contig sizes fix

metagenomics/bins_summary.py

@@ -6,7 +6,7 @@ import re
 import pandas as pd
 
 def parse_options():
-  usage = "\nbins_summary.py -b bins_folder -i idxstats_file -f flagstat_file -b busco_summaries_folder -c checkm_report -o output_summary.tsv"
+  usage = "\nbins_summary.py -b bins_folder -b busco_summaries_folder -c checkm_report -o output_summary.tsv"
   description = "Creates a summary of bins from different tools"
   parser = argparse.ArgumentParser(usage=usage, description=description)
 
@@ -29,20 +29,26 @@ def bin_stats(bin_dir,output_file):
     bin_file_path = bin_dir+"/"+bin
 
     # TOTAL SIZE, CONTIG SIZES, GC%
+    contig = ""
     total_size = 0
     contig_sizes = []
     contig_len = 0
     GC = 0
-    for line in open(bin_file_path,"r").readlines():
+    for i,line in enumerate(open(bin_file_path,"r").readlines()):
       line = line.strip()
-      if not line.startswith(">"):
+      if line.startswith(">"):
+        contig = bin+"\t"+line.strip(">")
+        bin_contigs_file.write(bin+"\t"+contig+"\n")
+        if i != 0:
+          contig_sizes.append(contig_len)
+          contig_len = 0
+      else:
         GC += line.count("G")+line.count("C")
         total_size += len(line.strip())
         contig_len += len(line.strip())
-      else:
-        bin_contigs_file.write(bin+"\t"+line.strip(">")+"\n")
-        contig_sizes.append(contig_len)
-        contig_len = 0
+
+    bin_contigs_file.write(bin+"\t"+contig+"\n")
+    contig_sizes.append(contig_len)
     
     GC = round(GC/total_size*100,1)
     contigs = len(contig_sizes)