use u128s for length to prevent under/overflow (?)

src/calculations.rs

@@ -1,4 +1,4 @@
-pub fn get_n(lengths: &Vec<u64>, nb_bases_total: u64, percentile: f64) -> u64 {
+pub fn get_n(lengths: &Vec<u128>, nb_bases_total: u128, percentile: f64) -> u128 {
     let mut acc = 0;
     for val in lengths.iter() {
         acc += *val;
@@ -20,7 +20,7 @@ pub fn median<T: Into<f64> + Copy>(array: &[T]) -> f64 {
     }
 }
 
-pub fn median_length(array: &[u64]) -> f64 {
+pub fn median_length(array: &[u128]) -> f64 {
     if (array.len() % 2) == 0 {
         let ind_left = array.len() / 2 - 1;
         let ind_right = array.len() / 2;

src/extract_from_bam.rs

@@ -3,7 +3,7 @@ use rust_htslib::bam::record::{Aux, Cigar};
 use rust_htslib::{bam, bam::Read, htslib};
 
 pub struct Data {
-    pub lengths: Option<Vec<u64>>,
+    pub lengths: Option<Vec<u128>>,
     pub all_counts: usize,
     pub identities: Option<Vec<f64>>,
     pub tids: Option<Vec<i32>>,
@@ -59,9 +59,9 @@ pub fn extract(args: &crate::Cli) -> Data {
         .filter(|read| filter_closure(read))
     {
         lengths.push(
-            read.seq_len() as u64
-                - read.cigar().leading_softclips() as u64
-                - read.cigar().trailing_softclips() as u64,
+            read.seq_len() as u128
+                - read.cigar().leading_softclips() as u128
+                - read.cigar().trailing_softclips() as u128,
         );
         if args.karyotype || args.phased {
             tids.push(read.tid());
@@ -80,10 +80,15 @@ pub fn extract(args: &crate::Cli) -> Data {
     }
     if let Some(s) = &args.arrow {
         match args.ubam {
-            true => crate::feather::save_as_arrow_ubam(s.to_string(), lengths.clone()),
-            false => {
-                crate::feather::save_as_arrow(s.to_string(), lengths.clone(), identities.clone())
-            }
+            true => crate::feather::save_as_arrow_ubam(
+                s.to_string(),
+                lengths.clone().iter().map(|x| *x as u64).collect(),
+            ),
+            false => crate::feather::save_as_arrow(
+                s.to_string(),
+                lengths.clone().iter().map(|x| *x as u64).collect(),
+                identities.clone(),
+            ),
         }
     }
     // sort vectors in descending order (required for N50/N75)

src/histograms.rs

@@ -11,8 +11,8 @@ use std::cmp::max;
 // as well as for future customizations
 // in principle it would be possible to enable the user to change the step size or max value, but I don't want to add too many options to the CLI
 
-pub fn make_histogram_lengths(array: &[u64]) {
-    let stepsize: u64 = 2000;
+pub fn make_histogram_lengths(array: &[u128]) {
+    let stepsize: u128 = 2000;
     let max_value = 60_000;
     let step_count = max_value / stepsize as usize;
     let mut counts = vec![0; step_count + 1];
@@ -34,7 +34,7 @@ pub fn make_histogram_lengths(array: &[u64]) {
             "{: >11} {}",
             format!(
                 "{}-{}",
-                index as u64 * stepsize,
+                index as u128 * stepsize,
                 (index + 1) * stepsize as usize
             ),
             "∎".repeat(entry / dotsize)

src/main.rs

@@ -142,7 +142,7 @@ fn metrics_from_bam(metrics: Data, args: Cli) {
 }
 
 fn generate_main_output(
-    lengths: &Vec<u64>,
+    lengths: &Vec<u128>,
     identities: Option<&Vec<f64>>,
     genome_size: u64,
     all_reads: usize,
@@ -152,7 +152,7 @@ fn generate_main_output(
         error!("Not enough reads to calculate metrics!");
         panic!();
     }
-    let data_yield: u64 = lengths.iter().sum::<u64>();
+    let data_yield: u128 = lengths.iter().sum::<u128>();
     println!("Number of alignments\t{num_reads}");
     println!(
         "% from total reads\t{:.2}",
@@ -163,12 +163,12 @@ fn generate_main_output(
         "Mean coverage\t{:.2}",
         data_yield as f64 / genome_size as f64
     );
-    let data_yield_long = lengths.iter().filter(|l| l > &&25000).sum::<u64>();
+    let data_yield_long = lengths.iter().filter(|l| l > &&25000).sum::<u128>();
     println!("Yield [Gb] (>25kb)\t{:.2}", data_yield_long as f64 / 1e9);
     println!("N50\t{}", calculations::get_n(lengths, data_yield, 0.50));
     println!("N75\t{}", calculations::get_n(lengths, data_yield, 0.75));
     println!("Median length\t{:.2}", calculations::median_length(lengths));
-    println!("Mean length\t{:.2}", data_yield / num_reads as u64);
+    println!("Mean length\t{:.2}", data_yield / num_reads as u128);
     if let Some(identities) = identities {
         println!("Median identity\t{:.2}", calculations::median(identities));
         println!(