xml_row_finder 0.1.0 → 0.3.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: cfc152e03ec51b624550b4b4bf031eae3e7205c42e1dccbdd4d45219d684860d
-  data.tar.gz: 77be8ae4da89aded4ec73a653e9fe1e3d9815a144a29a37a73158b97174da984
+  metadata.gz: 1407d0f852c5a904600ef49f43c93b7a8e5cc4a459bb88d9d2e0bc4c26928887
+  data.tar.gz: 3867560288f6ed062b2c59e78bea3e30b7ddaf5217d3a181f864b4fbf28a202f
 SHA512:
-  metadata.gz: e83c208f4dfeac7827b9f8ca453baba4f19ff4be9d4341bbd15f46695948afd7af42ae5ad0c71efb994b58832c46b9cf836fcb2d115ffb126a49ebfc26065772
-  data.tar.gz: 24335ada2416257aa9caebd29b687562e92fe2d0ca2ac71f59824448c98eee018e15f95ee60352ed326440a50cb5445d3fb5bd4cbb1a0c47eed58ffc85ff12ec
+  metadata.gz: c19790d1e1c125b0be88831e1a1ab870662b03ad35859fc68b2431390854c0f4e507a8629198fefb1bfd03ed460e75ceb78e278f5de4be43e9dd9a1ab4c03734
+  data.tar.gz: d9d5321b0f639868aeca083728f7ebaa6da2e1d101d609aa05076e9e96c0c7e33e2545db0e803e5efd53808514c028b2190a425e08267e36e42cd6b6118adb3a

data/lib/xml_row_finder.rb

@@ -2,19 +2,20 @@
 
 # file: xml_row_finder.rb
 
-require 'rexle'
-require 'rexml'
-include REXML
+require 'nokorexi'
 
 
 class XMLRowFinder
 
   attr_reader :to_a
 
-  def initialize(s, debug: false)
+  def initialize(raws, debug: false)
 
     @debug = debug
-    doc = Rexle.new(s)
+
+    doc = Nokorexi.new(raws, filter: true).to_doc
+
+    @doc = Rexle.new(doc.root.xml)
 
     a = []
 
@@ -23,51 +24,77 @@ class XMLRowFinder
       a << e.backtrack.to_xpath
     end
 
-    a2 = a.select{ |e| a.count(e) > 1 }.map {|x| x.split('/')}.uniq
+    @to_a = a2 = a.map {|e| [a.count(e), e] }.uniq
+    xpath = a2.max_by(&:first).last
 
-    # remove parent nodes on the same branch
-    #
-    a2.reject!.with_index do |x,i| 
-      next if i == a2.length-1
-      x == a2[i+1][0..-2]
-    end
+    a3 = xpath.split('/')
+    a4 = [xpath]
+    p1 = []
 
-    # remove elements from rows which only exist once in the document
-    #
-    a3 = a2.map do |row|
-      row.reject do |x|
-        found = doc.root.xpath('//' + x)
-        found.length < 2
-      end
+    until (a3.length < 1) do
+      p1 << a3.pop; a4 << a3.join('/') + "[%s]" % p1.reverse.join('/')
     end
 
-    # add parent node to the row as a reference for the xpath
+    # using Nokogiri since Rexle has a bug with xpath predicates
     #
-    a4 = a3.map.with_index do |row,i|
-      a2[i][-(row.length+1)..-1]
+    @doc2 = Nokogiri::XML(doc.root.xml)
+
+    a5 = a4[0..-2].map do |xpath2|
+      [@doc2.xpath(xpath2).length, xpath2]
     end
 
-    # find the parent node attributes
-    #
-    @to_a = a4.map do |col|
-
-      # currently using REXML for this XPath since there is a bug in 
-      #   Rexle when attempting the following
-      # 
-      doc2 = Document.new(s)
-      xpath = "//%s[%s]" % [col[0], col[1..-1].join('/')]
-      puts 'xpath: ' + xpath.inspect if @debug
-      r = XPath.first(doc2, xpath)
-      xpath_a = BacktrackXPath.new(r).to_xpath
-
-      if col.length >= 3
-        "%s/%s[%s]" % [xpath_a, col[1], col[2..-1].join('/')]
-      else
-        "%s/%s" % [xpath_a, col[1]]
-      end
+    puts 'a5: ' + a5.inspect if @debug
+    rows_xpath = a5.reverse.detect {|num, xpath2| num > 1}.last
+    doc3 = Document.new @doc.root.xml
+    @rows = XPath.match(doc3, rows_xpath)
+    @xpath = rows_xpath
+    #@xpath = BacktrackXPath.new(@rows.first).to_xpath.gsub("[@class='']",'')
+
+    last_row = XPath.match(doc3, @xpath).last
+    puts '@xpath: ' + @xpath.inspect
+
+    # find the container element
+    xpath = @xpath[/^[^\[]+/]
+    axpath = xpath.split('/')
+
+    e = XPath.first(doc3, xpath)
+    puts 'e: ' + e.to_s
+
+    until (e.nil? or e.to_s.include?(last_row.to_s)) do
+      axpath.pop
+      e = XPath.first(doc3, axpath.join('/'))
     end
-    
+
+    @cont_xpath = axpath.join('/')
+
   end
 
-end
+  # returns the container element for all rows
+  #
+  def body()
+    Rexle.new(@doc.element(@cont_xpath).xml)
+  end
+
+  # returns the xpath pointing to the container element for all rows
+  #
+  def body_xpath()
+    @cont_xpath
+  end
+
+  # returns rows
+  # object returned: An array of Nokogiri XML Element object
+  #
+  def rows()
+    @rows
+  end
+
+  # returns the xpath pointing to the rows
+  #
+  def to_xpath()
+    @xpath
+  end
 
+  alias rows_xpath to_xpath
+
+
+end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: xml_row_finder
 version: !ruby/object:Gem::Version
-  version: 0.1.0
+  version: 0.3.2
 platform: ruby
 authors:
 - James Robertson
@@ -35,28 +35,28 @@ cert_chain:
   a/pAsvo0jT6QTnSB7xzsx8LSFDT5tfHKR9Dcn1Y3R06fsh02JvwxaSAMgDBM2aFb
   2A7/BQ1hD7SU82VTxB1gFIHl
   -----END CERTIFICATE-----
-date: 2022-01-21 00:00:00.000000000 Z
+date: 2022-01-28 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
-  name: rexle
+  name: nokorexi
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - "~>"
-      - !ruby/object:Gem::Version
-        version: '1.5'
     - - ">="
       - !ruby/object:Gem::Version
-        version: 1.5.14
+        version: 0.6.0
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '0.6'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - "~>"
-      - !ruby/object:Gem::Version
-        version: '1.5'
     - - ">="
       - !ruby/object:Gem::Version
-        version: 1.5.14
+        version: 0.6.0
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '0.6'
 description: 
 email: digital.robertson@gmail.com
 executables: []
@@ -83,7 +83,8 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.2.22
+rubyforge_project: 
+rubygems_version: 2.7.10
 signing_key: 
 specification_version: 4
 summary: Attempts to find repeating rows in XHTML and returns the associated xpath.