rbbt-sources 0.4.0 → 1.0.0

data/share/install/lib/helpers.rb

@@ -1,9 +1,9 @@
-require 'rbbt-util'
+require 'rbbt'
 require 'rbbt/sources/biomart'
 require 'rbbt/sources/entrez'
 
 def tsv_file(url, native, extra, options = {})
-  options = Misc.add_defaults options, :persistence => false, :keep_empty => true
+  options = Misc.add_defaults options, :persist => false, :keep_empty => true
 
   case
   when Array === native
@@ -27,7 +27,7 @@ def tsv_file(url, native, extra, options = {})
     fields = nil
   end
 
-  tsv = TSV.new(Open.open(url), options)
+  tsv = TSV.open(Open.open(url), options)
   tsv.key_field ||= key_field
   tsv.fields    ||= fields
   tsv
@@ -35,13 +35,14 @@ end
 
 def merge_entrez(data, taxs, native, fix = nil, select = nil)
   entrez =  Entrez.entrez2native(taxs, :fix => fix, :select => select)
+  entrez.key_field = "Entrez Gene ID"
   entrez.fields = [native]
   entrez
 
-  data.smart_merge entrez, native
+  data.attach entrez, "Entrez Gene ID"
 end
 
 def merge_biomart(lexicon, db, native, other, match = nil)
   match ||= native.first
-  lexicon.smart_merge BioMart.tsv(db, native, other), match
+  lexicon.attach BioMart.tsv(db, native, other)
 end

data/test/rbbt/sources/test_biomart.rb

@@ -1,4 +1,7 @@
-require File.dirname(__FILE__) + '/../../test_helper'
+require File.expand_path(File.dirname(__FILE__) + '/../../test_helper')
+
+require 'rbbt/sources/pubmed'
+require 'test/unit'
 require 'rbbt/sources/biomart'
 require 'rbbt/util/tmpfile'
 require 'test/unit'
@@ -11,11 +14,11 @@ class TestBioMart < Test::Unit::TestCase
     end
 
     data = BioMart.get('scerevisiae_gene_ensembl','entrezgene', ['protein_id'],[], nil, :nocache => false, :merge => true, :wget_options => {:quiet => false})
-    tsv = TSV.new data, :double, :merge => true
+    tsv = TSV.open data, :double, :merge => true
     assert(tsv['852236'][0].include? 'CAA84864')
 
     data = BioMart.get('scerevisiae_gene_ensembl','entrezgene', ['external_gene_id'],[], data, :nocache => false, :wget_options => { :quiet => false} )
-    tsv = TSV.new data, :double, :merge => true
+    tsv = TSV.open data, :double, :merge => true
     assert(tsv['852236'][1].include? 'YBL044W')
   end
 
@@ -25,7 +28,7 @@ class TestBioMart < Test::Unit::TestCase
 
     TmpFile.with_file do |f|
       filename = BioMart.query('scerevisiae_gene_ensembl','entrezgene', ['protein_id','refseq_peptide','external_gene_id','ensembl_gene_id'], [], nil, :nocache => false, :wget_options => { :quiet => false}, :filename => f)
-      data = TSV.new Open.open(filename)
+      data = TSV.open Open.open(filename)
       assert(data['852236']['external_gene_id'].include? 'YBL044W')
     end
   end
@@ -38,7 +41,7 @@ class TestBioMart < Test::Unit::TestCase
 
     TmpFile.with_file do |f|
       filename = BioMart.tsv('scerevisiae_gene_ensembl',['Entrez Gene', 'entrezgene'], [['Protein ID', 'protein_id'],['RefSeq Peptide','refseq_peptide']], [], nil, :nocache => false, :wget_options => { :quiet => false}, :filename => f)
-      data = TSV.new Open.open(filename, :merge => true)
+      data = TSV.open Open.open(filename, :merge => true)
       assert(data['852236']['Protein ID'].include? 'CAA84864')
       assert_equal 'Entrez Gene', data.key_field
       assert_equal ['Protein ID', 'RefSeq Peptide'], data.fields

data/test/rbbt/sources/test_organism.rb

@@ -1,24 +1,28 @@
 require File.expand_path(File.dirname(__FILE__) + '/../../test_helper')
 require 'rbbt/sources/organism'
-require 'rbbt/sources/organism/sequence'
 require 'test/unit'
 
 class TestEntrez < Test::Unit::TestCase
+  def test_location
+    assert_equal "share/organisms/Sce/identifiers", Organism.identifiers('Sce')
+  end
+
+
   def test_identifiers
-    assert Organism.identifiers('Sce').tsv['S000006120']["Ensembl Gene ID"].include?('YPL199C')
-    assert Organism::Sce.identifiers.tsv['S000006120']["Ensembl Gene ID"].include?('YPL199C')
-    assert Organism.identifiers('Hsa').tsv(:key => "Entrez Gene ID")['1020']["Associated Gene Name"].include?('CDK5')
+    assert Organism.identifiers('Hsa').tsv(:key_field => "Entrez Gene ID", :persist => true)['1020']["Associated Gene Name"].include?('CDK5')
+    assert Organism.identifiers('Sce').tsv(:persist => true)['S000006120']["Ensembl Gene ID"].include?('YPL199C')
+    assert Organism::Sce.identifiers.tsv(:persist => true)['S000006120']["Ensembl Gene ID"].include?('YPL199C')
   end
 
   def test_lexicon
-    assert TSV.new(Organism.lexicon('Sce'))['S000006120'].flatten.include?('YPL199C')
+    assert TSV.open(Organism.lexicon('Sce'))['S000006120'].flatten.include?('YPL199C')
   end
 
   def test_guess_id
     ensembl = %w(YOL044W YDR289C YAL034C YGR246C ARS519 tH(GUG)E2 YDR218C YLR002C YGL224C)
     gene_name = %w(SNR64 MIP1 MRPS18 TFB2 JEN1 IVY1 TRS33 GAS3)
-    assert_equal "Ensembl Gene ID", Organism::Sce.guess_id(ensembl).first
     assert_equal "Associated Gene Name", Organism::Sce.guess_id(gene_name).first
+    assert_equal "Ensembl Gene ID", Organism::Sce.guess_id(ensembl).first
   end
 
   def test_organisms
@@ -27,7 +31,7 @@ class TestEntrez < Test::Unit::TestCase
   end
 
   def test_attach_translations
-    tsv = TSV.new({"1020" => []}, :list)
+    tsv = TSV.setup({"1020" => []}, :type => :list)
     tsv.key_field = "Entrez Gene ID"
     tsv.fields = []
     tsv.namespace = "Hsa"
@@ -38,20 +42,20 @@ class TestEntrez < Test::Unit::TestCase
     assert_equal "CDK5", tsv["1020"]["Associated Gene Name"]
   end
 
-  def test_genes_at_chromosome
-    pos = [12, 117799500]
-    assert_equal "ENSG00000089250", Organism::Hsa.genes_at_chromosome_positions(pos.first, pos.last)
-  end
+  #def test_genes_at_chromosome
+  #  pos = [12, 117799500]
+  #  assert_equal "ENSG00000089250", Organism::Hsa.genes_at_chromosome_positions(pos.first, pos.last)
+  #end
 
-  def test_genes_at_chromosome_array
-    pos = [12, [117799500, 106903900]]
-    assert_equal ["ENSG00000089250", "ENSG00000013503"], Organism::Hsa.genes_at_chromosome_positions(pos.first, pos.last)
-  end
+  #def test_genes_at_chromosome_array
+  #  pos = [12, [117799500, 106903900]]
+  #  assert_equal ["ENSG00000089250", "ENSG00000013503"], Organism::Hsa.genes_at_chromosome_positions(pos.first, pos.last)
+  #end
  
-  def test_genes_at_genomic_positions
-    pos = [[12, 117799500], [12, 106903900], [1, 115259500]]
-    assert_equal ["ENSG00000089250", "ENSG00000013503", "ENSG00000213281"], Organism::Hsa.genes_at_genomic_positions(pos)
-  end
+  #def test_genes_at_genomic_positions
+  #  pos = [[12, 117799500], [12, 106903900], [1, 115259500]]
+  #  assert_equal ["ENSG00000089250", "ENSG00000013503", "ENSG00000213281"], Organism::Hsa.genes_at_genomic_positions(pos)
+  #end
 
 end
 

metadata

@@ -1,13 +1,13 @@
 --- !ruby/object:Gem::Specification 
 name: rbbt-sources
 version: !ruby/object:Gem::Version 
-  hash: 15
-  prerelease: 
+  hash: 23
+  prerelease: false
   segments: 
+  - 1
   - 0
-  - 4
   - 0
-  version: 0.4.0
+  version: 1.0.0
 platform: ruby
 authors: 
 - Miguel Vazquez
@@ -15,7 +15,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2011-03-23 00:00:00 +01:00
+date: 2011-09-07 00:00:00 +02:00
 default_executable: 
 dependencies: 
 - !ruby/object:Gem::Dependency 
@@ -26,10 +26,12 @@ dependencies:
     requirements: 
     - - ">="
       - !ruby/object:Gem::Version 
-        hash: 3
+        hash: 63
         segments: 
+        - 4
         - 0
-        version: "0"
+        - 0
+        version: 4.0.0
   type: :runtime
   version_requirements: *id001
 - !ruby/object:Gem::Dependency 
@@ -74,6 +76,20 @@ dependencies:
         version: "0"
   type: :runtime
   version_requirements: *id004
+- !ruby/object:Gem::Dependency 
+  name: bio
+  prerelease: false
+  requirement: &id005 !ruby/object:Gem::Requirement 
+    none: false
+    requirements: 
+    - - ">="
+      - !ruby/object:Gem::Version 
+        hash: 3
+        segments: 
+        - 0
+        version: "0"
+  type: :runtime
+  version_requirements: *id005
 description: Data sources like PubMed, Entrez Gene, or Gene Ontology
 email: miguel.vazquez@fdi.ucm.es
 executables: []
@@ -83,28 +99,37 @@ extensions: []
 extra_rdoc_files: []
 
 files: 
+- etc/biomart/missing_in_archive
+- lib/rbbt/sources/COSMIC.rb
 - lib/rbbt/sources/COSTART.rb
 - lib/rbbt/sources/CTCAE.rb
+- lib/rbbt/sources/InterPro.rb
+- lib/rbbt/sources/NCI.rb
 - lib/rbbt/sources/bibtex.rb
 - lib/rbbt/sources/biomart.rb
 - lib/rbbt/sources/entrez.rb
 - lib/rbbt/sources/go.rb
 - lib/rbbt/sources/gscholar.rb
+- lib/rbbt/sources/jochem.rb
 - lib/rbbt/sources/organism.rb
 - lib/rbbt/sources/organism/sequence.rb
 - lib/rbbt/sources/polysearch.rb
 - lib/rbbt/sources/pubmed.rb
+- lib/rbbt/sources/wgEncodeBroadHmm.rb
+- share/install/InterPro/Rakefile
+- share/install/JoChem/Rakefile
+- share/install/NCI/Rakefile
 - share/install/Organism/Hsa/Rakefile
 - share/install/Organism/Rno/Rakefile
 - share/install/Organism/Sce/Rakefile
 - share/install/Organism/organism_helpers.rb
 - share/install/lib/helpers.rb
-- test/rbbt/sources/test_biomart.rb
+- test/test_helper.rb
 - test/rbbt/sources/test_entrez.rb
+- test/rbbt/sources/test_pubmed.rb
 - test/rbbt/sources/test_go.rb
+- test/rbbt/sources/test_biomart.rb
 - test/rbbt/sources/test_organism.rb
-- test/rbbt/sources/test_pubmed.rb
-- test/test_helper.rb
 has_rdoc: true
 homepage: http://github.com/mikisvaz/rbbt-sources
 licenses: []
@@ -135,14 +160,14 @@ required_rubygems_version: !ruby/object:Gem::Requirement
 requirements: []
 
 rubyforge_project: 
-rubygems_version: 1.4.2
+rubygems_version: 1.3.7
 signing_key: 
 specification_version: 3
 summary: Data sources for the Ruby Bioinformatics Toolkit (rbbt)
 test_files: 
-- test/rbbt/sources/test_biomart.rb
+- test/test_helper.rb
 - test/rbbt/sources/test_entrez.rb
+- test/rbbt/sources/test_pubmed.rb
 - test/rbbt/sources/test_go.rb
+- test/rbbt/sources/test_biomart.rb
 - test/rbbt/sources/test_organism.rb
-- test/rbbt/sources/test_pubmed.rb
-- test/test_helper.rb