rbbt-util 5.13.8 → 5.13.9

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: e7281b74a41b34bf9cc71d351b6ca175a40f8aeb
-  data.tar.gz: 6f31f788281e74e00b594e9ff0303ccf0cddb491
+  metadata.gz: 7d7f9b2c2641577f41749304b67d7a7df205ba84
+  data.tar.gz: 0698c8200f48d6f115fef5730865b7421c49e1b5
 SHA512:
-  metadata.gz: 7faad521cd49f6079fc8db95d49ad2d60dfda58f0ba99ee2f7249dcc519b3967df2c1a8cbe09f6de6af8d31c219233b3181ca7b70b859383885f335d7e13d80d
-  data.tar.gz: c218eb43aaaee50d7fb41c08d55b99b5615243e7c9f38bf6d1c499ee1766486c0f4e1cd21edd01389525747714f50669c35d6a120aab7e01c1011dd1410fd216
+  metadata.gz: 30ccf37acc2676ce5a28c8237d96447d16b00b97d662dbd73cde7b3a9b9558dc6c9f710463eeb6b291a2a942c9bf6de3b9e662c7ae651e29d1c8d094cfc7c9f8
+  data.tar.gz: a3d39bafa0638bd682f18e3e1993538bb2ab812ea4a96da5e8d77f1c84bd52b9811e1617acc8fd90b1758bf83ee590c58f69e204664d6b002436283fc4ef045f

data/lib/rbbt/tsv/util.rb

@@ -116,23 +116,28 @@ module TSV
   end
 
   def self.identify_field(key_field, fields, field)
-    case
-    when Integer === field
-      field
-    when (field.nil? or field == :key)
+    case field
+    when nil
       :key
-    when (String === field and not fields.nil?)
+    when Symbol
+      field == :key ? field : identify_field(key_field, fields, field.to_s)
+    when Integer
+      field
+    when (fields.nil? and String)
+      raise "No field information available and specified field not numeric: #{ field }" unless field =~ /^\d+$/
+      identify_field(key_field, fields, field.to_i)
+    when String
+      return key if key_field == field
       pos = fields.index field
       pos ||= :key if key_field == field
-      Log.medium "Field #{ field } was not found. Options: #{fields * ", "}" if pos.nil?
-      pos
-    when key_field == field
-      :key
+      return identify_field(key_field, fields, field.to_i) if field =~ /^\d+$/
+      raise "Field #{ field } was not found. Options: #{fields * ", "}" if pos.nil?
     else
-      raise "No fields specified in TSV.identify_field" if fields.nil?
-      Log.medium "Field #{ field } was not found. Options: (#{key_field}), #{fields * ", "}"
+      raise "Field #{ field } was not found. Options: (#{key_field || "NO_KEY_FIELD"}), #{(fields || ["NO_FIELDS"]) * ", "}"
     end
   end
+
+
   
   def self.header_lines(key_field, fields, entry_hash = {})
     sep = (Hash === entry_hash and entry_hash[:sep]) ? entry_hash[:sep] : "\t"

data/lib/rbbt/util/misc/objects.rb

@@ -52,7 +52,12 @@ module Misc
 
   def self.zip_fields(array)
     return [] if array.empty? or (first = array.first).nil?
-    first.zip(*array[1..-1])
+    max = array[1..-1].collect{|l| l.length}.max
+    rest = array[1..-1].collect{|v|
+      v.length == 1 & max > 1 ? v * max : v
+    }
+    first = first * max if first.length == 1 and max > 1
+    first.zip(*rest)
   end
 
   def self.field_position(fields, field, quiet = false)

data/share/rbbt_commands/tsv/unzip

@@ -15,8 +15,8 @@ Display summary information. Works with Tokyocabinet HDB and BDB as well.
 -tch--tokyocabinet File is a TC HDB
 -tcb--tokyocabinet_bd File is a TC BDB
 -hh--header_hash* Change the character used to mark the header line (defaults to #)
--m--merge* Merge lines
 -f--field* Field to unzip
+-m--merge Merge lines
 -h--help Help
 EOF
 

data/test/rbbt/util/test_misc.rb

@@ -6,7 +6,7 @@ require 'rbbt/entity'
 
 class TestMisc < Test::Unit::TestCase
 
-  def test_format_paragraph
+  def _test_format_paragraph
     p = <<-EOF
 Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod tempor
 incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis
@@ -26,7 +26,7 @@ in culpa qui officia deserunt mollit anim id est laborum.
     assert Misc.format_paragraph(p, 70, 10, 5) =~ /\n\s*\* two/sm
   end
 
-  def test_format_dl
+  def _test_format_dl
     p1 = <<-EOF
 Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod tempor
 incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis
@@ -56,13 +56,13 @@ eum fugiat quo voluptas nulla pariatur?"
     assert Misc.format_definition_list({:paragraph_first => p1, :paragraph_second => p2}) =~ /       /
   end
 
-  def test_parse_cmd_params
+  def _test_parse_cmd_params
     assert_equal ["workflow", "task", "Translation", "translate", "-f", "Associated Gene Name", "-l", "-"],
       Misc.parse_cmd_params("workflow task Translation translate -f 'Associated Gene Name' -l -")
   end
 
 
-  def test_fixutf8
+  def _test_fixutf8
     string = "abc\xffdef"
     string = string.force_encoding("UTF-8") if string.respond_to? :force_encoding
     assert(! string.valid_encoding?) if string.respond_to? :valid_encoding?
@@ -71,37 +71,37 @@ eum fugiat quo voluptas nulla pariatur?"
     assert( Misc.fixutf8(string).valid_encoding) if string.respond_to? :valid_encoding
   end
 
-  def test_colors_for
+  def _test_colors_for
     colors, used = Misc.colors_for([1,2,2,1,2,1,2,2,3,3,2,3,2])
     assert_equal Misc::COLOR_LIST[1], used[2]
   end
 
-  def test_total_length
+  def _test_total_length
     ranges = [(0..100), (50..150), (120..160)]
     ranges = [(0..100), (50..150), (120..160), (51..70)]
     assert_equal 161, Misc.total_length(ranges)
   end
 
-  def test_id_filename?
+  def _test_id_filename?
     TmpFile.with_file("") do |file|
       assert Misc.is_filename?(file)
       assert ! Misc.is_filename?("TEST STRING")
     end
   end
 
-  def test_merge_sorted_arrays
+  def _test_merge_sorted_arrays
     assert_equal [1,2,3,4], Misc.merge_sorted_arrays([1,3], [2,4])
   end
 
-  def test_intersect_sorted_arrays
+  def _test_intersect_sorted_arrays
     assert_equal [2,4], Misc.intersect_sorted_arrays([1,2,3,4], [2,4])
   end
 
-  def test_sorted_array_matches
+  def _test_sorted_array_matches
     assert_equal [1,3], Misc.sorted_array_hits(%w(a b c d e), %w(b d))
   end
 
-  def test_binary_include?
+  def _test_binary_include?
     a = %w(a b c d e).sort
     assert Misc.binary_include?(a, "a")
     assert(!Misc.binary_include?(a, "z"))
@@ -110,12 +110,12 @@ eum fugiat quo voluptas nulla pariatur?"
     assert(Misc.binary_include?(a, "d"))
   end
 
-  def test_process_to_hash
+  def _test_process_to_hash
     list = [1,2,3,4]
     assert_equal 4, Misc.process_to_hash(list){|l| l.collect{|e| e * 2}}[2]
   end
 
-  def test_pipe_fork
+  def _test_pipe_fork
     sout, sin = Misc.pipe
     pid = Process.fork do
       Misc.purge_pipes(sin)
@@ -127,7 +127,7 @@ eum fugiat quo voluptas nulla pariatur?"
     Process.kill :INT, pid
   end
 
-  def test_open_pipe
+  def _test_open_pipe
     t = 5
     stream = Misc.open_pipe do |sin|
       t.times do |i|
@@ -148,7 +148,7 @@ eum fugiat quo voluptas nulla pariatur?"
     assert_equal (0..t-1).to_a.collect{|i| "LINE #{ i }"}, lines
   end
 
-  def test_open_pipe_fork
+  def _test_open_pipe_fork
     t = 5
     stream = Misc.open_pipe(true) do |sin|
       t.times do |i|
@@ -169,7 +169,7 @@ eum fugiat quo voluptas nulla pariatur?"
     assert_equal (0..t-1).to_a.collect{|i| "LINE #{ i }"}, lines
   end
 
-  def test_open_pipe_fork_cascade
+  def _test_open_pipe_fork_cascade
     t = 500
     sleep_time = 2.0 / t
     time = Time.now
@@ -205,7 +205,7 @@ eum fugiat quo voluptas nulla pariatur?"
     assert_equal (0..t-1).to_a.collect{|i| "LINE #{ i }".reverse.downcase}, lines
   end
 
-  def test_tee_stream
+  def _test_tee_stream
     t = 500
     sleep_time = 2.0 / t
     time = Time.now
@@ -255,7 +255,7 @@ eum fugiat quo voluptas nulla pariatur?"
   end
 
 
-  def test_string2hash
+  def _test_string2hash
     assert(Misc.string2hash("--user-agent=firefox").include? "--user-agent")
     assert_equal(true, Misc.string2hash(":true")[:true])
     assert_equal(true, Misc.string2hash("true")["true"])
@@ -266,16 +266,16 @@ eum fugiat quo voluptas nulla pariatur?"
     assert_equal(:j, Misc.string2hash("a=b#c=d#:h=:j")[:h])
   end
   
-  def test_named_array
+  def _test_named_array
     a = NamedArray.setup([1,2,3,4], %w(a b c d))
     assert_equal(1, a['a'])
   end
 
-  def test_path_relative_to
+  def _test_path_relative_to
     assert_equal "test/foo", Misc.path_relative_to('/test', '/test/test/foo')
   end
 
-  def test_hash2string
+  def _test_hash2string
     hash = {}
     assert_equal hash, Misc.string2hash(Misc.hash2string(hash))
 
@@ -293,14 +293,14 @@ eum fugiat quo voluptas nulla pariatur?"
  
   end
 
-  def test_merge
+  def _test_merge
     a = [[1],[2]]
     a = NamedArray.setup a, %w(1 2)
     a.merge [3,4]
     assert_equal [1,3], a[0]
   end
 
-  def test_indiferent_hash
+  def _test_indiferent_hash
     a = {:a => 1, "b" => 2}
     a.extend IndiferentHash
 
@@ -310,7 +310,7 @@ eum fugiat quo voluptas nulla pariatur?"
     assert_equal 2, a[:b]
   end
 
-  def test_lockfile
+  def _test_lockfile
 
     TmpFile.with_file do |tmpfile|
       pids = []
@@ -334,7 +334,7 @@ eum fugiat quo voluptas nulla pariatur?"
     end
   end
 
-  def test_positions2hash
+  def _test_positions2hash
     inputs = Misc.positional2hash([:one, :two, :three], 1, :two => 2, :four => 4)
     assert_equal 1, inputs[:one]
     assert_equal 2, inputs[:two]
@@ -342,18 +342,24 @@ eum fugiat quo voluptas nulla pariatur?"
     assert_equal nil, inputs[:four]
   end
 
-  def test_mean
+  def _test_mean
     assert_equal 2, Misc.mean([1,2,3])
     assert_equal 3, Misc.mean([1,2,3,4,5])
   end
 
-  def test_zip_fields
+  def _test_zip_fields
     current = [[:a,1], [:b,2]]
     assert_equal [[:a, :b],[1,2]], Misc.zip_fields(current)
     assert_equal current, Misc.zip_fields(Misc.zip_fields(current))
   end
 
-  def test_add_zipped
+  def test_zip_fields_comp
+    current = [[:a,1], [:b,2], [:c]]
+    assert_equal [[:a, :b, :c],[1,2,nil]], Misc.zip_fields(current)
+    assert_equal current, Misc.zip_fields(Misc.zip_fields(current)).collect{|v| v.compact }
+  end
+
+  def _test_add_zipped
     current = [[:a,1], [:b,2]]
     new = %w(A B)
     Misc.append_zipped current, new
@@ -364,20 +370,20 @@ eum fugiat quo voluptas nulla pariatur?"
     assert_equal Math.sqrt(2), Misc.sd([1,3])
   end
 
-  def test_divide
+  def _test_divide
     assert_equal 2, Misc.divide(%w(1 2 3 4 5 6 7 8 9),2).length
   end
 
-  def test_ordered_divide
+  def _test_ordered_divide
     assert_equal 5, Misc.ordered_divide(%w(1 2 3 4 5 6 7 8 9),2).length
   end
 
-  def test_collapse_ranges
+  def _test_collapse_ranges
     ranges = [(0..100), (50..150), (51..61),(200..250), (300..324),(320..350)]
     assert_equal [(0..150),(200..250), (300..350)], Misc.collapse_ranges(ranges)
   end
 
-  def test_humanize
+  def _test_humanize
     str1 = "test_string"
     str2 = "TEST_string"
     str3 = "test"
@@ -389,22 +395,22 @@ eum fugiat quo voluptas nulla pariatur?"
     assert_equal "mutation_enrichment", Misc.snake_case("MutationEnrichment")
   end
 
-  def test_snake_case
+  def _test_snake_case
     str1 = "ACRONIMTest"
     str2 = "ACRONIM_test"
     assert_equal "ACRONIM_test", Misc.snake_case(str1)
     assert_equal "ACRONIM_test", Misc.snake_case(str2)
   end
 
-  def test_correct_vcf_mutations
+  def _test_correct_vcf_mutations
     assert_equal [737407, ["-----", "-----G", "-----GTTAAT"]], Misc.correct_vcf_mutation(737406, "GTTAAT", "G,GG,GGTTAAT")
   end
 
-  def test_fingerprint
+  def _test_fingerprint
     assert_equal '{a=>1}', Misc.fingerprint({:a => 1})
   end
 
-  def test_tarize
+  def _test_tarize
     path = File.expand_path('/home/mvazquezg/git/rbbt-util/lib')
     stream = Misc.tarize(path)
     TmpFile.with_file do |res|
@@ -414,7 +420,7 @@ eum fugiat quo voluptas nulla pariatur?"
     end
   end
 
-  def test_camel_case
+  def _test_camel_case
     assert_equal "DbSNP", Misc.camel_case("db_SNP")
     assert_equal "D3Js", Misc.camel_case("D3Js")
     assert_equal "Structure", Misc.camel_case("Structure")

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: rbbt-util
 version: !ruby/object:Gem::Version
-  version: 5.13.8
+  version: 5.13.9
 platform: ruby
 authors:
 - Miguel Vazquez
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2014-05-09 00:00:00.000000000 Z
+date: 2014-05-12 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rake