In generator project mix task, pass a UUID map

[josephjclark]

I am writing unit tests against the mix generate_workflow command. I am taking two projects, merging them, and comparing the results.

I've hit a fun problem on new nodes created in the source sandbox (ie, staging). Lightning generates new UUIDs for them in the target project. But I can't predict those UUIDs, so it's very hard to write compelling tests.

This PR allows me to explicitly pass a UUID mapping into the command. So if there's a new node called abc in the source, I want to assign it a UUID of xyz in the target:

mix lightning.generate_workflow staging.json main.json --uuid abc:xyz

Note that in tests I very rarely use actually UUIDs, so the command allows me to pass in fake UUIDs as strings or numbers.

I have deliberately not added a node to the changelog. I'm not sure it's relevant?

Closes #3964

AI Usage

Please disclose how you've used AI in this work (it's cool, we just want to know!):

Claude generated all the code over a series of commits, with some refactoring and design from myself.

• Code generation (copilot but not intellisense)
• Learning or fact checking
• Strategy / design
• Optimisation / refactoring
• Translation / spellchecking / doc gen
• Other
• I have not used AI

You can read more details in our Responsible AI Policy

[midigofrank]

@josephjclark this looks great. As it is, this looks like a fallback but I think it should be the first priority, almost as if the user specifies how they want their workflows/jobs to be mapped. What do you think?

[josephjclark]

Thanks @midigofrank. What are you envisioning by "make it priority"? What sort of changes would you imagine?

I think this is only needed when new nodes and edges are created in the source. Which is a common use-case but also not totally core.

If we want to broaden this up, we could explore a flag that enables seeded "uuids" which are just incrementing numbers. I use this approach all the time in the CLI, and it a least means the ids are predictable.

But tbh this approach here works great for me today and I wouldn't want to spend a lot of time anything more substantial. Not yet anyway :)

[midigofrank]

Aaah I see. I meant having it as the accumulator during the first step of mapping in the merge algorithm. Think of it as them being already mapped nodes

[josephjclark]

Maybe. I think I'd rather just merge it as-is. Given that it's a test util I think I prefer the idea that the implementation is out on the fringes. I also don't want to take a lot of resources on it, given that it works great from my point of view

[midigofrank]

@josephjclark I have made 2 changes:

1. I have renamed the variable to new_uuid_map because it's only used for new items
2. Fixed the typespec warnings

[midigofrank]

@josephjclark by the way why do we need to add both keys in here? And why are we only parsing the source and not the target?

[midigofrank]

Hmm, it's actually this part that's failing in the CI. I'll have a look tomorrow

[josephjclark]

This is probably AI silliness that has not been properly validated by this silly human

I do need string and integer types, although maybe there's a better way I can handle this (like maybe my test code should be casting numbers to strings)

[josephjclark]

Thank you so much @midigofrank 🙏

[codecov]

Codecov Report

❌ Patch coverage is 96.42857% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 88.67%. Comparing base (fcb972a) to head (3a9d740).
⚠️ Report is 11 commits behind head on main.

Files with missing lines	Patch %	Lines
lib/mix/tasks/merge_projects.ex	93.33%	1 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #3965   +/-   ##
=======================================
  Coverage   88.67%   88.67%           
=======================================
  Files         423      423           
  Lines       18926    18944   +18     
=======================================
+ Hits        16782    16798   +16     
- Misses       2144     2146    +2     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.