entityextractor 0.0.5 → 0.0.6

data/lib/entityextractor.rb

@@ -1,4 +1,5 @@
 require 'json'
+load 'extractdates.rb'
 
 class EntityExtractor
   def initialize(input, *extractfield)
@@ -109,6 +110,14 @@ class EntityExtractor
         i["extract"] = addlist
         @output.push(i)
 
+      # Extract dates
+      elsif type == "date"
+        @extractfield.each do |f|
+          d = ExtractDates.new(i[f])
+          outhash = d.chunk(i["path"])
+          @output.push(outhash)
+        end
+
       # Extract both set terms and ALLCAPS
       elsif type == "both"
         @extractfield.each do |f|
@@ -123,3 +132,4 @@ class EntityExtractor
     end 
  end
 end
+

data/lib/extractdates.rb

@@ -0,0 +1,58 @@
+require 'treat'
+include Treat::Core::DSL
+require 'date'
+require 'json'
+require 'american_date'
+
+class ExtractDates
+  def initialize(text)
+    @text = text
+    @output = Array.new
+  end
+  
+  def chunk(file)
+    if @text
+      begin
+        c = @text.chunk
+        c.each do |i|
+          s = paragraph(i).segment
+          dateExtract(s)
+          s.each do |j|
+            shash = Hash.new
+            shash[:date] = dateExtract(j)
+            if shash[:date]
+              shash[:file] = file
+              shash[:title] = j
+              shash[:description] = i
+              @output.push(shash)
+            end
+          end
+        end
+      rescue
+      end
+    end
+    return @output
+  end
+
+  def dateExtract(blob)
+    # Date formats-
+      # mm/dd/yy
+      # mm/dd/yyyy
+      # Month dd, yyyy
+      # Month ddth, yyyy
+      # Month yyyy
+
+    # TOADD: 
+      # Multiple dates
+      # Year detection
+      # Conditional American dates
+      # Time ranges
+      # Filtering
+
+    begin
+      return DateTime.parse(blob).to_s
+    rescue
+    end
+  end
+end
+

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: entityextractor
 version: !ruby/object:Gem::Version
-  version: 0.0.5
+  version: 0.0.6
   prerelease: 
 platform: ruby
 authors:
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2014-03-01 00:00:00.000000000 Z
+date: 2014-03-16 00:00:00.000000000 Z
 dependencies: []
 description: Extracts entities and terms from any JSON.
 email: shidash@shidash.com
@@ -17,6 +17,7 @@ executables: []
 extensions: []
 extra_rdoc_files: []
 files:
+- lib/extractdates.rb
 - lib/entityextractor.rb
 homepage: https://github.com/Shidash/EntityExtractor
 licenses: