miga-base 1.3.21.5 → 1.3.22.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 9522358be2993d435fcece0ff3a1916919c955754bce64f02ebd32e50c1cdc11
-  data.tar.gz: c3277224287f994d56fb1f0e27b0adc8d5dc2010ac00f48f124ca60ba55beee9
+  metadata.gz: 5e4cafb2b821ae172aa9537d5c71c1cafddd6ed73adfa993da1267876dddaedf
+  data.tar.gz: 9fe748cc53152b70fe03a3d85bcf56b517fdae772c7eb1a96329394253f5ca78
 SHA512:
-  metadata.gz: 6cb5214b223052e199de0185960fe8371e2de08d82cdcc97dcd3e1089da28bf25e05438b3d33c4d2bb3d3ae9be4950144db9c0c79d48cbc2091292952cc88a26
-  data.tar.gz: 1be4517ab0285d5bdbaf01b7aaa3514f459e623489c4887712a403a217268beb46b57f778d6bcd010d32cfb9716da76736c568b015b5ffc31f5b345d63ffd570
+  metadata.gz: 64c9508befc2bc9c7bf35041ab8ab2dc90aca74e38d80fc401dd6135e9a703135456a8cde5b88e61918007b6249b9e923b279ade15ade5c63ad2799d27b0350f
+  data.tar.gz: 3a74e3c63ca8332c16d5d3583890f6f4e47d8b0afd4cd178046877164083a73d918b37c04a8a8086c6c31d1b55efdf5bc913aa890869220ae5f2e68f8a976a1d

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

@@ -124,6 +124,13 @@ module MiGA::Cli::Action::Wf
       cli[:aai_p] = 'diamond'
       cli[:ani_p] = 'fastani'
     end
+    opt.on(
+      '--only-ani', 'Alias to: --haai-p no --aai-p no',
+      'Use only for collections with expected AAI around or above 85%'
+    ) do
+      cli[:haai_p] = 'no'
+      cli[:aai_p] = 'no'
+    end
     opt.on(
       '--haai-p STRING',
       'hAAI search engine. One of: blast+, fastaai, blat, diamond, fastaai, no',
@@ -131,7 +138,7 @@ module MiGA::Cli::Action::Wf
     ) { |v| cli[:haai_p] = v }
     opt.on(
       '--aai-p STRING',
-      'AAI search engine. One of: blast+, blat, diamond (default)'
+      'AAI search engine. One of: blast+, blat, diamond (default), no'
     ) { |v| cli[:aai_p] = v }
     opt.on(
       '--ani-p STRING',

data/lib/miga/daemon.rb

@@ -183,7 +183,7 @@ class MiGA::Daemon < MiGA::MiGA
   end
 
   ##
-  # Traverse datasets, and returns boolean indicating if at any reference
+  # Traverse datasets, and returns boolean indicating if any reference
   # datasets are incomplete
   def check_datasets
     l_say(2, 'Checking datasets')
@@ -196,11 +196,17 @@ class MiGA::Daemon < MiGA::MiGA
       queue_job(:d, ds)
     end
     unless show_log?
+      @_check_datasets_reported_done ||= false
       n = project.dataset_names.count
-      k = jobs_to_run.size + jobs_running.size
-      k -= 1 unless get_job(:maintenance).nil?
-      advance('Datasets:', n - k, n, false)
-      miga_say if k == 0
+      k = (jobs_to_run + jobs_running).select { |i| !i[:ds].nil? }.size
+      if k > 0
+        advance('Datasets:', n - k, n, false)
+        @_check_datasets_reported_done = false
+      elsif !@_check_datasets_reported_done
+        advance('Datasets:', n, n, false)
+        miga_say
+        @_check_datasets_reported_done = true
+      end
     end
     o
   end

data/lib/miga/project/base.rb

@@ -142,7 +142,7 @@ module MiGA::Project::Base
     },
     aai_p: {
       desc: 'Value of aai.rb -p on AAI', default: 'diamond', type: String,
-      in: %w[blast+ blast blat diamond]
+      in: %w[blast+ blast blat diamond no]
     },
     ani_p: {
       desc: 'Value of ani.rb -p on ANI', default: 'fastani', type: String,
@@ -154,7 +154,14 @@ module MiGA::Project::Base
     },
     aai_save_rbm: {
       desc: 'Should RBMs be saved for OGS analysis?',
-      default: proc { |project| project.clade? },
+      default: proc do |project|
+        project.clade? && project.option(:aai_p) != 'no'
+      end,
+      in: [true, false]
+    },
+    run_ogs: {
+      desc: 'Should orthologous groups be estimated?',
+      default: proc { |project| project.option(:aai_save_rbm) },
       in: [true, false]
     },
     ogs_identity: {

data/lib/miga/project/result.rb

@@ -31,9 +31,11 @@ module MiGA::Project::Result
   ##
   # Is this +task+ to be bypassed?
   def ignore_task?(task)
-    return true if metadata["run_#{task}"] == false
+    opt = "run_#{task}"
+    return true if metadata[opt] == false
+    return true if option?(opt) && option(opt) == false
 
-    !clade? && @@INCLADE_TASKS.include?(task) && metadata["run_#{task}"] != true
+    !clade? && @@INCLADE_TASKS.include?(task) && metadata[opt] != true
   end
 
   ##
@@ -72,20 +74,21 @@ module MiGA::Project::Result
   ##
   # Add result type +:clade_finding+ at +base+ (no +_opts+ supported).
   def add_result_clade_finding(base, _opts)
+    r = nil
     if result_files_exist?(base, %w[.empty])
       r = MiGA::Result.new("#{base}.json")
       r.add_file(:empty, 'miga-project.empty')
-      return r
-    end
-    return nil unless result_files_exist?(base, %w[.proposed-clades])
-    unless clade? ||
-           result_files_exist?(
-             base, %w[.pdf .classif .medoids .class.tsv .class.nwk]
-           )
-      return nil
+    else
+      return nil unless result_files_exist?(base, %w[.proposed-clades])
+      unless clade? ||
+             result_files_exist?(
+               base, %w[.pdf .classif .medoids .class.tsv .class.nwk]
+             )
+        return nil
+      end
+      r = add_result_iter_clades(base)
     end
 
-    r = add_result_iter_clades(base)
     r.add_file(:aai_dist_rds, 'miga-project.dist.rds')
     r.add_file(:aai_dist_rda, 'miga-project.dist.rda')
     r.add_file(:aai_tree,     'miga-project.aai.nwk')

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 = [1.3, 21, 5].freeze
+  VERSION = [1.3, 22, 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(2025, 1, 16)
+  VERSION_DATE = Date.new(2025, 2, 20)
 
   ##
   # References of MiGA

data/utils/distance/commands.rb

@@ -60,10 +60,15 @@ module MiGA::DistanceRunner::Commands
   # Note that ANI values may be returned for lower (or failing) AAIs if the
   # value is already stored in the database
   def ani_after_aai(targets, aai_limit = 85.0)
-    # Run AAI and select targets with AAI ≥ aai_limit
-    aai = aai(targets)
-    sbj = aai.each_with_index.map { |i, k| targets[k] if i&.> aai_limit }
-    sbj.compact!
+    sbj =
+      if opts[:aai_p] == 'no'
+        # If we skip AAI, run ANI for all targets
+        targets
+      else
+        # Otherwise, run AAI and select targets with AAI ≥ aai_limit
+        aai = aai(targets)
+        aai.each_with_index.map { |i, k| targets[k] if i&.> aai_limit }.compact
+      end
 
     # Run ANI
     ani(sbj) unless sbj.empty?
@@ -110,37 +115,37 @@ module MiGA::DistanceRunner::Commands
     f1 = tmp_file('largecontigs.fa')
     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)
-        if target_asm
-          fh.puts target_asm
-          empty = false
-        end
+    # Select targets with assemblies
+    target_asm =
+      targets.map do |target|
+        target&.result(:assembly)&.file_path(:largecontigs)
+      end.compact
+
+    # Run FastANI in batches of up to 100 genomes
+    until target_asm.empty?
+      File.open(f2 = tmp_file, 'w') do |fh|
+        target_asm.shift(100).each { |i| fh.puts i }
       end
-    end
-    return if empty
-    run_cmd <<~CMD
-              fastANI -q "#{f1}" --rl "#{f2}" -t #{opts[:thr]} \
-              -o "#{f3 = tmp_file}"
-            CMD
+      run_cmd <<~CMD
+                fastANI -q "#{f1}" --rl "#{f2}" -t #{opts[:thr]} \
+                -o "#{f3 = tmp_file}"
+              CMD
 
-    # Retrieve resulting data and save to DB
-    data = {}
-    File.open(f3, 'r') do |fh|
-      fh.each do |ln|
-        row = ln.chomp.split("\t")
-        n2 = File.basename(row[1], '.gz')
-        n2 = File.basename(n2, '.LargeContigs.fna')
-        data[n2] = [row[2].to_f, 0.0, row[3].to_i, row[4].to_i]
+      # Retrieve resulting data and save to DB
+      data = {}
+      File.open(f3, 'r') do |fh|
+        fh.each do |ln|
+          row = ln.chomp.split("\t")
+          n2 = File.basename(row[1], '.gz')
+          n2 = File.basename(n2, '.LargeContigs.fna')
+          data[n2] = [row[2].to_f, 0.0, row[3].to_i, row[4].to_i]
+        end
       end
-    end
-    batch_data_to_db(:ani, data)
+      batch_data_to_db(:ani, data)
 
-    # Cleanup
-    [f2, f3].each { |i| File.unlink(i) }
+      # Cleanup
+      [f2, f3].each { |i| File.unlink(i) }
+    end
   end
 
   ##

data/utils/subclades.R

@@ -226,8 +226,9 @@ generate_empty_files <- function (out_base) {
   plot(1, t = "n", axes = F)
   legend("center", "No data", bty = "n")
   dev.off()
-  file.create(paste(out_base, ".1.classif", sep = ""))
-  file.create(paste(out_base, ".1.medoids", sep = ""))
+  file.create(paste0(out_base, ".1.classif"))
+  file.create(paste0(out_base, ".1.medoids"))
+  file.create(paste0(out_base, ".empty"))
 }
 
 write_text_report <- function (out_base, ani.d, ani.medoids, ani.types) {

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: miga-base
 version: !ruby/object:Gem::Version
-  version: 1.3.21.5
+  version: 1.3.22.2
 platform: ruby
 authors:
 - Luis M. Rodriguez-R
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2025-01-16 00:00:00.000000000 Z
+date: 2025-02-20 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: daemons