rbbt-sources 3.1.47 → 3.1.48

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 6cecde427390fc444e5ef5b6bef42072bf87714b404647c6c086da92a1eaa13a
-  data.tar.gz: 673429c8eab433ec9854c81d52450f8885c57e4b3f8b6779ded44585e6f88343
+  metadata.gz: 92bb6e2852266c73666e0b07221484822441d19603b2ce9d4fe9ceb6faa44d37
+  data.tar.gz: 9eda4d907bb34ae5b5b95dbdb972b6c659111b042c5b68661d41cf3ccd93e3b8
 SHA512:
-  metadata.gz: e39a3a6418bfb514ec786da0128f6fb8437da9728224dacd41b837300680c6d34e821f08954820cb5b7fd064546fee21fd0557d179c469eda969485ffd5cd965
-  data.tar.gz: 9f332f35fcd9c66db7506875efcc7c5276b1f588dd3b70ca299aaa2a7f25c33a650e496ba207af4e12e3bb2a6dfbc9ec8b1f0729e3acedc354145572dd268836
+  metadata.gz: 4ac47f700cfc54144bfbc2dc7b3eeaf2e9dce885ccfc1e47c26d978a79dd02d99802a83a888bb1bd80de96ecf9bf023ae548041f3b1867308228c319777dabd4
+  data.tar.gz: 05b07ac77a194f0ad87a36b050eb97d2285490974d0bca039c2163d762c11461f4ccf6c4585df62b24298524bb9f02e4d1d65dae51418e491644b4725fbd864f

data/lib/rbbt/sources/biomart.rb

@@ -204,7 +204,6 @@ module BioMart
         missing+=v if Organism.compare_archives(current_archive, t) == -1
       elsif k=~ /^>(.*)/ 
         t = $1.strip
-        iii [current_archive, t, Organism.compare_archives(current_archive, t)]
         missing+=v if Organism.compare_archives(current_archive, t) == 1
       end
     end

data/lib/rbbt/sources/ensembl_ftp.rb

@@ -52,24 +52,35 @@ module Ensembl
       File.join("ftp://" + SERVER, ftp_directory_for(organism) )
     end
 
-    def self.url_for(organism, table)
-      "#{base_url(organism)}/#{table}.txt.gz.bz2"
+    def self.url_for(organism, table, extension)
+      File.join(base_url(organism), table) + ".#{extension}.gz"
+    end
+
+    def self._get_gz(url)
+      begin
+        CMD.cmd("wget '#{url}' -O  - | gunzip").read
+      rescue
+        CMD.cmd("wget '#{url}.bz2' -O  - | bunzip2 | gunzip").read
+      end
+    end
+
+    def self._get_file(organism, table, extension)
+      url = url_for(organism, table, extension)
+      self._get_gz(url)
     end
 
     def self.has_table?(organism, table)
-      sql_file = CMD.cmd("wget '#{base_url(organism)}/#{File.basename(base_url(organism))}.sql.gz.bz2' -O  -| bunzip2| gunzip").read
+      sql_file = _get_file(organism, File.basename(base_url(organism)), 'sql')
       ! sql_file.match(/^CREATE TABLE .#{table}. \((.*?)^\)/sm).nil?
     end
 
     def self.fields_for(organism, table)
-      sql_file = CMD.cmd("wget '#{base_url(organism)}/#{File.basename(base_url(organism))}.sql.gz.bz2' -O  -| bunzip2| gunzip").read
-      
+      sql_file = _get_file(organism, File.basename(base_url(organism)), 'sql')
       chunk = sql_file.match(/^CREATE TABLE .#{table}. \((.*?)^\)/sm)[1]
       chunk.scan(/^\s+`(.*?)`/).flatten
     end
 
     def self.ensembl_tsv(organism, table, key_field = nil, fields = nil, options = {})
-      url = url_for(organism, table)
       if key_field and fields
         all_fields = fields_for(organism, table)
         key_pos = all_fields.index key_field
@@ -78,7 +89,8 @@ module Ensembl
         options[:key_field] = key_pos
         options[:fields]    = field_pos
       end
-      tsv = TSV.open(CMD.cmd("wget '#{url}' -O - |bunzip2|gunzip", :pipe => true), options)
+
+      tsv = TSV.open(StringIO.new(_get_file(organism, table, "txt")), options)
       tsv.key_field = key_field
       tsv.fields = fields
       tsv

data/share/install/Organism/organism_helpers.rb

@@ -728,13 +728,18 @@ file 'protein_sequence' => ["transcripts", "transcript_5utr", "transcript_3utr",
   transcript_sequence.through do |transcript, sequence|
     protein = transcript_protein[transcript]
     next if protein.nil? or protein.empty?
+
     utr5 = transcript_5utr[transcript]
     utr3 = transcript_3utr[transcript]
     phase = transcript_phase[transcript] || 0
+
     if phase < 0
-      utr5 = - phase if utr5 == 0
+      if utr5.nil? || utr5 == 0 || utr5 == "0"
+        utr5 = 0
+      end
       phase = 0
     end
+
     psequence = Bio::Sequence::NA.new(("N" * phase) << sequence[utr5..sequence.length-utr3-1]).translate
     protein_sequence[protein]=psequence
   end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: rbbt-sources
 version: !ruby/object:Gem::Version
-  version: 3.1.47
+  version: 3.1.48
 platform: ruby
 authors:
 - Miguel Vazquez
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2020-07-14 00:00:00.000000000 Z
+date: 2020-08-12 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rbbt-util