ebook_tools 0.0.3 → 0.0.4

data/CHANGELOG

@@ -1,7 +1,7 @@
-0.0.1 2013.4.2
+0.0.3 2013.4.2
   add book id for docbook 
 
-0.0.1 2013.4.1
+0.0.2 2013.4.1
   change docbook format
 
 0.0.1 2013.4.1

data/bin/ebook_tools

@@ -82,10 +82,12 @@ def help(command=nil)
       batch_convert: 批量转换指定目录中的文件为epub格式文件，并存放到目标目录
       batch_extract: 批量提取指定目录中文件的书结构信息，并生成Docbook存放到目标目录
 
-    适用对象要求:
-      编码格式为utf-8
-
     具体命令的更多信息请通过'ebook_tools help <command>'查看。
+
+    其他命令:
+      sed -i 's/^\(\s*[0-9]\{1,3\}\)./\1 /' <file>     : 将 "1.xxx" 格式替换成 "1 xxx"
+      iconv --from-code GB2312 -t UTF-8 -c -o <output file> <file> : 将GB2312格式文件转换成UTF-8
+      sed -i 's/\s*第.\{1,3\}章.*[0-9]\{1,3\}$//g' <file>  : 将 "第xxx章 xxxx1" 格式清除。主要用于有些pdf文档转换过来时的页眉信息。
     EOF
   end
 

data/bin/para_import_mongo

@@ -0,0 +1,72 @@
+#!/usr/bin/env ruby
+# encoding: UTF-8
+require 'rubygems'
+require 'optparse'
+require File.join(File.expand_path('../../',__FILE__),'lib','utils')
+require File.join(File.expand_path('../../',__FILE__),'lib','paras_in_mongo')
+
+def help
+  puts <<-EOF
+    usage:
+      para_import_mongo [options] <docbook>
+
+    docbook: 已经标注重点段落的书文件或目录
+
+    options:
+      -H <host> , --host <host> : mongodb服务器,默认为localhost
+      -P <port> , --port <port> : mongodb服务器端口号,默认为27017 (Mongo默认端口号)
+      -D <database>, --database <database>  : 重点段落要存放的数据库
+      -C <collection>, --collection <collection> : 重点段落存放的集合 
+  EOF
+  exit
+end
+
+
+options = {:host=>'localhost',:port=>27017,:database=>'resource_development',:collection=>'paras'}
+opts = OptionParser.new do |opts|
+  opts.on('-H host','--host host') do |host|
+    options[:host] = host
+  end
+
+  opts.on('-P port','--port port') do |port|
+    options[:port] = port.to_i
+  end
+
+  opts.on('-D database','--database database') do |database|
+    options[:database] = database
+  end
+
+  opts.on('-C collection','--collection collection') do |collection|
+    options[:collection] = collection
+  end
+
+  opts.on('-h','--help') do
+    help
+  end
+end
+opts.parse ARGV
+
+docbook = ARGV[-1]
+
+if docbook.nil?
+  help
+end
+
+unless File.exists?(docbook)
+  raise "错误:指定的docbook不存在。"
+end
+
+files = if File.directory?(docbook)
+  Utils.scan_file_from_dir(docbook,{:format=>'.xml'})
+else
+  [docbook]
+end
+
+files.each do |file|
+  begin
+    ParasInMongo.file_in_mongo(file,options)
+  rescue
+    puts "error: #{file} import mongo failure!"
+  end
+end
+puts "success: #{docbook} in mongo successfully!"

data/ebook_tools.gemspec

@@ -2,11 +2,11 @@
 
 Gem::Specification.new do |s|
   s.name = %q{ebook_tools}
-  s.version = '0.0.3'
+  s.version = '0.0.4'
 
   s.required_rubygems_version = Gem::Requirement.new(">= 0") if s.respond_to? :required_rubygems_version=
   s.authors = ["Aaron"]
-  s.date = %q{2013-04-03}
+  s.date = %q{2013-04-05}
   s.description = %q{电子书工具集.}
   s.email = %q{aaron@nonobo.com}
   s.require_paths = ["lib"]
@@ -15,10 +15,12 @@ Gem::Specification.new do |s|
   s.has_rdoc = true
   s.rdoc_options = ["--charset=UTF-8"]
   s.executables << "ebook_tools"
+  s.executables << "para_import_mongo"
   s.files = [
     "README",
     "CHANGELOG",
     "bin/ebook_tools",
+    "bin/para_import_mongo",
     "lib/ebook_tools.rb",
     "lib/extract_book_struct.rb",
     "lib/header_detect.rb",
@@ -26,6 +28,7 @@ Gem::Specification.new do |s|
     "lib/txt.rb",
     "lib/epub.rb",
     "lib/utils.rb",
+    "lib/paras_in_mongo.rb",
     "ebook_tools.gemspec"
   ]
   s.add_dependency(%q<uuid>)
@@ -35,4 +38,5 @@ Gem::Specification.new do |s|
   s.add_dependency(%q<pdf-reader>)
   s.add_dependency(%q<nokogiri>)
   s.add_dependency(%q<levenshtein>)
+  s.add_dependency(%q<mongo>)
 end

data/lib/paras_in_mongo.rb

@@ -0,0 +1,42 @@
+# encoding: UTF-8
+require 'mongo'
+require 'nokogiri'
+
+include Mongo
+module ParasInMongo
+  extend self
+
+  def file_in_mongo(filename,options={})
+    client = MongoClient.new(options[:host], options[:port])
+    db     = client[options[:database]]
+    coll   = db[options[:collection]]   
+
+    doc = Nokogiri::XML(File.open(filename).read)
+
+    book_id = doc.search("book")[0]['id']
+    title = doc.search("book info title")[0].text
+    author = doc.search("book info author")[0].text
+    pubdate = doc.search("book info pubdate")[0].text
+    publisher = doc.search("book info publisher")[0].text
+
+    paras = doc.search("para[key=yes]")
+
+    source ={book: {title: title,book_id: book_id, author: author,pubdate: pubdate, publisher: publisher}}
+
+    paras.each do |para|
+      para_attrs = {'_id' => para['id']}
+      content = para.search("content")[0].text
+
+      keywords = []
+      para.search("keyword").each do |keyword|
+        keywords << {keyword: keyword.text, weight: keyword['weight'].to_i}
+      end
+
+      para_attrs = para_attrs.merge(keywords: keywords, content: content)
+      section = para.parent.search("info title").text
+      para_attrs = para_attrs.merge(source: source.merge(location: {section: section}))
+      
+      coll.insert(para_attrs)
+    end 
+  end
+end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: ebook_tools
 version: !ruby/object:Gem::Version
-  version: 0.0.3
+  version: 0.0.4
   prerelease: 
 platform: ruby
 authors:
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2013-04-03 00:00:00.000000000 Z
+date: 2013-04-05 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: uuid
@@ -123,16 +123,34 @@ dependencies:
     - - ! '>='
       - !ruby/object:Gem::Version
         version: '0'
+- !ruby/object:Gem::Dependency
+  name: mongo
+  requirement: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: '0'
 description: 电子书工具集.
 email: aaron@nonobo.com
 executables:
 - ebook_tools
+- para_import_mongo
 extensions: []
 extra_rdoc_files: []
 files:
 - README
 - CHANGELOG
 - bin/ebook_tools
+- bin/para_import_mongo
 - lib/ebook_tools.rb
 - lib/extract_book_struct.rb
 - lib/header_detect.rb
@@ -140,6 +158,7 @@ files:
 - lib/txt.rb
 - lib/epub.rb
 - lib/utils.rb
+- lib/paras_in_mongo.rb
 - ebook_tools.gemspec
 homepage: 
 licenses: []