cdb-crawlr 0.2.1 → 0.3.0

data/bin/cdb

@@ -1,94 +1,106 @@
-#!/usr/bin/env ruby
-
-require 'cdb-crawlr'
-# load 'lib/cdb-crawlr.rb'
-require 'optparse'
-
-$cli = CDB::CLI.new
-
-def print_help(opt = @global, error=nil)
-  puts(error.to_s+"\n\n") if error
-  puts opt
-  exit 1
-end
-
-@global = OptionParser.new do |opts|
-  opts.banner = "Usage: cdb [-h|--help] [-v|--version] <COMMAND> <TYPE> [<ARGS>]"
-
-  opts.on("-h", "--help", "Display this screen"){ print_help }
-  opts.on("-v", "--version", "Show version information") do
-    puts "cdb #{CDB::VERSION}"; exit
-  end
-
-  opts.separator "\nCOMMANDS:"
-  opts.separator "    search      Search for entries of a given TYPE matching QUERY"
-  opts.separator "    show        Show details of an entry using a CDB_ID obtained from search"
-end
-
-@search = OptionParser.new do |opts|
-  opts.banner = "Search for entries of a given TYPE matching QUERY\n"+
-                "Usage: cdb search [-h|--help] <TYPE> <QUERY>"
-
-  opts.on("-h", "--help", "Display this screen"){ print_help opts }
-
-  opts.separator "\nTYPES:"
-  opts.separator "    issue       Search comic issue names for given QUERY"
-  opts.separator "    series      Search comic series names for given QUERY"
-end
-
-@show = OptionParser.new do |opts|
-  opts.banner = "Show details of an entry using a CDB_ID obtained from search\n"+
-                "Usage: cdb show [-h|--help] <TYPE> <CDB_ID>"
-
-  opts.on("-h", "--help", "Display this screen"){ print_help opts }
-
-  opts.separator "\nTYPES:"
-  opts.separator "    series      Get all available details of a comic series"
-end
-
-@command_opts = {
-  'search' => @search,
-  'show' => @show
-}
-
-# Parse global flags
-begin
-  @global.order!
-rescue OptionParser::InvalidOption => e
-  puts e; print_help
-end
-
-# Pop and verify command
-begin
-  command = ARGV.shift
-  $cli[:command] = command
-  command_opt = @command_opts[$cli[:command]]
-rescue
-  error = "invalid COMMAND: #{command}" unless command.to_s.empty?
-  print_help @global, error
-end
-
-# Parse command flags
-begin
-  command_opt.order!
-rescue OptionParser::InvalidOption, OptionParser::MissingArgument => e
-  print_help command_opt, e
-end
-
-# Pop and verify type
-begin
-  type = ARGV.shift
-  $cli[:type] = type
-rescue
-  error = "invalid TYPE: #{type}" unless type.to_s.empty?
-  print_help command_opt, error
-end
-
-# Verify args
-begin
-  $cli[:args] = ARGV.join(' ')
-rescue => e
-  print_help command_opt, e
-end
-
-$cli.execute
+#!/usr/bin/env ruby
+
+require 'cdb-crawlr'
+# load 'lib/cdb-crawlr.rb'
+require 'optparse'
+
+$cli = CDB::CLI.new
+
+def print_help(opt = @global, error=nil)
+  puts "cdb: #{error}\n" if error && !error.to_s.empty?
+  puts opt
+  exit 1
+end
+
+@global = OptionParser.new do |opts|
+  opts.banner = "Usage: cdb [-v|--version] <COMMAND> [<ARGS>]"
+
+  opts.on("-h", "--help", "Display this screen"){ print_help }
+  opts.on("-v", "--version", "Show version information") do
+    puts "cdb #{CDB::VERSION}"; exit
+  end
+
+  opts.separator "\nCOMMANDS:"
+  opts.separator "    rename      Rename a directory of comics according to series data"
+  opts.separator "    search      Search for entries of a given TYPE matching QUERY"
+  opts.separator "    show        Show details of an entry using a CDB_ID obtained from search"
+end
+
+@search = OptionParser.new do |opts|
+  opts.banner = "Usage: cdb search <TYPE> <QUERY>"
+
+  opts.on("-h", "--help", "Display this screen"){ print_help opts }
+
+  opts.separator "\nTYPES:"
+  opts.separator "    issue       Search comic issue names for given QUERY"
+  opts.separator "    series      Search comic series names for given QUERY"
+end
+
+@show = OptionParser.new do |opts|
+  opts.banner = "Usage: cdb show <TYPE> <CDB_ID>"
+
+  opts.on("-h", "--help", "Display this screen"){ print_help opts }
+
+  opts.separator "\nTYPES:"
+  opts.separator "    series      Get all available details of a comic series"
+end
+
+@rename = OptionParser.new do |opts|
+  opts.banner = "Usage: cdb rename [-f|--force] <PATH> <CDB_ID>"
+
+  opts.on("-h", "--help", "Display this screen"){ print_help opts }
+  opts.on("-f", "--force", "Perform the rename without any confirmations"){ $cli[:force] = true }
+  opts.on("-i", "--ignore", "Ignore warnings about unknown and misformatted issue numbers"){ $cli[:ignore] = true }
+end
+
+@command_opts = {
+  'search' => @search,
+  'show' => @show,
+  'rename' => @rename
+}
+
+# Parse global flags
+begin
+  @global.order!
+rescue OptionParser::InvalidOption => e
+  puts e; print_help
+end
+
+# Pop and verify command
+begin
+  command = ARGV.shift
+  $cli[:command] = command
+  command_opt = @command_opts[$cli[:command]]
+rescue
+  error = "invalid COMMAND: #{command}" unless command.to_s.empty?
+  print_help @global, error
+end
+
+# Parse command flags
+begin
+  command_opt.order!
+rescue OptionParser::InvalidOption, OptionParser::MissingArgument => e
+  print_help command_opt, e
+end
+
+# Pop and verify third argument
+begin
+  next_arg = ARGV.shift
+  case $cli[:command]
+  when 'search, show'
+    $cli[:type] = next_arg
+  when 'rename'
+    $cli[:path] = next_arg
+  end
+rescue => e
+  print_help command_opt, e
+end
+
+# Verify args
+begin
+  $cli[:args] = ARGV.join(' ')
+rescue => e
+  print_help command_opt, e
+end
+
+$cli.execute

data/lib/cdb-crawlr.rb

@@ -1,51 +1,52 @@
-require 'json'
-require 'nokogiri'
-require 'open-uri'
-
-$:.unshift(File.dirname(__FILE__))
-
-require 'cdb/cli'
-require 'cdb/struct'
-require 'cdb/issue'
-require 'cdb/series'
-
-module CDB
-  VERSION = '0.2.1'
-
-  BASE_URL = 'http://www.comicbookdb.com'
-  REQUEST_HEADERS = {'Connection' => 'keep-alive'}
-  SEARCH_PATH = 'search.php'
-
-  class << self; attr
-
-    def search(query, type='FullSite')
-      data = URI.encode_www_form(
-        form_searchtype: type,
-        form_search: query
-      )
-      url = "#{BASE_URL}/#{SEARCH_PATH}?#{data}"
-      doc = read_page(url)
-      node = doc.css('h2:contains("Search Results")').first.parent
-      {
-        :series => CDB::Series.parse_results(node),
-        :issues => CDB::Issue.parse_results(node)
-      }
-    end
-
-    def show(id, type)
-      data = URI.encode_www_form('ID' => id)
-      url = "#{BASE_URL}/#{type::WEB_PATH}?#{data}"
-      page = read_page(url)
-      type.parse_data(id, page)
-    end
-
-  private
-
-    def read_page(url)
-      content = open(url, REQUEST_HEADERS).read
-      content.force_encoding('ISO-8859-1').encode!('UTF-8')
-      Nokogiri::HTML(content)
-    end
-
-  end
-end
+require 'json'
+require 'nokogiri'
+require 'open-uri'
+
+$:.unshift(File.dirname(__FILE__))
+
+require 'cdb/cli'
+require 'cdb/renamer'
+require 'cdb/struct'
+require 'cdb/issue'
+require 'cdb/series'
+
+module CDB
+  VERSION = '0.3.0'
+
+  BASE_URL = 'http://www.comicbookdb.com'
+  REQUEST_HEADERS = {'Connection' => 'keep-alive'}
+  SEARCH_PATH = 'search.php'
+
+  class << self; attr
+
+    def search(query, type='FullSite')
+      data = URI.encode_www_form(
+        form_searchtype: type,
+        form_search: query
+      )
+      url = "#{BASE_URL}/#{SEARCH_PATH}?#{data}"
+      doc = read_page(url)
+      node = doc.css('h2:contains("Search Results")').first.parent
+      {
+        :series => CDB::Series.parse_results(node),
+        :issues => CDB::Issue.parse_results(node)
+      }
+    end
+
+    def show(id, type)
+      data = URI.encode_www_form('ID' => id)
+      url = "#{BASE_URL}/#{type::WEB_PATH}?#{data}"
+      page = read_page(url)
+      type.parse_data(id, page)
+    end
+
+  private
+
+    def read_page(url)
+      content = open(url, REQUEST_HEADERS).read
+      content.force_encoding('ISO-8859-1').encode!('UTF-8')
+      Nokogiri::HTML(content)
+    end
+
+  end
+end

data/lib/cdb/cli.rb

@@ -1,61 +1,84 @@
-require 'pp'
-
-module CDB
-  class CLI
-    COMMANDS = %w[search show]
-    TYPES = %w[series issue issues]
-
-    def initialize(options={})
-      @options = options
-    end
-
-    def [](k)
-      @options[k]
-    end
-
-    def []=(k, v)
-      v = v.to_s.strip
-      case k
-      when :command
-        v = v.downcase
-        raise unless COMMANDS.include?(v)
-      when :type
-        v = v.downcase
-        if self[:command] == 'show'
-          # remove when "show issue" is supported
-          raise unless v == 'series'
-        else
-          raise unless TYPES.include?(v)
-        end
-      when :args
-        raise "invalid args" if v.empty?
-      end
-      @options[k] = v
-    end
-
-    def execute
-      send self[:command]
-    end
-
-  private
-
-    def search
-      case self[:type]
-      when 'series'
-        CDB::Series.search(self[:args]).each{|r| puts r.to_json}
-      when 'issue', 'issues'
-        CDB::Issue.search(self[:args]).each{|r| puts r.to_json}
-      end
-    end
-
-    def show
-      case self[:type]
-      when 'series'
-        res = CDB::Series.show(self[:args])
-        res.issues.each{|i| i.series=nil}
-        puts res.to_json(array_nl:"\n", object_nl:"\n", indent:'  ')
-      end
-    end
-
-  end
-end
+require 'pp'
+
+module CDB
+  class CLI
+    COMMANDS = %w[search show rename]
+    TYPES = %w[series issue issues]
+
+    def initialize(options={})
+      @options = options
+    end
+
+    def [](k)
+      @options[k]
+    end
+
+    def []=(k, v)
+      v = v.to_s.strip
+      begin
+        send("#{k}=", v)
+      rescue NoMethodError
+        @options[k] = v
+      end
+    end
+
+    def execute
+      send self[:command]
+    end
+
+  private
+
+    def search
+      case self[:type]
+      when 'series'
+        CDB::Series.search(self[:args]).each{|r| puts r.to_json}
+      when 'issue', 'issues'
+        CDB::Issue.search(self[:args]).each{|r| puts r.to_json}
+      end
+    end
+
+    def show
+      case self[:type]
+      when 'series'
+        res = CDB::Series.show(self[:args])
+        res.issues.each{|i| i.series=nil}
+        puts res.to_json(array_nl:"\n", object_nl:"\n", indent:'  ')
+      end
+    end
+
+    def rename
+      renamer = CDB::Renamer.new(@options)
+      renamer.execute
+    end
+
+    def args=(v)
+      raise "invalid args" if v.empty?
+      @options[:args] = v
+    end
+
+    def command=(v)
+      v = v.downcase
+      raise unless COMMANDS.include?(v)
+      @options[:command] = v
+    end
+
+    def type=(v)
+      v = v.downcase
+      error = "invalid TYPE: #{v}" unless v.empty?
+      if @options[:command] == 'show'
+        # remove when "show issue" is supported
+        raise error.to_s unless v == 'series'
+      else
+        raise error.to_s unless TYPES.include?(v)
+      end
+      @options[:type] = v
+    end
+
+    def path=(v)
+      error = "#{v}: No such directory" unless v.empty?
+      raise error.to_s unless File.directory?(v)
+      @options[:path] = v
+    end
+
+  end
+end

data/lib/cdb/renamer.rb

@@ -0,0 +1,117 @@
+module CDB
+  class Renamer
+    EXTENSIONS = %w[cbz cbr]
+    ISSUE_NUM = '[\d\.]+\w?'
+    INPUT_FORMAT = /#(#{ISSUE_NUM})/
+    OUTPUT_FORMAT  = "%{series} #%{padded_num} (%{cover_date})"
+
+    def initialize(options)
+      @path = options[:path]
+      @cdb_id = options[:args]
+      @force = options[:force]
+      @ignore = options[:ignore]
+    end
+
+    def execute
+      @rename_map =
+        files.each_with_object({}) do |filename, map|
+          map[filename]= transform(filename)
+        end.select{|k,v| v}
+
+      do_rename if verify_map
+    end
+
+  private
+
+    def do_rename
+      Dir.chdir(@path) do
+        @rename_map.each do |source, destination|
+          next if source == destination
+          puts "#{pad(source)} => #{destination}"
+          if @force
+            %x[ mv "#{source}" "#{destination}" ]
+          end
+        end
+      end
+    end
+
+    def verify_map
+      dups = @rename_map.select do |k,v|
+        @rename_map.values.count(v) > 1
+      end
+      dups.keys.uniq.each do |k|
+        padded = pad(k, dups.keys.map(&:length).max)
+        puts "ERROR: output name clash: #{padded} => #{dups[k]}"
+      end
+      dups.empty?
+    end
+
+    def transform(filename)
+      return unless num = parse_issue_num(filename)
+      if issue = issues[num]
+        generate_output(filename, issue)
+      else
+        puts "WARNING: #{filename}: unknown issue: #{num}"
+      end
+    end
+
+    def parse_issue_num(filename)
+      if match = filename.match(INPUT_FORMAT)
+        num = match[1].gsub(/^0+|\.$/,'')
+        num = '0' if num == ''
+        num
+      else
+        puts "WARNING: #{filename}: invalid input format" unless @ignore
+      end
+    end
+
+    def generate_output(filename, issue)
+      json = issue.as_json
+      json[:series] = issue.series.name
+      json[:padded_num] = pad_num(issue.num)
+      output = OUTPUT_FORMAT % json
+      sanitize(output + File.extname(filename))
+    end
+
+    def sanitize(filename)
+      filename.gsub(/[:]/, ' -')
+        .gsub(/[\/\\<>]/, '-')
+        .gsub(/[\?\*|"]/, '_')
+    end
+
+    def pad(file, max=nil)
+      max ||= files.map(&:length).max
+      file + (' '*(max-file.length))
+    end
+
+    def pad_num(num, max=nil)
+      max ||= max_num_length
+      '0'*(max-num.to_s.length)+num.to_s
+    end
+
+    def files
+      Dir.chdir(@path) do
+        @files ||= EXTENSIONS
+          .map{|e| Dir["*.#{e}"]}.flatten
+          .select{|f| File.file?(f)}.sort
+      end
+      @files
+    end
+
+    def max_num_length
+      @max_num ||= files.map{|f| parse_issue_num(f).length}.max
+    end
+
+    def series
+      @series ||= CDB::Series.show(@cdb_id)
+    end
+
+    def issues
+      # Only act on issues - not TPB, HC, or anything else
+      @issues ||= Hash[series.issues
+        .select{|i| i.num.match(/^#{ISSUE_NUM}$/)}
+        .map{|i| [i.num.to_s, i]}]
+    end
+
+  end
+end

data/lib/cdb/series.rb

@@ -30,7 +30,7 @@ module CDB
         start_d, end_d = dates.split('-').map(&:strip)
 
         series = new(
-          :cdb_id => id,
+          :cdb_id => id.to_i,
           :name => page.css('.page_headline').first.text.strip,
           :publisher => page.css('a[href^="publisher.php"]').first.text.strip,
           :imprint => (page.css('a[href^="imprint.php"]').first.text.strip rescue nil),

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: cdb-crawlr
 version: !ruby/object:Gem::Version
-  version: 0.2.1
+  version: 0.3.0
   prerelease: 
 platform: ruby
 authors:
@@ -9,11 +9,11 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2012-11-04 00:00:00.000000000Z
+date: 2012-11-15 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: nokogiri
-  requirement: &25288420 !ruby/object:Gem::Requirement
+  requirement: !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -21,7 +21,12 @@ dependencies:
         version: '0'
   type: :runtime
   prerelease: false
-  version_requirements: *25288420
+  version_requirements: !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: '0'
 description: cdb-crawlr is a Ruby gem and command-line tool for querying ComicBookDB.com
 email:
 - sgt.floydpepper@gmail.com
@@ -32,6 +37,7 @@ extra_rdoc_files: []
 files:
 - lib/cdb/cli.rb
 - lib/cdb/issue.rb
+- lib/cdb/renamer.rb
 - lib/cdb/series.rb
 - lib/cdb/struct.rb
 - lib/cdb-crawlr.rb
@@ -56,9 +62,8 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
 requirements: []
 rubyforge_project: 
-rubygems_version: 1.8.10
+rubygems_version: 1.8.24
 signing_key: 
 specification_version: 3
 summary: Ruby gem and command-line tool for querying ComicBookDB.com
 test_files: []
-has_rdoc: