rbbt-text 1.3.3 → 1.3.4

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: d25c6d473e1ee0a8ba79af357571181539b6e18e6b8d11e85fcca037069be3bf
-  data.tar.gz: dbc3621f7fbc0ab5569b9f98a527c20cbc4192c6db211504a904364452518caf
+  metadata.gz: 496288d7d3ff1215ded1fd210192d5887a6a071eea5f322295a669a5d648d77b
+  data.tar.gz: 47996496009cbcdaab38a9dc9bf6efbbe7fc0145f315b0a48bfab0f543742f94
 SHA512:
-  metadata.gz: 7a6568d91518fa0c4aedd748fe2b7c2db745a2997efb03c00993ebc24f6682422d209aa266912bcaca32c2033b6babbf9b14db2bf39973b4a33a69fa9ed07eca
-  data.tar.gz: 2c87eeccbb22e90c87611024429918e4a3fbdcf8212d6b6538e063d6e0116a457a6c855bafad4ac4621bfb7ae91ff18d2cea0cbcf56dfdff79c2ec88666cbf18
+  metadata.gz: 36e7415ad06207066844a30001c8541865f066d1e83a4a2ddc5182c54b704cd3d442cbccce219bd2114717a83656d07558c42725eca75597fea239b6e13244ab
+  data.tar.gz: 988eff4d242d0425910b96fac4188df079c8c53c3abea2825cc97d5af5118841680705fa33461a5b4cfa7b8d6b32a486465e44b75f20fad324e4623c6c8083d8

data/lib/rbbt/document.rb

@@ -24,7 +24,6 @@ module DocID
       docs = self.collect do |docid|
         text = self.corpus[docid]
         namespace, id, type = docid.split(":")
-        #Document.setup(text, namespace, id, type, :corpus => corpus)
         text
       end
       Document.setup(docs, :corpus => corpus)

data/lib/rbbt/document/corpus.rb

@@ -10,16 +10,16 @@ module Document::Corpus
 
   def add_document(document)
     docid = document.docid
-    return document if self.include?(docid)
+    return self[docid] if self.include?(docid)
     self.write_and_close do
       self[docid] = document
     end
   end
   
   def docids(prefix)
-    prefix += ":" unless prefix[-1] == ":"
+    prefix += ":" unless prefix == :all || prefix[-1] == ":"
     docids = self.read_and_close do
-      self.prefix(prefix)
+      prefix == :all ? self.keys : self.prefix(prefix)
     end
     DocID.setup(docids, :corpus => self)
   end
@@ -35,6 +35,7 @@ module Document::Corpus
       super(*args)
     end
     
+    res.force_encoding(Encoding.default_external) if res 
     return res if args.length > 1
 
     namespace, id, type  = docid.split(":")
@@ -45,6 +46,7 @@ module Document::Corpus
       end
     end
 
+    res.force_encoding(Encoding.default_external) if res 
     Document.setup(res, namespace, id, type, self) unless res.nil?
     
     res

data/lib/rbbt/segment/named_entity.rb

@@ -19,11 +19,14 @@ Score: #{score.inspect}
   end
 
   def html
+    title = code.nil? ? entity_type : [entity_type, code].compact * ":"
+
     text = <<-EOF
 <span class='Entity'\
 #{entity_type.nil? ? "" : " attr-entity-type='#{Array === entity_type ? entity_type * " " : entity_type}'"}\
 #{code.nil?  ? "" : " attr-entity-code='#{Array === code ? code * " " : code}'"}\
 #{score.nil? ? "" : " attr-entity-score='#{Array === score ? score * " " : score}'"}\
+#{title.nil? ? "" : " title='#{Array === title ? title * " " : title}'"}\
 >#{ self }</span>
     EOF
     text.chomp

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: rbbt-text
 version: !ruby/object:Gem::Version
-  version: 1.3.3
+  version: 1.3.4
 platform: ruby
 authors:
 - Miguel Vazquez
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2020-06-17 00:00:00.000000000 Z
+date: 2020-07-23 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rbbt-util