From 1fb94a3ca2169d343e071d558b70dbfcb47bd192 Mon Sep 17 00:00:00 2001
From: pdimens <pdimens@live.com>
Date: Mon, 13 May 2024 16:01:39 -0400
Subject: [PATCH] working on a better coverage calc

---
 src/harpy/scripts/depthWindows.py | 34 +++++++++++++++++++++++++++++++
 1 file changed, 34 insertions(+)
 create mode 100755 src/harpy/scripts/depthWindows.py

diff --git a/src/harpy/scripts/depthWindows.py b/src/harpy/scripts/depthWindows.py
new file mode 100755
index 000000000..0152c8166
--- /dev/null
+++ b/src/harpy/scripts/depthWindows.py
@@ -0,0 +1,34 @@
+#! /usr/bin/env python
+
+import sys
+import argparse
+
+parser = argparse.ArgumentParser(prog = 'depthWindows.py', description = 'Reads the output of samtools depth -a from stdin and calculates a windowed mean.')
+parser.add_argument('windowsize', type= int, help = "The window size to use to calcualte mean depth over (non-overlapping)")
+
+args = parser.parse_args()
+_sum = 0
+start = 1
+end = start + args.windowsize - 1
+lastcontig = None
+for line in sys.stdin:
+    # Remove the newline character at the end of the line
+    line = line.rstrip().split()
+    contig = line[0]
+    # the contig has changed, make the end position the last position, print output
+    if lastcontig and contig != lastcontig:
+        winsize = (position + 1) - start
+        print(f"{lastcontig}\t{position}\t{_sum / winsize}", file = sys.stdout)
+        # reset the window start/end and sum
+        start = 1
+        end = start + args.windowsize - 1
+        _sum = 0
+
+    position = int(line[1])
+    _sum += int(line[2])
+
+    if position == end:
+        print(f"{contig}\t{end}\t{_sum / args.windowsize}", file = sys.stdout)
+        start = end + 1
+        end = start + args.windowsize - 1
+    lastcontig = contig
\ No newline at end of file