echo_agent 0.2.0

Production-grade AI Agent framework for Rust — ReAct engine, multi-agent, memory, streaming, MCP, IM channels, workflows
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

//! Self-evolution engine — unified entry point for the entire improve loop.
//!
//! One switch to enable everything: eval with LLM grader, A/B comparison,
//! prompt regeneration, HTML report generation, and iterative improvement.
//!
//! # Usage
//!
//! ```rust,ignore
//! let mut evolution = SelfEvolution::new()
//!     .with_agent(agent)
//!     .with_eval_cases(cases)
//!     .with_grader(grader, grading_agent)
//!     .enable(); // ← single switch
//!
//! let result = evolution.run().await;
//! println!("Best score: {:.2} (iteration {})", result.best_score, result.best_iteration);
//! ```

use crate::eval::{EvalCase, generate_html};
use crate::improve::{ImprovementLoop, LoopResult};
use crate::trace::RunStore;
use std::sync::Arc;

/// Unified self-evolution engine. One `.enable()` turns everything on.
///
/// # Usage
///
/// ```rust,ignore
/// let result = SelfEvolution::new()
///     .with_eval_cases(cases)
///     .with_run_store(store)
///     .enable()
///     .run(agent_factory)
///     .await;
/// ```
pub struct SelfEvolution {
    cases: Vec<EvalCase>,
    run_store: Option<Arc<dyn RunStore>>,
    max_iterations: usize,
    report_dir: Option<String>,
    enabled: bool,
}

impl Default for SelfEvolution {
    fn default() -> Self {
        Self::new()
    }
}

impl SelfEvolution {
    pub fn new() -> Self {
        Self {
            cases: Vec::new(),
            run_store: None,
            max_iterations: 5,
            report_dir: None,
            enabled: false,
        }
    }

    pub fn with_eval_cases(mut self, cases: Vec<EvalCase>) -> Self {
        self.cases = cases;
        self
    }
    pub fn with_run_store(mut self, store: Arc<dyn RunStore>) -> Self {
        self.run_store = Some(store);
        self
    }
    pub fn max_iterations(mut self, max: usize) -> Self {
        self.max_iterations = max;
        self
    }
    pub fn with_report_dir(mut self, dir: &str) -> Self {
        self.report_dir = Some(dir.to_string());
        self
    }

    /// Enable self-evolution.
    pub fn enable(mut self) -> Self {
        self.enabled = true;
        self
    }
    pub fn is_enabled(&self) -> bool {
        self.enabled
    }

    /// Run the full self-evolution pipeline.
    pub async fn run<F>(&self, agent_factory: F) -> Option<LoopResult>
    where
        F: Fn() -> Box<dyn crate::agent::Agent>,
    {
        if !self.enabled || self.cases.is_empty() {
            return None;
        }

        let runner = ImprovementLoop::new();
        let result = runner
            .run(&self.cases, agent_factory, &self.run_store)
            .await;

        if let Some(ref dir) = self.report_dir {
            let _ = std::fs::create_dir_all(dir);
            for (i, iter) in result.iterations.iter().enumerate() {
                let html = generate_html(&iter.eval_report, &format!("Iteration {i}"));
                let _ = std::fs::write(format!("{dir}/iter_{i}.html"), html);
            }
            if let Some(last) = result.iterations.last() {
                let html = generate_html(&last.eval_report, "Final");
                let _ = std::fs::write(format!("{dir}/final.html"), html);
            }
        }
        Some(result)
    }
}