do-memory-mcp 0.1.29

Model Context Protocol (MCP) server for AI agents
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
139
140
141
142
143
144
145
146
147
148

use anyhow::Result;
use tracing::debug;

use super::ets_types::{ETSErrorType, ETSModel, ETSSeasonalType, ETSState, ETSTrendType};

impl super::engine::ForecastingEngine {
    /// Main ETS forecasting function
    pub(super) fn forecast_ets(
        &self,
        model: &ETSModel,
        state: &ETSState,
        horizon: usize,
    ) -> Result<Vec<f64>> {
        let mut forecasts = Vec::with_capacity(horizon);

        for h in 1..=horizon {
            // Forecast h steps ahead
            let trend_component = match model.trend {
                ETSTrendType::None => 0.0,
                ETSTrendType::Additive => state.trend * h as f64,
                ETSTrendType::AdditiveDamped => {
                    let mut phi_sum = 0.0;
                    for i in 1..=h {
                        phi_sum += model.phi.powi(i as i32);
                    }
                    state.trend * phi_sum
                }
            };

            let seasonal_component =
                if !state.seasonal.is_empty() && !matches!(model.seasonal, ETSSeasonalType::None) {
                    let seasonal_index = (state.n_obs + h - 1) % state.seasonal.len();
                    state.seasonal[seasonal_index]
                } else {
                    0.0
                };

            let forecast = match (model.error, model.seasonal) {
                (ETSErrorType::Additive, ETSSeasonalType::Additive) => {
                    state.level + trend_component + seasonal_component
                }
                (ETSErrorType::Additive, ETSSeasonalType::Multiplicative) => {
                    (state.level + trend_component) * (1.0 + seasonal_component)
                }
                (ETSErrorType::Multiplicative, ETSSeasonalType::Additive) => {
                    (state.level + trend_component) + seasonal_component
                }
                (ETSErrorType::Multiplicative, ETSSeasonalType::Multiplicative) => {
                    (state.level + trend_component) * (1.0 + seasonal_component)
                }
                _ => state.level + trend_component,
            };

            forecasts.push(forecast);
        }

        debug!("Generated {} forecasts", forecasts.len());
        Ok(forecasts)
    }

    pub(super) fn calculate_confidence_intervals(
        &self,
        _model: &ETSModel,
        forecasts: &[f64],
        _state: &ETSState,
        _confidence: f64,
    ) -> (Vec<f64>, Vec<f64>) {
        let residual_std = self.estimate_residual_variance(_state).sqrt();
        let z_score = 1.96; // 95% confidence

        let lower_bounds: Vec<f64> = forecasts
            .iter()
            .enumerate()
            .map(|(i, &f)| f - z_score * residual_std * ((i + 1) as f64).sqrt())
            .collect();

        let upper_bounds: Vec<f64> = forecasts
            .iter()
            .enumerate()
            .map(|(i, &f)| f + z_score * residual_std * ((i + 1) as f64).sqrt())
            .collect();

        (lower_bounds, upper_bounds)
    }

    pub(super) fn estimate_residual_variance(&self, _state: &ETSState) -> f64 {
        // Simplified - would use actual residuals in production
        0.1
    }

    pub(super) fn calculate_ets_fit_quality(
        &self,
        _actual: &[f64],
        _fitted: &[f64],
        _model: &ETSModel,
    ) -> f64 {
        // Calculate AIC or similar metric
        let n = _actual.len() as f64;
        let k = 3.0; // Number of parameters (simplified)

        // SSE
        let sse: f64 = _actual
            .iter()
            .zip(_fitted.iter())
            .map(|(a, f)| (a - f).powi(2))
            .sum();

        // AIC = n * ln(SSE/n) + 2k
        let aic = n * (sse / n).ln() + 2.0 * k;
        -aic // Return negative so higher is better
    }

    pub(super) fn calculate_model_metrics(
        &self,
        _actual: &[f64],
        _fitted: &[f64],
    ) -> (f64, f64, f64) {
        let n = _actual.len() as f64;

        // MAE
        let mae: f64 = _actual
            .iter()
            .zip(_fitted.iter())
            .map(|(a, f)| (a - f).abs())
            .sum::<f64>()
            / n;

        // RMSE
        let mse: f64 = _actual
            .iter()
            .zip(_fitted.iter())
            .map(|(a, f)| (a - f).powi(2))
            .sum::<f64>()
            / n;
        let rmse = mse.sqrt();

        // MAPE
        let mape: f64 = _actual
            .iter()
            .zip(_fitted.iter())
            .filter(|(a, _)| a.abs() > 1e-10)
            .map(|(a, f)| ((a - f).abs() / a.abs()) * 100.0)
            .sum::<f64>()
            / n;

        (mae, rmse, mape)
    }
}