rho-coding-agent 1.29.0

A lightweight agent harness inspired by Pi
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

use std::time::Duration;

use rho_sdk::model::{ContextUsage, ModelUsage};
use rho_sdk::ProviderStreamResetReason;

use crate::tui::{
    activity::ProviderRetryHint, app_state::SessionUiPhase, event_adapter::ViewModelEvent,
    tests::test_app,
};

fn model_call_metrics() -> rho_sdk::ModelCallMetrics {
    rho_sdk::ModelCallMetrics {
        output_tokens: Some(100),
        time_to_first_token: Some(Duration::from_millis(100)),
        generation_time: Some(Duration::from_millis(1_900)),
        total_latency: Duration::from_secs(2),
    }
}

#[test]
fn context_usage_event_is_tracked_separately_from_cumulative_usage() {
    let mut app = test_app();
    app.usage.cumulative_usage = Some(ModelUsage {
        input_tokens: Some(1_000),
        output_tokens: Some(500),
        ..ModelUsage::default()
    });

    assert!(app
        .record_agent_event(ViewModelEvent::ContextUsage(ContextUsage::estimated(
            250,
            Some(10_000),
        )))
        .is_none());

    assert_eq!(
        app.usage.current_context,
        Some(ContextUsage::estimated(250, Some(10_000)))
    );
    assert_eq!(
        app.usage
            .cumulative_usage
            .as_ref()
            .and_then(|usage| usage.input_tokens),
        Some(1_000)
    );
}

#[test]
fn completed_model_call_updates_the_active_model_average() {
    let mut app = test_app();
    let profile = app.info.runtime.model_call_profile();
    let metrics = model_call_metrics();

    app.record_agent_event(ViewModelEvent::ModelCallCompleted {
        profile: profile.clone(),
        metrics,
    });

    let summary = app.usage.model_performance.summary(&profile);
    assert_eq!(summary.latest_call, Some(metrics));
    assert_eq!(summary.average_output_tokens_per_second, Some(50.0));
    assert_eq!(summary.eligible_calls, 1);
}

#[test]
fn provider_stream_reset_preserves_completed_model_performance() {
    let mut app = test_app();
    let profile = app.info.runtime.model_call_profile();
    app.record_agent_event(ViewModelEvent::ModelCallCompleted {
        profile: profile.clone(),
        metrics: model_call_metrics(),
    });

    app.record_agent_event(ViewModelEvent::ProviderStreamReset(ProviderRetryHint {
        reason: ProviderStreamResetReason::InvalidResponse,
    }));

    let summary = app.usage.model_performance.summary(&profile);
    assert_eq!(summary.average_output_tokens_per_second, Some(50.0));
    assert_eq!(summary.eligible_calls, 1);
}

#[test]
fn step_started_clears_stream_state_without_clearing_model_performance() {
    let mut app = test_app();
    let profile = app.info.runtime.model_call_profile();
    app.streams.assistant_stream.push_delta("current");
    app.streams.reasoning_stream.push_delta("reasoning");
    app.record_agent_event(ViewModelEvent::ModelCallCompleted {
        profile: profile.clone(),
        metrics: model_call_metrics(),
    });

    assert!(app
        .record_agent_event(ViewModelEvent::StepStarted(2))
        .is_none());

    assert!(app.streams.assistant_stream.is_empty());
    assert!(app.streams.reasoning_stream.is_empty());
    let summary = app.usage.model_performance.summary(&profile);
    assert_eq!(summary.average_output_tokens_per_second, Some(50.0));
    assert_eq!(app.turn.session_ui(), SessionUiPhase::ProviderTurn);
    assert_eq!(app.status(), "running step 2");
}

#[test]
fn provider_retry_status_includes_rate_limit_reset_hint() {
    use rho_sdk::ProviderErrorKind;

    assert_eq!(
        ProviderRetryHint {
            reason: ProviderStreamResetReason::retryable_failure(
                ProviderErrorKind::RateLimit,
                Some(Duration::from_secs(12)),
            ),
        }
        .status_label(),
        "rate limited · retry in 12s"
    );
    assert_eq!(
        ProviderRetryHint {
            reason: ProviderStreamResetReason::retryable_failure(
                ProviderErrorKind::RateLimit,
                None
            ),
        }
        .status_label(),
        "rate limited · retrying"
    );
    assert_eq!(
        ProviderRetryHint {
            reason: ProviderStreamResetReason::InvalidResponse,
        }
        .status_label(),
        "retrying provider response"
    );
}