rbbt-text 1.3.5 → 1.3.6

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: '0846f900d745dd27df8006eecbc9d294f9f38a23dd76001de2a5dc0313db7e22'
-  data.tar.gz: 675985882a6c8b9813f620d7ef0a555efa5c148c7c2fe36e0030f84f3fd88cf0
+  metadata.gz: 110d05d205305e48b17ca63a91f8fef66ba6cb0d24639031614521c463dfde8f
+  data.tar.gz: 97df4097eb1a0cb645f4f6a0eeec9138f296760e09fcdc85a5ebebe6996640d0
 SHA512:
-  metadata.gz: dfd9c333b94181496134b825c63d6e93a0390f81d426526f79c00cf12556021b60004b29b57ca9b0b274141937027f7bc780552a60de007e5f790b19910354c0
-  data.tar.gz: 205beeb8829c8358fd29c0a18351522e566106e24220af3d7bec3676694d37d682b92243e4fd4cd495b542f9945a28cf8585e587342672d31779d0b21b53ae4e
+  metadata.gz: bc3c92ae7b5268f02b232f04d28a2b6491740a618c9a9f3d78a099d51ed03dee489ebb805485c82552fc38d8164a4eb75a8e19a6d1df53aaac536e99287009fb
+  data.tar.gz: 1bb1594b9cd5831455e1a7e2454932b57fdd7da4401bc2596146dc1669fa6783931df3042c9deed6a3cf032c59d8697500a470a8aeba859061868cd8ab8b4ab8

data/lib/rbbt/document/corpus.rb

@@ -6,6 +6,7 @@ module Document::Corpus
     corpus = Persist.open_tokyocabinet(corpus, true, :single, "BDB") if String === corpus
     corpus.extend Document::Corpus unless Document::Corpus === corpus
     corpus.extend Persist::TSVAdapter unless Persist::TSVAdapter === corpus
+    corpus.close
     corpus
   end
 

data/lib/rbbt/document/corpus/pubmed.rb

@@ -1,18 +1,26 @@
 require 'rbbt/sources/pubmed'
 
 module Document::Corpus
-  def add_pmid(pmid, type = nil)
+  PUBMED_NAMESPACE="PMID"
+  def add_pmid(pmid, type = nil, update = false)
+    type = :abstract if type.nil?
+    if update == false
+      id = [PUBMED_NAMESPACE, pmid, type].collect{|e| e.to_s}*":"
+      documents = self.documents(id)
+      return documents if documents.any?
+    end
+
     pmids = Array === pmid ? pmid : [pmid]
     type = nil if String === type and type.empty?
 
     res = PubMed.get_article(pmids).collect do |pmid, article|
-      document = if type.nil? || type.to_sym == :abstract
-                   Document.setup(article.abstract || "", "PMID", pmid, :abstract, self, :corpus => self)
+      document = if type.to_sym == :abstract
+                   Document.setup(article.abstract || "", PUBMED_NAMESPACE, pmid, :abstract, self, :corpus => self)
                  elsif type.to_sym == :title
-                   Document.setup(article.title, :PMID, pmid, :title, self)
+                   Document.setup(article.title, PUBMED_NAMESPACE, pmid, :title, self)
                  else
                    raise "No FullText available for #{ pmid }" if article.full_text.nil?
-                   Document.setup(article.full_text, :PMID, pmid, :fulltext, self, :corpus => self)
+                   Document.setup(article.full_text, PUBMED_NAMESPACE, pmid, :fulltext, self, :corpus => self)
                  end
       Log.debug "Loading pmid #{pmid}"
       add_document(document)

data/lib/rbbt/nlp/open_nlp/sentence_splitter.rb

@@ -7,7 +7,7 @@ module OpenNLP
   Rbbt.claim Rbbt.software.opt.OpenNLP, :install, Rbbt.share.install.software.OpenNLP.find
 
 
-  Rbbt.claim Rbbt.software.opt.OpenNLP.models["da-sent.bin"], :url, "http://opennlp.sourceforge.net/models-1.5/de-sent.bin"
+  Rbbt.claim Rbbt.software.opt.OpenNLP.models["da-sent.bin"], :url, "https://www.apache.org/dyn/closer.cgi/opennlp/models/ud-models-1.0/opennlp-en-ud-ewt-sentence-1.0-1.9.3.bin"
 
   MAX = 5
 

data/share/install/software/OpenNLP

@@ -1,12 +1,6 @@
 #!/bin/bash
 
 name='OpenNLP'
-url="http://apache.rediris.es/opennlp/opennlp-1.9.2/apache-opennlp-1.9.2-bin.tar.gz"
+url="http://apache.rediris.es/opennlp/opennlp-1.9.3/apache-opennlp-1.9.3-bin.tar.gz"
 
-get_src "$name" "$url"
-move_opt "$name" 
-
-
-ln -sf  "$OPT_DIR/$name/lib/"*.jar  "$OPT_JAR_DIR/"
-
-clean_build
+install_jar $name $url

data/test/rbbt/document/corpus/test_pubmed.rb

@@ -8,7 +8,6 @@ class TestCorpusPubmed < Test::Unit::TestCase
     corpus = Document::Corpus.setup({})
 
     document = corpus.add_pmid("33359141", :abstract).first
-    iii document.docid
     title = document.to(:title)
     assert title.include?("COVID-19")
   end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: rbbt-text
 version: !ruby/object:Gem::Version
-  version: 1.3.5
+  version: 1.3.6
 platform: ruby
 authors:
 - Miguel Vazquez
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2021-06-17 00:00:00.000000000 Z
+date: 2021-06-25 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rbbt-util