nautilus_serialization/arrow/instrument/
index_instrument.rs

1// -------------------------------------------------------------------------------------------------
2//  Copyright (C) 2015-2026 Nautech Systems Pty Ltd. All rights reserved.
3//  https://nautechsystems.io
4//
5//  Licensed under the GNU Lesser General Public License Version 3.0 (the "License");
6//  You may not use this file except in compliance with the License.
7//  You may obtain a copy of the License at https://www.gnu.org/licenses/lgpl-3.0.en.html
8//
9//  Unless required by applicable law or agreed to in writing, software
10//  distributed under the License is distributed on an "AS IS" BASIS,
11//  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
12//  See the License for the specific language governing permissions and
13//  limitations under the License.
14// -------------------------------------------------------------------------------------------------
15
16//! Arrow serialization for IndexInstrument instruments.
17
18use std::{collections::HashMap, str::FromStr, sync::Arc};
19
20use arrow::{
21    array::{
22        Array, BinaryArray, BinaryBuilder, StringArray, StringBuilder, UInt8Array, UInt64Array,
23    },
24    datatypes::{DataType, Field, Schema},
25    error::ArrowError,
26    record_batch::RecordBatch,
27};
28use nautilus_core::{Params, UnixNanos};
29use nautilus_model::{
30    identifiers::{InstrumentId, Symbol},
31    instruments::index_instrument::IndexInstrument,
32    types::{price::Price, quantity::Quantity},
33};
34
35use super::KEY_CLASS;
36use crate::arrow::{
37    ArrowSchemaProvider, EncodeToRecordBatch, EncodingError, KEY_INSTRUMENT_ID,
38    KEY_PRICE_PRECISION, extract_column, extract_column_by_name_or_index,
39    extract_optional_string_column_by_name, optional_ustr_value,
40};
41
42impl ArrowSchemaProvider for IndexInstrument {
43    fn get_schema(metadata: Option<HashMap<String, String>>) -> Schema {
44        let fields = vec![
45            Field::new("id", DataType::Utf8, false),
46            Field::new("raw_symbol", DataType::Utf8, false),
47            Field::new("currency", DataType::Utf8, false),
48            Field::new("price_precision", DataType::UInt8, false),
49            Field::new("price_increment", DataType::Utf8, false),
50            Field::new("size_precision", DataType::UInt8, false),
51            Field::new("size_increment", DataType::Utf8, false),
52            Field::new("tick_scheme", DataType::Utf8, true),
53            Field::new("info", DataType::Binary, true), // nullable
54            Field::new("ts_event", DataType::UInt64, false),
55            Field::new("ts_init", DataType::UInt64, false),
56        ];
57
58        let mut final_metadata = HashMap::new();
59        final_metadata.insert(KEY_CLASS.to_string(), "IndexInstrument".to_string());
60
61        if let Some(meta) = metadata {
62            final_metadata.extend(meta);
63        }
64
65        Schema::new_with_metadata(fields, final_metadata)
66    }
67}
68
69impl EncodeToRecordBatch for IndexInstrument {
70    fn encode_batch(
71        #[allow(unused)] metadata: &HashMap<String, String>,
72        data: &[Self],
73    ) -> Result<RecordBatch, ArrowError> {
74        let mut id_builder = StringBuilder::new();
75        let mut raw_symbol_builder = StringBuilder::new();
76        let mut currency_builder = StringBuilder::new();
77        let mut price_precision_builder = UInt8Array::builder(data.len());
78        let mut size_precision_builder = UInt8Array::builder(data.len());
79        let mut price_increment_builder = StringBuilder::new();
80        let mut size_increment_builder = StringBuilder::new();
81        let mut tick_scheme_builder = StringBuilder::new();
82        let mut info_builder = BinaryBuilder::new();
83        let mut ts_event_builder = UInt64Array::builder(data.len());
84        let mut ts_init_builder = UInt64Array::builder(data.len());
85
86        for index in data {
87            id_builder.append_value(index.id.to_string());
88            raw_symbol_builder.append_value(index.raw_symbol);
89            currency_builder.append_value(index.currency.to_string());
90            price_precision_builder.append_value(index.price_precision);
91            price_increment_builder.append_value(index.price_increment.to_string());
92            size_precision_builder.append_value(index.size_precision);
93            size_increment_builder.append_value(index.size_increment.to_string());
94
95            if let Some(tick_scheme) = index.tick_scheme {
96                tick_scheme_builder.append_value(tick_scheme);
97            } else {
98                tick_scheme_builder.append_null();
99            }
100
101            // Encode info dict as JSON bytes (matching Python's msgspec.json.encode)
102            if let Some(ref info) = index.info {
103                match serde_json::to_vec(info) {
104                    Ok(json_bytes) => {
105                        info_builder.append_value(json_bytes);
106                    }
107                    Err(e) => {
108                        return Err(ArrowError::InvalidArgumentError(format!(
109                            "Failed to serialize info dict to JSON: {e}"
110                        )));
111                    }
112                }
113            } else {
114                info_builder.append_null();
115            }
116
117            ts_event_builder.append_value(index.ts_event.as_u64());
118            ts_init_builder.append_value(index.ts_init.as_u64());
119        }
120
121        let mut final_metadata = metadata.clone();
122        final_metadata.insert(KEY_CLASS.to_string(), "IndexInstrument".to_string());
123
124        RecordBatch::try_new(
125            Self::get_schema(Some(final_metadata)).into(),
126            vec![
127                Arc::new(id_builder.finish()),
128                Arc::new(raw_symbol_builder.finish()),
129                Arc::new(currency_builder.finish()),
130                Arc::new(price_precision_builder.finish()),
131                Arc::new(price_increment_builder.finish()),
132                Arc::new(size_precision_builder.finish()),
133                Arc::new(size_increment_builder.finish()),
134                Arc::new(tick_scheme_builder.finish()),
135                Arc::new(info_builder.finish()),
136                Arc::new(ts_event_builder.finish()),
137                Arc::new(ts_init_builder.finish()),
138            ],
139        )
140    }
141
142    fn metadata(&self) -> HashMap<String, String> {
143        let mut metadata = HashMap::new();
144        metadata.insert(KEY_INSTRUMENT_ID.to_string(), self.id.to_string());
145        metadata.insert(
146            KEY_PRICE_PRECISION.to_string(),
147            self.price_precision.to_string(),
148        );
149        metadata
150    }
151}
152
153/// Decodes [`IndexInstrument`] instruments from a record batch.
154///
155/// Not a [`DecodeFromRecordBatch`] implementation because that trait requires `Into<Data>`.
156///
157/// # Errors
158///
159/// Returns an `EncodingError` if the record batch cannot be decoded.
160///
161/// [`DecodeFromRecordBatch`]: crate::arrow::DecodeFromRecordBatch
162pub fn decode_index_instrument_batch(
163    #[allow(unused)] metadata: &HashMap<String, String>,
164    record_batch: &RecordBatch,
165) -> Result<Vec<IndexInstrument>, EncodingError> {
166    let cols = record_batch.columns();
167    let num_rows = record_batch.num_rows();
168
169    let id_values = extract_column::<StringArray>(cols, "id", 0, DataType::Utf8)?;
170    let raw_symbol_values = extract_column::<StringArray>(cols, "raw_symbol", 1, DataType::Utf8)?;
171    let currency_values = extract_column::<StringArray>(cols, "currency", 2, DataType::Utf8)?;
172    let price_precision_values =
173        extract_column::<UInt8Array>(cols, "price_precision", 3, DataType::UInt8)?;
174    let price_increment_values =
175        extract_column::<StringArray>(cols, "price_increment", 4, DataType::Utf8)?;
176    let size_precision_values =
177        extract_column::<UInt8Array>(cols, "size_precision", 5, DataType::UInt8)?;
178    let size_increment_values =
179        extract_column::<StringArray>(cols, "size_increment", 6, DataType::Utf8)?;
180    let tick_scheme_values = extract_optional_string_column_by_name(record_batch, "tick_scheme")?;
181    let info_values =
182        extract_column_by_name_or_index::<BinaryArray>(record_batch, "info", 7, DataType::Binary)?;
183    let ts_event_values = extract_column_by_name_or_index::<UInt64Array>(
184        record_batch,
185        "ts_event",
186        8,
187        DataType::UInt64,
188    )?;
189    let ts_init_values = extract_column_by_name_or_index::<UInt64Array>(
190        record_batch,
191        "ts_init",
192        9,
193        DataType::UInt64,
194    )?;
195
196    let mut result = Vec::with_capacity(num_rows);
197
198    for i in 0..num_rows {
199        let id = InstrumentId::from_str(id_values.value(i))
200            .map_err(|e| EncodingError::ParseError("id", format!("row {i}: {e}")))?;
201        let raw_symbol = Symbol::from(raw_symbol_values.value(i));
202        let currency = super::decode_currency(
203            currency_values.value(i),
204            "currency",
205            "index_instrument.currency",
206            i,
207        )?;
208        let price_prec = price_precision_values.value(i);
209        let size_prec = size_precision_values.value(i);
210
211        let price_increment = Price::from_str(price_increment_values.value(i))
212            .map_err(|e| EncodingError::ParseError("price_increment", format!("row {i}: {e}")))?;
213        let size_increment = Quantity::from_str(size_increment_values.value(i))
214            .map_err(|e| EncodingError::ParseError("size_increment", format!("row {i}: {e}")))?;
215
216        // Decode info dict from JSON bytes (matching Python's msgspec.json.decode)
217        let info = if info_values.is_null(i) {
218            None
219        } else {
220            let info_bytes = info_values
221                .as_any()
222                .downcast_ref::<BinaryArray>()
223                .ok_or_else(|| EncodingError::ParseError("info", format!("row {i}: invalid type")))?
224                .value(i);
225
226            match serde_json::from_slice::<Params>(info_bytes) {
227                Ok(info_dict) => Some(info_dict),
228                Err(e) => {
229                    return Err(EncodingError::ParseError(
230                        "info",
231                        format!("row {i}: failed to deserialize JSON: {e}"),
232                    ));
233                }
234            }
235        };
236
237        let ts_event = UnixNanos::from(ts_event_values.value(i));
238        let ts_init = UnixNanos::from(ts_init_values.value(i));
239
240        let tick_scheme = optional_ustr_value(tick_scheme_values, i);
241
242        let index_instrument = IndexInstrument::builder()
243            .instrument_id(id)
244            .raw_symbol(raw_symbol)
245            .currency(currency)
246            .price_precision(price_prec)
247            .size_precision(size_prec)
248            .price_increment(price_increment)
249            .size_increment(size_increment)
250            .maybe_tick_scheme(tick_scheme)
251            .maybe_info(info)
252            .ts_event(ts_event)
253            .ts_init(ts_init)
254            .build()
255            .map_err(|e| super::instrument_validation_error::<IndexInstrument>(i, e))?;
256
257        result.push(index_instrument);
258    }
259
260    Ok(result)
261}
nautilus_serialization/arrow/instrument/index_instrument.rs

nautilus_serialization/arrow/instrument/
index_instrument.rs
