pdf-extract 0.0.3 → 0.0.4

data/lib/analysis/sections.rb

@@ -133,9 +133,10 @@ module PdfExtract
           # Score sections into categories based on their textual attributes.
           ideals = {
             :reference => {
-              :name_ratio => [0.2, 5],
+              :name_ratio => [0.2, 2],
               :letter_ratio => [0.25, 2],
-              :year_ratio => [0.05, 7]
+              :year_ratio => [0.05, 7],
+              :cap_ratio => [0.5, 5]
             },
             :body => {
               :name_ratio => [0.03, 1],

data/lib/font_metrics.rb

@@ -26,6 +26,10 @@ module PdfExtract
     attr_accessor :ascent, :descent, :bbox
 
     def initialize font
+      @ascent = 0
+      @descent = 0
+      @bbox = [0, 0, 0, 0]
+      
       base_font = font.basefont.to_s
       if @@base_fonts.key? base_font
         @ascent = @@base_fonts[base_font][:Ascent]
@@ -41,8 +45,10 @@ module PdfExtract
         @glyph_width_lookup = proc { |c| font.glyph_width c }
       end
 
-      @ascent = @bbox[3] if @ascent.zero?
-      @descent = @bbox[1] if @descent.zero?
+      if not @bbox.nil?
+        @ascent = @bbox[3] if @ascent.nil? || @ascent.zero?
+        @descent = @bbox[1] if @descent.nil? || @descent.zero?
+      end
     end
 
     def glyph_width c

data/lib/names.rb

@@ -7,7 +7,7 @@ require_relative "pdf-extract"
 module PdfExtract::Names
 
   class NamesDatabase
-    @@ambiguous_weighting = 0.1
+    @@ambiguous_weighting = 0.0
     @@unambiguous_weighting = 1.0
 
     def self.path_to_data data_filename

data/lib/references/references.rb

@@ -3,7 +3,7 @@ require_relative "../spatial"
 module PdfExtract
   module References
  
-    Settings.default :min_score, 6.4
+    Settings.default :min_score, 8.0
     Settings.default :min_sequence_count, 3
     Settings.default :max_reference_order, 1000
 

data/lib/spatial.rb

@@ -165,9 +165,15 @@ module PdfExtract
           items.each do |item|
             diff = (item[var_name] - ideals[name][var_name][0]).abs
             if diff.zero?
-              diff = Float::MIN
+              scores << 1.0
+            else
+              s = 1.0 / diff
+              if not s.finite?
+                scores << 0.0
+              else
+                scores << s
+              end
             end
-            scores << 1.0 / diff
           end
 
           score_max = scores.max

metadata

@@ -5,8 +5,8 @@ version: !ruby/object:Gem::Version
   segments: 
   - 0
   - 0
-  - 3
-  version: 0.0.3
+  - 4
+  version: 0.0.4
 platform: ruby
 authors: 
 - Karl Jonathan Ward