rbbt-sources 3.1.14 → 3.1.15

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: cc90d49b8322a4c9257bb4bf83610511b30a3a73
-  data.tar.gz: 53ea5b6bfbb4ae19ddd3b4323225c554bc624c9b
+  metadata.gz: 968b6c9ab9794fa92209ba2c8f36537bfc942711
+  data.tar.gz: 6a5501945048ef2a52872d54e35a85d8637b25a0
 SHA512:
-  metadata.gz: 03a06569443e9e0b196fd1917c282bfb48f879fca7ecfb93c5139893540a8bfa6f661f742a840dec4b97e4891ee2566f855b3c85148e87e8468b0162f64757b3
-  data.tar.gz: aafbcc3a8a02b2e1c32912a4fab1ffc6e897e74acdf01ca3252057a28c83798ba8e92c6348f9458d248a9adfedb2b0aa44c2910e0ea2b467b3b42f099208b10b
+  metadata.gz: 8ded417d4f6a04e6ff5ba21da1e454d54a0cf73aab95783a1a60ada25ea39faad2f8d4edcae462ae44f03de0ff488ea99737dc235d4472493d309342debe3391
+  data.tar.gz: 91b3fbe0852496b15cae6b512f150e5cf86c1742d57d41104b2247df3e75661dc5dce5aa4bae6e3eb86a3bd427addb842f97617a32a6f977a31fc0385daded3a

data/etc/allowed_biomart_archives

@@ -5,3 +5,4 @@ feb2014
 dec2013
 dec2014
 dec2015
+oct2016

data/lib/rbbt/sources/pubmed.rb

@@ -227,14 +227,15 @@ module PubMed
     result_files = FileCache.cache_online_elements(pmids, 'pubmed-{ID}.xml') do |ids|
       result = {}
       values = []
-      Misc.divide(ids, (ids.length / 100) + 1).each do |list|
+      Misc.divide(ids, (ids.length / 20) + 1).each do |list|
         begin
           Misc.try3times do
             url = "http://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi" 
 
             postdata = "db=pubmed&retmode=xml&id=#{list* ","}"
             xml = TmpFile.with_file(postdata) do |postfile|
-              Open.read(url, :quiet => true, :nocache => true, :nice => @@pubmed_lag, :nice_key => "PubMed", "--post-file=" => postfile)
+              #Open.read(url, :quiet => true, :nocache => true, :nice => @@pubmed_lag, :nice_key => "PubMed", "--post-file=" => postfile)
+              Open.read(url+'?'+postdata, :quiet => true, :nocache => true, :nice => @@pubmed_lag, :nice_key => "PubMed", "--__post-file=" => postfile)
             end
 
             values += xml.scan(/(<PubmedArticle>.*?<\/PubmedArticle>)/smu).flatten

data/lib/rbbt/sources/tfacts.rb

@@ -47,6 +47,81 @@ module TFacts
     TFacts.targets.tsv.reorder("Transcription Factor (Associated Gene Name)").to_s
   end
 
+  TFacts.claim TFacts.tf_tg, :proc do
+    require 'spreadsheet'
+    book = Spreadsheet.open TFacts[".source"]["Catalogues.xls"].produce.find
+
+    tsv = TSV.setup({}, :key_field => "Transcription Factor (Associated Gene Name)", :fields => ["Target Gene (Associated Gene Name)", "Sign", "Species", "Source", "PMID"], :namespace => "Hsa", :type => :double)
+
+    sheet = book.worksheet 1
+    sheet.each do |row|
+      target, tf, sign, source, species, pmids = row.values_at(0, 1, 2, 3, 4, 5).collect{|v| v.to_s.gsub(/\s/, ' ') }
+      pmids = "" if pmids.nil? or pmids == "0"
+      next if tf =~ /OFFICIAL_/
+
+      species = species.split(";").compact.reject{|s| s.empty?}.collect{|s| 
+        case s
+        when /Homo sap/i
+          "Homo sapiens"
+        when /muscul/i
+          "Mus musculus"
+        else
+          s
+        end
+      }*";"
+      source = source.split(";").compact.reject{|s| s.empty?}*";"
+      pmids = pmids.split(";").compact.reject{|s| s.empty?}.collect{|s| s.sub(/\s*pmid:\s*/,'').strip}*";"
+
+      values = [target, sign, species, source, pmids]
+      values = values.collect{|v| [v]}
+      tsv.zip_new(tf, values)
+    end
+
+    sheet = book.worksheet 0
+    sheet.each do |row|
+      target, tf, source, species, pmids = row.values_at(0, 1, 2, 3, 4, 5).collect{|v| v.to_s.gsub(/\s/, ' ') }
+      next if tf =~ /OFFICIAL_/
+
+      species = species.split(";").compact.reject{|s| s.empty?}.collect{|s| 
+        case s
+        when /Homo sap/i
+          "Homo sapiens"
+        when /muscul/i
+          "Mus musculus"
+        else
+          s
+        end
+      }*";"
+      source = source.split(";").compact.reject{|s| s.empty?}*";"
+      pmids = "" if pmids.nil? or pmids == "0"
+      pmids = pmids.split(";").compact.reject{|s| s.empty?}.collect{|s| s.sub(/\s*pmid:\s*/,'').strip}*";"
+
+      current = tsv[tf]
+      if current and current[0].include? target
+        new = []
+        Misc.zip_fields(current).each do |ntarget,nsign,nspecies,nsource,npmids|
+          if target == ntarget
+            if species != nspecies
+              nspecies = (nspecies.split(";") + species.split(";")).uniq * ";"
+            end
+            npmids = (npmids.split(";") + pmids.split(";")).uniq * ";"
+          end
+          new << [ntarget, nsign, nspecies, nsource, npmids]
+        end
+        values = Misc.zip_fields(new)
+        tsv[tf] = values
+      else
+        pmids = "" if pmids.nil?
+
+        values = [target, "", species, source, pmids]
+        tsv.zip_new(tf, values)
+      end
+
+    end
+
+    tsv.to_s
+  end
+
 end
 
 if defined? Entity and defined? Gene and Entity === Gene

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: rbbt-sources
 version: !ruby/object:Gem::Version
-  version: 3.1.14
+  version: 3.1.15
 platform: ruby
 authors:
 - Miguel Vazquez
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2017-03-06 00:00:00.000000000 Z
+date: 2017-04-18 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rbbt-util
@@ -181,19 +181,19 @@ signing_key:
 specification_version: 4
 summary: Data sources for the Ruby Bioinformatics Toolkit (rbbt)
 test_files:
-- test/test_helper.rb
 - test/rbbt/sources/test_pubmed.rb
-- test/rbbt/sources/test_string.rb
-- test/rbbt/sources/test_gscholar.rb
+- test/rbbt/sources/test_pharmagkb.rb
 - test/rbbt/sources/test_biomart.rb
-- test/rbbt/sources/test_go.rb
+- test/rbbt/sources/test_gscholar.rb
+- test/rbbt/sources/test_kegg.rb
 - test/rbbt/sources/test_pina.rb
 - test/rbbt/sources/test_entrez.rb
-- test/rbbt/sources/test_pharmagkb.rb
-- test/rbbt/sources/test_organism.rb
 - test/rbbt/sources/test_matador.rb
-- test/rbbt/sources/test_tfacts.rb
-- test/rbbt/sources/test_kegg.rb
+- test/rbbt/sources/test_HPRD.rb
 - test/rbbt/sources/test_synapse.rb
+- test/rbbt/sources/test_go.rb
 - test/rbbt/sources/test_stitch.rb
-- test/rbbt/sources/test_HPRD.rb
+- test/rbbt/sources/test_organism.rb
+- test/rbbt/sources/test_string.rb
+- test/rbbt/sources/test_tfacts.rb
+- test/test_helper.rb