validate-website 0.5.7 → 0.6.0

data/lib/validate_website/option_parser.rb

@@ -0,0 +1,133 @@
+# encoding: utf-8
+require 'optparse'
+
+module ValidateWebsite
+  class Parser
+    DEFAULT_OPTS_CRAWL = {
+      :site              => 'http://localhost:3000/',
+      :markup_validation => true,
+      :exclude           => nil,
+      :file              => nil,
+      # log not found url (404 status code)
+      :not_found         => false,
+      # internal verbose for ValidateWebsite
+      :validate_verbose  => false,
+      :quiet             => false,
+
+      # Anemone options see anemone/lib/anemone/core.rb
+      :verbose           => false,
+      :cookies           => nil,
+      :accept_cookies    => true,
+      :redirect_limit    => 0,
+    }
+
+    DEFAULT_OPTS_STATIC = {
+      :pattern           => '**/*.html',
+      :file              => nil,
+      :validate_verbose  => false,
+      :quiet             => false,
+    }
+
+    def self.parse(options, type)
+      if const_defined?("DEFAULT_OPTS_#{type.to_s.upcase}")
+        @@default_opts = const_get("DEFAULT_OPTS_#{type.to_s.upcase}")
+        if Array === options
+          send("command_line_parse_#{type}", options)
+        else
+          @@default_opts.merge(options)
+        end
+      else
+        raise ArgumentError, "Unknown options type : #{type}"
+      end
+    end
+
+    def self.command_line_parse_crawl(args)
+      options = {}
+      opts = OptionParser.new do |o|
+        o.set_summary_indent('  ')
+        o.banner =    'Usage: validate-website [OPTIONS]'
+        o.define_head 'validate-website - Web crawler for checking the ' +
+          'validity of your documents'
+        o.separator   ''
+
+        o.on("-s", "--site 'SITE'", String,
+             "Website to crawl (Default: #{@@default_opts[:site]})") { |v|
+          options[:site] = v
+        }
+        o.on("-u", "--user-agent 'USERAGENT'", String,
+             "Change user agent") { |v|
+          options[:user_agent] = v
+        }
+        o.on("-e", "--exclude 'EXCLUDE'", String,
+             "Url to exclude (ex: 'redirect|news')") { |v|
+          options[:exclude] = v
+        }
+        o.on("-f", "--file 'FILE'", String,
+             "Save not well formed or not found urls") { |v|
+          options[:file] = v
+        }
+
+        o.on("-c", "--cookies 'COOKIES'", String,
+             "Set defaults cookies") { |v|
+          options[:cookies] = v
+        }
+
+        o.on("-m", "--[no-]markup-validation",
+             "Markup validation (Default: #{@@default_opts[:markup_validation]})") { |v|
+          options[:markup_validation] = v
+        }
+        o.on("-n", "--not-found",
+             "Log not found url (Default: #{@@default_opts[:not_found]})") { |v|
+          options[:not_found] = v
+        }
+        o.on("-v", "--verbose",
+             "Show validator errors (Default: #{@@default_opts[:validate_verbose]})") { |v|
+          options[:validate_verbose] = v
+        }
+        o.on("-q", "--quiet",
+             "Only report errors (Default: #{@@default_opts[:quiet]})") { |v|
+          options[:quiet] = v
+        }
+        o.on("-d", "--debug",
+             "Show anemone log (Default: #{@@default_opts[:verbose]})") { |v|
+          options[:verbose] = v
+        }
+
+        o.separator ""
+        o.on_tail("-h", "--help", "Show this help message.") { puts o; exit }
+      end
+      opts.parse!(args)
+      @@default_opts.merge(options)
+    end
+
+    def self.command_line_parse_static(args)
+      options = {}
+      opts = OptionParser.new do |o|
+        o.set_summary_indent('  ')
+        o.banner =    'Usage: validate-website-static [OPTIONS]'
+        o.define_head 'validate-website-static - check the validity of ' +
+          'your documents'
+        o.separator   ''
+
+        o.on("-p", "--pattern 'PATTERN'", String,
+             "Change filenames pattern (Default: #{@@default_opts[:pattern]})") { |v|
+          options[:pattern] = v.strip
+        }
+        o.on("-f", "--file 'FILE'", String,
+             "Save not well formed urls") { |v|
+          options[:file] = v
+        }
+        o.on("-v", "--verbose",
+             "Show validator errors (Default: #{@@default_opts[:validate_verbose]})") { |v|
+          options[:validate_verbose] = v
+        }
+        o.on("-q", "--quiet",
+             "Only report errors (Default: #{@@default_opts[:quiet]})") { |v|
+          options[:quiet] = v
+        }
+      end
+      opts.parse!(args)
+      @@default_opts.merge(options)
+    end
+  end
+end

data/lib/validate_website/runner.rb

@@ -0,0 +1,35 @@
+require 'validate_website/core'
+
+module ValidateWebsite
+  class Runner
+    def self.trap_interrupt
+      trap('INT') do
+        STDERR.puts "\nExiting..."
+        exit!(1)
+      end
+    end
+
+    def self.run_crawl(args)
+      trap_interrupt
+      validate_website = ValidateWebsite::Core.new(args, :crawl)
+      validate_website.crawl
+      validate_website.exit_status
+    end
+
+    def self.run_static(args)
+      trap_interrupt
+      validate_website = ValidateWebsite::Core.new(args, :static)
+
+      files = Dir.glob(validate_website.options[:pattern])
+      files.each do |f|
+        next unless File.file?(f)
+
+        body = open(f).read
+        doc = Nokogiri::HTML(body)
+
+        validate_website.validate(doc, body, f)
+      end
+      validate_website.exit_status
+    end
+  end
+end

data/lib/validate_website/validator.rb

@@ -0,0 +1,69 @@
+# encoding: utf-8
+
+module ValidateWebsite
+  class Validator
+    XHTML_PATH = File.join(File.dirname(__FILE__), '..', '..', 'share', 'schemas')
+
+    attr_reader :original_doc, :body, :dtd, :doc, :namespace, :xsd, :errors
+
+    def initialize(original_doc, body)
+      @original_doc = original_doc
+      @body = body
+      @dtd = @original_doc.internal_subset
+      init_namespace(@dtd)
+      @errors = []
+      @errors << 'Unknown document' if @namespace.nil?
+
+      if @errors.empty?
+        if @dtd_uri && @body.match(@dtd_uri.to_s)
+          document = @body.sub(@dtd_uri.to_s, @namespace + '.dtd')
+        else
+          document = @body
+        end
+        @doc = Dir.chdir(XHTML_PATH) do
+          Nokogiri::XML(document) { |cfg|
+            cfg.noent.dtdload.dtdvalid
+          }
+        end
+
+        # http://www.w3.org/TR/xhtml1-schema/
+        @xsd = Dir.chdir(XHTML_PATH) do
+          if File.exists?(@namespace + '.xsd')
+            Nokogiri::XML::Schema(File.read(@namespace + '.xsd'))
+          end
+        end
+
+        if @xsd
+          # have the xsd so use it
+          @errors = @xsd.validate(@doc)
+        else
+          # dont have xsd fall back to dtd
+          @doc = Dir.chdir(XHTML_PATH) do
+            Nokogiri::HTML.parse(document)
+          end
+          @errors = @doc.errors
+        end
+      end
+
+    rescue Nokogiri::XML::SyntaxError => e
+      # http://nokogiri.org/tutorials/ensuring_well_formed_markup.html
+      @errors << e
+    end
+
+    def valid?
+      @errors.length == 0
+    end
+
+    private
+    def init_namespace(dtd)
+      if dtd.system_id
+        dtd_uri = URI.parse(dtd.system_id)
+        if dtd.system_id && dtd_uri.path
+          @dtd_uri = dtd_uri
+          # http://www.w3.org/TR/xhtml1/#dtds
+          @namespace = File.basename(@dtd_uri.path, '.dtd')
+        end
+      end
+    end
+  end
+end

data/man/man1/validate-website-static.1

@@ -0,0 +1,82 @@
+'\" t
+.\"     Title: validate-website-static
+.\"    Author: [see the "AUTHOR" section]
+.\" Generator: DocBook XSL Stylesheets v1.75.2 <http://docbook.sf.net/>
+.\"      Date: 12/24/2010
+.\"    Manual: \ \&
+.\"    Source: \ \&
+.\"  Language: English
+.\"
+.TH "VALIDATE\-WEBSITE\-S" "1" "12/24/2010" "\ \&" "\ \&"
+.\" -----------------------------------------------------------------
+.\" * Define some portability stuff
+.\" -----------------------------------------------------------------
+.\" ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+.\" http://bugs.debian.org/507673
+.\" http://lists.gnu.org/archive/html/groff/2009-02/msg00013.html
+.\" ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+.ie \n(.g .ds Aq \(aq
+.el       .ds Aq '
+.\" -----------------------------------------------------------------
+.\" * set default formatting
+.\" -----------------------------------------------------------------
+.\" disable hyphenation
+.nh
+.\" disable justification (adjust text to left margin only)
+.ad l
+.\" -----------------------------------------------------------------
+.\" * MAIN CONTENT STARTS HERE *
+.\" -----------------------------------------------------------------
+.SH "NAME"
+validate-website-static \- check the validity of your documents
+.SH "SYNOPSIS"
+.sp
+\fBvalidate\-website\-static\fR [\fIOPTIONS\fR]
+.SH "DESCRIPTION"
+.sp
+validate\-website\-static check the markup validity of your local documents\&.
+.SH "OPTIONS"
+.PP
+\fB\-p\fR, \fB\-\-pattern\fR \fIPATTERN\fR
+.RS 4
+Change filenames pattern (Default: **/*\&.html)
+.RE
+.PP
+\fB\-f\fR, \fB\-\-file\fR \fIFILE\fR
+.RS 4
+Save not well formed urls
+.RE
+.PP
+\fB\-v\fR, \fB\-\-verbose\fR
+.RS 4
+Show detail of validator errors (Default: false)\&.
+.RE
+.PP
+\fB\-q\fR, \fB\-\-quiet\fR
+.RS 4
+Only report errors (Default: false)\&.
+.RE
+.PP
+\fB\-h\fR, \fB\-\-help\fR
+.RS 4
+Show help message and exit\&.
+.RE
+.SH "EXIT STATUS"
+.PP
+0
+.RS 4
+Markup is valid\&.
+.RE
+.PP
+64
+.RS 4
+Not valid markup found\&.
+.RE
+.SH "AUTHOR"
+.sp
+Laurent Arnoud <laurent@spkdev\&.net>
+.SH "LICENSE"
+.sp
+The MIT License
+.sp
+Copyright (c) 2009\-2010 Laurent Arnoud <laurent@spkdev\&.net>