bibletools 0.2.0 → 0.2.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: c6ae238a57f75318c4238835d64567683523fdcf106e3a4eba8007e833de88ac
-  data.tar.gz: bd3471c321d7caf26d360bc66ee44eda5b3276d808de12e4ec450044a5f62515
+  metadata.gz: e10c0e0f8c335ebc195c2086d59c3d27a9a4f3d2c19e42ba85bca938365a4c2b
+  data.tar.gz: bf5259405950f58998288d7d17f783a6672600c053617e56e34a5363c44f938f
 SHA512:
-  metadata.gz: 5dddc1de4ef4dc2439a3f3ca8d4e49a7973dd89c06362ac1761be02ed2240a09367b622b3abdac5ad919f23c49ef3e72ee63dbac13bd4818aa3abe57b19dddb7
-  data.tar.gz: b39d63a2795fc9468e176ee44e674a79042e2f5608c64099b9604ce8e8d75f75376cd26fd6040760f5c113579fba00cfb1c8ee5d1c345576c1a064d54d9443a4
+  metadata.gz: f8e7b6f0a5246ab59b5db9732c8ea29e34e354c687f6a6bdae7971a99f89947a53252acad367d0a006984fbd1bf7c8349f09cf21b2b7f9b4e66d42a25a53e087
+  data.tar.gz: 9fd82593908e3f1a97b0a65f5d7b77bc34b8fc3e343abe1d4b81e66d4f8bb152dab81efcf3a0424a306d8df4621c7bff5a93473e5cf3b9a4e70ae8950ed897b8

data/lib/bibletools.rb

@@ -93,17 +93,21 @@ module BibleTools
   
   class Scraper
 
-    def initialize(book: 'Exodus', url: '')
+    def initialize(book: 'Exodus', url: '', debug: false)
       
       raise ScraperErr, 'Please specify the URL' if url.empty?
       
       baseurl = url
 
       id = if book.is_a? String then
-        (BOOKS.index(book) + 1).to_s
+        i = BOOKS.index(book.gsub(/\b./){|x| x.upcase})
+        puts 'book *' + book + '* not found' unless i
+        (i + 1).to_s
       else
         book.to_s
       end
+      
+      puts 'id: ' + id.inspect if @debug
 
       url2 = baseurl + '?id=' + id
       doc = Nokorexi.new(url2).to_doc
@@ -163,8 +167,8 @@ module BibleTools
       @debug = debug
       
       if bible_obj then
+        
         @doc = bible_obj.to_doc
-        @verses = @doc.root.xpath('chapter/verse')
         @check = EnglishSpellcheck.new debug: false #verbose: false
         
         if tts then
@@ -180,9 +184,28 @@ module BibleTools
     #   Automatically select verses; verses are selected based upon word 
     # frequency 
     #
-    def asr()
+    def asr(chapters: [])
+      
+      r = if chapters.any? then
+        
+        doc = Rexle.new('<book/>')
+                
+        chapters.each do |n|
+          e2 = @doc.root.element("chapter[@no='#{n.to_s}']")
+          doc.root.add e2 if e2
+        end
+        
+        return doc
+        
+        keyword = wordtally(doc: doc).keys.first
+        #return keyword
+        #get the wordstally for the chapters
+        assoc_r(keyword, doc)
+        
+      else
+        assoc_r()
+      end
       
-      r = assoc_r()
       e = r[1].root.element('chapter/verse')
       verse_no = e.attributes[:no]
       chptr_no = e.parent.attributes[:no]
@@ -190,16 +213,18 @@ module BibleTools
       # we are backtracking the results to hopefully find a new branch to explore
       trail = r.first
 
+      #return r
       a, trail2 = mine_words(r, trail, level: 2)
       a << [chptr_no, verse_no,  e.text, 1]
-
+      puts 'a: ' + a.inspect
+      #return a
       h = a.group_by(&:first)
       a2 = h.sort_by {|x, _| x.to_i}
-      a3 = a2.map do |chapter, verses|
+      @verses = a2.map do |chapter, verses|
         [chapter, verses.sort_by {|_,x| x.to_i}]
       end
         
-      [a3, trail2]
+      [@verses, trail2]
     end    
     
     # find the words associated with a given keyword
@@ -266,9 +291,12 @@ module BibleTools
     def search(keyword, doc=@doc)
       
       verses = doc.root.xpath('chapter/verse')
+      puts 'verses: ' + verses.inspect
       a = verses.select {|x| x.text =~ /#{keyword}/i}
       
       a2 = a.map.with_index do |verse, i|
+        
+        #puts 'verse: ' + verse.xml.inspect
         txt = verse.text
         
         puts 'txt: ' + txt.inspect if @debug
@@ -278,9 +306,11 @@ module BibleTools
         
         until txt.rstrip[/[\.\?!]\W*$/] or n > 2 do
           n +=1
+          puts 'n: ' + n.inspect
           nverse = read(chapter, verse.attributes[:no].to_i + n)[0]
+          puts 'nverse: ' + nverse.inspect
           puts 'nverse: ' + nverse&.text.inspect if @debug
-          txt += nverse.text
+          txt += nverse.text if nverse
         end 
         
         if @debug then
@@ -321,6 +351,52 @@ module BibleTools
       
     end
     
+    def verses(level: nil, html: false, title: nil)                  
+
+      a = @verses.map {|_, body| body}.flatten(1)
+      a1 = level ? a.select {|c,v,t,l| l <= level} : a.map {|c,v,t,l| [c,v,t]}
+      a2 = a1.map {|c,v,t,l| [c,v,t]}.uniq
+      h = a2.group_by(&:first)
+      
+      puts 'html: ' + html.inspect
+      return h unless html == true
+      
+      doc = Rexle.new('<html/>')
+
+
+      head = Rexle::Element.new('head')
+      head.add(Rexle::Element.new('h1').add_text(title) ) if title
+
+      style = Rexle::Element.new('style')
+      style.add_text '     ins {font-size: 0.8em;padding: 0.8em;}'
+      head.add style
+
+      doc.root.add head
+      body = Rexle::Element.new('body')
+
+      h.each do |chptr, verses|
+
+        h2 = Rexle::Element.new('h2').add_text chptr
+        body.add h2
+
+        verses.each do |cno, vno, text|
+
+          ins = Rexle::Element.new('ins').add_text vno
+          para = Rexle::Element.new('p')
+          para.add ins
+          para.add_text text
+          body.add para
+          
+        end
+        
+      end
+      
+      doc.root.add body
+
+      doc.root.xml pretty: true
+      
+    end
+    
     private
 
     def mine_words(r, trail, verses=[], level: 0)
@@ -329,28 +405,32 @@ module BibleTools
       found = []
       (i-=1; found = r.last[i][0].keys - trail) until found.any?
       r2 = r.last[i][0].keys - trail
+      
       a3 = r2.map do |x|
+        
         _, doc_verse = assoc_r x, r.last[i][1]
         chaptr_no = doc_verse.root.element('chapter').attributes[:no]
         everse = doc_verse.root.element('chapter/verse')
         verse_no = everse.attributes[:no]
+        
         [chaptr_no, verse_no, everse.text, level]
+        
       end.uniq
 
-      #a3b = a3.group_by {|x| x[0]}.map(&:last).flatten(1).sort_by {|x| x[0].to_i}
-      #a3b.length
-
       puts 'verses: ' + verses.inspect if @debug
+      
       verses.concat a3
       verses.uniq!
       trail.concat r2
+      
       puts 'trail.length: ' + trail.length.inspect if @debug
       puts 'verses.length: ' + verses.length.inspect if @debug
 
-      if verses.length < 30 or trail.length < 100
-        mine_words(r, trail, verses, level: level+1)
-      else
+      if verses.length > 30 or trail.length > 100
         [verses, trail]
+      else
+
+        mine_words(r, trail, verses, level: level+1)        
       end
     end    
     

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: bibletools
 version: !ruby/object:Gem::Version
-  version: 0.2.0
+  version: 0.2.2
 platform: ruby
 authors:
 - James Robertson
@@ -36,7 +36,7 @@ cert_chain:
   hIlyeyH9vsaJLam1TYsSqh3KUBmRCIvc3XELg2MAlggTrHuCjBPdYXqjIOFvO6HZ
   Z1XDnMf4RhcEPF4AU3Q+Fefsw4qKRWH3YUw=
   -----END CERTIFICATE-----
-date: 2022-11-05 00:00:00.000000000 Z
+date: 2023-01-19 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: nokorexi
@@ -144,7 +144,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.3.7
+rubygems_version: 3.4.4
 signing_key: 
 specification_version: 4
 summary: Performs a word count within a specified book. Returns the verses within