rough parser for a ASG contamination file

filter_merged.cr

#!/bin/env crystal

require "option_parser"

phylum="Arthropoda|insect"
dir="20230226_qqAmaFero1.20230225.haplotigs.fa_asg_cobiont_check_run/collected_tables/"
OptionParser.parse do |parser|
	parser.banner = "Usage: filter_merged --phylum xyz --infile <in.merged>"
	parser.on("-p PHYLUM","--phylum=PHYLUM","Specifies the phylum(s) of the host separated by | [default=#{phylum}]"){|p|phylum=p}
	parser.on("-d directory","--directory=DIR","merged ASG directory[default=#{dir}]"){|d|dir=d}
	parser.on("-h", "--help", "Show this help") do
		puts parser
		exit
	end
  	parser.invalid_option do |flag|
    		STDERR.puts "ERROR: #{flag} is not a valid option."
		STDERR.puts parser
		exit(1)
  	end
end


Dir.glob("#{dir}/*vecscreen_contamination").each{|file|
  hits = [] of String 
  File.each_line(file) do |line|
	hits << line if /VecScreen/.match(line)
  end
  if hits.size > 0
      hits << ""
      puts "========== EUKARYOTE ADAPTOR SCREEN =========="
      hits.each{|h|puts h}
  end
}


Dir.glob("#{dir}/*contamination_check_merged_table.csv").each{|file|
  header=0
  categories = {} of String => Array(String)

  File.each_line(file) do |line|
	header += 1
	next if header == 1

	next if /#{phylum}/.match(line)

	columns = line.split(',')
	next if /no-hit/.match(columns[-1])

	categories[columns[-1]]||=[] of String
	categories[columns[-1]] << columns[0]
  end

  categories.each{|k,v|
	puts "##{k}"
	v.each{|scaffold|
		puts "REMOVE\t#{scaffold}"
	}
	puts ""
  }
}

filter_merged.rb

#!/bin/env ruby

require "optionparser"

phylum="Arthropoda|insect"
dir="20230226_qqAmaFero1.20230225.haplotigs.fa_asg_cobiont_check_run/collected_tables/"
OptionParser.new do |parser|
	parser.banner = "Usage: filter_merged --phylum xyz --infile <in.merged>"
	parser.on("-p PHYLUM","--phylum=PHYLUM","Specifies the phylum(s) of the host separated by | [default=#{phylum}]"){|p|phylum=p}
	parser.on("-d DIR","--directory=DIR","merged ASG infile [default=#{file}]"){|d|dir=d}
	parser.on("-h", "--help", "Show this help") do
		puts parser
		exit
	end
end.parse!



Dir.glob("#{dir}/*vecscreen_contamination").each{|file|
  hits = []
  File.open(file).each_line do |line|
      hits << line if /VecScreen/.match(line)
  end
  if hits.size > 0
      hits << ""
      puts "========== EUKARYOTE ADAPTOR SCREEN =========="
      hits.each{|h|puts h}
  end  
}

Dir.glob("#{dir}/*contamination_check_merged_table.csv").each{|file|
	header=0
	categories= {}
	
	File.open(file).each_line do |line|
		header += 1
		next if header == 1
	
		next if /#{phylum}/.match(line)
	
		columns = line.split(',')
		next if /no-hit/.match(columns[-1])
	
		categories[columns[-1]]||=[]
		categories[columns[-1]] << columns[0]
	end
	
	categories.each{|k,v|
		puts "##{k}"
		v.each{|scaffold|
			puts "REMOVE\t#{scaffold}"
		}
		puts ""
	}
}

it skips all lines that match the phyla and the ones with no-hit in as classifier, then pretty-prints the merged classifier.