rbbt-sources 3.1.43 → 3.1.45

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: cb929c25537c93976e03cac4193b7bcbde8c6b8fa1975506779290cc9579d2cb
-  data.tar.gz: d316a40f9ac9c00997dbecf35246290ee5c87704bf42fab5b687b1e595f803c3
+  metadata.gz: ecb6a979911fd4e4a4988ab9abee186cb181eab335b41505d71e3405f1cac454
+  data.tar.gz: 5b6ba6e080cff6d6bd864dc468f418b1d74d9ee6fe916df0b9edc2a8ba2e5471
 SHA512:
-  metadata.gz: 93b9260a58ed0bfdfa2be1d63160fae4a70a71df3f9c8510e42b99ac740d2c90b1433c8b2aae28e2978ddccfb8abb822bc106341bf8dcef31092db768b5caf19
-  data.tar.gz: 7931ac20e016779659aca2bf6966abdefe21eec7dc9ca6bf843ddd09004f6fda3c022a74ed48e6d269d2b702aa4bf77cfb22ae760ea65d80db916efeb6998ebd
+  metadata.gz: 72cc59fa3439ffbb4b850723c4b6c5403f65301a70ddf61f6672e52646ee1e90644ac125234cff1a51702a100d7ea2731940cfb33a49d0c3531b5fd8877d1d24
+  data.tar.gz: fdf336019c660cf2fb0eea5af696edae0dbb3d31adf7c063060626c44d2ecf6864f1c0dd614479a1166583ad575cb288ce50e4291f90781f26d06568d296c3e7

data/lib/rbbt/sources/entrez.rb

@@ -14,7 +14,7 @@ module Entrez
     options = Misc.add_defaults options, :key_field => 1, :fields => [5], :persist => true, :merge => true
 
     taxs = [taxs] unless Array === taxs
-    options.merge! :grep => taxs.collect{|t| "^" + t.to_s}
+    options.merge! :grep => taxs.collect{|t| "^" + t.to_s}, :fixed_grep => false
     
     tsv = Rbbt.share.databases.entrez.gene_info.tsv :flat, options
     tsv.key_field = "Entrez Gene ID"
@@ -26,7 +26,7 @@ module Entrez
     options = Misc.add_defaults options, :key_field => 1, :fields => [2], :persist => true, :merge => true
 
     taxs = [taxs] unless Array === taxs
-    options.merge! :grep => taxs.collect{|t| "^" + t.to_s}
+    options.merge! :grep => taxs.collect{|t| "^" + t.to_s}, :fixed_grep => false
     
     tsv = Rbbt.share.databases.entrez.gene_info.tsv :flat, options
     tsv.key_field = "Entrez Gene ID"
@@ -39,7 +39,7 @@ module Entrez
     options = {:key_field => 1, :fields => [2], :persist => true, :merge => true}
 
     taxs = [taxs] unless taxs.is_a?(Array)
-    options.merge! :grep => taxs.collect{|t| "^" + t.to_s}
+    options.merge! :grep => taxs.collect{|t| "^" + t.to_s}, :fixed_grep => false
 
     Rbbt.share.databases.entrez.gene2pubmed.tsv :flat, options
   end

data/lib/rbbt/sources/go.rb

@@ -27,7 +27,7 @@ module GO
   def self.init
     Persist.persist_tsv(nil, 'gene_ontology', {}, :persist => true) do |info|
       info.serializer = :marshal if info.respond_to? :serializer
-      Rbbt.share.databases.GO.gene_ontology.read.split(/\[Term\]/).each{|term| 
+      Rbbt.share.databases.GO.gene_ontology.produce.read.split(/\[Term\]/).each{|term| 
         term_info = {}
 
         term.split(/\n/). select{|l| l =~ /:/}.each{|l| 

data/lib/rbbt/sources/pubmed.rb

@@ -109,6 +109,14 @@ module PubMed
       end
     end
 
+    def pmc_full_xml
+      begin
+        Open.read("https://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi?db=pmc&id=#{pmid}")
+      rescue
+        nil
+      end
+    end
+
     def pdf_url
       return pmc_pdf if pmc_pdf
       @gscholar_pdf ||= begin
@@ -121,18 +129,22 @@ module PubMed
     end
 
     def full_text
-      return nil if pdf_url.nil?
-
-      text = nil
-      TmpFile.with_file do |pdf|
-
-        # Change user-agent, oh well...
-        `wget --user-agent=firefox #{ pdf_url } -O #{ pdf } -t 3`
-        TmpFile.with_file do |txt|
-          `pdftotext #{ pdf } #{ txt }`
-          text = Open.read(txt) if File.exists? txt
-        end
-      end
+      text = if pdf_url
+               text = nil
+               TmpFile.with_file do |pdf|
+                 # Change user-agent, oh well...
+                 `wget --user-agent=firefox #{ pdf_url } -O #{ pdf } -t 3`
+                 TmpFile.with_file do |txt|
+                   `pdftotext #{ pdf } #{ txt }`
+                   text = Open.read(txt) if File.exists? txt
+                 end
+               end
+               text
+             elsif pmc_full_xml
+               pmc_full_xml
+             else
+               nil
+             end
 
       Misc.fixutf8(text)
     end

data/test/rbbt/sources/test_entrez.rb

@@ -17,7 +17,10 @@ class TestEntrez < Test::Unit::TestCase
   def test_entrez2pubmed
     tax   = $yeast_tax
 
+    Log.severity = 0
     data = Entrez.entrez2pubmed(tax)
+    data.read
+    Log.tsv data
     assert(data['850320'].include? '1574125') 
   end
 

data/test/rbbt/sources/test_pubmed.rb

@@ -17,6 +17,12 @@ class TestPubMed < Test::Unit::TestCase
     pmid = '16438716'
     assert(PubMed.get_article(pmid).full_text =~ /Discovering/)
   end
+
+  def test_pmc_full_xml
+    pmid = '4304705'
+    assert PubMed.get_article(pmid).pmc_full_xml.include?("HBV antigen")
+  end
+ 
  
   def test_query
     assert(PubMed.query('chagoyen[All Fields] AND ("loattrfull text"[sb] AND hasabstract[text])').include? '16438716')

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: rbbt-sources
 version: !ruby/object:Gem::Version
-  version: 3.1.43
+  version: 3.1.45
 platform: ruby
 authors:
 - Miguel Vazquez
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2020-04-01 00:00:00.000000000 Z
+date: 2020-04-13 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rbbt-util