entityextractor 0.0.10 → 0.0.11

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 0f99308791264c4e567c495fe9964e68ad932ac5
-  data.tar.gz: ec717ca535969ad20c646bb4559763c69a973aee
+  metadata.gz: 6d73c2f909fbe862625da5f0952032715d696de2
+  data.tar.gz: c33ee22a6a5cbb3eb37f649c6db2df26df366c75
 SHA512:
-  metadata.gz: 2d7acd5e1645f5de5c1b52b2fb3cd14eb739c8444f962c34d54e914d1161aa1cfa0e17ba3cb10afdeb7100b9cbbb4d1a266b6078f2ebdde125d72a897a71e78c
-  data.tar.gz: 97cfa5c16df408a6dfd23b4a51bda99ef50caa997042ffc4370a15549b01106eff089889a430284377833e68edb608091079bd40c0129dbc36e08ddaf394e554
+  metadata.gz: f98e67a81133a74c666d474b5d22bd0018e5d941ec7bc5f875a319ea1ee3677d93ff3917591f23fdbfdb0c0e2cc14c4f866c4b1289aa8a908acf83f6f8c07791
+  data.tar.gz: a5a40b53bcb12550bf096cea1b4f72f63be2b254d079df89c322619041f67dfb51f3d9da9d479f422c058acbe4257e6b17dd1e8eb0834d08ba2874d4cf31c2bd

data/lib/entityextractor.rb

@@ -1,24 +1,32 @@
 require 'json'
 load 'extractdates.rb'
+load 'handleinput.rb'
 require 'uploadconvert'
 
 class EntityExtractor
-  def initialize(input, *extractfield)
+  def initialize(input, fieldoutname, *extractfield)
     @input = JSON.parse(input)
+    @fieldoutname = fieldoutname
     @extractfield = *extractfield
     @output = Array.new
   end
 
   # Extract terms input from preset list
-  def extractTerms(*terms, i, addlist, field)
+  def extractTerms(extractlist, i, addlist, field)
     count = 0
+    downcased = i[field].to_s.downcase
       
     # Check the item for each term
-    terms.each do |t|
+    extractlist.each do |t, c|
       count+=1
-   
-      if i[field].to_s.include? t
-        addlist.push(t.upcase)
+      if c == true
+        if i[field].to_s.include? t
+          addlist.push(t)
+        end
+      else
+        if downcased.include? t.downcase
+          addlist.push(t)
+        end
       end
     end
   end
@@ -70,7 +78,7 @@ class EntityExtractor
     
     # Generate hash of all extracted terms
     @output.each do |i|
-      i["extract"].each do |e|
+      i[@fieldoutname].each do |e|
         if extracthash.has_key? e
           extracthash[e] += 1
         else
@@ -88,8 +96,12 @@ class EntityExtractor
     JSON.pretty_generate(@output)
   end
 
-  def extract(type, minchar, ignoreterms, *terms)
+  def extract(type, minchar, ignoreterms, terms, ignorefields, caseinfo, mapto)
     flag = 0
+   
+    h = HandleInput.new(terms, ignorefields, caseinfo)
+    extractlist = h.detecttype
+
     @input.each do |i|
       if i.length == 2
         i = @input
@@ -101,10 +113,14 @@ class EntityExtractor
       # Generate set terms list
       if type == "set"
         @extractfield.each do |f|
-          extractTerms(*terms, i, addlist, f)
+          extractTerms(extractlist, i, addlist, f)
+        end
+        
+        if mapto
+          i[@fieldoutname] = h.mapout(addlist, mapto)
+        else
+          i[@fieldoutname] = addlist
         end
-
-        i["extract"] = addlist
         @output.push(i)
       
       # Generate ALLCAPS terms list
@@ -114,7 +130,7 @@ class EntityExtractor
           parseALLCAPS(i[f].to_s, i, minchar, addlist, ignoreterms, savefield, f)
         end
         
-        i["extract"] = addlist
+        i[@fieldoutname] = addlist
         @output.push(i)
 
       # Extract dates
@@ -130,10 +146,15 @@ class EntityExtractor
         @extractfield.each do |f|
           savefield = i[f].to_s + " "
           parseALLCAPS(i[f].to_s, i, minchar, addlist, ignoreterms, savefield, f)
-          extractTerms(*terms, i, addlist, f)
+          extractTerms(extractlist, i, addlist, f)
+        end
+
+        if mapto
+          i[@fieldoutname] = h.mapout(addlist, mapto)
+        else
+          i[@fieldoutname] = addlist
         end
 
-        i["extract"] = addlist
         @output.push(i)
       end
 
@@ -143,3 +164,4 @@ class EntityExtractor
     end 
  end
 end
+

data/lib/handleinput.rb

@@ -0,0 +1,138 @@
+require 'json'
+
+class HandleInput
+  def initialize(json, ignorefields, caseinfo)
+    @json = json
+    if @ignorefields != nil
+      @ignorefields = ignorefields
+    else
+      @ignorefields = Array.new
+    end
+    
+    @caseinfo = caseinfo
+    @output = Array.new
+    @outhash = Hash.new
+  end
+
+  # Map output to value
+  def mapout(addlist, mapto)
+    outarr = Array.new
+    
+    addlist.each do |a|
+      if mapto == "key"
+        @json.each do |k, v|
+          # If it's a nested hash
+          if v.is_a? Hash
+            # Go through all values
+            v.each do |z, w|
+              # Check if k is already included
+              if !outarr.include? k
+                if w == a
+                  outarr.push(k)
+                end
+              end
+            end
+          else
+            # Map for dictionaries
+            if !outarr.include? k
+              if v == a || k == a
+                outarr.push(k)
+              end
+            end
+          end
+        end
+      else
+        @json.each do |k, v|
+          v.each do |z, w|
+            # Only map if not already matched
+            if !outarr.include? v[mapto]
+              # Check if vals match
+              if w == a
+                outarr.push(v[mapto])
+              end
+            end
+          end
+        end
+      end
+    end
+
+    return outarr
+  end
+
+  # Figure out which type of input it is: array, hash, hash with hash values
+  def detecttype
+    if @json.is_a? Array
+      @output = @json
+      checkCase
+    elsif @json.is_a? Hash
+      @json.each do |k, v|
+        if v.is_a? Hash
+          parseValHash
+          break
+        else
+          parseDictionary
+          break
+        end
+      end
+    end
+
+    return @outhash
+  end
+
+  # Adds case sensitive preferences
+  def checkCase
+    if @caseinfo == "casesensitive"
+      @output.each do |i|
+        @outhash[i] = true
+      end
+    elsif @caseinfo == "noncasesensitive"
+      @output.each do |i|
+        @outhash[i] = false
+      end
+    end
+  end
+
+  # Handle hashes where the values are a hash
+  def parseValHash
+    @json.each do |k, v|
+      if !@ignorefields.include? "hashkey"
+        if @caseinfo.include? "hashkey"
+          @outhash[k] = false
+        else
+          @outhash[k] = true
+        end
+      end
+
+      v.each do |i, j|
+        if !@ignorefields.include? i
+          if @caseinfo.include? i
+            @outhash[j] = false
+          else
+            @outhash[j] = true
+          end
+        end
+      end
+    end
+  end
+
+  # Handle hashes
+  def parseDictionary
+    @json.each do |k, v|
+      if !@ignorefields.include? "hashkey"
+        if @caseinfo.include? "hashkey"
+          @outhash[k] = false
+        else
+          @outhash[k] = true
+        end
+      end
+
+      if !@ignorefields.include? "hashval"
+        if @caseinfo.include? "hashval"
+          @outhash[v] = false
+        else
+          @outhash[v] = true
+        end
+      end
+    end
+  end
+end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: entityextractor
 version: !ruby/object:Gem::Version
-  version: 0.0.10
+  version: 0.0.11
 platform: ruby
 authors:
 - M. C. McGrath
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2014-04-11 00:00:00.000000000 Z
+date: 2014-08-11 00:00:00.000000000 Z
 dependencies: []
 description: Extracts entities and terms from any JSON.
 email: shidash@shidash.com
@@ -17,6 +17,7 @@ extensions: []
 extra_rdoc_files: []
 files:
 - lib/extractdates.rb
+- lib/handleinput.rb
 - lib/entityextractor.rb
 homepage: https://github.com/Shidash/EntityExtractor
 licenses:
@@ -43,3 +44,4 @@ signing_key:
 specification_version: 4
 summary: Extracts entities and terms
 test_files: []
+has_rdoc: