miga-base 0.3.2.0 → 0.3.2.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
-SHA1:
-  metadata.gz: c395d2565cacafe425c91a277c03cdeaa3ac9ece
-  data.tar.gz: 1978e8a2df4d0646bc884ce9921bf8de2bc13a25
+SHA256:
+  metadata.gz: 82b444c1fd986a4bc95f5cca2e2485b2fb0dd29b82d7424ef7615e41d9f901d3
+  data.tar.gz: 8bc59e799f8af6e13dc1a3e0653743490c44189e056fbe8a9f86dc53b029a6bf
 SHA512:
-  metadata.gz: fff79ff971c5fc9e0e0684585b9b35abe5377cff0742c772fc1a7c58005b224ea18b36e7fe489b78c6103f6c4b9ffc90be75e7a385d5fce59a917e2d68c3765a
-  data.tar.gz: cda85047eabd8ba1c76bb5fe690247a3e71583e034f3c5dbef9178776d4c5da02bd3075cfc0e3154e1e97aa69da8ddea85a00ef99efa065e867950c10a860a71
+  metadata.gz: a3aabd7fa5fd037b4e427b131c311a66a103f68f5ff0aaf4907abf4663ed285e7010d89a10d2443e72f4d9c1f0a032ba112faf961972da34bc44b449a3531332
+  data.tar.gz: 4b1f302f219addc604f66526f72171f711c1e89252a7f152d68e839d4235ab723a43811f16664b81dee77b91c149fd38e06e54cca7d8412bf84594f26de36dcb

data/actions/doctor.rb

@@ -6,7 +6,8 @@
 require "sqlite3"
 
 o = {q:true, ld:false,
-  db: true, dist: true, files: true, ess: true, mts: true, tax: true}
+  db: true, dist: true, files: true,
+  ess: true, mts: true, start: true, tax: true}
 OptionParser.new do |opt|
   opt_banner(opt)
   opt_object(opt, o, [:project])
@@ -22,6 +23,8 @@ OptionParser.new do |opt|
     "Do not check unarchived essential genes."){ |v| o[:ess]=!v }
   opt.on("--ignore-mytaxa-scan",
     "Do not check unarchived MyTaxa scan."){ |v| o[:mts]=!v }
+  opt.on("--ignore-start",
+    "Do not check lingering legacy .start files."){ |v| o[:start]=!v }
   opt.on("--ignore-taxonomy",
     "Do not check taxonomy consistency."){ |v| o[:tax]=!v }
   opt_common(opt, o)
@@ -174,6 +177,18 @@ if o[:mts]
   end
 end
 
+if o[:start]
+  $stderr.puts "o Looking for legacy .start files lingering." unless o[:q]
+  p.each_dataset do |d|
+    d.each_result do |r_k, r|
+      if File.exist? r.path(:start)
+        $stderr.puts "    > Registering again #{d.name}:#{r_k}" if o[:ld]
+        r.save
+      end
+    end
+  end
+end
+
 if o[:tax]
   #$stderr.puts "o Checking for taxonomy/distances consistency" unless o[:q]
   # TODO: Find 95%ANI clusters with entries from different species

data/bin/miga

@@ -171,9 +171,13 @@ def add_metadata(o, obj)
   o[:metadata].split(",").each do |pair|
     (k,v) = pair.split("=")
     case v
-      when "true";  v = true
-      when "false"; v = false
-      when "nil";   v = nil
+      when 'true';  v = true
+      when 'false'; v = false
+      when 'nil';   v = nil
+    end
+    if k=='_step'
+      obj.metadata["_try_#{v}"] ||= 0
+      obj.metadata["_try_#{v}"]  += 1
     end
     obj.metadata[k] = v
   end unless o[:metadata].nil?

data/lib/miga/common/format.rb

@@ -3,18 +3,17 @@ require 'tempfile'
 require 'zlib'
 
 module MiGA::Common::Format
-
   ##
   # Tabulates an +values+, and Array of Arrays, all with the same number of
   # entries as +header+. Returns an Array of String, one per line.
   def tabulate(header, values)
     fields = [header.map(&:to_s)]
-    fields << fields.first.map{ |h| h.gsub(/\S/, '-') }
-    fields += values.map{ |row| row.map{ |cell| cell.nil? ? '?' : cell.to_s } }
-    clen = fields.map{ |row| row.map(&:length) }.transpose.map(&:max)
-    fields.map do |row|
-      (0 .. clen.size-1).map do |col_n|
-        col_n==0 ? row[col_n].rjust(clen[col_n]) : row[col_n].ljust(clen[col_n])
+    fields << fields.first.map { |h| h.gsub(/\S/, '-') }
+    fields += values.map { |r| r.map { |cell| cell.nil? ? '?' : cell.to_s } }
+    clen = fields.map { |r| r.map(&:length) }.transpose.map(&:max)
+    fields.map do |r|
+      (0 .. clen.size - 1).map do |col_n|
+        col_n == 0 ? r[col_n].rjust(clen[col_n]) : r[col_n].ljust(clen[col_n])
       end.join('  ')
     end
   end
@@ -40,7 +39,7 @@ module MiGA::Common::Format
           (id, df) = [$1, $2]
           tmp_fh.print buffer.wrap_width(80)
           buffer = ''
-          tmp_fh.puts ">#{id.gsub(/[^A-Za-z0-9_\|\.]/, "_")}#{df}"
+          tmp_fh.puts ">#{id.gsub(/[^A-Za-z0-9_\|\.]/, '_')}#{df}"
         else
           buffer << ln.gsub(/[^A-Za-z\.\-]/, '')
         end
@@ -65,30 +64,30 @@ module MiGA::Common::Format
   # controlled via the +opts+ Hash. Supported options include:
   # - +:n50+: If true, it also returns the N50 and the median (in bp).
   # - +gc+: If true, it also returns the G+C content (in %).
-  def seqs_length(file, format, opts={})
+  def seqs_length(file, format, opts = {})
     fh = (file =~ /\.gz/) ? Zlib::GzipReader.open(file) : File.open(file, 'r')
     l = []
     gc = 0
     i = 0 # <- Zlib::GzipReader doesn't set $.
     fh.each_line do |ln|
       i += 1
-      if (format==:fasta and ln =~ /^>/) or (format==:fastq and (i % 4)==1)
+      if (format == :fasta and ln =~ /^>/) or (format == :fastq and (i % 4)==1)
         l << 0
-      elsif format==:fasta or (i % 4)==2
+      elsif format == :fasta or (i % 4) == 2
         l[l.size-1] += ln.chomp.size
         gc += ln.scan(/[GCgc]/).count if opts[:gc]
       end
     end
     fh.close
-    
+
     o = { n: l.size, tot: l.inject(:+) }
-    o[:avg] = o[:tot].to_f/l.size
-    o[:var] = l.map{ |a| a ** 2 }.inject(:+).to_f/l.size - o[:avg]**2
+    o[:avg] = o[:tot].to_f / l.size
+    o[:var] = l.map { |a| a**2 }.inject(:+).to_f / l.size - o[:avg]**2
     o[:sd]  = Math.sqrt o[:var]
-    o[:gc]  = 100.0*gc/o[:tot] if opts[:gc]
+    o[:gc]  = 100.0 * gc / o[:tot] if opts[:gc]
     if opts[:n50]
       l.sort!
-      thr = o[:tot]/2
+      thr = o[:tot] / 2
       pos = 0
       l.each do |a|
         pos += a
@@ -96,7 +95,7 @@ module MiGA::Common::Format
         break if pos >= thr
       end
       o[:med] = o[:n].even? ?
-        0.5*l[o[:n]/2-1,2].inject(:+) : l[(o[:n]-1)/2]
+            0.5 * l[o[:n] / 2 - 1, 2].inject(:+) : l[(o[:n] - 1) / 2]
     end
     o
   end
@@ -105,7 +104,6 @@ end
 ##
 # MiGA extensions to the String class.
 class String
-
   ##
   # Replace any character not allowed in a MiGA name for underscore (_). This
   # results in a MiGA-compliant name EXCEPT for empty strings, that results in
@@ -123,13 +121,13 @@ class String
   ##
   # Replace underscores by spaces or dots (depending on context).
   def unmiga_name
-    gsub(/_(str|sp|subsp|pv)__/,"_\\1._").tr('_', ' ')
+    gsub(/_(str|sp|subsp|pv)__/, '_\\1._').tr('_', ' ')
   end
 
   ##
   # Wraps the string with fixed Integer +width+.
   def wrap_width(width)
-    gsub(/([^\n\r]{1,#{width}})/,"\\1\n")
+    gsub(/([^\n\r]{1,#{width}})/, "\\1\n")
   end
 end
 

data/lib/miga/dataset.rb

@@ -76,6 +76,20 @@ class MiGA::Dataset < MiGA::MiGA
     self.results.each{ |r| r.remove! }
     self.metadata.remove!
   end
+
+  ##
+  # Inactivate a dataset. This halts automated processing by the daemon.
+  def inactivate!
+    self.metadata[:inactive] = true
+    self.metadata.save
+  end
+
+  ##
+  # Activate a dataset. This removes the +:inactive+ flag.
+  def activate!
+    self.metadata[:inactive] = nil
+    self.metadata.save
+  end
   
   ##
   # Get standard metadata values for the dataset as Array.
@@ -96,22 +110,27 @@ class MiGA::Dataset < MiGA::MiGA
   ##
   # Is this dataset known to be multi-organism?
   def is_multi?
-    return false if metadata[:type].nil? or
-      @@KNOWN_TYPES[type].nil?
+    return false if metadata[:type].nil? or @@KNOWN_TYPES[type].nil?
     @@KNOWN_TYPES[type][:multi]
   end
   
   ##
   # Is this dataset known to be single-organism?
   def is_nonmulti?
-    return false if metadata[:type].nil? or
-      @@KNOWN_TYPES[type].nil?
+    return false if metadata[:type].nil? or @@KNOWN_TYPES[type].nil?
     !@@KNOWN_TYPES[type][:multi]
   end
+
+  ##
+  # Is this dataset active?
+  def is_active?
+    metadata[:inactive].nil? or !metadata[:inactive]
+  end
   
   ##
   # Should I ignore +task+ for this dataset?
   def ignore_task?(task)
+    return true unless is_active?
     return !metadata["run_#{task}"] unless metadata["run_#{task}"].nil?
     return true if task==:taxonomy and project.metadata[:ref_project].nil?
     pattern = [true, false]

data/lib/miga/dataset/result.rb

@@ -78,7 +78,13 @@ module MiGA::Dataset::Result
     return nil if first.nil?
     @@PREPROCESSING_TASKS.each do |t|
       next if ignore_task? t
-      return t if after_first and add_result(t, save).nil?
+      if after_first and add_result(t, save).nil?
+        if (metadata["_try_#{t}"] || 0) > (project.metadata[:max_try] || 10)
+          inactivate!
+          return nil
+        end
+        return t
+      end
       after_first = (after_first or (t==first))
     end
     nil

data/lib/miga/version.rb

@@ -10,7 +10,7 @@ module MiGA
   # - Float representing the major.minor version.
   # - Integer representing gem releases of the current version.
   # - Integer representing minor changes that require new version number.
-  VERSION = [0.3, 2, 0]
+  VERSION = [0.3, 2, 1]
   
   ##
   # Nickname for the current major.minor version.

data/scripts/miga.bash

@@ -14,6 +14,9 @@ for i in $(miga plugins -P "$PROJECT") ; do
   source "$i/scripts-plugin.bash"
 done
 
+[[ -n $DATASET ]] \
+      && miga add -P "$PROJECT" -D "$DATASET" -m "_step=$SCRIPT" --update
+
 #if [[ "$RUNTYPE" == "qsub" ]] ; then
 #elif [[ "$RUNTYPE" == "msub" ]] ; then
 #fi

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: miga-base
 version: !ruby/object:Gem::Version
-  version: 0.3.2.0
+  version: 0.3.2.1
 platform: ruby
 authors:
 - Luis M. Rodriguez-R
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2018-06-22 00:00:00.000000000 Z
+date: 2018-06-28 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rest-client
@@ -503,7 +503,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
 requirements: []
 rubyforge_project: 
-rubygems_version: 2.6.13
+rubygems_version: 2.7.6
 signing_key: 
 specification_version: 4
 summary: MiGA