miga-base 1.2.2.0 → 1.2.3.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: ab62d2e6d20aa126a5f455bd3b8159549d5c39b7b2e747587a0b0444a30d935a
-  data.tar.gz: 273671304affb6a21d1e4dc0374c4ff17edf39a96cba8893d2a666640ab1c7c9
+  metadata.gz: 94374ec5cc66632861a18159a8adf69591ce9d3d5249de5baacd636d7519bb9e
+  data.tar.gz: '0395e8ac95c1770b2ee859a4bc59d0ba806f1298cc271f4ff0b64a4c7363897d'
 SHA512:
-  metadata.gz: a157f0c4c52cfe4729eb5bda27900150153606b80f816a169ef18d6603b3fcd0ac03574cdbafbfcfe73e56a55ffbbf168a8c4316f066674ff6aebf6cef168500
-  data.tar.gz: 1a7b42d7dc75c5cac04f3408e6fba0e989d46ac63e926602c020e15340e24e8dc4694bf07b5f75ecc3a707db9d110d4a379841c426f7eb4a5670da5bcfa9642e
+  metadata.gz: 69667e9595a04b972865911b5c2e4204f879d8c8f7fee6cd68e5c4e7a84662d9c9d4018a4f533cbbb33eada796fc26faa4797b3c0c016d4ef34fc8c1d13138e2
+  data.tar.gz: 86aae6d82404516d2cbd8812eef539e00690fdf47ccecf7205bd4e4f301dc11f415f630111a40f22459a448579b4251698906e7e0380eb1631a1fae05915e183

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/doctor/base.rb

@@ -31,6 +31,7 @@ module MiGA::Cli::Action::Doctor::Base
         file = File.join(base, dir, file_db)
         blk[file, metric, result, rank] if File.exist? file
       end
+      # Query databases for reference databases refer to taxonomy runs
       base = File.join(base, '05.taxonomy')
       result = :taxonomy
     end
@@ -137,10 +138,11 @@ module MiGA::Cli::Action::Doctor::Base
   def save_bidirectional(a, dist)
     each_database_file(a) do |db_file, metric, result, rank|
       next if rank == :haai # No need for hAAI to be bidirectional
+      next if result == :taxonomy # Taxonomy is never bidirectional
 
       b2a = dist[rank].map { |b_name, v| b_name if v[a.name] }.compact
       a2b = dist[rank][a.name]&.keys || []
-      SQLite3::Database.new(db_file) do |db|
+      MiGA::SQLite.new(db_file).run do |db|
         sql = <<~SQL
           insert into #{metric}(seq1, seq2, #{metric}, sd, n, omega) \
           values(?, ?, ?, ?, ?, ?);

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/wf.rb

@@ -69,15 +69,20 @@ module MiGA::Cli::Action::Wf
       'By default: ~/.miga_daemon.json'
     ) { |v| cli[:daemon_json] = v }
     opt.on(
-      '-j', '--jobs INT',
+      '-j', '--jobs INT', Integer,
       'Number of parallel jobs to execute',
       'By default controlled by the daemon configuration (maxjobs)'
-    ) { |v| cli[:jobs] = v.to_i }
+    ) { |v| cli[:jobs] = v }
     opt.on(
-      '-t', '--threads INT',
+      '-t', '--threads INT', Integer,
       'Number of CPUs to use per job',
       'By default controlled by the daemon configuration (ppn)'
-    ) { |v| cli[:threads] = v.to_i }
+    ) { |v| cli[:threads] = v }
+    opt.on(
+      '--threads-project INT', Integer,
+      'Number of CPUs to use per project-wide job',
+      'By default controlled by the daemon configuration (ppn_project or ppn)'
+    ) { |v| cli[:threads_project] = v }
   end
 
   def opts_for_wf_distances(opt)
@@ -184,9 +189,10 @@ module MiGA::Cli::Action::Wf
 
   def run_daemon
     cmd  = ['daemon', 'run', '-P', cli[:outdir], '--shutdown-when-done']
-    cmd += ['--json', cli[:daemon_json]] unless cli[:daemon_json].nil?
-    cmd += ['--max-jobs', cli[:jobs]] unless cli[:jobs].nil?
-    cmd += ['--ppn', cli[:threads]] unless cli[:threads].nil?
+    cmd += ['--json', cli[:daemon_json]] if cli[:daemon_json]
+    cmd += ['--max-jobs', cli[:jobs]] if cli[:jobs]
+    cmd += ['--ppn', cli[:threads]] if cli[:threads]
+    cmd += ['--ppn-project', cli[:threads_project]] if cli[:threads_project]
     cmd += ['--debug', MiGA::MiGA.debug_trace? ? '2' : '1'] if MiGA::MiGA.debug?
     cwd = Dir.pwd
     call_cli(cmd)

data/lib/miga/common/with_result.rb

@@ -26,9 +26,7 @@ module MiGA::Common::WithResult
     task = task.to_sym
     return nil if result_dirs[task].nil?
 
-    base = File.join(
-      project.path, "data/#{result_dirs[task]}/#{result_base}"
-    )
+    base = File.join(project.path, 'data', result_dirs[task], result_base)
     json = "#{base}.json"
     return MiGA::Result.load(json) unless save
 

data/lib/miga/common.rb

@@ -43,9 +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/sqlite.rb

@@ -31,16 +31,30 @@ class MiGA::SQLite < MiGA::MiGA
   def initialize(path, opts = {})
     @opts = MiGA::SQLite.default_opts(opts)
     @path = File.absolute_path(path)
+    MiGA::MiGA.DEBUG("Accessing database: #{path}")
   end
 
   ##
-  # Executes +cmd+ and returns the result
+  # Executes +cmd+ and returns the result. Alternatively, if a block is
+  # passed, the commands are ignored and the block is executed with a
+  # single parameter of the database connection
   def run(*cmd)
+    if block_given?
+      run_block { |conn| yield(conn) }
+    else
+      y = nil
+      run_block { |conn| y = conn.execute(*cmd) }
+      y
+    end
+  end
+
+  ##
+  # Executes +blk+ that accepts a single parameter for the database
+  # connection
+  def run_block(&blk)
     busy_attempts ||= 0
     io_attempts ||= 0
-    y = nil
-    SQLite3::Database.new(path) { |conn| y = conn.execute(*cmd) }
-    y
+    SQLite3::Database.new(path) { |conn| blk[conn] }
   rescue SQLite3::BusyException => e
     busy_attempts += 1
     raise "Database busy #{path}: #{e.message}" if busy_attempts >= 3

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, 2, 0].freeze
+  VERSION = [1.2, 3, 0].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, 12)
+  VERSION_DATE = Date.new(2022, 2, 17)
 
   ##
   # References of MiGA

data/utils/core-pan-plot.R

@@ -1,29 +1,48 @@
 #!/usr/bin/env Rscript
 
-argv <- commandArgs(trailingOnly=T)
+argv <- commandArgs(trailingOnly = TRUE)
 
-plot_core_pan <- function(core_pan, pdf){
-  a <- read.table(core_pan, sep="\t", header=TRUE)
+plot_core_pan <- function (core_pan, pdf) {
+  a <- read.table(core_pan, sep = "\t", header = TRUE)
+  col <- c(
+    rgb(96, 11, 64, max = 255),
+    rgb(0, 121, 166, max = 255),
+    rgb(0.5, 0.5, 0.5, 1/3),
+    rep(rgb(0.5, 0.5, 0.5), 2)
+  )
   pdf(pdf, 7, 5)
-  plot(1, type="n", xlim=c(0, max(a$genomes)*1.05), xaxs="i", yaxs="i",
-    ylim=c(0, max(a$pan_q3)*1.05), xlab="Genomes", ylab="Orthologous Groups")
+  plot(
+    1, type = "n", xlim = c(0, max(a$genomes) * 1.05), xaxs = "i", yaxs = "i",
+    ylim = c(0, max(a$pan_q3) * 1.05)/1e3, xlab = "Genomes",
+    ylab = "Orthologous Groups (thousands)"
+  )
+
   # Core
-  polygon(c(a$genomes, rev(a$genomes)), c(a$core_q1, rev(a$core_q3)),
-    border=NA, col=rgb(0, 121, 166, 128/2, max=255))
-  lines(a$genomes, a$core_avg, col=rgb(0,121,166,max=255), lty=2)
-  lines(a$genomes, a$core_q2, col=rgb(0,121,166,max=255), lty=1)
+  polygon(
+    c(a$genomes, rev(a$genomes)), c(a$core_q1, rev(a$core_q3))/1e3,
+    border = NA, col = rgb(0, 121, 166, 256/4, max = 255)
+  )
+  lines(a$genomes, a$core_avg/1e3, col = col[2], lty = 2)
+  lines(a$genomes, a$core_q2/1e3, col = col[2], lty = 1)
+
   # Pan
-  polygon(c(a$genomes, rev(a$genomes)), c(a$pan_q1, rev(a$pan_q3)),
-    border=NA, col=rgb(96, 11, 64, 128/2, max=255))
-  lines(a$genomes, a$pan_avg, col=rgb(96,11,64,max=255), lty=2)
-  lines(a$genomes, a$pan_q2, col=rgb(96,11,64,max=255), lty=1)
+  polygon(
+    c(a$genomes, rev(a$genomes)), c(a$pan_q1, rev(a$pan_q3))/1e3,
+    border = NA, col = rgb(96, 11, 64, 256/4, max = 255)
+  )
+  lines(a$genomes, a$pan_avg/1e3, col = col[1], lty = 2)
+  lines(a$genomes, a$pan_q2/1e3, col = col[1], lty = 1)
+
   # Legend
-  legend("topleft",
-    legend=c("pangenome","core genome","Inter-Quartile","Median","Average"),
-    pch=c(16,16,15,NA,NA),lty=c(NA,NA,NA,1,2), pt.cex=c(1,1,2,NA,NA),
-    col=c(rgb(96,11,64,max=255), rgb(0,121,166,max=255),
-      rgb(0.5,0.5,0.5,166/255), rep(rgb(0.5,0.5,0.5),2)), bty="n")
+  legend <- c("pangenome", "core genome", "Inter-Quartile", "Median", "Average")
+  legend(
+    "topleft", legend = legend, col = col, bty = "n",
+    pch = c(16, 16, 15, NA, NA),
+    lty = c(NA, NA, NA, 1, 2),
+    pt.cex = c(1, 1, 2, NA, NA),
+  )
   dev.off()
 }
 
 plot_core_pan(argv[1], argv[2])
+

data/utils/index_metadata.rb

@@ -19,6 +19,7 @@ end
 p.each_dataset do |d|
   next unless d.ref? && d.active?
 
+  db.execute('BEGIN TRANSACTION')
   searchable(db, d, :name, d.name)
   d.metadata.each do |k, v|
     next if [:created, :updated].include? k
@@ -26,4 +27,5 @@ p.each_dataset do |d|
     v = v.sorted_ranks.map { |r| r[1] }.join(' ') if k == :tax
     searchable(db, d, k, v)
   end
+  db.execute('COMMIT')
 end

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: 1.2.2.0
+  version: 1.2.3.0
 platform: ruby
 authors:
 - Luis M. Rodriguez-R
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2021-12-12 00:00:00.000000000 Z
+date: 2022-02-17 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: daemons