ai-dispatch 10.52.1

Multi-AI CLI team orchestrator
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

// One-source category rules, CLI/effort boundaries, and independent scaling.
// Covers all categories using captured relay models and exact harness rows.
// Deps: evidence_from, rescale, test_support; no network.

use super::test_support::{live, sample};
use super::*;

const OPUS: &str = "anthropic/claude-opus-5.5";

#[test]
fn category_rules_use_one_board_and_never_blend() {
    let feed = live();
    for category in [
        TaskCategory::SimpleEdit,
        TaskCategory::ComplexImpl,
        TaskCategory::Debugging,
        TaskCategory::Testing,
        TaskCategory::Refactoring,
    ] {
        let evidence = evidence_from(&feed, OPUS, "claude", Some("max"), category);
        assert!((evidence.capability.expect("accuracy") - 6.485).abs() < 1e-10);
        assert_eq!(
            evidence.selected.expect("selected").source,
            "terminal_bench"
        );
    }
    for category in [TaskCategory::Research, TaskCategory::Documentation] {
        let evidence = evidence_from(&feed, OPUS, "claude", Some("max"), category);
        assert_eq!(evidence.capability, Some(10.0), "{category:?}");
        assert_eq!(evidence.selected.expect("selected").board, "eci");
    }
    let frontend = evidence_from(&feed, OPUS, "claude", Some("max"), TaskCategory::Frontend);
    assert_eq!(frontend.capability, Some(10.0));
    assert_eq!(frontend.selected.expect("selected").source, "lmarena");
}

#[test]
fn terminal_bench_requires_exact_cli_model_and_effort_else_harness_unmeasured() {
    let feed = live();
    for (cli, effort) in [
        ("codex", Some("max")),
        ("claude", Some("low")),
        ("claude", None),
    ] {
        let evidence = evidence_from(&feed, OPUS, cli, effort, TaskCategory::Testing);
        assert_eq!(evidence.capability, Some(10.0));
        assert_eq!(evidence.harness, "harness unmeasured");
        assert_eq!(evidence.selected.expect("fallback").source, "epoch");
    }
    let other = evidence_from(
        &feed,
        "openai/gpt-4o-mini",
        "claude",
        Some("max"),
        TaskCategory::Testing,
    );
    assert_eq!(other.capability, Some(0.0));
    assert_eq!(other.harness, "harness unmeasured");
    let exact = evidence_from(&feed, OPUS, "claude", Some("max"), TaskCategory::Testing);
    assert!((exact.capability.expect("accuracy") - 6.485).abs() < 1e-10);
    assert_eq!(exact.harness, "measured");
}

#[test]
fn unrelated_board_failed_source_unmapped_or_absent_model_is_unknown() {
    let mut feed = sample();
    for id in ["302ai/gemini-3-pro-preview", "Baichuan 2-7B", "missing"] {
        assert_eq!(
            evidence_from(&feed, id, "codex", None, TaskCategory::Research).capability,
            None
        );
    }
    feed.sources.get_mut("epoch").expect("epoch").ok = false;
    assert_eq!(
        evidence_from(
            &feed,
            "302ai/kimi-k2-thinking",
            "codex",
            None,
            TaskCategory::Research
        )
        .capability,
        None
    );
    assert_eq!(
        evidence_from(
            &feed,
            "302ai/kimi-k2-thinking",
            "codex",
            None,
            TaskCategory::Frontend
        )
        .capability,
        None
    );
}

#[test]
fn degenerate_source_range_is_unknown_and_accuracy_uses_its_own_range() {
    let mut feed = sample();
    feed.models[3].scores.clear();
    assert_eq!(
        evidence_from(
            &feed,
            "302ai/kimi-k2-thinking",
            "codex",
            None,
            TaskCategory::Research
        )
        .capability,
        None
    );
    let mut score = live().models[0].scores[0].clone();
    score.value = 0.72;
    score.unit = "rate".into();
    assert_eq!(rescale(&feed, &score), Some(7.199999999999999));
    score.value = 1.1;
    assert_eq!(rescale(&feed, &score), None);
}

#[test]
fn webdev_prefers_configured_effort_else_highest_raw_value() {
    let feed = live();
    let id = "anthropic/claude-sonnet-5.5";
    for (effort, selected_effort) in [
        (Some("high"), "high"),
        (None, "xhigh"),
        (Some("unmeasured"), "xhigh"),
    ] {
        let evidence = evidence_from(&feed, id, "cursor", effort, TaskCategory::Frontend);
        let score = evidence.selected.expect("webdev");
        let captured = feed.models[1]
            .scores
            .iter()
            .find(|row| {
                row.source == "lmarena"
                    && row.board == "webdev"
                    && row.effort.as_deref() == Some(selected_effort)
            })
            .expect("captured webdev row");
        assert_eq!(
            &score, captured,
            "preserve the exact captured value and metadata"
        );
        let expected =
            10.0 * (score.value - 1446.5703748835622) / (1813.6003351106383 - 1446.5703748835622);
        assert!((evidence.capability.expect("webdev range") - expected).abs() < 1e-10);
        assert_eq!(evidence.harness, "harness unmeasured");
    }
}