lemma-engine 0.10.2

A pure, declarative language for business rules.
Documentation
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
//! Behaviour of unless chains that planning folds into an ordered dispatch table.
//!
//! The fold is invisible by design: these cases pin the observable behaviour it must
//! not change — selected result, veto propagation, released data and narration.
//! Plan-shape assertions live in `engine/src/tests/transitive_normalization_plan_shape.rs`.

use indexmap::IndexMap;
use lemma::{format_explanation, DateTimeValue, Engine, EvaluateDataValue, Response, RuleResult};

fn engine_for(code: &str) -> Engine {
    let mut engine = Engine::new();
    engine
        .load([(lemma::SourceType::Volatile, code.to_string())])
        .expect("spec must load");
    engine
}

fn run(engine: &Engine, spec: &str, bindings: &[(&str, &str)], explain: bool) -> Response {
    let now = DateTimeValue::now();
    let data: IndexMap<String, EvaluateDataValue> = bindings
        .iter()
        .map(|(key, value)| ((*key).to_string(), (*value).into()))
        .collect();
    engine
        .evaluate(None, spec, Some(&now), data, None, explain)
        .expect("evaluation must succeed")
}

fn rule<'a>(response: &'a Response, name: &str) -> &'a RuleResult {
    response.results.get(name).unwrap_or_else(|| {
        panic!(
            "rule '{name}' missing from results: {:?}",
            response.results.keys().collect::<Vec<_>>()
        )
    })
}

fn display(response: &Response, name: &str) -> String {
    rule(response, name)
        .result()
        .map(|d| d.to_string())
        .unwrap_or_else(|| {
            panic!(
                "rule '{name}' produced no value: {:?}",
                rule(response, name)
            )
        })
}

const LOOKUP: &str = r#"
spec lookup
data code: text
  -> option "NL"
  -> option "BE"
  -> option "DE"
rule name: veto "unknown code"
  unless code is "NL" then "Netherlands"
  unless code is "BE" then "Belgium"
  unless code is "DE" then "Germany"
"#;

/// Same exclusive is-chain without option constraints so an unknown code reaches
/// the veto default instead of failing input validation.
const LOOKUP_OPEN: &str = r#"
spec lookup
data code: text
rule name: veto "unknown code"
  unless code is "NL" then "Netherlands"
  unless code is "BE" then "Belgium"
  unless code is "DE" then "Germany"
"#;

const TIERS: &str = r#"
spec tiers
data quantity: number
rule discount: 0
  unless quantity >= 10 then 5
  unless quantity >= 100 then 10
  unless quantity is 50 then 7
"#;

/// `tip` borrows the Kind of `alias`, which borrows the Kind of the rule that
/// owns the fold: the Piecewise pre-image sits two references away, not one.
const REFERENCE_CHAIN: &str = r#"
spec chain
data code: text
rule country: "other"
  unless code is "NL" then "Netherlands"
  unless code is "BE" then "Belgium"
rule alias: country
rule tip: alias
"#;

/// The dispatch rule is reached through a `uses` binding, so inside the
/// consumer's plan the scrutinee is a rule reference, not a data leaf.
const BOUND_SCRUTINEE: &str = r#"
spec zones
data dest: text
rule zone_of: 2
  unless dest is "100" then 5
  unless dest is "200" then 8

spec rates
data zone: number
data weight: number
rule rate: 0
  unless zone is 5 and weight >= 1 then 14.20
  unless zone is 8 and weight >= 1 then 20.00

spec quote
uses z: zones
  -> with dest: dest
uses r: rates
  -> with zone: zone
  -> with weight: weight
data dest: text
data weight: number
rule zone: z.zone_of
rule total: r.rate
"#;

const RATE_CARD: &str = r#"
spec rates
data zone: number
data weight: number
rule rate: 0
  unless zone is 2 and weight >= 1 then 10
  unless zone is 2 and weight >= 5 then 20
  unless zone is 3 and weight >= 1 then 30
  unless zone is 3 and weight >= 5 then 40
"#;

#[test]
fn conjunctive_rate_card_selects_by_zone_and_weight() {
    let engine = engine_for(RATE_CARD);
    for (zone, weight, expected) in [
        ("1", "10", "0"),
        ("2", "0", "0"),
        ("2", "1", "10"),
        ("2", "4", "10"),
        ("2", "5", "20"),
        ("3", "1", "30"),
        ("3", "5", "40"),
        ("9", "100", "0"),
    ] {
        assert_eq!(
            display(
                &run(
                    &engine,
                    "rates",
                    &[("zone", zone), ("weight", weight)],
                    false
                ),
                "rate"
            ),
            expected,
            "zone={zone} weight={weight}"
        );
    }
}

#[test]
fn conjunctive_rate_card_explain_matches_value_mode() {
    let engine = engine_for(RATE_CARD);
    for (zone, weight) in [("2", "5"), ("3", "1"), ("1", "10")] {
        let without = display(
            &run(
                &engine,
                "rates",
                &[("zone", zone), ("weight", weight)],
                false,
            ),
            "rate",
        );
        let with = display(
            &run(
                &engine,
                "rates",
                &[("zone", zone), ("weight", weight)],
                true,
            ),
            "rate",
        );
        assert_eq!(
            without, with,
            "zone={zone} weight={weight}: explain must match value mode"
        );
    }
}

#[test]
fn conjunctive_rate_card_releases_weight_when_zone_misses() {
    let engine = engine_for(RATE_CARD);
    let response = run(&engine, "rates", &[("zone", "9")], false);
    assert!(
        rule(&response, "rate").missing_data().is_empty(),
        "no zone matches, so weight is not needed: {:?}",
        rule(&response, "rate").missing_data()
    );
}

#[test]
fn conjunctive_rate_card_still_needs_weight_when_zone_hits() {
    let engine = engine_for(RATE_CARD);
    let response = run(&engine, "rates", &[("zone", "2")], false);
    assert_eq!(
        rule(&response, "rate").missing_data(),
        vec!["weight".to_string()],
        "a matching zone still needs the weight residual"
    );
}

#[test]
fn equality_lookup_selects_the_matching_arm() {
    let engine = engine_for(LOOKUP);
    for (code, expected) in [("NL", "Netherlands"), ("BE", "Belgium"), ("DE", "Germany")] {
        assert_eq!(
            display(&run(&engine, "lookup", &[("code", code)], false), "name"),
            expected
        );
    }
}

#[test]
fn ordering_chain_holds_at_below_and_above_every_boundary() {
    let engine = engine_for(TIERS);
    // 50 is an exact hit on the last arm, which wins over the `>= 10` arm below it.
    for (quantity, expected) in [
        ("0", "0"),
        ("9", "0"),
        ("10", "5"),
        ("49", "5"),
        ("50", "7"),
        ("51", "5"),
        ("99", "5"),
        ("100", "10"),
        ("1000", "10"),
    ] {
        assert_eq!(
            display(
                &run(&engine, "tiers", &[("quantity", quantity)], false),
                "discount"
            ),
            expected,
            "quantity {quantity}"
        );
    }
}

#[test]
fn a_negative_scrutinee_lands_below_every_boundary() {
    let engine = engine_for(TIERS);
    assert_eq!(
        display(
            &run(&engine, "tiers", &[("quantity", "-1000")], false),
            "discount"
        ),
        "0"
    );
}

#[test]
fn an_unmatched_scrutinee_falls_through_to_the_default() {
    let engine = engine_for(
        r#"
spec lookup
data code: text
rule name: veto "unknown code"
  unless code is "NL" then "Netherlands"
"#,
    );
    let response = run(&engine, "lookup", &[("code", "ZZ")], false);
    assert!(
        rule(&response, "name").vetoed,
        "an unmatched code must reach the default veto: {:?}",
        rule(&response, "name")
    );
}

#[test]
fn an_unbound_scrutinee_propagates_as_missing_data() {
    let engine = engine_for(LOOKUP);
    let response = run(&engine, "lookup", &[], false);
    let result = rule(&response, "name");
    assert!(
        result.vetoed,
        "an unbound scrutinee must veto, got {result:?}"
    );
    assert_eq!(
        result.missing_data(),
        vec!["code".to_string()],
        "the scrutinee is the only thing the rule still needs"
    );
}

/// Each reference borrows the Kind of the one below it and records no pre-image
/// of its own, so the walk that reports missing data has to follow the chain
/// down to the rule that owns the fold.
#[test]
fn missing_data_travels_through_a_chain_of_references_to_a_dispatch_rule() {
    let engine = engine_for(REFERENCE_CHAIN);
    let response = run(&engine, "chain", &[], false);
    for name in ["country", "alias", "tip"] {
        let result = rule(&response, name);
        assert!(
            result.vetoed,
            "'{name}' has no value without a code: {result:?}"
        );
        assert_eq!(
            result.missing_data(),
            vec!["code".to_string()],
            "'{name}' still needs the scrutinee the references lead to"
        );
    }
}

#[test]
fn a_reference_chain_carries_the_value_it_borrows() {
    let engine = engine_for(REFERENCE_CHAIN);
    let response = run(&engine, "chain", &[("code", "BE")], false);
    for name in ["country", "alias", "tip"] {
        assert_eq!(display(&response, name), "Belgium", "rule '{name}'");
    }
}

#[test]
fn a_bound_rule_scrutinee_reports_the_data_behind_it() {
    let engine = engine_for(BOUND_SCRUTINEE);

    let response = run(&engine, "quote", &[("weight", "1")], false);
    let result = rule(&response, "total");
    assert!(
        result.vetoed,
        "the rule bound to the scrutinee has no value: {result:?}"
    );
    assert_eq!(result.missing_data(), vec!["dest".to_string()]);

    let response = run(&engine, "quote", &[("dest", "100"), ("weight", "1")], false);
    assert_eq!(display(&response, "total"), "14.2");
}

/// The regions the dispatch did not select must release their data, exactly as the
/// untaken arms of a Piecewise do.
#[test]
fn losing_regions_release_their_data() {
    let engine = engine_for(
        r#"
spec routing
data code: text
data dutch_rate: number
data belgian_rate: number
rule rate: 0
  unless code is "NL" then dutch_rate
  unless code is "BE" then belgian_rate
"#,
    );
    let response = run(&engine, "routing", &[("code", "NL")], false);
    let missing = rule(&response, "rate").missing_data();
    assert_eq!(
        missing,
        ["dutch_rate".to_string()].as_slice(),
        "only the selected region's data is still needed, got {missing:?}"
    );
}

#[test]
fn every_region_is_released_when_the_default_wins() {
    let engine = engine_for(
        r#"
spec routing
data code: text
data dutch_rate: number
data belgian_rate: number
rule rate: 0
  unless code is "NL" then dutch_rate
  unless code is "BE" then belgian_rate
"#,
    );
    let response = run(&engine, "routing", &[("code", "ZZ")], false);
    assert!(
        rule(&response, "rate").missing_data().is_empty(),
        "no arm can win, so neither rate is needed: {:?}",
        rule(&response, "rate").missing_data()
    );
}

/// A result shared by several regions stays live when one of them is selected.
#[test]
fn a_result_reachable_from_two_regions_is_not_released() {
    let engine = engine_for(
        r#"
spec shared
data code: text
data special: number
rule rate: 0
  unless code is "NL" then special
  unless code is "BE" then special
"#,
    );
    let response = run(&engine, "shared", &[("code", "BE")], false);
    assert_eq!(
        rule(&response, "rate").missing_data(),
        vec!["special".to_string()],
        "the selected region needs it, so it must not be released"
    );
}

#[test]
fn explain_and_value_modes_agree_at_every_tier_region() {
    let engine = engine_for(TIERS);
    for quantity in ["0", "9", "10", "49", "50", "51", "99", "100", "1000"] {
        let without = display(
            &run(&engine, "tiers", &[("quantity", quantity)], false),
            "discount",
        );
        let with = display(
            &run(&engine, "tiers", &[("quantity", quantity)], true),
            "discount",
        );
        assert_eq!(
            without, with,
            "quantity {quantity}: explain result must match value mode"
        );
    }
}

#[test]
fn explanation_states_the_matched_condition_with_its_data() {
    let engine = engine_for(TIERS);
    let response = run(&engine, "tiers", &[("quantity", "100")], true);
    let explanation = rule(&response, "discount")
        .explanation
        .as_ref()
        .expect("explanation built");

    let matched = explanation
        .causes
        .iter()
        .find(|cause| cause.condition == "quantity >= 100")
        .unwrap_or_else(|| panic!("matched condition missing from {:?}", explanation.causes));
    assert_eq!(matched.value, "true");
    assert!(
        matched
            .children
            .iter()
            .any(|child| format!("{child:?}").contains("100")),
        "the cause must carry the value that drove it, got {:?}",
        matched.children
    );
}

#[test]
fn explanation_states_untaken_conditions_as_the_flipped_fact() {
    let engine = engine_for(TIERS);
    let response = run(&engine, "tiers", &[("quantity", "10")], true);
    let explanation = rule(&response, "discount")
        .explanation
        .as_ref()
        .expect("explanation built");
    let conditions: Vec<&str> = explanation
        .causes
        .iter()
        .map(|cause| cause.condition.as_str())
        .collect();
    assert_eq!(
        conditions,
        vec!["quantity < 100", "quantity is not 50", "quantity >= 10"],
        "later last-wins failures then the winner"
    );
}

#[test]
fn exclusive_point_lookup_mid_hit_states_only_the_winner() {
    let engine = engine_for(LOOKUP);
    let response = run(&engine, "lookup", &[("code", "DE")], true);
    let explanation = rule(&response, "name")
        .explanation
        .as_ref()
        .expect("explanation built");
    let conditions: Vec<&str> = explanation
        .causes
        .iter()
        .map(|cause| cause.condition.as_str())
        .collect();
    assert_eq!(
        conditions,
        vec!["code is DE"],
        "exclusive is-chain keeps only the winning cause, got {:?}",
        explanation.causes
    );
    assert!(
        explanation.causes[0].children.is_empty(),
        "held data-vs-literal is omits restated Data child, got {:?}",
        explanation.causes[0].children
    );
    let ascii = format_explanation(explanation);
    assert!(
        ascii.contains("code is DE"),
        "ASCII must state the winning cause, got {ascii}"
    );
    assert!(
        !ascii.contains("code: DE"),
        "ASCII must not restate the binding under a held is-line, got {ascii}"
    );
}

#[test]
fn inequality_matched_cause_still_expands_data_in_ascii() {
    let engine = engine_for(TIERS);
    let response = run(&engine, "tiers", &[("quantity", "100")], true);
    let explanation = rule(&response, "discount")
        .explanation
        .as_ref()
        .expect("explanation built");
    let ascii = format_explanation(explanation);
    assert!(
        ascii.contains("quantity >= 100"),
        "ASCII must state the matched inequality, got {ascii}"
    );
    assert!(
        ascii.contains("quantity: 100"),
        "ASCII must expand exact magnitude under inequality, got {ascii}"
    );
}

#[test]
fn exclusive_point_lookup_default_attaches_scrutinee_without_is_not_dump() {
    let engine = engine_for(LOOKUP_OPEN);
    let response = run(&engine, "lookup", &[("code", "FR")], true);
    let result = rule(&response, "name");
    assert!(result.vetoed, "unknown code must veto");
    let explanation = result.explanation.as_ref().expect("explanation built");
    assert!(
        explanation.causes.is_empty(),
        "exclusive default must not dump is-not causes, got {:?}",
        explanation.causes
    );
    assert!(
        explanation.children.iter().any(|child| {
            matches!(
                child,
                lemma::ExplanationNode::Data { name, result }
                    if name.input_key() == "code"
                        && result.result.as_deref() == Some("FR")
            ) || matches!(
                child,
                lemma::ExplanationNode::Compose { operands, .. }
                    if operands.iter().any(|op| {
                        matches!(
                            op,
                            lemma::ExplanationNode::Data { name, result }
                                if name.input_key() == "code"
                                    && result.result.as_deref() == Some("FR")
                        )
                    })
            )
        }),
        "exclusive default must attach scrutinee data, got {:?}",
        explanation.children
    );
}

/// Narration walks the pre-image chain, so a vetoed scrutinee still reports the
/// condition it was evaluating — the dispatch table has no conditions to name.
#[test]
fn a_vetoed_scrutinee_narrates_a_condition_from_the_pre_image() {
    let engine = engine_for(LOOKUP);
    let response = run(&engine, "lookup", &[], true);
    let explanation = rule(&response, "name")
        .explanation
        .as_ref()
        .expect("explanation built");
    assert_eq!(
        explanation.body, "code is DE",
        "the reverse scan starts at the last arm and propagates its veto"
    );
}

/// Issue #92: a residual that is a bare boolean data leaf stays a Piecewise inside
/// each key region. Every arm body is the literal `true`, so the three arms share
/// one body id.
const ISSUE_NEGATED: &str = r#"
spec repro
data model: text
data quantity: number
data remote_negated: boolean
data hybrid_negated: boolean
data onsite_negated: boolean
rule negated: false
  unless model is "remote" and remote_negated then true
  unless model is "hybrid" and hybrid_negated then true
  unless model is "onsite" and onsite_negated then true
rule accepted: quantity is 1 and not negated
"#;

#[test]
fn issue_92_negated_matrix_agrees_in_value_and_explain_modes() {
    let engine = engine_for(ISSUE_NEGATED);
    for model in ["remote", "hybrid", "onsite", "other"] {
        let flag = format!("{model}_negated");
        for value in ["false", "true"] {
            let mut bindings: Vec<(&str, &str)> = vec![
                ("model", model),
                ("quantity", "1"),
                ("remote_negated", "false"),
                ("hybrid_negated", "false"),
                ("onsite_negated", "false"),
            ];
            if model != "other" {
                bindings
                    .iter_mut()
                    .find(|(key, _)| *key == flag.as_str())
                    .expect("flag bound")
                    .1 = value;
            }
            let expected_negated = value == "true" && model != "other";
            for explain in [false, true] {
                let response = run(&engine, "repro", &bindings, explain);
                assert_eq!(
                    display(&response, "negated"),
                    expected_negated.to_string(),
                    "model={model} {flag}={value} explain={explain}"
                );
                assert_eq!(
                    display(&response, "accepted"),
                    (!expected_negated).to_string(),
                    "model={model} {flag}={value} explain={explain}"
                );
                if explain && expected_negated {
                    let explanation = rule(&response, "negated")
                        .explanation
                        .as_ref()
                        .expect("explanation built");
                    let ascii = format_explanation(explanation);
                    assert!(
                        ascii.contains(&format!("model is {model}")),
                        "explanation must name the winning unless, got {ascii}"
                    );
                }
            }
        }
    }
}

#[test]
fn issue_92_missing_residual_reports_only_the_selected_flag() {
    let engine = engine_for(ISSUE_NEGATED);
    let response = run(
        &engine,
        "repro",
        &[("model", "remote"), ("quantity", "1")],
        false,
    );
    let negated = rule(&response, "negated");
    assert!(negated.vetoed, "remote_negated is needed: {negated:?}");
    assert_eq!(
        negated.missing_data(),
        vec!["remote_negated".to_string()],
        "only the selected region's residual is needed"
    );
}

/// Residual `n > m` compares two data, so it does not fold into a nested dispatch.
const TWO_DATA_RESIDUAL: &str = r#"
spec residual
data model: text
data n: number
data m: number
rule r: 0
  unless model is "a" and n > m then 1
  unless model is "b" and n > m then 2
"#;

#[test]
fn issue_92_two_data_residual_agrees_in_both_modes() {
    let engine = engine_for(TWO_DATA_RESIDUAL);
    for (model, n, m, expected) in [
        ("a", "5", "3", "1"),
        ("a", "1", "3", "0"),
        ("b", "5", "3", "2"),
        ("b", "1", "3", "0"),
        ("c", "5", "3", "0"),
    ] {
        for explain in [false, true] {
            assert_eq!(
                display(
                    &run(
                        &engine,
                        "residual",
                        &[("model", model), ("n", n), ("m", m)],
                        explain
                    ),
                    "r"
                ),
                expected,
                "model={model} n={n} m={m} explain={explain}"
            );
        }
    }
}

/// Boolean residuals with numeric bodies; no residual is a comparison.
const BOOLEAN_RESIDUAL: &str = r#"
spec flags
data model: text
data flag: boolean
data flag2: boolean
rule r: 0
  unless model is "a" and flag then 1
  unless model is "b" and flag2 then 2
"#;

#[test]
fn issue_92_boolean_residual_with_numeric_bodies_agrees_in_both_modes() {
    let engine = engine_for(BOOLEAN_RESIDUAL);
    for (model, flag, flag2, expected) in [
        ("a", "true", "false", "1"),
        ("a", "false", "false", "0"),
        ("b", "false", "true", "2"),
        ("b", "false", "false", "0"),
        ("c", "true", "true", "0"),
    ] {
        for explain in [false, true] {
            assert_eq!(
                display(
                    &run(
                        &engine,
                        "flags",
                        &[("model", model), ("flag", flag), ("flag2", flag2)],
                        explain
                    ),
                    "r"
                ),
                expected,
                "model={model} flag={flag} flag2={flag2} explain={explain}"
            );
        }
    }
}

/// The `b` region decides from `extra` alone; the `c` arm sits above it and its
/// residual `flag_c` is never read by the evaluator for `model is b`.
const RESIDUAL_ABOVE_WINNER: &str = r#"
spec above
data model: text
data extra: boolean
data amount: number
data flag_c: boolean
rule r: 0
  unless model is "b" and extra then amount
  unless flag_c and model is "c" then 3
"#;

#[test]
fn issue_92_missing_data_ignores_residuals_of_regions_the_evaluator_does_not_read() {
    let engine = engine_for(RESIDUAL_ABOVE_WINNER);
    let response = run(
        &engine,
        "above",
        &[("model", "b"), ("extra", "true")],
        false,
    );
    let result = rule(&response, "r");
    assert!(result.vetoed, "amount is needed: {result:?}");
    assert_eq!(
        result.missing_data(),
        vec!["amount".to_string()],
        "flag_c belongs to the c region, which the b selection never reads"
    );
}