Rescue start and versions on result overwrites

lib/miga/cli/action/doctor/base.rb

@@ -73,7 +73,9 @@ def saved_targets(dataset)
 
   ##
   # Reads all the distance estimates in +a+ -> * for +metric+ and
-  # returns them as a hash +{"b_name" => [val, sd, ...], ...}+
+  # returns them as a hash +{"b_name" => [val, sd, ...], ...}+ for
+  # rows with values other than the metric, or +{"b_name" => val}+ for
+  # rows with the metric only
   def read_bidirectional(a, metric)
     db_file = a.result(:distances)&.file_path("#{metric}_db") or return {}
     sql = "select seq2, #{metric}, sd, n, omega from #{metric}"

lib/miga/cli/action/doctor/distances.rb

@@ -70,10 +70,10 @@ def check_dist(cli)
   #---- Auxuliary functions -----
 
   ##
-  # Make a temporal directory holding partial bidirectionality reports (one per thread)
-  # in a custom multi-JSON format. Requires a MiGA::Project +project+ and the iterator of
-  # the reference datasets +ref_ds+. Returns the path to the temporal directory created.
-  # Used by +check_bidir+
+  # Make a temporal directory holding partial bidirectionality reports (one per
+  # thread) in a custom multi-JSON format. Requires a MiGA::Project +project+
+  # and the iterator of the reference datasets +ref_ds+. Returns the path to the
+  # temporal directory created. Used by +check_bidir+
   def partial_bidir_tmp(project, ref_ds)
     n = ref_ds.size
 
@@ -106,9 +106,9 @@ def partial_bidir_tmp(project, ref_ds)
   end
 
   ##
-  # Read partial temporal reports of bidirectionality (located in +tmp+), and return
-  # a two-deep hash with the final missingness report by metric (first key) and
-  # dataset name (second key). Used by +check_bidir+
+  # Read partial temporal reports of bidirectionality (located in +tmp+), and
+  # return a two-deep hash with the final missingness report by metric (first
+  # key) and dataset name (second key). Used by +check_bidir+
   def merge_bidir_tmp(tmp)
     dist = { aai: {}, ani: {} }
     cli[:threads].times do |i|

lib/miga/cli/action/doctor/operations.rb

@@ -58,7 +58,7 @@ def check_files(cli)
         unless ok
           cli.say "  > Registering again #{d.name}:#{r_k}   "
           d.add_result(r_k, true, force: true)
-          sr = d.result(:stats) and sr.remove!
+          d.result(:stats)&.remove!
         end
       end
     end

lib/miga/result.rb

@@ -33,14 +33,26 @@ def load(path)
     ##
     # Check if +path+ describes a result and otherwise create
     # it using the passed block. If +force+, ignore existing
-    # JSON in +path+ if any.
+    # JSON in +path+ if any, but it can rescue the versions 
+    # field from the old one if it exists and the new one doesn't
+    # contain such field.
     def create(path, force = false)
-      FileUtils.rm(path) if force && File.exist?(path)
-      r_pre = load(path)
-      return r_pre unless r_pre.nil?
+      # Deal with old results first
+      r_old = load(path)
+      return r_old if r_old && !force
 
+      # Create the new result using the block passed
+      FileUtils.rm(path) if r_old
       yield
-      load(path)
+
+      # Load and return
+      load(path).tap do |r_new|
+        # Rescue versions and start (if any and if necessary)
+        if r_old
+          %i[versions started].each { |i| r_new[i] ||= r_old[i] }
+          r_new[:versions] = (r_old[:versions] || {}).merge(r_new[:versions])
+        end
+      end
     end
   end
 

lib/miga/version.rb

@@ -12,15 +12,15 @@ module MiGA
   # - String indicating release status:
   #   - rc* release candidate, not released as gem
   #   - [0-9]+ stable release, released as gem
-  VERSION = [1.3, 13, 7].freeze
+  VERSION = [1.3, 13, 8].freeze
 
   ##
   # Nickname for the current major.minor version.
   VERSION_NAME = 'mezzotint'
 
   ##
   # Date of the current gem relese.
-  VERSION_DATE = Date.new(2024, 3, 22)
+  VERSION_DATE = Date.new(2024, 3, 23)
 
   ##
   # References of MiGA

manual/part2/requirements/conda.md

@@ -29,15 +29,15 @@ curl -Lso miga.yml \
 conda env create -f miga.yml
 rm miga.yml
 
-# Fix conda-forge Ruby issue (see https://github.com/bio-miga/miga/issues/168)
-mkdir -p "$GEM_HOME/bin"
-ln -s "$CONDA_PREFIX/bin/ruby" "$GEM_HOME/bin/ruby"
-
 # Tell MiGA to activate the proper conda environment
 echo 'eval "$(conda shell.bash hook)" && conda activate miga' > ~/.miga_modules
 
 # Activate the environment
 . ~/.miga_modules
+
+# Fix conda-forge Ruby issue (see https://github.com/bio-miga/miga/issues/168)
+mkdir -p "$GEM_HOME/bin"
+ln -s "$CONDA_PREFIX/bin/ruby" "$GEM_HOME/bin/ruby"
 ```
 
 ## MyTaxa utils

test/result_stats_test.rb

@@ -23,12 +23,12 @@ def test_single_raw_reads
     File.open(fq, 'w') { |fh| fh.puts '@1', 'ACTAC', '+', '####' }
     touch_done(dir)
     r = dataset.add_result(:raw_reads)
-    assert_equal({}, r[:stats])
+    assert_equal({}, r.stats)
     r.compute_stats
-    assert_not_empty(r[:stats])
-    assert_equal(Hash, r[:stats].class)
-    assert_equal(1, r[:stats][:reads])
-    assert_equal([40.0, '%'], r[:stats][:g_c_content])
+    assert_not_empty(r.stats)
+    assert_equal(Hash, r.stats.class)
+    assert_equal(1, r.stats[:reads])
+    assert_equal([40.0, '%'], r.stats[:g_c_content])
   end
 
   def test_coupled_raw_reads
@@ -40,20 +40,20 @@ def test_coupled_raw_reads
     touch_done(dir)
     r = dataset.add_result(:raw_reads)
     r.compute_stats
-    assert_not_empty(r[:stats])
-    assert_nil(r[:stats][:reads])
-    assert_equal(1, r[:stats][:read_pairs])
-    assert_equal([40.0, '%'], r[:stats][:reverse_g_c_content])
+    assert_not_empty(r.stats)
+    assert_nil(r.stats[:reads])
+    assert_equal(1, r.stats[:read_pairs])
+    assert_equal([40.0, '%'], r.stats[:reverse_g_c_content])
   end
 
   def test_trimmed_reads
     dir = 'data/02.trimmed_reads'
     FileUtils.touch(file_path(dir, '.1.clipped.fastq'))
     touch_done(dir)
     r = dataset.add_result(:trimmed_reads)
-    assert_equal({}, r[:stats])
+    assert_equal({}, r.stats)
     r.compute_stats
-    assert_equal({}, r[:stats])
+    assert_equal({}, r.stats)
   end
 
   def test_read_quality
@@ -62,9 +62,9 @@ def test_read_quality
     Dir.mkdir(file_path(dir, '.fastqc'))
     touch_done(dir)
     r = dataset.add_result(:read_quality)
-    assert_equal({}, r[:stats])
+    assert_equal({}, r.stats)
     r.compute_stats
-    assert_equal({}, r[:stats])
+    assert_equal({}, r.stats)
   end
 
   def test_trimmed_fasta
@@ -133,10 +133,10 @@ def test_taxonomy
     r = dataset.add_result(:taxonomy)
 
     # Test assertions
-    assert_nil(r[:stats][:closest_relative])
+    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])
+    assert_equal('dad', r.stats[:closest_relative])
+    assert_equal([100.0, '%'], r.stats[:aai])
+    assert_equal(0.0, r.stats[:phylum_pvalue])
   end
 end

test/result_test.rb

@@ -26,6 +26,30 @@ def test_add_result
     assert_instance_of(MiGA::Result, r)
   end
 
+  def test_overwrite_result
+    r = dataset.add_result(:trimmed_reads)
+    r[:ephemeral] = '@'
+    r[:versions] = { you: 'best' }
+    r.save
+
+    # Before reloading
+    assert_equal('@', r[:ephemeral])
+    assert_equal('best', r[:versions][:you])
+    assert_nil(r[:versions][:MiGA])
+
+    # After reloading (without forcing)
+    r = dataset.add_result(:trimmed_reads, true, force: false)
+    assert_equal('@', r[:ephemeral])
+    assert_equal('best', r[:versions][:you])
+    assert_nil(r[:versions][:MiGA])
+
+    # After reloading (with forcing)
+    r = dataset.add_result(:trimmed_reads, true, force: true)
+    assert_nil(r[:ephemeral])
+    assert_equal('best', r[:versions][:you])
+    assert_not_nil(r[:versions][:MiGA])
+  end
+
   def test_unlink
     r = project.add_result(:clade_finding)
     path = r.path