openml 0.1.2

A rust interface to [OpenML](http://openml.org/).
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

//! implementations to convert the API's JSON responses into corresponding Rust structures
use arff;
use arff::dynamic::DataSet as ArffDataSet;
use serde_json;

use dataset::DataSet;
use error::Result;
use procedures::{Fold, FrozenSets};
use tasks::{SupervisedClassification, SupervisedRegression};

use super::api_types::{CrossValItem, GenericResponse, TrainTest};
use super::web_access::get_cached;

impl DataSet {
    fn from_json(item: &serde_json::Value) -> Self {
        let v = &item["data_set"];
        let id = v["data_set_id"].as_str().unwrap();
        let target = v["target_feature"].as_str();

        let info_url = format!("https://www.openml.org/api/v1/json/data/{}", id);
        let info: GenericResponse = serde_json::from_str(&get_cached(&info_url).unwrap()).unwrap();

        let default_target = info.look_up("/data_set_description/default_target_attribute")
            .and_then(|v| v.as_str());

        let target = match (default_target, target) {
            (Some(s), None) | (_, Some(s)) => Some(s.to_owned()),
            (None, None) => None,
        };

        let dset_url = info.look_up("/data_set_description/url")
            .unwrap()
            .as_str()
            .unwrap();
        let dset_str = get_cached(&dset_url).unwrap();
        let dset = ArffDataSet::from_str(&dset_str).unwrap();

        DataSet { arff: dset, target }
    }
}

impl SupervisedClassification {
    pub fn from_json(task_json: &serde_json::Value) -> Self {
        let mut source_data = None;
        let mut estimation_procedure = None;
        //let mut cost_matrix = None;

        for input_item in task_json["input"].as_array().unwrap() {
            match input_item["name"].as_str() {
                Some("source_data") => source_data = Some(DataSet::from_json(input_item)),
                Some("estimation_procedure") => {
                    estimation_procedure = Some(Box::new(FrozenSets::from_json(input_item)))
                }
                //Some("cost_matrix") => cost_matrix = Some(input_item.into()),
                Some(_) => {}
                None => panic!("/task/input/name is not a string"),
            }
        }

        SupervisedClassification {
            id: task_json["task_id"].as_str().unwrap().to_owned(),
            name: task_json["task_name"].as_str().unwrap().to_owned(),
            source_data: source_data.unwrap(),
            estimation_procedure: estimation_procedure.unwrap(),
            //cost_matrix: cost_matrix.unwrap(),
        }
    }
}

impl SupervisedRegression {
    pub fn from_json(task_json: &serde_json::Value) -> Self {
        let mut source_data = None;
        let mut estimation_procedure = None;

        for input_item in task_json["input"].as_array().unwrap() {
            match input_item["name"].as_str() {
                Some("source_data") => source_data = Some(DataSet::from_json(input_item)),
                Some("estimation_procedure") => {
                    estimation_procedure = Some(Box::new(FrozenSets::from_json(input_item)))
                }
                Some(_) => {}
                None => panic!("/task/input/name is not a string"),
            }
        }

        SupervisedRegression {
            id: task_json["task_id"].as_str().unwrap().to_owned(),
            name: task_json["task_name"].as_str().unwrap().to_owned(),
            source_data: source_data.unwrap(),
            estimation_procedure: estimation_procedure.unwrap(),
        }
    }
}

impl FrozenSets {
    fn from_json(item: &serde_json::Value) -> Self {
        let v = &item["estimation_procedure"];
        let typ = v["type"].as_str();
        let splits = v["data_splits_url"].as_str();

        match (typ, splits) {
            (_, Some(url)) => FrozenSets::from_url(url).unwrap(),
            _ => unimplemented!(),
        }
    }

    fn from_url(url: &str) -> Result<Self> {
        let raw = get_cached(url)?;
        let data: Vec<CrossValItem> = arff::from_str(&raw)?;

        let mut folds = vec![];
        for item in data {
            if item.repeat >= folds.len() {
                folds.resize(item.repeat + 1, vec![]);
            }
            let mut rep = &mut folds[item.repeat];

            if item.fold >= rep.len() {
                rep.resize(item.fold + 1, Fold::new());
            }
            let mut fold = &mut rep[item.fold];

            match item.purpose {
                TrainTest::Train => fold.trainset.push(item.rowid),
                TrainTest::Test => fold.testset.push(item.rowid),
            }
        }

        Ok(FrozenSets { folds })
    }
}