rbbt-text 1.3.7 → 1.3.8

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 8dfc374254fcbe88c8be6bfffd9a3cfabf6e23c953c11ecd2f61cf41027ff3d6
-  data.tar.gz: 3d3211f41cfecea05862505d1508a4b7b76eecb3c90b3b0000194eb08033715e
+  metadata.gz: e1b07b2646ecdc51599e2a2356fd18708e88d819944910a07930f67ec3fc012d
+  data.tar.gz: 03bcbe61f41d830668b50fcfc253fa2b43285774040f61fb3fb0a58f80e9dfd3
 SHA512:
-  metadata.gz: 7ed870e46bae2c113d0885697bfbade6064732a89477833c640eaf4ee8bdb2c0fbf52f69f456af5eb30a82e56a7f0aeb37e71127f884430c3d315202a07fa3cb
-  data.tar.gz: e31853e816321a5ead788036b5f67eecaca179c75168c0bb2804be1f18ae844031ab808a4e3c9d67e1f9a52f94ca478949798b8101e164eba32481c0182a1f58
+  metadata.gz: ae6de2dd809642ca38276ff82e243efeb193cf432bc78aea92e772ab21ff489f23224b9e93de726dcacdb06910716f1107171433cc39e7b022ba14ee4ed284f6
+  data.tar.gz: 82768060a28248d459031030b6ba49b500b63a9d3ae2199ccdf1417fd3b1f66ce0d962db17875615ee36bb3b5879d8ccbbdec892942f544fa08481b4551a1003

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

@@ -4,10 +4,11 @@ module Document::Corpus
   PUBMED_NAMESPACE="PMID"
   def add_pmid(pmid, type = nil, update = false)
     type = :abstract if type.nil?
-    if update == false
+
+    if ! (update || Array === pmid)
       id = [PUBMED_NAMESPACE, pmid, type].collect{|e| e.to_s}*":"
       documents = self.documents(id)
-      return documents if documents.any?
+      return documents.first if documents.any?
     end
 
     pmids = Array === pmid ? pmid : [pmid]
@@ -27,7 +28,14 @@ module Document::Corpus
       document
     end
 
-    Document.setup(res)
+    if Array === pmid
+      corpus = res.first.corpus if res.first
+      Document.setup(res, :corpus => corpus)
+    else
+      res = res.first
+    end
+
+    res
   end
   
   def add_pubmed_query(query, max = 3000, type = nil)
@@ -35,8 +43,8 @@ module Document::Corpus
     add_pmid(pmids, type)
   end
 
-  self.claim "PMID" do |id, type|
+  self.claim "PMID" do |id,type,update|
     Log.debug "Claiming #{id}"
-    self.add_pmid(id, type).first
+    self.add_pmid(id, type,update)
   end
 end

data/lib/rbbt/nlp/spaCy.rb

@@ -153,8 +153,8 @@ module SpaCy
 
     chunk_index = Segment.index(SpaCy.chunk_segments(text, lang))
 
-    source_id = chunk_index[source.offset].first || source.segid
-    target_id = chunk_index[target.offset].first || target.segid
+    source_id = chunk_index[source.offset.to_i].first || source.segid
+    target_id = chunk_index[target.offset.to_i].first || target.segid
 
     path = Paths.dijkstra(graph, source_id, [target_id])
 

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

@@ -7,7 +7,7 @@ class TestCorpusPubmed < Test::Unit::TestCase
   def test_add_pmid
     corpus = Document::Corpus.setup({})
 
-    document = corpus.add_pmid("33359141", :abstract).first
+    document = corpus.add_pmid("33359141", :abstract, true)
     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: 1.3.7
+  version: 1.3.8
 platform: ruby
 authors:
 - Miguel Vazquez
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2022-06-21 00:00:00.000000000 Z
+date: 2022-11-08 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rbbt-util