From e041be35ba4045f319ac72d3e8badbfab1fb9ffc Mon Sep 17 00:00:00 2001
From: wdecoster <decosterwouter@gmail.com>
Date: Wed, 6 Sep 2023 09:42:35 +0200
Subject: [PATCH] fix histgram dotsize for very small datasets

---
 src/histograms.rs | 7 ++++---
 src/main.rs       | 6 +++---
 2 files changed, 7 insertions(+), 6 deletions(-)

diff --git a/src/histograms.rs b/src/histograms.rs
index 6f6407e..8917b08 100644
--- a/src/histograms.rs
+++ b/src/histograms.rs
@@ -1,4 +1,5 @@
 use itertools::Itertools;
+use std::cmp::max;
 
 // the histograms below are fully defined by the step size and the maximum value
 // the step size is the size of each bin
@@ -25,7 +26,7 @@ pub fn make_histogram_lengths(array: &[u64]) {
     counts[step_count] = array.len() - counts.iter().sum::<usize>();
     // the dotsize variable determines how many reads are represented by a single dot
     // I either have to set this dynamically or experiment with it further
-    let dotsize = array.len() / 500;
+    let dotsize = max(array.len() / 500, 1);
     println!("\n\n# Histogram for read lengths:");
     // print every entry in the vector, except the last one which is done separately
     for (index, entry) in counts.iter().dropping_back(1).enumerate() {
@@ -60,7 +61,7 @@ pub fn make_histogram_identities(array: &[f64]) {
     counts[step_count] = array.len() - counts.iter().sum::<usize>();
     // the dotsize variable determines how many reads are represented by a single dot
     // I either have to set this dynamically or experiment with it further
-    let dotsize = array.len() / 500;
+    let dotsize = max(array.len() / 500, 1);
     println!("\n\n# Histogram for Phred-scaled accuracies:");
     // print every entry in the vector, except the last one which is done separately
     for (index, entry) in counts.iter().dropping_back(1).enumerate() {
@@ -100,7 +101,7 @@ pub fn make_histogram_phaseblocks(array: &[i64]) {
     counts[step_count] = array.len() - counts.iter().sum::<usize>();
     // the dotsize variable determines how many reads are represented by a single dot
     // I either have to set this dynamically or experiment with it further
-    let dotsize = array.len() / 500;
+    let dotsize = max(array.len() / 500, 1);
     println!("\n\n# Histogram for phaseblock lengths:");
     // print every entry in the vector, except the last one which is done separately
     for (index, entry) in counts.iter().dropping_back(1).enumerate() {
diff --git a/src/main.rs b/src/main.rs
index 58c9ef7..c6289cc 100644
--- a/src/main.rs
+++ b/src/main.rs
@@ -203,9 +203,9 @@ fn extract_ubam() {
         hist: true,
         checksum: false,
         arrow: Some("test.feather".to_string()),
-        karyotype: true,
-        phased: true,
-        spliced: true,
+        karyotype: false,
+        phased: false,
+        spliced: false,
         ubam: true,
     };
     let metrics = extract_from_bam::extract(&args);