Function get_unique_groups

src/memory/processing/diffexp/mod.rs:739–797

Extract unique group labels from a categorical or string column in the observation metadata. This function handles different data types commonly used for group labels: - String columns - Integer columns (converted to strings) - Categorical columns (with proper ordering preservation) # Arguments `adata` - The AnnData object `groupby` - Column name in `adata.obs` containing group labels # Returns

(adata: &IMAnnData, groupby: &str)

/// # Returns
/// Sorted vector of unique group names as strings
fn get_unique_groups(adata: &IMAnnData, groupby: &str) -> anyhow::Result<Vec<String>> {
    let group_col = adata.obs().get_column_from_df(groupby)?;

    let mut all_groups: Vec<String> = match group_col.dtype() {
        DataType::String => {
            let string_col = group_col.str()?;
            let mut unique_groups = std::collections::HashSet::new();

            for i in 0..string_col.len() {
                if let Some(value) = string_col.get(i) {
                    unique_groups.insert(value.to_string());
                }
            }

            unique_groups.into_iter().collect()
        }
        DataType::Int8
        | DataType::Int16
        | DataType::Int32
        | DataType::Int64
        | DataType::UInt8
        | DataType::UInt16
        | DataType::UInt32
        | DataType::UInt64 => {
            let int_col = group_col.i64()?;
            let mut unique_groups = std::collections::HashSet::new();

            for i in 0..int_col.len() {
                if let Some(value) = int_col.get(i) {
                    unique_groups.insert(value.to_string());
                }
            }

            unique_groups.into_iter().collect()
        }
        DataType::Categorical(_, _) => {
            let string_col = group_col.cast(&DataType::String)?;
            let string_col = string_col.str()?;
            let mut unique_groups = std::collections::HashSet::new();

            for i in 0..string_col.len() {
                if let Some(value) = string_col.get(i) {
                    unique_groups.insert(value.to_string());
                }
            }

            unique_groups.into_iter().collect()
        }
        other => {
            return Err(anyhow::anyhow!(
                "Unsupported data type for groupby column: {:?}",
                other
            ));
        }
    };

    all_groups.sort();
    Ok(all_groups)