ClawFoxyVision 0.2.0

Advanced financial time series forecasting library using LSTM, GRU, and CNN-LSTM neural networks for price prediction with Rust and Burn
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
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173

// External crates
use polars::prelude::*;

// Use rustalib for technical indicators
use rustalib::indicators::moving_averages::{calculate_ema, calculate_sma};
use rustalib::indicators::oscillators::{calculate_macd, calculate_rsi};
use rustalib::indicators::volatility::{calculate_atr, calculate_bollinger_bands};

/// Calculates lagged features for a given column
pub fn calculate_lagged_features(
    df: &DataFrame,
    column: &str,
    lags: &[usize],
) -> PolarsResult<Vec<Series>> {
    let series = df.column(column)?.as_materialized_series().clone();
    let mut result = Vec::with_capacity(lags.len());

    for &lag in lags {
        let lagged = series.shift(lag as i64);
        let name = format!("{}_lag_{}", column, lag);
        result.push(lagged.with_name(name.into()));
    }

    Ok(result)
}

/// Calculates returns over different periods
pub fn calculate_period_returns(df: &DataFrame, periods: &[usize]) -> PolarsResult<Vec<Series>> {
    let close = df.column("close")?.f64()?.clone().into_series();
    let mut result = Vec::with_capacity(periods.len());

    for &period in periods {
        let shifted = close.shift(period as i64);
        // Manual calculation of return
        let mut returns = Vec::with_capacity(df.height());

        let close_values: Vec<_> = close.f64()?.into_iter().collect();
        let shifted_values: Vec<_> = shifted.f64()?.into_iter().collect();

        for i in 0..close_values.len() {
            if let (Some(curr), Some(prev)) = (close_values[i], shifted_values[i]) {
                if prev != 0.0 {
                    returns.push(Some((curr - prev) / prev));
                } else {
                    returns.push(Some(0.0));
                }
            } else {
                returns.push(None);
            }
        }

        let name = format!("returns_{}min", period);
        result.push(Series::new(name.into(), returns));
    }
    Ok(result)
}

/// Adds all technical indicators to the DataFrame
pub fn add_technical_indicators(df: &mut DataFrame) -> PolarsResult<DataFrame> {
    // Clone the DataFrame to avoid borrowing issues
    let mut result_df = df.clone();
    let df_height = result_df.height();

    // Ensure all price columns are Float64
    for col_name in ["open", "high", "low", "close", "volume"].iter() {
        if result_df.column(col_name).is_ok() {
            // Convert to Float64 if needed
            let series = result_df.column(col_name)?;
            let f64_series = series.cast(&DataType::Float64)?;
            result_df.with_column(f64_series)?;
        }
    }

    // Helper function to ensure all series have the same length as the DataFrame
    fn ensure_same_length(series: Series, df_height: usize) -> Series {
        if series.len() < df_height {
            // Pad with nulls at the beginning to match DataFrame height
            let missing = df_height - series.len();
            let mut padded = vec![None; missing];
            // Collect the series values
            let values: Vec<Option<f64>> = series.f64().unwrap().into_iter().collect();
            // Append the non-null values
            padded.extend(values);
            Series::new(series.name().to_string().into(), padded)
        } else {
            series
        }
    }

    // Add SMA using rustalib
    let sma_20 = calculate_sma(&result_df, "close", 20)?;
    let sma_20 = ensure_same_length(sma_20.with_name("sma_20".into()), df_height);
    result_df.with_column(sma_20)?;
    
    let sma_50 = calculate_sma(&result_df, "close", 50)?;
    let sma_50 = ensure_same_length(sma_50.with_name("sma_50".into()), df_height);
    result_df.with_column(sma_50)?;

    // Add EMA using rustalib
    let ema_20 = calculate_ema(&result_df, "close", 20)?;
    let ema_20 = ensure_same_length(ema_20.with_name("ema_20".into()), df_height);
    result_df.with_column(ema_20)?;

    // Calculate returns
    let close_series = result_df.column("close")?.clone();
    let close_vals: Vec<Option<f64>> = close_series.f64()?.into_iter().collect();
    let mut returns = Vec::with_capacity(close_vals.len());

    // Manual calculation to avoid operator issues
    for i in 1..close_vals.len() {
        if let (Some(curr), Some(prev)) = (close_vals[i], close_vals[i - 1]) {
            if prev != 0.0 {
                returns.push(Some((curr - prev) / prev));
            } else {
                returns.push(Some(0.0));
            }
        } else {
            returns.push(None);
        }
    }

    // Add leading value to match length
    returns.insert(0, None);
    result_df.with_column(Series::new("returns".into(), returns))?;

    // Calculate price range (High - Low) / Close using manual calculation
    let high_vals: Vec<Option<f64>> = result_df.column("high")?.f64()?.into_iter().collect();
    let low_vals: Vec<Option<f64>> = result_df.column("low")?.f64()?.into_iter().collect();
    let close_vals: Vec<Option<f64>> = result_df.column("close")?.f64()?.into_iter().collect();

    let mut price_range = Vec::with_capacity(close_vals.len());
    for i in 0..close_vals.len() {
        if let (Some(h), Some(l), Some(c)) = (high_vals[i], low_vals[i], close_vals[i]) {
            if c != 0.0 {
                price_range.push(Some((h - l) / c));
            } else {
                price_range.push(Some(0.0));
            }
        } else {
            price_range.push(None);
        }
    }

    result_df.with_column(Series::new("price_range".into(), price_range))?;

    // Add RSI using rustalib
    let rsi_14 = calculate_rsi(&result_df, 14, "close")?;
    let rsi_14 = ensure_same_length(rsi_14.with_name("rsi_14".into()), df_height);
    result_df.with_column(rsi_14)?;

    // Add MACD using rustalib
    let (macd_series, signal_series) = calculate_macd(&result_df, 12, 26, 9, "close")?;
    let macd_series = ensure_same_length(macd_series.with_name("macd".into()), df_height);
    let signal_series = ensure_same_length(signal_series.with_name("macd_signal".into()), df_height);
    result_df.with_column(macd_series)?;
    result_df.with_column(signal_series)?;

    // Add Bollinger Bands using rustalib
    let (bb_middle, bb_upper, bb_lower) = calculate_bollinger_bands(&result_df, 20, 2.0, "close")?;
    let bb_middle = ensure_same_length(bb_middle.with_name("bb_middle".into()), df_height);
    let bb_upper = ensure_same_length(bb_upper.with_name("bb_upper".into()), df_height);
    let bb_lower = ensure_same_length(bb_lower.with_name("bb_lower".into()), df_height);
    result_df.with_column(bb_middle)?;
    result_df.with_column(bb_upper)?;
    result_df.with_column(bb_lower)?;

    // Add ATR using rustalib
    let atr_14 = calculate_atr(&result_df, 14)?;
    let atr_14 = ensure_same_length(atr_14.with_name("atr_14".into()), df_height);
    result_df.with_column(atr_14)?;

    Ok(result_df)
}