arrolio 0.1.6 → 0.1.7

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 057f22d28911ebd5fb250d0a3ee3aae065a493c2c8aee0a5ff329a138da38674
-  data.tar.gz: 8477b14736601c87ee24c690aa867ef175de94568ef866b48d260948c3d98116
+  metadata.gz: 6d2fb34e355fe255920ca92e339c7c2529db69b9668130d2f5c9f36319f2aa13
+  data.tar.gz: a591ae1001ae525c96d19a7321eee6840a105e3d9e6404a5b2a2454e68bd7b46
 SHA512:
-  metadata.gz: 2aef88dd6df44e8aef2be218f34f53de0e332ecfb94bc2ef5aa8a87f0221fab634ac05ad0a900997fc1ebb870529626a464eeaa302c7f3e19d1b05664bfd7e90
-  data.tar.gz: b984b3dae46396305d96e5c6977a24992c4ddae317be8a4682fafed9de46a35a7e2442dbc418147c77afb943ba7981c6b8be28b562b85723fe5624b3242b6c7c
+  metadata.gz: 3587a3bf6d2bebc2d66b4f990bccc365f8508a08ee75d191636c737a5d4c6eb31ff1cd8069175e010a40a38f17d2b98a8c788da3bb164dbb182aa632eb85fd99
+  data.tar.gz: deed9c49dc8e31bc26854c8dcb92b4f46214318da53a067786bdc1ac86329e6403d53bce132826a314b7e6876d28c51bd45ce6089ad8eb1d07eb61d0fb92e0b2

data/lib/arrolio/flowables/table_flowable.rb

@@ -99,20 +99,15 @@ module Arrolio
         header, body = partition_rows
         header_count = header.length
 
-        head_groups = []
-        tail_groups = []
         # The head part also carries the caption and the repeated
         # header rows — they consume the same budget as body rows.
-        used = caption_height(width) + heights[0, header_count].sum
-        body_row_groups.each do |group|
-          group_h = heights[group.first, group.length].sum
-          if used + group_h <= remaining_height
-            head_groups << group
-            used += group_h
-          else
-            tail_groups << group
-          end
-        end
+        budget = remaining_height - caption_height(width) -
+                 heights[0, header_count].sum
+        result = Table::SplitPolicy.new(groups: body_row_groups,
+                                        group_heights: heights,
+                                        budget: budget).call
+        head_groups = result.head_groups
+        tail_groups = result.tail_groups
 
         head_rows = head_groups.flat_map { |g| body[g.first - header_count, g.length] }
         tail_rows = tail_groups.flat_map { |g| body[g.first - header_count, g.length] }

data/lib/arrolio/harness/text_extractor.rb

@@ -51,7 +51,11 @@ module Arrolio
       def split_by_page(text)
         return [text] unless text.include?("\f")
 
-        text.split("\f", -1)
+        pages = text.split("\f", -1)
+        # pdftotext ends output with a trailing form-feed for some
+        # writers; the empty string after it is not a page.
+        pages.pop if pages.last.to_s.strip.empty? && pages.length > 1
+        pages
       end
     end
   end

data/lib/arrolio/table/split_policy.rb

@@ -0,0 +1,52 @@
+# frozen_string_literal: true
+
+module Arrolio
+  module Table
+    # Pure table-split decision: which welded row groups fit the
+    # page budget. The budget is the remaining page height minus
+    # what travels with every part (the caption and the repeated
+    # header rows). Groups are greedily packed while they fit; an
+    # empty head means whole-move (the engine moves the entire
+    # table to the next page).
+    class SplitPolicy
+      Result = Struct.new(:head_groups, :tail_groups, keyword_init: true) do
+        def whole_move?
+          head_groups.empty?
+        end
+      end
+
+      # +groups+ Array of Arrays of row indexes (welded by rowspan).
+      # +group_heights+ parallel Array of Float heights per group.
+      # +budget+ Float height available for body groups.
+      def initialize(groups:, group_heights:, budget:)
+        @groups = groups
+        @group_heights = group_heights
+        @budget = budget.to_f
+      end
+
+      # Rows split IN ORDER: once a group does not fit, every
+      # later group goes to the tail (a fitting later group must
+      # never jump ahead of an overflowing earlier one - the
+      # original greedy loop allowed exactly that reordering).
+      def call
+        head = []
+        tail = []
+        used = 0.0
+        overflowing = false
+        @groups.each do |group|
+          height = group_height_for(group)
+          overflowing ||= used + height > @budget
+          (overflowing ? tail : head) << group
+          used += height unless overflowing
+        end
+        Result.new(head_groups: head, tail_groups: tail)
+      end
+
+      private
+
+      def group_height_for(group)
+        @group_heights[group.first, group.length].sum
+      end
+    end
+  end
+end

data/lib/arrolio/table.rb

@@ -6,5 +6,6 @@ module Arrolio
   module Table
     autoload :AutoLayout, 'arrolio/table/auto_layout'
     autoload :Grid, 'arrolio/table/grid'
+    autoload :SplitPolicy, 'arrolio/table/split_policy'
   end
 end

data/lib/arrolio/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Arrolio
-  VERSION = '0.1.6'
+  VERSION = '0.1.7'
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: arrolio
 version: !ruby/object:Gem::Version
-  version: 0.1.6
+  version: 0.1.7
 platform: ruby
 authors:
 - Ribose Inc.
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2026-08-31 00:00:00.000000000 Z
+date: 2026-09-02 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: fontisan
@@ -432,6 +432,7 @@ files:
 - lib/arrolio/table.rb
 - lib/arrolio/table/auto_layout.rb
 - lib/arrolio/table/grid.rb
+- lib/arrolio/table/split_policy.rb
 - lib/arrolio/text_direction.rb
 - lib/arrolio/text_layout.rb
 - lib/arrolio/text_layout/break_opportunity.rb