Extend cases for WGS fetching

bio-miga · Apr 1, 2024 · e18c651 · e18c651
1 parent be0a10e
commit e18c651
Show file tree

Hide file tree

Showing 4 changed files with 27 additions and 5 deletions.
diff --git a/lib/miga/remote_dataset.rb b/lib/miga/remote_dataset.rb
@@ -300,6 +300,16 @@ def ncbi_asm_json_doc
       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
 

diff --git a/lib/miga/remote_dataset/base.rb b/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 },

diff --git a/lib/miga/remote_dataset/download.rb b/lib/miga/remote_dataset/download.rb
@@ -107,9 +107,12 @@ def ncbi_gb_get(opts)
 
       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
 

diff --git a/lib/miga/version.rb b/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, 9].freeze
+  VERSION = [1.3, 13, 10].freeze
 
   ##
   # Nickname for the current major.minor version.
   VERSION_NAME = 'mezzotint'
 
   ##
   # Date of the current gem relese.
-  VERSION_DATE = Date.new(2024, 3, 31)
+  VERSION_DATE = Date.new(2024, 4, 1)
 
   ##
   # References of MiGA