Review onboarding experience and get started

[shcheklein]

We've done a lot of things recently greatly simplifying user experience when they first come and try to use DVC through VS Code. Namely, we introduced the onboarding flow where we share DVCLive snippet, we introduced a way to migrate to pipelines (dvc.yaml generation). we create a way to connect Studio and VS Code.

We need to spend some time reviewing everything one more time and update messages, documentation, Gif if needed.

• "Add pipeline button" and its message.
• dvc.yaml that we generate. Do we need to put outputs, etc. More comments.
• Settings screen Consolidate onboarding, connect to Studio, settings & walkthrough #3434
• Get started documentation
• Animation in the docs
• Connect Studio page Simplify Studio onboarding screen and workflow #3864
• Looking through the Studio onboarding from VS Code, do we need the "sign in" button? The "get token" button will take them to the same place if not signed in by @dberenbaum here Simplify Studio onboarding screen and workflow #3864

[dberenbaum]

From https://iterativeai.slack.com/archives/CUSNDR35K/p1686138538724399?thread_ts=1686068465.203069&cid=CUSNDR35K and related to https://github.com/iterative/studio/issues/5933#issuecomment-1580736124, it might be good to explain under experiments that the experiments are saved in Git and the user doesn't need to commit each experiment themselves.

[julieg18]

Reviewed our "pipeline" process and reviewed a couple things that could possibly be improved:

"Create Pipeline" message

• Is just mentioning experiment reproduction enough info to explain what the command is doing? Should we expand a bit more on what a pipeline is or is it most likely clear enough for users new to DVC?

Run Experiment Button

When you click the "Run Experiment" button without dvc.yaml, it runs the create a pipeline command:

Screen.Recording.2023-06-26.at.7.10.30.PM.mov

• Similar to the previous comment, is it obvious enough that creating a pipeline is how we run the experiment? Otherwise, it might seem odd that "Run Experiment" opens a quick pick for setting up a pipeline.

Generated dvc.yaml

# Read about DVC pipeline configuration (https://dvc.org/doc/user-guide/project-structure/dvcyaml-files#stages)
# to customize your stages even more
stages:
  train:
    cmd: python requirements.txt
    deps:
      - requirements.txt

• There's an empty line on top of the file. Minor but the file will look a little cleaner without it.

[julieg18]

After a discussion with Ivan, going to close this issue as most of the tasks are finished and dvc.yaml related tasks are in #4194.