miga-base 0.7.2.0 → 0.7.3.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: bcf3f0505e7890ec5e74d7656d86514f989223a358271f0d4167898ae97f2fdf
-  data.tar.gz: ee0fab9ecd1166f23b26e5da3939e4cfd048d5193fb2ddac4537c380530e3b04
+  metadata.gz: 5b233f892ba1294bd0959433c443944f267ff9b8c7ec4d220dc4bbacaca985a6
+  data.tar.gz: bdc51401c6680d63872e7aab594eab50dbc500e6662d244a6fa04f6b6ea2587d
 SHA512:
-  metadata.gz: 8c812ee39a83e5aa8817884a5a480f786102f587cd14e3c8940253401713c4a677019ddcded616c8bbc0adf70b27ce165dcefe8c0ef0f6364d15add9d3beb078
-  data.tar.gz: 97efc31ea93ee3558bdbe8a85db0617349143a4f6c9b845abe688bff68e7420ed8c53c59b5b7bfba0baa878983d697c91933acf573ad3befa3b884b4dc286aea
+  metadata.gz: 96bc61749ae2964656a9d82a2b5b0c74691513af237837960dfe146482a5691dadf7ea8aa958fb5ff35abd1e8ac829c447e219fdf330095e151efd4448470d73
+  data.tar.gz: 299a4806eea3364a0a64d86aa0eabfa5798f27801d7892241e13bc8ec36e0d3325cddc4eb321ec5b0826b498bb86bdf98494e1384e42cfa6441d97940c690b4f

data/lib/miga/common/with_daemon_class.rb

@@ -27,6 +27,8 @@ module MiGA::Common::WithDaemonClass
     f = alive_file(path)
     f = terminated_file(path) unless File.exist? f
     return nil unless File.exist? f
-    Time.parse(File.read(f))
+    c = File.read(f)
+    return nil if c.nil? || c.empty?
+    Time.parse(c)
   end
 end

data/lib/miga/dataset.rb

@@ -3,6 +3,7 @@
 
 require 'miga/metadata'
 require 'miga/dataset/result'
+require 'miga/dataset/status'
 require 'miga/dataset/hooks'
 require 'sqlite3'
 
@@ -10,15 +11,15 @@ require 'sqlite3'
 # Dataset representation in MiGA.
 class MiGA::Dataset < MiGA::MiGA
   include MiGA::Dataset::Result
+  include MiGA::Dataset::Status
   include MiGA::Dataset::Hooks
 
   # Class-level
   class << self
-
     ##
     # Does the +project+ already have a dataset with that +name+?
     def exist?(project, name)
-      not project.dataset_names_hash[name].nil?
+      !project.dataset_names_hash[name].nil?
     end
 
     ##
@@ -26,7 +27,6 @@ class MiGA::Dataset < MiGA::MiGA
     def INFO_FIELDS
       %w(name created updated type ref user description comments)
     end
-
   end
 
   # Instance-level
@@ -57,7 +57,10 @@ class MiGA::Dataset < MiGA::MiGA
       File.expand_path("metadata/#{name}.json", project.path),
       metadata
     ]
-    save unless File.exist? @metadata_future[0]
+    unless File.exist? @metadata_future[0]
+      save
+      pull_hook :on_create
+    end
   end
 
   ##
@@ -117,32 +120,52 @@ class MiGA::Dataset < MiGA::MiGA
 
   ##
   # Is this dataset a reference?
-  def is_ref? ; !!metadata[:ref] ; end
+  def ref? ; !!metadata[:ref] ; end
 
   ##
   # Is this dataset a query (non-reference)?
-  def is_query? ; !metadata[:ref] ; end
+  def query? ; !metadata[:ref] ; end
 
   ##
   # Is this dataset known to be multi-organism?
-  def is_multi?
+  def multi?
     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?
+  def nonmulti?
     return false if metadata[:type].nil? or @@KNOWN_TYPES[type].nil?
     !@@KNOWN_TYPES[type][:multi]
   end
 
   ##
   # Is this dataset active?
-  def is_active?
+  def active?
     metadata[:inactive].nil? or !metadata[:inactive]
   end
 
+  ##
+  # Same as +ref?+ for backwards-compatibility
+  alias is_ref? ref?
+
+  ##
+  # Same as +query?+ for backwards-compatibility
+  alias is_query? query?
+
+  ##
+  # Same as +multi?+ for backwards-compatibility
+  alias is_multi? multi?
+
+  ##
+  # Same as +is_nonmulti?+ for backwards-compatibility
+  alias is_nonmulti? nonmulti?
+
+  ##
+  # Same as +active?+ for backwards-compatibility
+  alias is_active? active?
+
   ##
   # Returns an Array of +how_many+ duples (Arrays) sorted by AAI:
   # - +0+: A String with the name(s) of the reference dataset.
@@ -151,7 +174,7 @@ class MiGA::Dataset < MiGA::MiGA
   # +ref_project+ is false (default), and only for reference dataset when
   # +ref_project+ is true. It returns +nil+ if this analysis is not supported.
   def closest_relatives(how_many = 1, ref_project = false)
-    return nil if (is_ref? != ref_project) or is_multi?
+    return nil if (ref? != ref_project) or multi?
     r = result(ref_project ? :taxonomy : :distances)
     return nil if r.nil?
     db = SQLite3::Database.new(r.file_path :aai_db)
@@ -159,6 +182,4 @@ class MiGA::Dataset < MiGA::MiGA
       'SELECT seq2, aai FROM aai WHERE seq2 != ? ' \
       'GROUP BY seq2 ORDER BY aai DESC LIMIT ?', [name, how_many])
   end
-
 end
-

data/lib/miga/dataset/hooks.rb

@@ -4,29 +4,34 @@ require 'miga/common/hooks'
 ##
 # Helper module including specific functions to handle dataset hooks.
 # Supported events:
+# - on_create(): When first created
 # - on_load(): When loaded
 # - on_save(): When saved
 # - on_remove(): When removed
-# - on_inactivate(): When inactivated
 # - on_activate(): When activated
+# - on_inactivate(): When inactivated
 # - on_result_ready(result): When any result is ready, with key +result+
 # - on_result_ready_{result}(): When +result+ is ready
 # - on_preprocessing_ready(): When preprocessing is complete
 # Supported hooks:
 # - run_lambda(lambda, args...)
+# - recalculate_status()
 # - clear_run_counts()
 # - run_cmd(cmd)
 # Internal hooks:
 # - _pull_result_hooks()
 module MiGA::Dataset::Hooks 
-
   include MiGA::Common::Hooks
 
+  ##
+  # Dataset hooks triggered by default
   def default_hooks
     {
-      on_preprocessing_ready: [[:clear_run_counts]],
+      on_create: [[:recalculate_status]],
+      on_activate: [[:clear_run_counts], [:recalculate_status]],
+      on_inactivate: [[:recalculate_status]],
       on_result_ready: [[:_pull_result_hooks]],
-      on_activate: [[:clear_run_counts]]
+      on_preprocessing_ready: [[:clear_run_counts], [:recalculate_status]],
     }
   end
 
@@ -40,6 +45,12 @@ module MiGA::Dataset::Hooks
     save
   end
 
+  ##
+  # Recalculate the dataset status and save in metadata
+  def hook_recalculate_status(_hook_args, _event_args)
+    recalculate_status
+  end
+
   ##
   # Run +cmd+ in the command-line with {{variables}}: dataset, project, miga,
   # object (as defined for the event, if any)

data/lib/miga/dataset/result.rb

@@ -33,17 +33,17 @@ module MiGA::Dataset::Result
   # - nonmulti: incompatible dataset, only for nonmulti
   # - execute: do not ignore, execute the task
   def why_ignore(task)
-    if !is_active?
+    if !active?
       :inactive
     elsif !metadata["run_#{task}"].nil?
       metadata["run_#{task}"] ? :execute : :force
     elsif task == :taxonomy && project.metadata[:ref_project].nil?
       :project
-    elsif @@_EXCLUDE_NOREF_TASKS_H[task] && !is_ref?
+    elsif @@_EXCLUDE_NOREF_TASKS_H[task] && !ref?
       :noref
-    elsif @@_ONLY_MULTI_TASKS_H[task] && !is_multi?
+    elsif @@_ONLY_MULTI_TASKS_H[task] && !multi?
       :multi
-    elsif @@_ONLY_NONMULTI_TASKS_H[task] && !is_nonmulti?
+    elsif @@_ONLY_NONMULTI_TASKS_H[task] && !nonmulti?
       :nonmulti
     else
       :execute
@@ -139,7 +139,7 @@ module MiGA::Dataset::Result
   # the project as reference datasets.
   def cleanup_distances!
     r = get_result(:distances)
-    ref = project.datasets.select(&:is_ref?).select(&:is_active?).map(&:name)
+    ref = project.datasets.select(&:ref?).select(&:active?).map(&:name)
     return if r.nil?
     %i[haai_db aai_db ani_db].each do |db_type|
       db = r.file_path(db_type)
@@ -294,7 +294,7 @@ module MiGA::Dataset::Result
   ##
   # Add result type +:mytaxa+ at +base+ (no +_opts+ supported)
   def add_result_mytaxa(base, _opts)
-    if is_multi?
+    if multi?
       return nil unless
         result_files_exist?(base, '.mytaxa') ||
         result_files_exist?(base, '.nomytaxa.txt')
@@ -319,7 +319,7 @@ module MiGA::Dataset::Result
   ##
   # Add result type +:mytaxa_scan+ at +base+ (no +_opts+ supported)
   def add_result_mytaxa_scan(base, _opts)
-    if is_nonmulti?
+    if nonmulti?
       return nil unless
         result_files_exist?(base, %w[.pdf .mytaxa]) ||
         result_files_exist?(base, '.nomytaxa.txt')
@@ -345,8 +345,8 @@ module MiGA::Dataset::Result
   ##
   # Add result type +:distances+ at +base+ (no +_opts+ supported)
   def add_result_distances(base, _opts)
-    if is_nonmulti?
-      if is_ref?
+    if nonmulti?
+      if ref?
         add_result_distances_ref(base)
       else
         add_result_distances_nonref(base)

data/lib/miga/dataset/status.rb

@@ -0,0 +1,25 @@
+##
+# Helper module including specific functions for dataset status
+module MiGA::Dataset::Status
+  ##
+  # Returns the status of the dataset. If the status is not yet defined,
+  # it recalculates it and, if +save+ is true, saves it in metadata.
+  # Return values are:
+  # - +:inactive+ The dataset is currently inactive
+  # - +:incomplete+ The dataset is not yet fully processed
+  # - +:complete+ The dataset is fully processed
+  def status(save = false)
+    recalculate_status(save) if metadata[:status].nil?
+    metadata[:status].to_sym
+  end
+
+  ##
+  # Identify the current status instead of relying on metadata, and save
+  # it if +save+ is true. Return codes are the same as +status+.
+  def recalculate_status(save = true)
+    metadata[:status] =
+      !active? ? :inactive : done_preprocessing? ? :complete : :incomplete
+    self.save if save
+    metadata[:status].to_sym
+  end
+end

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 = [0.7, 2, 0]
+  VERSION = [0.7, 3, 0]
 
   ##
   # Nickname for the current major.minor version.
@@ -18,7 +18,7 @@ module MiGA
 
   ##
   # Date of the current gem release.
-  VERSION_DATE = Date.new(2020, 4, 20)
+  VERSION_DATE = Date.new(2020, 4, 22)
 
   ##
   # Reference of MiGA.

data/test/dataset_test.rb

@@ -39,6 +39,7 @@ class DatasetTest < Test::Unit::TestCase
     assert_equal('dataset1', $d1.name)
     assert_predicate($d1, :is_ref?)
     assert_equal(MiGA::Metadata, $d1.metadata.class)
+    assert_equal(:incomplete, $d1.status)
   end
 
   def test_save
@@ -181,4 +182,15 @@ class DatasetTest < Test::Unit::TestCase
     end
   end
 
+  def test_activate
+    d = $d1
+    assert_equal(:incomplete, d.status)
+    assert_predicate(d, :active?)
+    d.inactivate!
+    assert_equal(:inactive, d.status)
+    assert_not_predicate(d, :active?)
+    d.activate!
+    assert_equal(:incomplete, d.status)
+    assert_predicate(d, :active?)
+  end
 end

data/test/result_stats_test.rb

@@ -1,5 +1,6 @@
 require 'test_helper'
 require 'miga/project'
+require 'zlib'
 
 class ResultStatsTest < Test::Unit::TestCase
 
@@ -88,11 +89,14 @@ class ResultStatsTest < Test::Unit::TestCase
   end
 
   def test_assembly
+    # Prepare result
     dir = 'data/05.assembly'
     fa = file_path(dir, '.LargeContigs.fna')
     File.open(fa, 'w') { |fh| fh.puts '>1','ACTAC' }
     touch_done(dir)
     r = $d.add_result(:assembly)
+
+    # Test assertions
     assert_equal({}, r[:stats])
     r.compute_stats
     assert_equal(1, r[:stats][:contigs])
@@ -101,11 +105,18 @@ class ResultStatsTest < Test::Unit::TestCase
   end
 
   def test_cds
+    # Prepare result
     dir = 'data/06.cds'
     fa = file_path(dir, '.faa')
     File.open(fa, 'w') { |fh| fh.puts '>1','M' }
+    gff = file_path(dir, '.gff3.gz')
+    Zlib::GzipWriter.open(gff) do |fh|
+      fh.puts '# Model Data: a=b;transl_table=11;'
+    end
     touch_done(dir)
     r = $d.add_result(:cds)
+
+    # Test assertions
     assert_equal({}, r[:stats])
     r.compute_stats
     assert_equal(1, r[:stats][:predicted_proteins])
@@ -114,6 +125,28 @@ class ResultStatsTest < Test::Unit::TestCase
     test_assembly
     r.compute_stats
     assert_equal([60.0, '%'], r[:stats][:coding_density])
+    assert_equal('11', r[:stats][:codon_table])
+  end
+
+  def test_taxonomy
+    # Prepare result
+    dir = 'data/09.distances/05.taxonomy'
+    FileUtils.touch(file_path(dir, '.aai-medoids.tsv'))
+    FileUtils.touch(file_path(dir, '.aai.db'))
+    File.open(file_path(dir, '.intax.txt'), 'w') do |fh|
+      fh.puts 'Closest relative: dad with AAI: 100.0.'
+      3.times { fh.puts '' }
+      fh.puts ' phylum Abc  0.0  **** '
+    end
+    touch_done(dir)
+    r = $d.add_result(:taxonomy)
+
+    # Test assertions
+    assert_nil(r[:stats][:closest_relative])
+    r.compute_stats
+    assert_equal('dad', r[:stats][:closest_relative])
+    assert_equal([100.0, '%'], r[:stats][:aai])
+    assert_equal(0.0, r[:stats][:phylum_pvalue])
   end
 
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: miga-base
 version: !ruby/object:Gem::Version
-  version: 0.7.2.0
+  version: 0.7.3.0
 platform: ruby
 authors:
 - Luis M. Rodriguez-R
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2020-04-20 00:00:00.000000000 Z
+date: 2020-04-22 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: daemons
@@ -167,6 +167,7 @@ files:
 - lib/miga/dataset/base.rb
 - lib/miga/dataset/hooks.rb
 - lib/miga/dataset/result.rb
+- lib/miga/dataset/status.rb
 - lib/miga/json.rb
 - lib/miga/lair.rb
 - lib/miga/metadata.rb