Function describe_obs

src/memory/statistics/qc.rs:117–212

Calculate cell-level (observation) quality control metrics. Computes comprehensive QC statistics for each cell including expression totals, gene detection rates, and specialized metrics like mitochondrial gene percentages. ## Parameters `adata` - AnnData object containing the dataset `x` - Expression matrix to analyze `expr_type` - Label for expression type (e.g., "counts", "UMI") `var_type` - L

(
    adata: &IMAnnData,
    x: &IMArrayElement,
    expr_type: &str,
    var_type: &str,
    qc_vars: &[&str],
    percent_top: &[usize],
    log1p: bool,
)

/// - Expression concentration in top N genes
/// - Optional log1p transformations for all count metrics
fn describe_obs(
    adata: &IMAnnData,
    x: &IMArrayElement,
    expr_type: &str,
    var_type: &str,
    qc_vars: &[&str],
    percent_top: &[usize],
    log1p: bool,
) -> anyhow::Result<DataFrame> {
    let n_obs = adata.n_obs();
    let n_vars = adata.n_vars();

    let n_genes_by_counts: Vec<u32> = x.nonzero_whole(&Direction::ROW)?;
    let total_counts: Vec<f64> = x.sum_whole(&Direction::ROW)?;

    let mut columns = vec![];

    let col_name = format!("n_{}_by_{}", var_type, expr_type);
    columns.push(Column::new(col_name.into(), n_genes_by_counts.clone()));

    if log1p {
        let log_values: Vec<f64> = n_genes_by_counts
            .iter()
            .map(|&x| (x as f64 + 1.0).ln())
            .collect();
        let col_name = format!("log1p_n_{}_by_{}", var_type, expr_type);
        columns.push(Column::new(col_name.into(), log_values));
    }

    let col_name = format!("total_{}", expr_type);
    columns.push(Column::new(col_name.into(), total_counts.clone()));

    if log1p {
        let log_values: Vec<f64> = total_counts.iter().map(|&x| (x + 1.0).ln()).collect();
        let col_name = format!("log1p_total_{}", expr_type);
        columns.push(Column::new(col_name.into(), log_values));
    }

    if !percent_top.is_empty() {
        use crate::shared::statistics::ComputeTopSegmentProportions;
        let proportions = x.top_segment_proportions(&Direction::ROW, percent_top)?;
        for (i, &n) in percent_top.iter().enumerate() {
            let values: Vec<f64> = proportions.column(i).iter().map(|&x| x * 100.0).collect();
            let col_name = format!("pct_{}_in_top_{}_{}", expr_type, n, var_type);
            columns.push(Column::new(col_name.into(), values));
        }
    }

    for qc_var in qc_vars {
        let var_mask = adata
            .var()
            .get_column_from_df(qc_var)?
            .bool()?
            .into_iter()
            .map(|x| x.unwrap_or(false))
            .collect::<Vec<bool>>();

        let qc_var_indices: Vec<usize> = var_mask