remem-ai 0.4.9

Persistent memory for Claude Code and Codex — single binary, automatic context
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

use std::fmt::{Display, Formatter, Result as FmtResult};

use super::types::{GoldenEvalReport, MetricAverages};

impl Display for GoldenEvalReport {
    fn fmt(&self, f: &mut Formatter<'_>) -> FmtResult {
        writeln!(
            f,
            "remem eval — {} queries, k={}, rank_k={}",
            self.total_queries, self.k, self.rank_k
        )?;
        if let Some(version) = self.version.as_deref() {
            writeln!(f, "schema: {version}")?;
        }
        if let Some(description) = self.description.as_deref() {
            writeln!(f, "{description}")?;
        }
        writeln!(f)?;

        for query in &self.queries {
            if let Some(metrics) = query.metrics.as_ref() {
                writeln!(
                    f,
                    "  [{}] {:>4} | H@{}={:.2} MRR@10={:.2} P@{}={:.2} R@{}={:.2} nDCG@10={:.2} evidence@{}={:.2} | {} | {}",
                    query.id,
                    query.status.label(),
                    self.k,
                    metrics.hit_at_k,
                    metrics.mrr_at_10,
                    self.k,
                    metrics.precision_at_k,
                    self.k,
                    metrics.recall_at_k,
                    metrics.ndcg_at_10,
                    self.k,
                    metrics.evidence_recall_at_k,
                    query.category,
                    query.query
                )?;
            } else {
                writeln!(
                    f,
                    "  [{}] {:>4} | results={} expected_refs={} matched_refs={} | {} | {}",
                    query.id,
                    query.status.label(),
                    query.result_count,
                    query.expected_refs,
                    query.matched_refs,
                    query.category,
                    query.query
                )?;
            }
        }

        writeln!(f)?;
        writeln!(
            f,
            "--- Overall ({} scored, {} abstention, {} skipped) ---",
            self.scored_queries, self.abstention_queries, self.skipped_queries
        )?;
        if let Some(metrics) = self.overall.as_ref() {
            write_metrics(f, "all", metrics, self.k)?;
        }
        if self.abstention_queries > 0 {
            writeln!(
                f,
                "  Abstention: {}/{}",
                self.abstention_passed, self.abstention_queries
            )?;
        }

        if !self.by_category.is_empty() {
            writeln!(f)?;
            writeln!(f, "--- By Category ---")?;
            for (category, evaluation) in &self.by_category {
                if let Some(metrics) = evaluation.metrics.as_ref() {
                    write_metrics(f, category, metrics, self.k)?;
                } else {
                    writeln!(
                        f,
                        "  {}: scored=0 abstention={}/{} total={}",
                        category,
                        evaluation.abstention_passed,
                        evaluation.abstention_queries,
                        evaluation.total_queries
                    )?;
                }
            }
        }
        Ok(())
    }
}

fn write_metrics(
    f: &mut Formatter<'_>,
    label: &str,
    metrics: &MetricAverages,
    k: usize,
) -> FmtResult {
    writeln!(
        f,
        "  {}: n={} H@{}={:.3} MRR@10={:.3} P@{}={:.3} R@{}={:.3} nDCG@10={:.3} evidence@{}={:.3}",
        label,
        metrics.count,
        k,
        metrics.hit_at_k,
        metrics.mrr_at_10,
        k,
        metrics.precision_at_k,
        k,
        metrics.recall_at_k,
        metrics.ndcg_at_10,
        k,
        metrics.evidence_recall_at_k
    )
}