feat(domain_debugging): add a pretty output option for domain of expressions

[ca1umac]

Description

Adds expr_domains as an option to the --output-format= argument in the pretty subcommand.
Prints out a JSON of expressions, detailing:

1. a nice format of what the expression is (e.g., a string of the essence syntax)
2. the domain of that expression
3. the children of that expression

Related issues

helping toward #1547 by adding the data source for testing.

Key changes

Adding a new output format

How to test/review

Run cargo run pretty <file> --output-format=expr_domains and manually check that the result is as expected.

[conjure-bot]

Documentation Coverage

Report: https://conjure-cp.github.io/conjure-oxide-reports/pr/1624/coverage-docs/index.html

This PR: 43.48% documented, 4.97% with examples (8/70/161)
Main: 43.48% documented, 4.97% with examples (8/70/161)
Delta: docs +0.00 pp, examples +0.00 pp

[conjure-bot]

Code Coverage

Report: https://conjure-cp.github.io/conjure-oxide-reports/pr/1624/coverage-code/index.html
Diff report: https://conjure-cp.github.io/conjure-oxide-reports/pr/1624/coverage-code/diff-coverage.html

This PR: 74.63% lines (16739/22428)
Diff: 67.00% lines (23/34 measured; 239 changed)
Main: 72.21% lines (16711/23143)
Delta vs main: +2.43 pp

[ozgurakgun]

I like the CLI UI for this! Nitpick: can it be called expression-domains instead of expr_domains please?

[ca1umac]

@ozgurakgun I think this is ready to go. I think we may as well keep the nested structure since it keeps more information than not. I think it'd be more helpful for debugging if nothing else.

[ozgurakgun]

how do you like this? I kept yours as expression-domains-json but I am tempted to keep just this.

[ozgurakgun]

to be clear indentation indicates nesting

[ozgurakgun]

and :: is meant to mean domain-of. like how : is type-of.

[ozgurakgun]

btw

find x,y,a,b : set (maxSize 2) of int(1..3)
such that x union y subsetEq a intersect b

prints

((x union y) subsetEq (a intersect b)) :: bool
 (x union y) :: set  of set  of int(1..3)
  x :: set (maxSize(2)) of int(1..3)
  y :: set (maxSize(2)) of int(1..3)
 (a intersect b) :: set  of set  of int(1..3)
  a :: set (maxSize(2)) of int(1..3)
  b :: set (maxSize(2)) of int(1..3)

which has several issues in it.

[ca1umac]

I think yours is significantly more readable, it's brilliant. I don't see any reason to keep the json one, it'll just be confusing.

I've gotten rid of the superfluous brackets in the attributes, I'm assuming that was one of the issues. I am also guessing that (x union y) should just be a set and not a set of a set

[ozgurakgun]

I am also guessing that (x union y) should just be a set and not a set of a set

I was more worried about this one, yes.

[ozgurakgun]

I think yours is significantly more readable, it's brilliant. I don't see any reason to keep the json one, it'll just be confusing.

Great - do you want to remove the son version then, please?

[ca1umac]

I am also guessing that (x union y) should just be a set and not a set of a set

I was more worried about this one, yes.

indeed...

[ca1umac]

I think yours is significantly more readable, it's brilliant. I don't see any reason to keep the json one, it'll just be confusing.

Great - do you want to remove the son version then, please?

Can do!

[ozgurakgun]

I think yours is significantly more readable, it's brilliant. I don't see any reason to keep the json one, it'll just be confusing.

Great - do you want to remove the son version then, please?

Can do!

I don't want to leave this unmerged for a long time, so just removed the json option. Will merge soon.