Function data_clean_ohlcv

crates/pyopenquant/src/data.rs:10–57

(
    py: Python<'_>,
    timestamps_us: Vec<i64>,
    symbols: Vec<String>,
    open: Vec<f64>,
    high: Vec<f64>,
    low: Vec<f64>,
    close: Vec<f64>,
    volume: Vec<f64>,
    adj_close: Vec<f6

use crate::helpers::{build_ohlcv_columns, report_to_pydict, to_py_err};

#[pyfunction(name = "clean_ohlcv")]
fn data_clean_ohlcv(
    py: Python<'_>,
    timestamps_us: Vec<i64>,
    symbols: Vec<String>,
    open: Vec<f64>,
    high: Vec<f64>,
    low: Vec<f64>,
    close: Vec<f64>,
    volume: Vec<f64>,
    adj_close: Vec<f64>,
    dedupe_keep_last: bool,
) -> PyResult<(
    Vec<i64>,
    Vec<String>,
    Vec<f64>,
    Vec<f64>,
    Vec<f64>,
    Vec<f64>,
    Vec<f64>,
    Vec<f64>,
    PyObject,
)> {
    let cols =
        build_ohlcv_columns(timestamps_us, symbols, open, high, low, close, volume, adj_close)?;
    let (clean, report) = clean_ohlcv_columns(&cols, dedupe_keep_last).map_err(to_py_err)?;

    let out_report = PyDict::new(py);
    out_report.set_item("row_count", report.row_count)?;
    out_report.set_item("symbol_count", report.symbol_count)?;
    out_report.set_item("duplicate_key_count", report.duplicate_key_count)?;
    out_report.set_item("gap_interval_count", report.gap_interval_count)?;
    out_report
        .set_item("ts_min", report.ts_min.map(|v| v.format("%Y-%m-%d %H:%M:%S").to_string()))?;
    out_report
        .set_item("ts_max", report.ts_max.map(|v| v.format("%Y-%m-%d %H:%M:%S").to_string()))?;
    out_report.set_item("rows_removed_by_deduplication", report.rows_removed_by_deduplication)?;
    Ok((
        clean.timestamps_us,
        clean.symbols,
        clean.open,
        clean.high,
        clean.low,
        clean.close,
        clean.volume,
        clean.adj_close,
        out_report.into_pyobject(py).unwrap().into_any().unbind(),
    ))
}