miga-base 1.3.13.9 → 1.3.14.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 843f35d4fa4d3543cc030db30b32ba87e40e21c967c4ecf3832973857e58dc05
-  data.tar.gz: e6f121106bce201cba4d6da053c1a07d977b12f5ee4d02d84c63ca204e4afa29
+  metadata.gz: 2459aa0f06af71628701bc0a71c1830bd21fc24d5c3e0999c8f59b6bce8b6cf6
+  data.tar.gz: 1590b03b5dfbe42241dd943a61388ff2e22a4956c6e9169d321a9bb857b9713f
 SHA512:
-  metadata.gz: de88c20be5deffa734c0fc692bde8ecc7f87c1bea571c125915b032232e1c9a54e41eb37f3c666f54a2adfd1be64adfb12ec61000553a3b1ace9ba70c7a3d048
-  data.tar.gz: d2ff4a70266159ccca8c1cff3dbaf0c52316e5608096014c04523e309f6ce768e1daa4f779275ef0293cb83d14af982d9c8a5412e1d433f25425f76adcac8261
+  metadata.gz: cf4da3f0519c77fd290e92650a6a53b871d678516eda891ef15272e0fa5a8b4aa518a4531211b0ee9d78ca2e42798f6421b41a3980c3f6e64a159f78bf105e9a
+  data.tar.gz: e771646155d87a7f8e7896b5cfa121f94a60dbce78cd5be0231faf42e7f4aa2b75de4c3fdf043dadb33d07a67bca180d12a12871d1066bff73a454eecdf565a5

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

@@ -179,7 +179,7 @@ class MiGA::Cli::Action::Browse < MiGA::Cli::Action
     str
       .to_s.unmiga_name
       .sub(/^./, &:upcase)
-      .gsub(/(Aai|Ani|Ogs|Cds|Ssu|Rds|ani95|aai90| db$| ssu )/, &:upcase)
+      .gsub(/(Aai|Ani|Ogs|Cds|Ssu|Rds|Rda|ani95|aai90| db$| ssu )/, &:upcase)
       .sub(/Haai/, 'hAAI')
       .sub(/Mytaxa/, 'MyTaxa')
       .sub(/ pvalue$/, ' p-value')

data/lib/miga/project/result.rb

@@ -55,10 +55,13 @@ module MiGA::Project::Result
   ##
   # Add result of any type +:*_distances+ at +base+ (no +_opts+ supported).
   def add_result_distances(base, _opts)
-    return nil unless result_files_exist?(base, %w[.rds .txt])
+    return nil unless result_files_exist?(base, ['.txt']) &&
+      (result_files_exist?(base, ['.rds']) ||
+       result_files_exist?(base, ['.rda']))
 
     r = MiGA::Result.new("#{base}.json")
     r.add_file(:rds,    'miga-project.rds')
+    r.add_file(:rda,    'miga-project.rda')
     r.add_file(:rdata,  'miga-project.Rdata') # Legacy file
     r.add_file(:matrix, 'miga-project.txt')
     r.add_file(:log,    'miga-project.log') # Legacy file
@@ -84,6 +87,7 @@ module MiGA::Project::Result
 
     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')
     r.add_file(:proposal,     'miga-project.proposed-clades')
     r.add_file(:clades_aai90, 'miga-project.aai90-clades')
@@ -108,6 +112,7 @@ module MiGA::Project::Result
     r = add_result_iter_clades(base)
     r.add_file(:ani_tree, 'miga-project.ani.nwk')
     r.add_file(:ani_dist_rds, 'miga-project.dist.rds')
+    r.add_file(:ani_dist_rda, 'miga-project.dist.rda')
     r
   end
 

data/lib/miga/remote_dataset/base.rb

@@ -134,6 +134,15 @@ module MiGA::RemoteDataset::Base
       end,
       method: :get
     },
+    ncbi_fetch: {
+      dbs: { nuccore: { stage: :metadata, format: :gb } },
+      uri: lambda do |opts|
+        @@_EUTILS_BUILD[:efetch,
+          db: opts[:db], id: opts[:ids], rettype: opts[:format], retmode: :text
+        ]
+      end,
+      method: :get
+    },
     ncbi_search: {
       dbs: {
         assembly: { stage: :metadata, format: :json },

data/lib/miga/remote_dataset/download.rb

@@ -107,9 +107,12 @@ class MiGA::RemoteDataset
 
       MiGA::MiGA.DEBUG 'Empty sequence, attempting download as WGS records'
       a, b = opts[:obj].metadata[:ncbi_wgs].split('-', 2)
-      pref = longest_common_prefix([a, b])
-      rang = a[pref.size .. -1].to_i .. b[pref.size .. -1].to_i
-      ids  = rang.map { |k| "%s%0#{a.size - pref.size}i" % [pref, k] }
+      ids  = [a]
+      unless b.nil?
+        pref = longest_common_prefix([a, b])
+        rang = a[pref.size .. -1].to_i .. b[pref.size .. -1].to_i
+        ids  = rang.map { |k| "%s%0#{a.size - pref.size}i" % [pref, k] }
+      end
       download_rest(opts.merge(universe: :ncbi, db: :nuccore, ids: ids))
     end
 

data/lib/miga/remote_dataset.rb

@@ -300,6 +300,16 @@ class MiGA::RemoteDataset < MiGA::MiGA
       metadata[:web_assembly_gz] ||=
         '%s/%s_genomic.fna.gz' % [url_dir, File.basename(url_dir)]
     end
+
+    # If all conditions are right, try getting the WGS range
+    if @_ncbi_asm_json_doc['wgs'] && !@_ncbi_asm_json_doc['wgs'].empty? &&
+         metadata[:ncbi_nuccore] && !metadata[:ncbi_wgs]
+      doc = self.class.download(:ncbi_fetch, :nuccore, metadata[:ncbi_nuccore], :gb).split(/\n/)
+      ln  = doc.grep(/^WGS\s+\S+-\S+/).first
+      wgs = ln&.gsub(/^WGS\s+(\S+-\S+).*/, '\1')
+      metadata[:ncbi_wgs] = wgs if wgs
+    end
+
     @_ncbi_asm_json_doc
   end
 

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, 13, 9].freeze
+  VERSION = [1.3, 14, 1].freeze
 
   ##
   # Nickname for the current major.minor version.
@@ -20,7 +20,7 @@ module MiGA
 
   ##
   # Date of the current gem relese.
-  VERSION_DATE = Date.new(2024, 3, 31)
+  VERSION_DATE = Date.new(2024, 4, 1)
 
   ##
   # References of MiGA

data/scripts/aai_distances.bash

@@ -40,15 +40,16 @@ rm "miga-project.txt.lno"
 # R-ify
 cat <<R | R --vanilla
 file <- gzfile("miga-project.txt.gz")
-aai <- read.table(
-  file, sep = "\t", header = TRUE, as.is = TRUE, quote = "",
-  stringsAsFactors = FALSE, comment.char = "", nrows = $LNO,
-  colClasses = c("character", "character",
-                 "numeric", "numeric", "integer", "integer")
-)
-saveRDS(aai, file = "miga-project.rds")
-if(sum(aai[, "a"] != aai[, "b"]) > 0) {
-  h <- hist(aai[aai[, "a"] != aai[, "b"], "value"], breaks = 100, plot = FALSE)
+text <- readLines(file, n = $LNO + 1, ok = FALSE)
+list <- strsplit(text[-1], "\t", fixed = TRUE)
+a <- sapply(list, function(x) x[1])
+b <- sapply(list, function(x) x[2])
+d <- sapply(list, function(x) 1 - (as.numeric(x[3]) / 100))
+save(a, b, d, file = "miga-project.rda")
+
+non_self <- a != b
+if(sum(non_self) > 0) {
+  h <- hist((1 - d[non_self]) * 100, breaks = 100, plot = FALSE)
   len <- length(h[["breaks"]])
   write.table(
     cbind(h[["breaks"]][-len], h[["breaks"]][-1], h[["counts"]]),

data/scripts/ani_distances.bash

@@ -34,15 +34,16 @@ rm "miga-project.txt.lno"
 # R-ify
 cat <<R | R --vanilla
 file <- gzfile("miga-project.txt.gz")
-ani <- read.table(
-  file, sep = "\t", header = TRUE, as.is = TRUE, quote = "",
-  stringsAsFactors = FALSE, comment.char = "", nrows = $LNO,
-  colClasses = c("character", "character",
-                 "numeric", "numeric", "integer", "integer")
-)
-saveRDS(ani, file = "miga-project.rds")
-if(sum(ani[, "a"] != ani[, "b"]) > 0) {
-  h <- hist(ani[ani[, "a"] != ani[, "b"], "value"], breaks = 100, plot = FALSE)
+text <- readLines(file, n = $LNO + 1, ok = FALSE)
+list <- strsplit(text[-1], "\t", fixed = TRUE)
+a <- sapply(list, function(x) x[1])
+b <- sapply(list, function(x) x[2])
+d <- sapply(list, function(x) 1 - (as.numeric(x[3]) / 100))
+save(a, b, d, file = "miga-project.rda")
+
+non_self <- a != b
+if(sum(non_self) > 0) {
+  h <- hist((1 - d[non_self]) * 100, breaks = 100, plot = FALSE)
   len <- length(h[["breaks"]])
   write.table(
     cbind(h[["breaks"]][-len], h[["breaks"]][-1], h[["counts"]]),

data/utils/find-medoid.R

@@ -16,15 +16,14 @@ if(Sys.getenv("MIGA") == ""){
   ))
 }
 
-find_medoids <- function (ani.df, out, clades) {
-  if(nrow(ani.df) == 0) return(NULL)
-  ani.df$d <- 1 - (ani.df$value/100)
-  dist <- enve.df2dist(ani.df, "a", "b", "d", default.d = max(ani.df$d) * 1.2)
+find_medoids <- function (a, b, d, out, clades) {
+  if (length(d) == 0) return(NULL)
+  dist <- enve.df2dist(cbind(a, b, d), "a", "b", "d", default.d = max(d) * 1.2)
   dist <- as.matrix(dist)
   cl <- read.table(clades, header = FALSE, sep = "\t", as.is = TRUE)[,1]
   cl.s <- c()
   medoids <- c()
-  for(i in cl){
+  for (i in cl) {
     lab <- strsplit(i, ",")[[1]]
     if(length(lab) == 1) {
       lab.s <- lab
@@ -44,6 +43,12 @@ find_medoids <- function (ani.df, out, clades) {
 
 #= Main
 cat("Finding Medoids\n")
-ani <- readRDS(argv[1])
-find_medoids(ani.df = ani, out = argv[2], clades = argv[3])
+if (grepl("\\.rds$", argv[1])) {
+  ani <- readRDS(argv[1])
+  find_medoids(ani$a, ani$b, 1 - (ani$value / 100),
+    out = argv[2], clades = argv[3])
+} else {
+  load(argv[1]) # assume .rda
+  find_medoids(a, b, d, out = argv[2], clades = argv[3])
+}
 

data/utils/subclade/pipeline.rb

@@ -48,9 +48,10 @@ module MiGA::SubcladeRunner::Pipeline
     # Find genomospecies medoids
     src = File.expand_path('utils/find-medoid.R', MiGA::MiGA.root_path)
     dir = opts[:gsp_metric] == 'aai' ? '02.aai' : '03.ani'
+    dat = "../../09.distances/#{dir}/miga-project.rda"
+    dat = "../../09.distances/#{dir}/miga-project.rds" unless File.exist?(dat)
     run_cmd([
-      'Rscript', src, "../../09.distances/#{dir}/miga-project.rds",
-      'miga-project.gsp-medoids', 'miga-project.gsp-clades'
+      'Rscript', src, dat, 'miga-project.gsp-medoids', 'miga-project.gsp-clades'
     ])
     if File.exist? 'miga-project.gsp-clades.sorted'
       File.rename 'miga-project.gsp-clades.sorted', 'miga-project.gsp-clades'

data/utils/subclades.R

@@ -338,18 +338,25 @@ ggplotColours <- function (n = 6, h = c(0, 360) + 15, alpha = 1) {
 }
 
 ani_distance <- function (ani_file, sel) {
-  # Try to locate rds, otherwise read gzipped table
-  rds <- gsub("\\.txt\\.gz$", ".rds", ani_file)
-  if (file.exists(rds)) {
-    sim <- readRDS(rds)
+  # Try to locate rda, then rds, and otherwise read gzipped table
+  rda <- gsub("\\.txt\\.gz$", ".rda", ani_file)
+  if (file.exists(rda)) {
+    load(rda) # Should already contain `a`, `b`, and `d` as vectors
   } else {
-    sim <- read.table(gzfile(ani_file), sep = "\t", header = TRUE, as.is = TRUE)
-  }
+    rds <- gsub("\\.txt\\.gz$", ".rds", ani_file)
+    if (file.exists(rds)) {
+      sim <- readRDS(rds)
+    } else {
+      sim <- read.table(
+        gzfile(ani_file), sep = "\t", header = TRUE, as.is = TRUE
+      )
+    }
 
-  # Extract individual variables to deal with very large matrices
-  a <- sim$a
-  b <- sim$b
-  d <- 1 - (sim$value / 100)
+    # Extract individual variables to deal with very large matrices
+    a <- sim$a
+    b <- sim$b
+    d <- 1 - (sim$value / 100)
+  }
   
   # If there is no data, end process
   if (length(a) == 0) return(NULL)
@@ -359,7 +366,7 @@ ani_distance <- function (ani_file, sel) {
   if (!is.na(sel) && file.exists(sel)) {
     say("Filter selection")
     ids <- read.table(sel, sep = "\t", head = FALSE, as.is = TRUE)[,1]
-    sel.idx <- which(sim$a %in% ids & sim$b %in% ids)
+    sel.idx <- which(a %in% ids & b %in% ids)
     a <- a[sel.idx]
     b <- b[sel.idx]
     d <- d[sel.idx]

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: miga-base
 version: !ruby/object:Gem::Version
-  version: 1.3.13.9
+  version: 1.3.14.1
 platform: ruby
 authors:
 - Luis M. Rodriguez-R
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2024-03-31 00:00:00.000000000 Z
+date: 2024-04-01 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: daemons