Merge pull request #456 from tabulapdf/new-detection-algorithm

Implement new table detector
tabulapdf · Feb 17, 2016 · 6c1856b · 6c1856b
2 parents 9fcb8fd + 7de1120
commit 6c1856b
Show file tree

Hide file tree

Showing 3 changed files with 3 additions and 4 deletions.
diff --git a/...-0.7.4-SNAPSHOT-jar-with-dependencies.jar → ...rs/tabula-0.8.0-jar-with-dependencies.jar b/...-0.7.4-SNAPSHOT-jar-with-dependencies.jar → ...rs/tabula-0.8.0-jar-with-dependencies.jar
diff --git a/lib/tabula_job_executor/jobs/detect_tables.rb b/lib/tabula_job_executor/jobs/detect_tables.rb
@@ -14,15 +14,14 @@ def perform
     begin
       extractor = Tabula::Extraction::ObjectExtractor.new(filepath, :all)
       page_count = extractor.page_count
-      sea = Java::TechnologyTabulaExtractors::SpreadsheetExtractionAlgorithm.new
+      nda = Java::TechnologyTabulaDetectors::NurminenDetectionAlgorithm.new
       extractor.extract.each do |page|
         page_index = page.getPageNumber
 
         at( (page_count + page_index) / 2, page_count, "auto-detecting tables...") #starting at 50%...
         changed
 
-        cells = Java::TechnologyTabulaExtractors::SpreadsheetExtractionAlgorithm.findCells(page.getHorizontalRulings, page.getVerticalRulings)
-        areas = sea.findSpreadsheetsFromCells(cells)
+        areas = nda.detect(page)
         page_areas_by_page << areas.map { |rect|
           [ rect.getLeft,
             rect.getTop,

diff --git a/webapp/tabula_web.rb b/webapp/tabula_web.rb
@@ -9,7 +9,7 @@
 require 'fileutils'
 require 'securerandom'
 
-require_relative '../lib/jars/tabula-extractor-0.7.4-SNAPSHOT-jar-with-dependencies.jar'
+require_relative '../lib/jars/tabula-0.8.0-jar-with-dependencies.jar'
 
 require_relative '../lib/tabula_java_wrapper.rb'
 java_import 'java.io.ByteArrayOutputStream'