aether-agent-core 0.6.24

A minimal Rust library for building AI agents with MCP tool integration
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

use aether_core::events::{AgentEvent, ContextEvent, ContextUsage};
use aether_core::testing::{TestScenario, test_agent};
use llm::ModelSettings;
use llm::testing::{FakeLlmProvider, llm_response};

fn context_usage(events: &[AgentEvent]) -> ContextUsage {
    events
        .iter()
        .find_map(|event| match event {
            AgentEvent::Context(ContextEvent::UsageUpdated { usage }) => Some(usage.clone()),
            _ => None,
        })
        .expect("agent should emit context usage")
}

#[tokio::test]
async fn context_window_override_supplies_unknown_provider_limit() {
    let events = test_agent()
        .llm_responses(&[llm_response("msg").usage(100_000, 10).build()])
        .without_mcp()
        .provider_context_window(None)
        .context_window_override(200_000)
        .user_text("hello")
        .run()
        .await
        .unwrap();

    let usage = context_usage(&events);
    assert_eq!(usage.context_limit, Some(200_000));
    assert_eq!(usage.usage_ratio, Some(0.5));
    assert_eq!(usage.input_tokens, 100_000);
}

#[tokio::test]
async fn context_window_override_beats_provider_limit() {
    let events = test_agent()
        .llm_responses(&[llm_response("msg").usage(100_000, 10).build()])
        .without_mcp()
        .provider_context_window(Some(128_000))
        .context_window_override(200_000)
        .user_text("hello")
        .run()
        .await
        .unwrap();

    let usage = context_usage(&events);
    assert_eq!(usage.context_limit, Some(200_000));
    assert_eq!(usage.usage_ratio, Some(0.5));
}

/// Requires sequential messaging (`SwitchModel` after initial setup).
#[tokio::test]
async fn context_window_override_survives_model_switch() {
    let events = test_agent()
        .without_mcp()
        .provider_context_window(Some(128_000))
        .context_window_override(200_000)
        .scenario(
            TestScenario::new()
                .switch_model(
                    FakeLlmProvider::new(vec![]).with_display_name("new fake").with_context_window(Some(32_000)),
                )
                .wait_for(|event| matches!(event, AgentEvent::Context(ContextEvent::UsageUpdated { .. }))),
        )
        .run()
        .await
        .unwrap();

    let usage = context_usage(&events);
    assert_eq!(usage.context_limit, Some(200_000));
    assert_eq!(usage.usage_ratio, Some(0.0));
}

#[tokio::test]
async fn spawn_applies_model_settings_to_context() {
    let settings = ModelSettings { temperature: Some(0.0), max_tokens: Some(64), ..Default::default() };

    let result = test_agent()
        .llm_responses(&[llm_response("msg").usage(10, 10).build()])
        .model_settings(settings.clone())
        .user_text("hello")
        .run_with_context()
        .await
        .unwrap();

    let contexts = result.captured_contexts.lock().unwrap();
    assert_eq!(contexts[0].model_settings(), &settings);
}