. |
Beta Was this translation helpful? Give feedback.
2 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
So, I am starting a discussion for having a general guideline for DVC's output. There are two schools of thoughts:
grep-parseable
).It feels like we try to opt for 2, but a few commands seem to be going for human-readable first:
Making outputs human-readable by default helps us improve and experiment with different output formats. Machine-readable formats could be through some flags (
--plain
,--show-json
, etc.). Though, you lose the ability to grep things by default. (Humans first, Machine second).On the other hand,
grep-parseable
means that the output will be grepable, so machine readable. And, we should also try to make it human-readable. We lose the ability to change output format here though, every change in the output format is going to backward-incompatible and we won't be able to improve/experiment much. (Machine first, human second?)If I take the example of
git
, Git seems to be also mixed, but generally it seems to follow 1) and tries to provide more hints and current state.Git instead provides a
--short
it parseable.On the other hand, heroku's cli opts for being grep-parseable by default and provides a
--json
additionally.I'd prefer 1, as it is flexible for us, but I do see benefits of 2 for DVC. In the end, what matters is consistency (or, consistency with certain exceptions). Let's discuss and try to fix output formats and try to make it consistent.
@dberenbaum @shcheklein @dmpetrov
Beta Was this translation helpful? Give feedback.
All reactions