gfa 0.5.0 → 0.6.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 9eaaaf8bcac372e7d6b63d96aee94493faa51d78f5cc2b7edb525c3fbd7efaa2
-  data.tar.gz: 950e9ffa9cd07cdea8ef5cc80b6c162bf009ff3bef5ab757dcda51c443925043
+  metadata.gz: cf18e4a80f49fc9c0ae4574c5c7bbc3c97fd41eaf5c7aff9fdd7991b7d562ccf
+  data.tar.gz: ac6999ffbbbeba45f9b81c6081ac16339ce4bb55502a77b116c77f91870af991
 SHA512:
-  metadata.gz: 92e53b9d6c09b1814f3feda86b67dcea6192bd935e3c04dcca7b6ee9c0e7eed919c711941985313c2918f02772f3598883f037638e6e1b7cc9cf8897fec80caf
-  data.tar.gz: 4c14ce2bd91582b7b81bc6123443bd5b0cffe8c83550fb85dc48c8da030d35dafaf393506d537c5df7550b25b87ab10003a938c407bd6210983e5f8841552317
+  metadata.gz: c8ae1b273166574d8520aa1f747ac0dbd08e8f31846754ca6c3fa7f0432d5f31583aa915fcebb04a220e90fe14aeeb4f06dec2a5c23527571c9bb337de0ef9a5
+  data.tar.gz: f5fce0959cace1fe961d92fab2946bf8b67571ee6ceef3cd6e7047d8c46f0855b28e6b703ffbfe327ef8d09e1174b94ee98423ae36bb858bab76615849e0bd94

data/bin/gfa-add-gaf

@@ -18,11 +18,13 @@ unless degree
     <input-gaf>  Input GAF file to read
     <output>     Output GFA file to write
     <degree>     Maximum degree of separation between the segment set in the GAF
-                 and any other included segments. If 0, only segments are
-                 included. If 1, only the target segments, records linking to
-                 them, and segments linked by those records. Any integer > 1
-                 includes additional expansion rounds for those linked segments.
-                 Use -1 to include the complete original GAF without subsetting.
+                 and any other segments included in the output GFA.
+                 - -1: include the complete original GAF without subsetting.
+                 - 0: only segments in the GAF are included.
+                 - 1: only the target segments in the GAF, records linking to
+                   them, and segments linked by those records are included.
+                 - Any integer > 1: include additional expansion rounds for
+                   those linked segments.
     <pref>       A prefix to name all recorded paths
                  By default: Based on the GAF file name
     <threads>    If passed, parallelize process with these many threads

data/bin/gfa-mean-depth

@@ -0,0 +1,37 @@
+#!/usr/bin/env ruby
+
+# @package MiGA
+# @license Artistic-2.0
+
+$LOAD_PATH.push File.expand_path('../../lib', __FILE__)
+$LOAD_PATH.push File.expand_path('../../lib', File.realpath(__FILE__))
+
+require 'gfa'
+
+input, threads = ARGV
+
+unless input
+  $stderr.puts <<~HELP
+    Calculate the average sequencing depth of all segments in the GFA
+    weighted by the segment lengths
+
+    gfa-mean-depth <input> [<threads>]
+
+    <input>      Input GFA file to read
+    <threads>    If passed, parallelize process with these many threads
+  HELP
+  exit(1)
+end
+
+$stderr.puts "Loading GFA: #{input}"
+gfa = GFA.load_parallel(input, (threads || 1).to_i)
+
+$stderr.puts 'Calculating average depth'
+n = gfa.total_length
+avg =
+  gfa.segments.set.map do |segment|
+    raise "Some segments are missing depth data" unless segment.DP
+    segment.DP.value * segment.length / n
+  end.inject(:+)
+puts avg
+

data/bin/gfa-merge

@@ -0,0 +1,34 @@
+#!/usr/bin/env ruby
+
+# @package MiGA
+# @license Artistic-2.0
+
+$LOAD_PATH.push File.expand_path('../../lib', __FILE__)
+$LOAD_PATH.push File.expand_path('../../lib', File.realpath(__FILE__))
+
+require 'gfa'
+
+output = ARGV.shift
+input  = ARGV
+
+if input.empty?
+  $stderr.puts <<~HELP
+    Combine several GFAs into a single GFA. Requires uniqueness of element names
+
+    gfa-merge <output> <input...>
+
+    <output>     Output GFA file to be created
+    <input...>   List of input GFA files to read
+  HELP
+  exit(1)
+end
+
+gfa = GFA.new
+input.each do |i|
+  $stderr.puts "Merging GFA: #{i}"
+  gfa.merge! GFA.load(i)
+end
+
+$stderr.puts "Saving GFA: #{output}"
+gfa.save(output)
+

data/bin/gfa-paths-to-fasta

@@ -0,0 +1,29 @@
+#!/usr/bin/env ruby
+
+# @package MiGA
+# @license Artistic-2.0
+
+$LOAD_PATH.push File.expand_path('../../lib', __FILE__)
+$LOAD_PATH.push File.expand_path('../../lib', File.realpath(__FILE__))
+
+require 'gfa'
+
+input, output, threads = ARGV
+
+unless output
+  $stderr.puts <<~HELP
+    Extract the sequences of the paths from a GFA to FastA file
+
+    gfa-merge <input> <output> [<threads>]
+
+    <input>      Input GFA file to read
+    <output>     Output FastA file to be created
+    <threads>    If passed, parallelize process with these many threads
+  HELP
+  exit(1)
+end
+
+$stderr.puts "Loading GFA: #{input}"
+gfa = GFA.load_parallel(input, (threads || 1).to_i)
+
+

data/lib/gfa/common.rb

@@ -67,4 +67,17 @@ class GFA
   def total_length
     segments.total_length
   end
+
+  ##
+  # Adds the entrie of +gfa+ to itself
+  def merge!(gfa)
+    records.each { |k, v| v.merge!(gfa.records[k]) }
+    self
+  end
+
+  ##
+  # Creates a new GFA based on itself and appends all entries in +gfa+
+  def merge(gfa)
+    GFA.new(opts).merge!(self).merge!(gfa)
+  end
 end

data/lib/gfa/graph.rb

@@ -146,7 +146,7 @@ class GFA
 
   def internally_linking_records(segments, edges)
     $stderr.puts '- Gathering internally linking records'
-    segments = Hash[segments.set.map { |i| [i.name.value, true]}]
+    segments = Hash[segments.map { |i| [i.name.value, true]}]
     edges.select { |record| record.segment_names_a.all? { |s| segments[s] } }
   end
 

data/lib/gfa/version.rb

@@ -1,5 +1,5 @@
 class GFA
-  VERSION       = '0.5.0'
+  VERSION       = '0.6.1'
   VERSION_ARRAY = VERSION.split(/\./).map { |x| x.to_i } # :nodoc:
   VERSION_MAJOR = VERSION_ARRAY[0] # :nodoc:
   VERSION_MINOR = VERSION_ARRAY[1] # :nodoc:

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: gfa
 version: !ruby/object:Gem::Version
-  version: 0.5.0
+  version: 0.6.1
 platform: ruby
 authors:
 - Luis M. Rodriguez-R
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2023-03-26 00:00:00.000000000 Z
+date: 2023-10-05 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rgl
@@ -65,6 +65,9 @@ files:
 - Rakefile
 - bin/gfa-add-gaf
 - bin/gfa-greedy-modules
+- bin/gfa-mean-depth
+- bin/gfa-merge
+- bin/gfa-paths-to-fasta
 - bin/gfa-subgraph
 - lib/gfa.rb
 - lib/gfa/common.rb