miga-base 1.2.0.1 → 1.2.2.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 27068ded9d2ff63c3575a5aafd3fa68374f4bcf8aef924354e0c9cbde7dc1f6d
-  data.tar.gz: ef438e828180d304a9e3449d49551ace933a5e5a46128970b3c8bc385c836d58
+  metadata.gz: 4b1fb93676b3b4c48d2204f48a106575274ee50ab294bbf34a3790e7c9adb94b
+  data.tar.gz: 629d74fd9c861b9afc99c9764b1990e8ea8c9b89e5d10780201828fd68ab3c7c
 SHA512:
-  metadata.gz: 894cbeca2aa4f5e29c6bd6e1c23c9c358b9f3b497960998116169fa62d7122da78417e357789dc9b9b7bfdfad824f835b3daa035d1320cb948e5cdc4b1d74bb3
-  data.tar.gz: 8bea4585b36855195f3a2d76005a6ff664133c1df639a677402907c88fd8f6b0a25fb34cc4bd0e9e38bc7666dbe3a382644137dcf5cddd4005c93256d2c75c05
+  metadata.gz: 46d9ad97e0033b0c99e06476594148baaf2bc630f59c4775d763a2ab322a2ad449229fa9f342509fdfe6ca9a0949405f50b17c719fc817eb426f32a9dd287f54
+  data.tar.gz: 4e6ae0abee96441ac65eb2e8e886f4c2f3d00cc9716205936563022d7858f0d966d28d110c3b18ecdaa8aa328ef58b3b53410e9c32560aa0368b88e7ea6604b8

data/lib/miga/cli/action/add.rb

@@ -57,6 +57,7 @@ class MiGA::Cli::Action::Add < MiGA::Cli::Action
 
   def perform
     p = cli.load_project
+    cli.ensure_par(type: '-t')
     files, file_type = get_files_and_type
 
     paired = cli[:input_type].to_s.include?('_paired')
@@ -113,7 +114,7 @@ class MiGA::Cli::Action::Add < MiGA::Cli::Action
       cli.ensure_type(Dataset)
       files = [nil]
     else
-      raise 'Please specify input type (-i).' if cli[:input_type].nil?
+      cli.ensure_par({ input_type: '-i' }, 'Please specify input type (-i)')
 
       file_type = self.class.INPUT_TYPES[cli[:input_type]]
       raise "Unrecognized input type: #{cli[:input_type]}." if file_type.nil?

data/lib/miga/cli/action/init/files_helper.rb

@@ -112,7 +112,7 @@ module MiGA::Cli::Action::Init::FilesHelper
           && tar zxf #{arch.shellescape} \
           && rm #{arch.shellescape}
       CMD
-      run_cmd(cmd, source: nil)
+      run_cmd(cli, cmd, source: nil)
       cli.puts
     end
   end

data/lib/miga/cli/action/ls.rb

@@ -1,5 +1,4 @@
-# @package MiGA
-# @license Artistic-2.0
+# frozen_string_literal: true
 
 require 'miga/cli/action'
 
@@ -42,33 +41,53 @@ class MiGA::Cli::Action::Ls < MiGA::Cli::Action
         '-s', '--silent',
         'No output and exit with non-zero status if the dataset list is empty'
       ) { |v| cli[:silent] = v }
+      opt.on(
+        '--exec CMD',
+        'Command to execute per dataset, with the following token variables:',
+        '~ {{dataset}}: Name of the dataset',
+        '~ {{project}}: Path to the project'
+      ) { |v| cli[:exec] = v }
     end
   end
 
   def perform
     ds = cli.load_and_filter_datasets(cli[:silent])
+    p  = cli.load_project
     exit(ds.empty? ? 1 : 0) if cli[:silent]
-    if !cli[:datum].nil?
+
+    head = nil
+    fun  = nil
+    if cli[:datum]
       cli[:tabular] = true
-      format_table(ds, [nil, nil]) { |d| [d.name, d.metadata[cli[:datum]]] }
-    elsif !cli[:fields].nil?
-      format_table(ds, [:name] + cli[:fields]) do |d|
-        [d.name] + cli[:fields].map { |f| d.metadata[f] }
-      end
+      head = [nil, nil]
+      fun  = proc { |d| [d.name, d.metadata[cli[:datum]]] }
+    elsif cli[:fields]
+      head = [:name] + cli[:fields]
+      fun  = proc { |d| [d.name] + cli[:fields].map { |f| d.metadata[f] } }
     elsif cli[:info]
-      format_table(ds, Dataset.INFO_FIELDS) { |d| d.info }
+      head = Dataset.INFO_FIELDS
+      fun  = proc(&:info)
     elsif cli[:processing]
-      comp = %w[- done queued]
-      format_table(ds, [:name] + MiGA::Dataset.PREPROCESSING_TASKS) do |d|
-        [d.name] + d.profile_advance.map { |i| comp[i] }
+      head = [:name] + MiGA::Dataset.PREPROCESSING_TASKS
+      fun  = proc do |d|
+        [d.name] + d.profile_advance.map { |i| %w[- done queued][i] }
       end
     elsif cli[:taskstatus]
-      format_table(ds, [:name] + MiGA::Dataset.PREPROCESSING_TASKS) do |d|
-        [d.name] + d.results_status.values
-      end
+      head = [:name] + MiGA::Dataset.PREPROCESSING_TASKS
+      fun  = proc { |d| [d.name] + d.results_status.values }
     else
       cli[:tabular] = true
-      format_table(ds, [nil]) { |d| [d.name] }
+      head = [nil]
+      fun  = proc { |d| [d.name] }
+    end
+
+    format_table(ds, head) do |d|
+      if cli[:exec]
+        MiGA::MiGA.run_cmd(
+          cli[:exec].miga_variables(dataset: d.name, project: p.path)
+        )
+      end
+      fun[d]
     end
   end
 

data/lib/miga/cli/base.rb

@@ -1,5 +1,4 @@
-# @package MiGA
-# @license Artistic-2.0
+# frozen_string_literal: true
 
 module MiGA::Cli::Base
   @@TASK_DESC = {
@@ -20,10 +19,10 @@ module MiGA::Cli::Base
     add: 'Create a dataset in a MiGA project',
     get: 'Download a dataset from public databases into a MiGA project',
     ncbi_get: 'Download all genomes in a taxon from NCBI into a MiGA project',
-    rm: 'Remove a dataset from an MiGA project',
+    rm: 'Remove a dataset from a MiGA project',
     find: 'Find unregistered datasets based on result files',
     ln: 'Link datasets (including results) from one project to another',
-    ls: 'List all registered datasets in an MiGA project',
+    ls: 'List all registered datasets in a MiGA project',
     archive: 'Generate a tar-ball with all files from select datasets',
     # Results
     add_result: 'Register a result',

data/lib/miga/common/format.rb

@@ -1,3 +1,5 @@
+# frozen_string_literal: true
+
 require 'tempfile'
 
 ##
@@ -37,13 +39,13 @@ module MiGA::Common::Format
         tmp_path = tmp_fh.path
         fh = File.open(file, 'r')
       end
-      buffer = ''
+      buffer = ''.dup
       fh.each_line do |ln|
         ln.chomp!
         if ln =~ /^>\s*(\S+)(.*)/
           id, df = $1, $2
           tmp_fh.print buffer.wrap_width(80)
-          buffer = ''
+          buffer = ''.dup
           tmp_fh.puts ">#{id.gsub(/[^A-Za-z0-9_\|\.]/, '_')}#{df}"
         else
           buffer << ln.gsub(/[^A-Za-z\.\-]/, '')
@@ -165,7 +167,7 @@ class String
   ##
   # Replace {{variables}} using the +vars+ hash
   def miga_variables(vars)
-    o = "#{self}"
+    o = self.dup
     vars.each { |k, v| o.gsub!("{{#{k}}}", v.to_s) }
     o
   end

data/lib/miga/common.rb

@@ -1,5 +1,4 @@
-# @package MiGA
-# @license Artistic-2.0
+# frozen_string_literal: true
 
 require 'zlib'
 require 'stringio'
@@ -44,9 +43,8 @@ class MiGA::MiGA
   # +ext+ values (Array of String).
   def result_files_exist?(base, ext)
     ext = [ext] unless ext.is_a? Array
-    ext.all? do |f|
-      File.exist?(base + f) or File.exist?("#{base}#{f}.gz")
-    end
+    MiGA::MiGA.DEBUG("Assserting files for result: #{ext}")
+    ext.all? { |f| File.exist?(base + f) or File.exist?("#{base}#{f}.gz") }
   end
 
   ##

data/lib/miga/daemon.rb

@@ -66,6 +66,12 @@ class MiGA::Daemon < MiGA::MiGA
     "MiGA:#{project.name}"
   end
 
+  ##
+  # Alias to +project.path+ for compatibility with lairs
+  def path
+    project.path
+  end
+
   ##
   # Run only in the first loop
   def daemon_first_loop

data/lib/miga/version.rb

@@ -12,7 +12,7 @@ module MiGA
   # - String indicating release status:
   #   - rc* release candidate, not released as gem
   #   - [0-9]+ stable release, released as gem
-  VERSION = [1.2, 0, 1].freeze
+  VERSION = [1.2, 2, 2].freeze
 
   ##
   # Nickname for the current major.minor version.
@@ -20,7 +20,7 @@ module MiGA
 
   ##
   # Date of the current gem relese.
-  VERSION_DATE = Date.new(2021, 12, 11)
+  VERSION_DATE = Date.new(2022, 2, 2)
 
   ##
   # References of MiGA

data/test/format_test.rb

@@ -58,6 +58,9 @@ class FormatTest < Test::Unit::TestCase
     assert_equal(50.0, o[:gc])
     assert_equal(5, o[:n50])
     assert_equal(4.0, o[:med])
+    o = MiGA::MiGA.seqs_length(f, :fasta, skew: true)
+    assert_equal(-50.0, o[:at_skew])
+    assert_equal(-25.0, o[:gc_skew])
   end
 
   def test_seqs_length_fastq
@@ -77,4 +80,22 @@ class FormatTest < Test::Unit::TestCase
     assert_equal('123  45', tab[2])
     assert_equal('678  90', tab[3])
   end
+
+  def test_miga_name
+    assert_not('a-bad-name'.miga_name?)
+    assert('a_good_one'.miga_name?)
+
+    assert('After_it_s_fixed_', 'After it\'s fixed!'.miga_name)
+
+    assert('A sp.', 'A_sp_'.unmiga_name)
+    assert('B str. C', 'B_sp_C'.unmiga_name)
+    assert('The X content', 'The_x_content'.unmiga_name)
+  end
+
+  def test_miga_variables
+    assert_equal(
+      '1 a box!',
+      '{{n}} {{my}} {{secret}}!'.miga_variables(my: 'a', secret: :box, n: 1)
+    )
+  end
 end

data/utils/distance/commands.rb

@@ -111,12 +111,17 @@ module MiGA::DistanceRunner::Commands
     return unless File.size?(f1)
 
     # Run FastANI
+    empty = true
     File.open(f2 = tmp_file, 'w') do |fh|
       targets.each do |target|
         target_asm = target&.result(:assembly)&.file_path(:largecontigs)
-        fh.puts target_asm if target_asm
+        if target_asm
+          fh.puts target_asm
+          empty = false
+        end
       end
     end
+    return if empty
     run_cmd <<~CMD
               fastANI -q "#{f1}" --rl "#{f2}" -t #{opts[:thr]} \
               -o "#{f3 = tmp_file}"

data/utils/distance/pipeline.rb

@@ -33,6 +33,10 @@ module MiGA::DistanceRunner::Pipeline
 
   # Builds a tree with all visited medoids from any classification level
   def build_medoids_tree(metric)
+    # TODO Reduce impact and bring back the medoids tree
+    $stderr.puts 'Bypassing medoids tree'
+    return
+
     $stderr.puts "Building medoids tree (metric = #{metric})"
     db = query_db(metric)
     return unless File.size? db

data/utils/ref-tree.R

@@ -5,38 +5,43 @@
 #
 
 #= Load stuff
-argv <- commandArgs(trailingOnly=T)
+argv <- commandArgs(trailingOnly = TRUE)
 suppressPackageStartupMessages(library(ape))
-if(Sys.getenv('MIGA') == ''){
+if (Sys.getenv("MIGA") == "") {
   suppressPackageStartupMessages(library(enveomics.R))
-}else{
-  source(file.path(Sys.getenv('MIGA'),
-    'utils', 'enveomics', 'enveomics.R', 'R', 'df2dist.R'))
+} else {
+  source(
+    file.path(
+      Sys.getenv("MIGA"), "utils", "enveomics", "enveomics.R", "R", "df2dist.R"
+    )
+  )
 }
 inst <- c("phangorn", "phytools") %in% rownames(installed.packages())
-if(inst[1]){
+if (inst[1]) {
   suppressPackageStartupMessages(library(phangorn))
   reroot.fun <- midpoint
-}else if(inst[2]){
+} else if (inst[2]) {
   suppressPackageStartupMessages(library(phytools))
   reroot.fun <- midpoint.root
-}else{
+} else {
   reroot.fun <- function(x) return(x)
 }
 
 #= Main function
 ref_tree <- function(ani_file, out_base, q_dataset) {
-  a <- read.table(ani_file, sep="\t", header=TRUE, as.is=TRUE)
-  ani.d <- enve.df2dist(a[,1:3], default.d=0.9, max.sim=100)
+  a <- read.table(ani_file, sep = "\t", header = TRUE, as.is = TRUE)
+  ani.d <- enve.df2dist(a[, 1:3], default.d = 0.9, max.sim = 100)
   ani.ph <- reroot.fun(bionj(ani.d))
-  write.tree(ani.ph, paste(out_base, ".nwk", sep=""))
-  pdf(paste(out_base, ".nwk.pdf", sep=""), 7, 7)
-  plot(ani.ph, cex=1/3, type='fan',
-    tip.color=c('red', 'black')[ifelse(ani.ph$tip.label==q_dataset, 1, 2)])
+  write.tree(ani.ph, paste(out_base, ".nwk", sep = ""))
+  pdf(paste(out_base, ".nwk.pdf", sep = ""), 7, 7)
+  plot(
+    ani.ph, cex = 1/3, type = "fan",
+    tip.color = c("red", "black")[ifelse(ani.ph$tip.label == q_dataset, 1, 2)]
+  )
   add.scale.bar()
   dev.off()
 }
 
 #= Main
-ref_tree(ani_file=argv[1], out_base=argv[2], q_dataset=argv[3])
+ref_tree(ani_file = argv[1], out_base = argv[2], q_dataset = argv[3])
 

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: miga-base
 version: !ruby/object:Gem::Version
-  version: 1.2.0.1
+  version: 1.2.2.2
 platform: ruby
 authors:
 - Luis M. Rodriguez-R
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2021-12-11 00:00:00.000000000 Z
+date: 2022-02-02 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: daemons