nettan 0.4.0

nettan
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
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
174
175
176

use polars::error::PolarsResult;
use polars::frame::DataFrame;
use polars::lazy::dsl::Expr;
use polars::prelude::*;

#[derive(Clone, Debug, Default)]
pub struct FeatureOperator {
    feature_expressions: Vec<Expr>,
    derived_feature_expressions: Vec<Expr>,
}

impl FeatureOperator {
    pub fn new() -> FeatureOperator {
        FeatureOperator {
            feature_expressions: Vec::new(),
            derived_feature_expressions: Vec::new(),
        }
    }

    pub fn append_expression(&mut self, expression: Expr) -> &mut FeatureOperator {
        self.feature_expressions.push(expression);

        self
    }

    pub fn append_derived_expression(&mut self, expression: Expr) -> &mut FeatureOperator {
        self.derived_feature_expressions.push(expression);

        self
    }

    pub fn collect_lazy_expressions(
        &self,
        df: DataFrame,
        n_rows: Option<u32>,
        exclude_columns: Option<Vec<String>>,
    ) -> PolarsResult<DataFrame> {
        let exclude_columns = exclude_columns.unwrap_or_default();

        match n_rows {
            Some(n_rows) => df
                .lazy()
                .with_columns(&self.feature_expressions)
                .with_columns(&self.derived_feature_expressions)
                .tail(n_rows)
                .select([all().exclude(exclude_columns)])
                .collect(),
            // TODO: Optionally use collect_concurrently?
            // .collect_concurrently(),
            None => df
                .lazy()
                .with_columns(&self.feature_expressions)
                .with_columns(&self.derived_feature_expressions)
                .select([all().exclude(exclude_columns)])
                .collect(),
            // TODO: Optionally use collect_concurrently?
            // .collect_concurrently(),
        }
    }

    pub fn feature_expressions_count(&self) -> usize {
        self.feature_expressions.len()
    }
}

#[cfg(test)]
mod tests {
    use super::*;

    fn create_feature_operator() -> FeatureOperator {
        let mut feature_operator = FeatureOperator::new();

        let addition_expression = (col("high") + col("low")).alias("addition");
        let subtraction_expression = (col("high") - col("low")).alias("subtraction");
        let division_expression = (col("high") / col("low")).alias("division");
        let multiplication_expression = (col("high") * col("low")).alias("multiplication");

        feature_operator.append_expression(addition_expression);
        feature_operator.append_expression(subtraction_expression);
        feature_operator.append_expression(division_expression);
        feature_operator.append_expression(multiplication_expression);

        feature_operator
    }

    fn create_df() -> DataFrame {
        df!(
            "ticker" => ["AAPL", "NVDA", "MSFT", "GOOG", "AMZN"],
            "price" => [229.9, 138.93, 420.56, 166.41, 188.4],
            "high" => [231.31, 139.6, 424.04, 167.62, 189.83],
            "low" => [228.6, 136.3, 417.52, 164.78, 188.44],
        )
        .unwrap()
    }

    #[test]
    fn count_features() {
        let feature_operator = create_feature_operator();

        assert_eq!(feature_operator.feature_expressions_count(), 4);
    }

    #[test]
    fn collect_expressions() {
        let df = create_df();
        let (height, width) = df.shape();
        let feature_operator = create_feature_operator();
        let feature_count = feature_operator.feature_expressions_count();

        let df = feature_operator
            .collect_lazy_expressions(df, None, None)
            .unwrap();

        dbg!(&df);

        assert_eq!(df.shape(), (height, width + feature_count));
        assert_eq!(
            df.get_column_names(),
            &[
                "ticker",
                "price",
                "high",
                "low",
                "addition",
                "subtraction",
                "division",
                "multiplication"
            ]
        );

        let exclude_columns = vec![String::from("addition"), String::from("subtraction")];
        let df = feature_operator
            .collect_lazy_expressions(df, Some(5), Some(exclude_columns.clone()))
            .unwrap();

        assert_eq!(
            df.shape(),
            (height, width + feature_count - exclude_columns.len())
        );
        assert_eq!(
            df.get_column_names(),
            &[
                "ticker",
                "price",
                "high",
                "low",
                "division",
                "multiplication"
            ]
        );
    }

    #[test]
    fn rolling_window_expression() {
        let df = create_df();
        let mut feature_operator = FeatureOperator::new();

        let rolling_window_expression = (col("price").rolling_mean(RollingOptionsFixedWindow {
            window_size: 3,
            min_periods: 3,
            weights: None,
            center: false,
            fn_params: Default::default(),
        }))
        .alias("rolling_mean");

        feature_operator.append_expression(rolling_window_expression);
        let df = feature_operator
            .collect_lazy_expressions(df, None, None)
            .unwrap();

        dbg!(&df);

        assert!(df.get_column_names_str().to_vec().contains(&"rolling_mean"));
    }
}