iguvium 0.8.4 → 0.9.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 622824e3031b9d70a8385ea50d8283c375c6e15924dee4ce5b03cc282415de48
-  data.tar.gz: daa366102b6815f9df955d10396f665347529d036531829dbd63ec7d61307f83
+  metadata.gz: cf00782b6d5ec4d4e06d12ce386ff8ece314f8ba896bb70117232ec100ec5cdb
+  data.tar.gz: 544dbce4b3cfacd303f711e774284cd497b9b3dcdc8de0590b4825f84ff0e399
 SHA512:
-  metadata.gz: da44bc6dd6806fd66f74d388b8ab3334cde54a0efda764bf28564f97377f0aac0da82c69473e1ce73429b3642f547f426581b4bbd39c94f9fa6865147911a021
-  data.tar.gz: 9685fbe9866ac87931d978e3f588a38ea9bf47c7ffaf5f58e823d3deae0608ed6a0e5db2596da52e46e511644b6aa7df3be12a441a82b64451c118d33c7ffb7d
+  metadata.gz: b6340302943733d007a37e16206337c712646199ab9133699d328cb44c4a1ea6c1e4f65fb61c559467b018d1232ccd35c03e48b760f67645292b66dbc2640ed1
+  data.tar.gz: 0b9dcce49fc880a1fa7a7d0b512e3037512800dd4531caad230d788cb5297647190591fd27057e0216d92ed32465b9b60bbbc38eb34d68c6a324f5f06edb6d43

data/CHANGELOG.md

@@ -2,6 +2,16 @@
 
 ## [Unreleased]
 
+## [0.9.0] - 2018-12-07
+### Added
+- Open cells rendering added. Tables like this are now processed correctly:
+```
+__|____|_______|_____|
+__|____|_______|_____|
+__|____|_______|_____|
+```
+
+
 ## [0.8.4] - 2018-11-24
 ### Changed
 - Render phrases before cell assembly option of Iguvium::Table#to_a method is now true by default. 

data/README.md

@@ -24,7 +24,8 @@ Get this table:
 
 * Characters extraction is done by [PDF::Reader gem](https://github.com/yob/pdf-reader). Some PDFs are so messed up it can't extract meaningful text from them. If so, so does Iguvium.
 
-* Current version extracts regular (with constant number of rows per column and vise versa) tables with explicit lines formatting, like this:
+* Current version extracts regular (with constant number of rows per column and vise versa) 
+tables with explicit lines formatting, like this:
 
 ```
 .__________________.
@@ -32,7 +33,15 @@ Get this table:
 |____|_______|_____|
 |____|_______|_____|
 ```
-  Merged cells content is split as if cells were not merged.
+And, after version 0.9.0, like this:
+```
+__|____|_______|_____|
+__|____|_______|_____|
+__|____|_______|_____|
+```
+
+
+  Merged cells content is split as if cells were not merged unless you use `:phrases` option.
 
 * Performance: considering the fact it has computer vision under the hood, the gem is reasonably fast. Full page extraction takes up to 1 second on modern CPUs and up to 2 seconds on the older ones.
 
@@ -107,15 +116,7 @@ Initially inspired by [camelot](https://github.com/socialcopsdev/camelot/) idea
 
 ## Roadmap
 
-The next version will deal with open-edged tables like
-
-```
-__|____|_______|_____|
-__|____|_______|_____|
-__|____|_______|_____|
-```
-
-It also will keep open-edged rows metadata ('floorless' and 'roofless') for the needs of multipage tables merger.
+The next version will keep open-edged rows metadata ('floorless' and 'roofless') for the needs of multipage tables merger.
 
 The final one will recognize tables with merged cells.
 

data/exe/iguvium

@@ -18,6 +18,7 @@ opts = Slop.parse { |o|
   end
   o.on '-h', '--help', 'show help' do
     puts o.to_s.gsub(/(usage:).+(iguvium)/, '\1 \2 filename.pdf')
+    puts Iguvium::VERSION
     exit
   end
 }

data/lib/iguvium/cv.rb

@@ -39,11 +39,11 @@ module Iguvium
     # Prepares image for recognition: initial blur
     # @param image [ChunkyPNG::Image] from {Iguvium::Image.read}
     def initialize(image)
-      @image = blur image
+      @blurred = blur(image)
+      @image = to_narray(image).to_a
     end
 
-    # @return [Array] 8-bit representation of an image
-    attr_reader :image
+    attr_reader :image, :blurred
 
     # @return [Recognized]
     #   lines most probably forming table cells and tables' outer borders as boxes
@@ -62,9 +62,8 @@ module Iguvium
         {
           vertical: Labeler.new(verticals)
                            .lines
-                           .map { |line| flip_line line }
-                           .sort_by { |x, yrange| [yrange.begin, x] },
-          horizontal: Labeler.new(horizontals).lines.map { |line| flip_line line }.sort_by { |_, y| [y] }
+                           .map { |line| flip_line line },
+          horizontal: Labeler.new(horizontals).lines.map { |line| flip_line line }
         }
     end
 
@@ -80,14 +79,14 @@ module Iguvium
 
     def verticals(threshold = 3)
       Matrix
-        .rows(convolve(NArray[*horizontal_scan(image)], VERTICAL, 0).to_a)
+        .rows(convolve(NArray[*horizontal_scan(blurred)], VERTICAL, 0).to_a)
         .map { |pix| pix < threshold ? nil : pix }
         .to_a
     end
 
     def horizontals(threshold = 3)
       Matrix
-        .rows(convolve(NArray[*vertical_scan(image)], HORIZONTAL, 0).to_a)
+        .rows(convolve(NArray[*vertical_scan(blurred)], HORIZONTAL, 0).to_a)
         .map { |pix| pix < threshold ? nil : pix }
         .to_a
     end
@@ -186,8 +185,6 @@ module Iguvium
     def box(coord_array)
       ax, bx = coord_array.map(&:last).minmax
       ay, by = coord_array.map(&:first).minmax
-      # additional pixels removed from the box definition
-      # [ax - 1..bx + 1, ay - 1..by + 1]
       [ax..bx, flip_range(ay..by)]
     end
   end

data/lib/iguvium/row.rb

@@ -0,0 +1,32 @@
+# frozen_string_literal: true
+
+module Iguvium
+
+  class Row
+    # gets characters limited by yrange and set of column ranges
+    def initialize(columns, characters, phrases: true)
+      @columns = columns
+      if phrases
+        characters =
+          characters
+          .sort
+          .chunk_while { |a, b| a.mergable?(b) }
+          .map { |chunk| chunk.inject(:+) }
+      end
+      @characters = characters
+    end
+
+    def cells
+      @columns.map { |range|
+        @characters.select { |character| range.cover?(character.x) }
+      }
+    end
+
+    # @return rendered row array
+    def render(newlines: false)
+    end
+
+    def merge(other)
+    end
+  end
+end

data/lib/iguvium/table.rb

@@ -16,6 +16,8 @@ module Iguvium
       @box = box
       @lines = page.lines
       @page = page
+      grid
+      heal
     end
 
     # Renders the table into an array of strings.
@@ -30,24 +32,74 @@ module Iguvium
     # @return [Array] 2D array of strings (content of table's cells)
     #
     def to_a(newlines: false, phrases: true)
-      grid[:rows]
+      @to_a ||=
+        grid[:rows]
         .reverse
         .map { |row|
-        grid[:columns].map do |column|
-          render(
-            phrases ? words_inside(column, row) : chars_inside(column, row),
-            newlines: newlines
-          )
-        end
-      }
+          grid[:columns].map do |column|
+            render(
+              phrases ? words_inside(column, row) : chars_inside(column, row),
+              newlines: newlines
+            )
+          end
+        }
     end
 
+    # def width
+    #   grid[:columns].count
+    # end
+
+    # def mergeable?(other)
+    #   width == other.width
+    # end
+
+    # def roofless?
+    #   @roofless
+    # end
+
+    # def floorless?
+    #   @floorless
+    # end
+
     private
 
     attr_reader :page, :lines, :box
 
-    def enhancer(grid)
-      # @todo write grid enhancer to detect cells between outer grid lines and box borders
+    # Looks if there are characters inside the box but outside of already detected cells
+    # and adds rows and/or columns if necessary.
+    # @return [Iguvium::Table] with added open-cell rows and columns
+    def heal
+      heal_rows
+      heal_cols
+      self
+    end
+
+    def wide_box
+      @wide_box ||= [
+        box.first.begin - 2..box.first.end + 2,
+        box.last.begin - 2..box.last.end + 2
+      ]
+    end
+
+    def heal_cols
+      leftcol = box.first.begin..grid[:columns].first.begin
+      rightcol = grid[:columns].last.end..box.first.end
+      @grid[:columns].unshift(leftcol) if chars_inside(leftcol, box.last).any?
+      @grid[:columns].append(rightcol) if chars_inside(rightcol, box.last).any?
+    end
+
+    def heal_rows
+      # TODO: shrink box (like `box.last.end - 2`)
+      roofrow = box.last.begin..grid[:rows].first.begin
+      floorrow = grid[:rows].last.end..box.last.end
+      if chars_inside(box.first, roofrow).any?
+        @grid[:rows].unshift(roofrow)
+        @roofless = true
+      end
+      if chars_inside(box.first, floorrow).any?
+        @grid[:rows].append(floorrow)
+        @floorless = true
+      end
     end
 
     def characters
@@ -74,7 +126,9 @@ module Iguvium
     end
 
     def grid
-      @grid ||=
+      return @grid if @grid
+
+      @grid =
         {
           rows: lines_to_ranges(lines[:horizontal]),
           columns: lines_to_ranges(lines[:vertical])
@@ -82,7 +136,8 @@ module Iguvium
     end
 
     def lines_to_ranges(lines)
-      lines.select { |line| line_in_box?(line, box) }
+      # TODO: extend box for the sake of lines select
+      lines.select { |line| line_in_box?(line, wide_box) }
            .map { |line| line.first.is_a?(Numeric) ? line.first : line.last }
            .sort
            .uniq

data/lib/iguvium/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Iguvium
-  VERSION = '0.8.4'
+  VERSION = '0.9.0'
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: iguvium
 version: !ruby/object:Gem::Version
-  version: 0.8.4
+  version: 0.9.0
 platform: ruby
 authors:
 - Dima Ermilov
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2018-11-24 00:00:00.000000000 Z
+date: 2018-12-07 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: pdf-reader
@@ -135,6 +135,7 @@ files:
 - lib/iguvium/image.rb
 - lib/iguvium/labeler.rb
 - lib/iguvium/page.rb
+- lib/iguvium/row.rb
 - lib/iguvium/table.rb
 - lib/iguvium/version.rb
 homepage: https://github.com/adworse/iguvium