diabible-parser 0.1.6 → 0.3.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 4890da70174e919b49e385c291425fc5c2941230f410107b41aede336ca931ef
-  data.tar.gz: 5f97266c8dcd64e27a8a208fca5d8239abfd446c91a3a4cf6d12bfa5a83df79c
+  metadata.gz: c8b00b4ed3517b91ce76c3d217e46ab5e3a7a43c95776517d75edd2102661f46
+  data.tar.gz: 1ce6691990f539cc0fd2841400e03d29f6d32019f9f7ecd3f0560180896281a9
 SHA512:
-  metadata.gz: dabfdfaaf2e7a256d9cdc94d29a68973602f9a254a55206dd35cc5ffcd710af27b3b6d0461c119ca4b2cad82cecdce85e01a385c931d9b9bda10ba09ff1d3ede
-  data.tar.gz: c8545b7ffa872bb1161917be990d0c50297018512c30730e97f039bc920689cd1b9f5d6ec7d62e9b697b132a33bf87ac2185a047592973ee0535aacc26cefb51
+  metadata.gz: 4f6abc6a44871efb78474f513c9d79d217e26baec0033c89a51fca26c248df10c3d0ab6401c67cca656ccecdd8aad2f3d511df56135828e6593b755ebf2ecbcc
+  data.tar.gz: cab40cfebe6c0c7a3a03bb74c82bd4ce28caeb14ad820c20fcc252c5812c00c9eb3d654ab02b9e41b0e0fa6135031cd7b3bf7bdafafd44f2eae6b8ba15366064

data/lib/diabible/parser/document.rb

@@ -29,11 +29,11 @@ module Diabible
 
     class Document
       include Helpers::Info
-      include Helpers::Notes
+      include Helpers::Metadata
       include Helpers::Text
 
       # default constructor
-      def initialize(document, info: true, notes: true, text: true)
+      def initialize(document, info: true, text: true, limit: 0)
         Zip::File.open(URI.open(document)) do |zip_file|
           # open content xml
           @document = Nokogiri::XML(zip_file.read("word/document.xml"))
@@ -43,9 +43,9 @@ module Diabible
         # defaults
         @text = ''
         @metadata = []
+        @limit = limit
         # parse
         parse_info if info
-        parse_notes if notes
         parse_text if text
       end
 
@@ -70,19 +70,9 @@ module Diabible
 
       private
 
-      def info_count
-        @info_count ||= 0
-        @info_count += 1
-      end
-
-      def note_count
-        @note_count ||= 0
-        @note_count += 1
-      end
-
-      def text_count
-        @text_count ||= 0
-        @text_count += 1
+      def count
+        @count ||= 0
+        @count += 1
       end
 
       # get text content of all nodes matched by xpath
@@ -96,18 +86,6 @@ module Diabible
         # return container
         content
       end
-
-      # add meta hash
-      def add_metadata(name, type, value, position = nil)
-        # firstly check if the last field is not a same type
-        if @metadata.last and @metadata.last[:type] != PARAGRAPH and @metadata.last[:type] == type
-          # merge fields together
-          @metadata.last[:value] += value
-        else
-          # normally add a new meta hash with uniqe namee
-          @metadata << { name: "#{name}_#{send("#{name}_count")}_#{type}", type: type, value: value, position: position }
-        end
-      end
     end
   end
 end

data/lib/diabible/parser/helpers.rb

@@ -19,8 +19,9 @@
 # Authors: Michal Mocnak <michal@marigan.net>
 #
 
+require 'diabible/parser/helpers/constants'
 require 'diabible/parser/helpers/info'
-require 'diabible/parser/helpers/notes'
+require 'diabible/parser/helpers/metadata'
 require 'diabible/parser/helpers/text'
 
 module Diabible

data/lib/diabible/parser/helpers/{notes.rb → constants.rb}

@@ -22,28 +22,10 @@
 module Diabible
   module Parser
     module Helpers
-      module Notes
+      module Constants
 
-        private
-
-        # parse notes box
-        def parse_notes
-          # iterate through all paragraphs with style
-          @document.xpath("//body/p").each do |p|
-            # store pStyle value
-            pStyle = p.xpath(".//pStyle/@val")
-            # process only those with pstyle
-            unless pStyle.empty? or ['Hlavicka', 'Incipit', 'Podnadpis', 'Podnadpis1'].include?(pStyle.first.value)
-              # iterate over r blocks
-              p.xpath(".//r").each do |r|
-                name = pStyle.first.value
-                value = content(r, ".//t")
-                # add into container
-                add_metadata("note", name, value)
-              end
-            end
-          end
-        end
+        NONE = 'none'
+        PARAGRAPH = 'paragraph'
       end
     end
   end

data/lib/diabible/parser/helpers/metadata.rb

@@ -0,0 +1,71 @@
+#
+# Copyright (C) 2021 diabible.com
+#
+# This file is part of Diabible Parser.
+#
+# Diabible Parser is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# Diabible Parser is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with Diabible Parser. If not, see <http://www.gnu.org/licenses/>.
+#
+# Authors: Michal Mocnak <michal@marigan.net>
+#
+
+module Diabible
+  module Parser
+    module Helpers
+      module Metadata
+
+        # add meta hash
+        def add_metadata(group, type, value, position = nil)
+          # get index
+          index = count
+          # normally add a new meta hash with uniqe name
+          @metadata << { name: "#{index}_#{group}_#{type}", group: group, type: type, index: index, value: value, position: position }
+        end
+
+        # get metadata array per group or type
+        def get_metadata(group: nil, type: nil, filter: :exclusive, filters: [])
+          if group and type
+            return metadata_query(filter: filter, filters: filters) do |m|
+              m[:group] == group and m[:type] == type
+            end
+          elsif group
+            return metadata_query(filter: filter, filters: filters) do |m|
+              m[:group] == group
+            end
+          elsif type
+            return metadata_query(filter: filter, filters: filters) do |m|
+              m[:type] == type
+            end
+          else
+            return metadata_query(filter: filter, filters: filters) do |m|
+              true
+            end
+          end
+        end
+
+        def metadata_query(filter: :exclusive, filters: [])
+          return @metadata.select do |m|
+            if filters.empty?
+              yield m
+            else
+              # get filters intersection
+              filtered = filters & [m[:type], m[:group]]
+              # resolution
+              (yield m) and (filter == :exclusive ? filtered.empty? : !filtered.empty?)
+            end
+          end
+        end
+      end
+    end
+  end
+end

data/lib/diabible/parser/helpers/text.rb

@@ -19,47 +19,55 @@
 # Authors: Michal Mocnak <michal@marigan.net>
 #
 
+require 'diabible/parser/helpers/constants'
+
 module Diabible
   module Parser
     module Helpers
       module Text
+        include Diabible::Parser::Helpers::Constants
 
         private
 
-        # paragraph type id
-        PARAGRAPH = 'paragraph'
-
         # parse main text entries
         def parse_text
           # iterate through all paragraphs without style
           @document.xpath("//body/p").each do |p|
             # store pStyle value
             pStyle = p.xpath(".//pStyle/@val")
-            # process only those without pstyle
-            if pStyle.empty? or (pStyle.size and ['Incipit', 'Podnadpis', 'Podnadpis1'].include?(pStyle.first.value))
-              # position in for paragraph
-              paragraph_in = @text.length
-              # iterate over r blocks
-              p.xpath(".//r").each do |r|
-                # store rstyle value
-                rStyle = r.xpath(".//rStyle/@val")
-                # if empty then it's a plain text without style
-                if rStyle.empty?
-                  # update text object
-                  @text += content(r, ".//t")
-                else
-                  # otherwise create metadata field
-                  name = rStyle.first.value
-                  value = content(r, ".//t")
-                  # add into container
-                  add_metadata("text", name, value, { in: @text.length })
-                end
+            # prepare group
+            group = pStyle.first.value unless pStyle.empty?
+            group ||= NONE
+            # position in for paragraph against plain text
+            paragraph_in = @text.length
+            # iterate over r blocks
+            p.xpath(".//r").each do |r|
+              # store rstyle value
+              rStyle = r.xpath(".//rStyle/@val")
+              # if empty then it's a plain text without style
+              if (rStyle.empty? or rStyle.first.value === 'Text') and pStyle.empty?
+                # update text object
+                @text += content(r, ".//t")
+              else
+                # otherwise create metadata field
+                # prepare type
+                type = rStyle.first.value unless rStyle.empty?
+                type ||= NONE
+                # prepare value
+                value = content(r, ".//t")
+                # add into container
+                add_metadata(group, type, value, { in: @text.length })
               end
-              # position out for paragraph
-              paragraph_out = @text.length
-              # filter paragraphs without content
-              if paragraph_in != paragraph_out
-                add_metadata("text", PARAGRAPH, nil, { in: paragraph_in, out: paragraph_out })
+            end
+            # position out for paragraph
+            paragraph_out = @text.length
+            # filter paragraphs without content
+            add_metadata(PARAGRAPH, group, NONE, { in: paragraph_in, out: paragraph_out })
+            # limit check
+            if @limit > 0
+              if @metadata.size >= @limit
+                # when meets limit break
+                break
               end
             end
           end

data/lib/diabible/parser/version.rb

@@ -21,6 +21,6 @@
 
 module Diabible
   module Parser
-    VERSION = "0.1.6"
+    VERSION = "0.3.1"
   end
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: diabible-parser
 version: !ruby/object:Gem::Version
-  version: 0.1.6
+  version: 0.3.1
 platform: ruby
 authors:
 - Michal Mocnak
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2021-03-01 00:00:00.000000000 Z
+date: 2021-03-08 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: nokogiri
@@ -65,8 +65,9 @@ files:
 - lib/diabible/parser.rb
 - lib/diabible/parser/document.rb
 - lib/diabible/parser/helpers.rb
+- lib/diabible/parser/helpers/constants.rb
 - lib/diabible/parser/helpers/info.rb
-- lib/diabible/parser/helpers/notes.rb
+- lib/diabible/parser/helpers/metadata.rb
 - lib/diabible/parser/helpers/text.rb
 - lib/diabible/parser/version.rb
 - spec/spec_helper.rb