Add create-testnet-data command

[smelc]

Changelog

- description: |
    Add `create-testnet-data` command, that generates files for starting a testnet, with a UI that is more conveninent than `create-staked`
# uncomment types applicable to the change:
  type:
  - feature        # introduces a new feature
  # - breaking       # the API has changed in a breaking way
  - compatible     # the API has changed but is non-breaking
  # - optimisation   # measurable performance improvements
  # - improvement    # QoL changes e.g. refactoring
  # - bugfix         # fixes a defect
  # - test           # fixes/modifies tests
  # - maintenance    # not directly related to the code
  # - release        # related to a new release preparation
  # - documentation  # change in code docs, haddocks...

Context

Fixes #473

How to trust this PR

• Code that is moved from Run/Genesis.hs to Run/CreateTestnetData.hs is not modified
• Hence, existing behavior of Run/Genesis.hs is unchanged.
• The structure of the files created by create-testnet-data is in a golden file, and it is consistent with what we said we would do.

Checklist

• Commit sequence broadly makes sense and commits have useful messages
• New tests are added if needed and existing tests are updated. See Running tests for more details
• Self-reviewed the diff

[smelc]

A question to reviewers:

• Do we want the new create-testnet-data command to be available in all eras? Right now it is, I followed the same pattern as create-staked.

Comments on design choices:

• This is v1 version. For example I intentionally did not tackle this comment nor this one. This can be done in later PRs and I wanted this one to be merged sooner than later. Long-running PRs are harder for authors (damn I should get this merged!) and for reviewers (this PR again!?).
• I chose to share code between create-staked and this command when there was no change to it. Sharing was easier for me and is better for the code health in the long run, whether we ultimately delete create-staked or not. To be sure that the code that was moved was unchanged, run git diff --color-moved main..smelc/new-create-staked and look at the yellow and purple lines.

[carlhammann]

I left quite a few comments, but mostly nitpicks. This is still a messy bit of code, and there are a few things I still feel uneasy about (for example readGenDelegsMap: the way it's implemented to compute the intersection and differences of a few maps, and then apply another transformation to the result, the fact that we even need it (going from numbers to file names seems brittle), the weird error message containing lists of meaningless (to the user) numbers,...) .

Nonetheless, I think this is an improvement. Approved, thank you!

[Jimbo4350]

Maximum supply of lovelace right? A comment would be useful here too.

[smelc]

It's the maximum supply of non-delegated lovelace, not the total supply. I propose to discuss the final API we want on Slack, and follow-up with another PR.

[Jimbo4350]

It would be clearer to factor this out IMO and have a little comment as to what the map will be used for. Could be done in a follow up PR.

[smelc]

Done 👍

[Jimbo4350]

So we use ListMap for performance reasons. You'll need to ask @newhoggy if this is OK.

[newhoggy]

This could be a problem. Please avoid conversion from List to Set and back again, which I think is what happens. Please also ensure that the list can be evaluated lazily.

[newhoggy]

Also double-check that you can create 30 million UTxOs on a 64G laptop: IntersectMBO/cardano-node#3938

[smelc]

@newhoggy> I removed the conversion and will benchmark soon 👍

[carbolymer]

LGTM