Group by rule

[moritz-gross]

feature request by Marthe: group by rule first, and then subgroup by issue type.

also added more unit testing, including for full rich output: audit_translations/tests/golden/rich/cli_calculus_verbose.golden

---

╭──────────────────────────────────────────────────────────────────────────────╮
│ MathCAT Translation Audit: ES                                                │
╰──────────────────────────────────────────────────────────────────────────────╯

  Comparing against English (en) reference files
  Files to check: 1

────────────────────────────────────────────────────────────────────────────────
⚠ SharedRules/calculus.yaml
  English: 4 rules  →  Translated: 3 rules
────────────────────────────────────────────────────────────────────────────────

  ≠ Rule Issues [13] (grouped by rule and issue type)
      • laplacian (laplacian)
          Missing in Translation [1]
              • (line 4 in English)
      • divergence (divergence)
          Untranslated Text [3]
              • (line 10 tr) "divergence"
              • (line 11 tr) "div"
              • (line 12 tr) "of"
          Match Pattern Differences [1]
              • (line 22 en, 6 tr)
                  Match pattern differs
                  en: count(*) = 1
                  tr: .
          Condition Differences [1]
              • (line 25 en, 9 tr)
                  Conditions differ
                  en: $Verbosity='Terse', not(IsNode(*[1], 'leaf'))
                  tr: $Verbosity='Verbose', not(IsNode(*[1], 'leaf'))
      • curl (curl)
          Untranslated Text [1]
              • (line 22 tr) "curl of"
          Match Pattern Differences [1]
              • (line 35 en, 20 tr)
                  Match pattern differs
                  en: count(*) = 1
                  tr: .
          Condition Differences [1]
              • (line 39 en, 24 tr)
                  Conditions differ
                  en: $Verbosity!='Terse', not(IsNode(*[1], 'leaf'))
                  tr: not(IsNode(*[1], 'leaf'))
          Structure Differences [1]
              • (line 38 en, 18 tr)
                  Rule structure differs (test/if/then/else blocks)
                  en: replace: test: if: then: test: if: then:
                  tr: replace: test: if: then:
      • gradient (gradient)
          Untranslated Text [2]
              • (line 34 tr) "gradient of"
              • (line 35 tr) "del"
          Match Pattern Differences [1]
              • (line 48 en, 30 tr)
                  Match pattern differs
                  en: count(*) = 1
                  tr: .
╭──────────────────────────────────────────────────────────────────────────────╮
│                 SUMMARY                                                      │
│   Files checked                     1                                        │
│   Files with issues                 1                                        │
│   Files OK                          0                                        │
│   Missing rules                     1                                        │
│   Untranslated text                 6                                        │
│   Rule differences                  6                                        │
│   Extra rules                       0                                        │
╰──────────────────────────────────────────────────────────────────────────────╯

[moritz-gross]

I also found out we could use pytest-xdist to have multiple workers with uv run pytest -n <#workers>, but as of so far the overhead makes it actually take longer.