#
# File:: data_string_analyser.rb
# Description::
#
# Author:: David Muir <david.muir@rockstarnorth.com>
# Date:: 11 November 2008
#

#----------------------------------------------------------------------------
# Uses
#----------------------------------------------------------------------------
require 'pipeline/os/path'
require 'pipeline/util/string'

#----------------------------------------------------------------------------
# Implementation
#----------------------------------------------------------------------------
module Pipeline
module ProjectUtil

	#
	# == Description
	# The main code dependency analyser entry-point.  This function takes a
	# list of files to analyse using the registered StringAnalyser classes
	# and invokes the user-block for each dependency it finds.
	#
	# The user-block should take two arguments, the first being the absolute
	# path of the file being analysed, the second being the string recognised
	# by the analyser as the dependency.
	#
	# The dependencies are also returned in a Hash with filenames as the
	# key and an Array of dependency String objects.
	#
	# === Example Usage
	#
	#   filelist = OS::find_files_recurse( root_path, '*.*' )
	#   ProjectUtil::code_dependency_analyser( filelist ) do |file, dependency|
	#		## User-code
	#   end
	#
	def ProjectUtil::code_dependency_analyser( files, analyser, &block )
		throw ArgumentError.new( "Invalid StringAnalyser specified (#{analyser.class})." ) \
			unless ( analyser.is_a?( StringAnalyser ) )
		
		dependencies = {}
		files.each do |filename|
			# Skip file if our analyser doesn't know about it.
			extension = OS::Path::get_extension( filename )
			next unless ( analyser.extensions.include?( extension ) )
			
			# Read all file data as text.
			data = []
			File.open( filename, 'r' ) do |fp|
				data = fp.readlines( )
			end
			
			# Loop through file content line by line.
			data.each do |line|
				# Check each pattern.
				analyser.patterns.each_pair do |regexp, index|
					md = line.match( regexp )
					next if ( md.nil? )
					
					# We have a match so we invoke the user block.
					yield( filename, md[index] ) if ( block_given? )
					
					# Complete Hash being returned.
					dependencies[filename] = [] unless ( dependencies.has_key?( filename ) )
					dependencies[filename] << md[index]
				end
			end
		end
		dependencies
	end	
	
	#
	# == Description
	# String analyser base class.  The constructor takes a file extension
	# argument and a Hash object mapping regular expressions to integer
	# indices within the matchdata Array for the (typically filename) 
	# dependency.
	#
	# In general, the index Hash value will be 1 unless you are reusing the
	# Regexp object from another script.
	#
	class StringAnalyser
		attr_reader :extensions
		attr_reader :patterns
		
		def initialize( extensions, patterns )
			throw ArgumentError.new( "Invalid Array of extensions (#{extensions.class})." ) \
				unless ( extensions.is_a?( Array ) )	
			throw ArgumentError.new( "Invalid Hash of patterns (#{patterns.class})." ) \
				unless ( patterns.is_a?( Hash ) )	
			@extensions = extensions
			@patterns = patterns
		end
		
		def to_s()
			"#{self.class}: #{@extensions.join( ' ' )} #{@patterns.join( '; ' )}"
		end
		
		def StringAnalyser::registered_analysers( )
			@@registered_analysers
		end
		
		def StringAnalyser::inherited( subclass )
			@@registered_analysers = [] if ( @registered_analysers.nil? )
			@@registered_analysers << subclass
		end
	end
	
	#
	# == Description
	# Ruby file string dependency analyser.
	#
	class RubyFileDependencyAnalyser < StringAnalyser
		
		def initialize( )
			super( EXTENSIONS, PATTERNS )
		end
		
		#----------------------------------------------------------------------
		# Private Data
		#----------------------------------------------------------------------
	private
		EXTENSIONS = [ 'rb', 'rbw' ]
		
		# Define some useful Regexp objects to fill in our patterns Hash.
		REGEXP_REQUIRE_DQ = /^require \"(.*)\"/i
		REGEXP_REQUIRE_SQ = /^require \'(.*)\'/i
		REGEXP_GEM_DQ = /^gem \"(.*)\"/i
		REGEXP_GEM_SQ = /^gem \'(.*)\'/i
		
		# Define our patterns hash, using our regexp constants above.
		PATTERNS = {
			REGEXP_REQUIRE_DQ => 1,
			REGEXP_REQUIRE_SQ => 1,
			REGEXP_GEM_DQ => 1,
			REGEXP_GEM_SQ => 1,
		}
	end
	
	#
	# == Description
	# MaxScript file string dependency analyser.
	#
	class MaxScriptFileDependencyAnalyser < StringAnalyser
		
		def initialize( )
			super( EXTENSIONS, PATTERNS )
		end
	
		#----------------------------------------------------------------------
		# Private Data
		#----------------------------------------------------------------------
	private
		EXTENSIONS = [ 'ms', 'mcr' ]
	
		# Define some useful Regexp objects to fill in our patterns Hash.
		REGEXP_FILEIN = /^filein \"(.*)\"/i
		REGEXP_INCLUDE = /^include \"(.*)\"/i
		
		# Define our patterns hash, using our regexp constants above.
		PATTERNS = {
			REGEXP_FILEIN => 1,
			REGEXP_INCLUDE => 1,
		}
	end
		
end # ProjectUtil module
end # Pipeline module

if ( __FILE__ == $0 ) then
	
	begin
		require 'pipeline/os/file'
		include Pipeline
		
		files = OS::FindEx::find_files_recurse( 'x:/pipedev/tools/dcc/current/max2009/scripts/*.ms' )
		analyser = ProjectUtil::MaxScriptFileDependencyAnalyser.new( )
		dependencies = ProjectUtil::code_dependency_analyser( files, analyser )
		dependencies.each do |filename, deps|
			puts "#{filename} dependencies:"
			deps.each do |dep|
				puts "\t#{dep}"
			end
		end
		puts "#{dependencies.size} files with dependencies."
		
	rescue Exception => ex
		
		puts "Unhandled exception: #{ex.message}"
		puts ex.backtrace.join( "\n" )
	end
end

# code_dependency_analyser.rb