liboxen 0.7.2

Oxen is a fast, unstructured data version control, to help version large machine learning datasets written in Rust.
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

use crate::constants::{FIELDS_DIR, HISTORY_DIR, INDICES_DIR};
use crate::core::db;
use crate::error::OxenError;
use crate::model::{schema::Field, Commit, Schema};
use crate::util;

use rocksdb::{DBWithThreadMode, MultiThreaded};
use std::path::PathBuf;
use std::str;

use crate::model::LocalRepository;

pub struct SchemaFieldValIndex {
    field_indices_db: DBWithThreadMode<MultiThreaded>,
}

impl SchemaFieldValIndex {
    pub fn db_dir(
        repo: &LocalRepository,
        commit: &Commit,
        schema: &Schema,
        field: &Field,
    ) -> PathBuf {
        let key_hash = util::hasher::hash_str(&field.name);
        // .oxen/history/COMMIT_ID/indices/SCHEMA_HASH/fields/FIELD_NAME_HASH
        util::fs::oxen_hidden_dir(&repo.path)
            .join(HISTORY_DIR)
            .join(&commit.id)
            .join(INDICES_DIR)
            .join(&schema.hash)
            .join(FIELDS_DIR)
            .join(key_hash)
    }

    pub fn new(
        repository: &LocalRepository,
        commit: &Commit,
        schema: &Schema,
        field: &Field,
    ) -> Result<SchemaFieldValIndex, OxenError> {
        let db_path = SchemaFieldValIndex::db_dir(repository, commit, schema, field);
        log::debug!("SchemaFieldValIndex db {:?}", db_path);
        let opts = db::opts::default();
        if !db_path.exists() {
            std::fs::create_dir_all(&db_path)?;
        }

        Ok(SchemaFieldValIndex {
            field_indices_db: DBWithThreadMode::open(&opts, dunce::simplified(&db_path))?,
        })
    }

    pub fn insert_index<S: AsRef<str>>(&self, key: S, indices: Vec<u32>) -> Result<(), OxenError> {
        db::index_db::insert_indices(&self.field_indices_db, key, indices)
    }

    pub fn get_indices<S: AsRef<str>>(&self, key: S) -> Result<Option<Vec<u32>>, OxenError> {
        db::index_db::get_indices(&self.field_indices_db, key)
    }
}

#[cfg(test)]
mod tests {
    use crate::core::index::SchemaFieldValIndex;
    use crate::error::OxenError;
    use crate::model::schema;
    use crate::test;
    use crate::{api, command};

    #[test]
    fn test_list_empty_indices() -> Result<(), OxenError> {
        test::run_training_data_repo_test_fully_committed(|repo| {
            let history = api::local::commits::list(&repo)?;
            let last_commit = history.first().unwrap();
            let schemas = command::schemas::list(&repo, Some(&last_commit.id))?;
            let schema = schemas
                .iter()
                .find(|s| s.name.as_ref().unwrap() == "bounding_box")
                .unwrap();

            // As if we aggregated up the "label" field and "dog" was a value
            let field = schema::Field {
                name: String::from("label"),
                dtype: String::from("str"),
            };

            let reader = SchemaFieldValIndex::new(&repo, last_commit, schema, &field)?;
            let indices = reader.get_indices("cat")?;

            assert!(indices.is_none());

            Ok(())
        })
    }

    #[test]
    fn test_insert_index() -> Result<(), OxenError> {
        test::run_training_data_repo_test_fully_committed(|repo| {
            let history = api::local::commits::list(&repo)?;
            let last_commit = history.first().unwrap();
            let schemas = command::schemas::list(&repo, Some(&last_commit.id))?;
            let schema = schemas
                .iter()
                .find(|s| s.name.as_ref().unwrap() == "bounding_box")
                .unwrap();

            // As if we aggregated up the "label" field and "dog" was a value
            let field = schema::Field {
                name: String::from("label"),
                dtype: String::from("str"),
            };

            let index = SchemaFieldValIndex::new(&repo, last_commit, schema, &field)?;

            let indices: Vec<u32> = vec![4, 3, 2, 5, 1];
            index.insert_index("cat", indices)?;

            let indices = index.get_indices("cat")?.unwrap();
            assert_eq!(indices.len(), 5);

            Ok(())
        })
    }

    #[test]
    fn test_insert_large_index() -> Result<(), OxenError> {
        test::run_training_data_repo_test_fully_committed(|repo| {
            let history = api::local::commits::list(&repo)?;
            let last_commit = history.first().unwrap();
            let schemas = command::schemas::list(&repo, Some(&last_commit.id))?;
            let schema = schemas
                .iter()
                .find(|s| s.name.as_ref().unwrap() == "bounding_box")
                .unwrap();

            // As if we aggregated up the "label" field and "dog" was a value
            let field = schema::Field {
                name: String::from("label"),
                dtype: String::from("str"),
            };

            let index = SchemaFieldValIndex::new(&repo, last_commit, schema, &field)?;

            // Large insert failed when working with the bytevec lib, so had to write some custom code
            let num_indices: u32 = 1000000;
            let mut indices: Vec<u32> = vec![];
            for i in 0..num_indices {
                indices.push(i);
            }
            index.insert_index("cat", indices)?;

            let indices = index.get_indices("cat")?.unwrap();
            assert_eq!(indices.len(), num_indices as usize);

            Ok(())
        })
    }
}