mannie-taverna-scufl 0.6.0

data/lib/scufl/dot.rb

@@ -0,0 +1,248 @@
+module Scufl
+  
+  # This class enables you to write the script will will be used by dot
+  # (which is part of GraphViz[http://www.graphviz.org/Download.php])
+  # to generate the image showing the structure of a given model.
+  # To get started quickly, you could try:
+  #   out_file = File.new("path/to/file/you/want/the/dot/script/to/be/written", "w+")
+  #   workflow = File.new("path/to/workflow/file", "r").read
+  #   model = Scufl::Parser.new.parse(workflow)
+  #   Scufl::Dot.new.write_dot(out_file, model)
+  #   `dot -Tpng -o"path/to/the/output/image" #{out_file.path}`
+  class Dot
+
+    @@processor_colours = {
+      'apiconsumer' => 'palegreen',
+      'beanshell' => 'burlywood2',
+      'biomart' => 'lightcyan2',                      
+      'local' => 'mediumorchid2',
+      'biomobywsdl' => 'darkgoldenrod1',
+      'biomobyobject' => 'gold',
+      'biomobyparser' => 'white',
+      'inferno' => 'violetred1',
+      'notification' => 'mediumorchid2',
+      'rdfgenerator' => 'purple',
+      'rserv' => 'lightgoldenrodyellow',
+      'seqhound' => '#836fff',
+      'soaplabwsdl' => 'lightgoldenrodyellow',
+      'stringconstant' => 'lightsteelblue',
+      'talisman' => 'plum2',
+      'bsf' => 'burlywood2',
+      'abstractprocessor' => 'lightgoldenrodyellow',
+      'rshell' => 'lightgoldenrodyellow',
+      'arbitrarywsdl' => 'darkolivegreen3',
+      'workflow' => 'crimson'}
+    
+    @@fill_colours = %w{white aliceblue antiquewhite beige}
+    
+    @@ranksep = '0.22'
+    @@nodesep = '0.05'
+    
+    # Creates a new dot object for interaction.
+    def initialize
+      # @port_style IS CURRENTLY UNUSED. IGNORE!!!
+      @port_style = 'none' # 'all', 'bound' or 'none'
+    end
+    
+    # Writes to the given stream (File, StringIO, etc) the script to generate
+    # the image showing the internals of the given workflow model.  
+    # === Usage
+    #   stream = File.new("path/to/file/you/want/the/dot/script/to/be/written", "w+")
+    #   workflow = .......
+    #   model = Scufl::Parser.new.parse(workflow)
+    #   Scufl::Dot.new.write_dot(stream, model)
+    def write_dot(stream, model)
+      stream.puts 'digraph scufl_graph {'
+      stream.puts ' graph ['
+      stream.puts '  style=""'
+      stream.puts '  labeljust="left"'
+      stream.puts '  clusterrank="local"'
+      stream.puts "  ranksep=\"#@@ranksep\""
+      stream.puts "  nodesep=\"#@@nodesep\""
+      stream.puts ' ]'
+      stream.puts
+      stream.puts ' node ['
+      stream.puts '  fontname="Helvetica",'
+      stream.puts '  fontsize="10",'
+      stream.puts '  fontcolor="black", '
+      stream.puts '  shape="box",'
+      stream.puts '  height="0",'
+      stream.puts '  width="0",'
+      stream.puts '  color="black",'
+      stream.puts '  fillcolor="lightgoldenrodyellow",'
+      stream.puts '  style="filled"'
+      stream.puts ' ];'
+      stream.puts
+      stream.puts ' edge ['
+      stream.puts '  fontname="Helvetica",'
+      stream.puts '  fontsize="8",'
+      stream.puts '  fontcolor="black",'
+      stream.puts '  color="black"'
+      stream.puts ' ];'
+      write_workflow(stream, model)
+      stream.puts '}'
+      
+      stream.flush
+    end
+    
+    def write_workflow(stream, model, prefix="", name="", depth=0) # :nodoc:
+      if name != ""
+        stream.puts "subgraph cluster_#{prefix}#{name} {"
+        stream.puts " label=\"#{name}\""
+        stream.puts ' fontname="Helvetica"'
+        stream.puts ' fontsize="10"'
+        stream.puts ' fontcolor="black"'
+        stream.puts ' clusterrank="local"'
+        stream.puts " fillcolor=\"#{@@fill_colours[depth % @@fill_colours.length]}\""
+        stream.puts ' style="filled"'
+      end
+      model.processors.each {|processor| write_processor(stream, processor, prefix, depth)}
+      write_source_cluster(stream, model.sources, prefix)
+      write_sink_cluster(stream, model.sinks, prefix)
+      model.links.each {|link| write_link(stream, link, model, prefix)}
+      model.coordinations.each {|coordination| write_coordination(stream, coordination, model, prefix)}
+      if name != ""
+        stream.puts '}'
+      end
+    end
+    
+    def write_processor(stream, processor, prefix, depth) # :nodoc:
+      # nested workflows
+      if processor.model
+        write_workflow(stream, processor.model, prefix + processor.name, processor.name, depth.next)
+      else
+        stream.puts " \"#{prefix}#{processor.name}\" ["
+        stream.puts "  fillcolor=\"#{get_colour processor.type}\","
+        stream.puts '  shape="box",'
+        stream.puts '  style="filled",'
+        stream.puts '  height="0",'
+        stream.puts '  width="0",'
+        stream.puts "  label=\"#{processor.name}\""
+        stream.puts ' ];'
+      end
+    end
+    
+    def write_source_cluster(stream, sources, prefix) # :nodoc:
+      if sources.length > 0
+        stream.puts " subgraph cluster_#{prefix}sources {"
+        stream.puts '  style="dotted"'
+        stream.puts '  label="Workflow Inputs"'
+        stream.puts '  fontname="Helvetica"'
+        stream.puts '  fontsize="10"'
+        stream.puts '  fontcolor="black"'
+        stream.puts '  rank="same"'
+        stream.puts " \"#{prefix}WORKFLOWINTERNALSOURCECONTROL\" ["
+        stream.puts '  shape="triangle",'
+        stream.puts '  width="0.2",'
+        stream.puts '  height="0.2",'
+        stream.puts '  fillcolor="brown1"'
+        stream.puts '  label=""'
+        stream.puts ' ]'
+        sources.each {|source| write_source(stream, source, prefix)}
+        stream.puts ' }'
+      end
+    end
+    
+    def write_source(stream, source, prefix) # :nodoc:
+      stream.puts " \"#{prefix}WORKFLOWINTERNALSOURCE_#{source.name}\" ["
+      stream.puts '   shape="box",'
+      stream.puts "   label=\"#{source.name}\""
+      stream.puts '   width="0",'
+      stream.puts '   height="0",'
+      stream.puts '   fillcolor="skyblue"'
+      stream.puts ' ]' 
+    end
+    
+    def write_sink_cluster(stream, sinks, prefix) # :nodoc:
+      if sinks.length > 0
+        stream.puts " subgraph cluster_#{prefix}sinks {"
+        stream.puts '  style="dotted"'
+        stream.puts '  label="Workflow Outputs"'
+        stream.puts '  fontname="Helvetica"'
+        stream.puts '  fontsize="10"'
+        stream.puts '  fontcolor="black"'
+        stream.puts '  rank="same"'
+        stream.puts " \"#{prefix}WORKFLOWINTERNALSINKCONTROL\" ["
+        stream.puts '  shape="invtriangle",'
+        stream.puts '  width="0.2",'
+        stream.puts '  height="0.2",'
+        stream.puts '  fillcolor="chartreuse3"'
+        stream.puts '  label=""'
+        stream.puts ' ]'
+        sinks.each {|sink| write_sink(stream, sink, prefix)}
+        stream.puts ' }'
+      end
+    end
+    
+    def write_sink(stream, sink, prefix) # :nodoc:
+      stream.puts " \"#{prefix}WORKFLOWINTERNALSINK_#{sink.name}\" ["
+      stream.puts '   shape="box",'
+      stream.puts "   label=\"#{sink.name}\""
+      stream.puts '   width="0",'
+      stream.puts '   height="0",'
+      stream.puts '   fillcolor="lightsteelblue2"'
+      stream.puts ' ]'     
+    end
+    
+    def write_link(stream, link, model, prefix) # :nodoc:
+      if model.sources.select{|s| s.name == link.source} != []
+        stream.write " \"#{prefix}WORKFLOWINTERNALSOURCE_#{link.source}\""
+      else 
+        processor = model.processors.select{|p| p.name == link.source.split(':')[0]}[0]
+        if processor.model
+          stream.write " \"#{prefix}#{processor.name}WORKFLOWINTERNALSINK_#{link.source.split(':')[1]}\""
+        else
+          stream.write " \"#{prefix}#{processor.name}\""
+        end
+      end
+      stream.write '->'
+      if model.sinks.select{|s| s.name == link.sink} != []
+        stream.write "\"#{prefix}WORKFLOWINTERNALSINK_#{link.sink}\""
+      else 
+        processor = model.processors.select{|p| p.name == link.sink.split(':')[0]}[0]
+        if processor.model
+          stream.write "\"#{prefix}#{processor.name}WORKFLOWINTERNALSOURCE_#{link.sink.split(':')[1]}\""
+        else
+          stream.write "\"#{prefix}#{processor.name}\""
+        end
+      end
+      stream.puts ' ['
+      stream.puts ' ];'
+    end
+    
+    def write_coordination(stream, coordination, model, prefix) # :nodoc:
+      stream.write " \"#{prefix}#{coordination.controller}"
+      processor = model.processors.select{|p| p.name == coordination.controller}[0]
+      if processor.model
+        stream.write 'WORKFLOWINTERNALSINKCONTROL'
+      end
+      stream.write '"->"'
+      stream.write "#{prefix}#{coordination.target}\""
+      processor = model.processors.select{|p| p.name == coordination.target}[0]
+      if processor.model
+        stream.write 'WORKFLOWINTERNALSOURCECONTROL'
+      end
+      stream.puts ' ['
+      stream.puts '  color="gray",'
+      stream.puts '  arrowhead="odot",'
+      stream.puts '  arrowtail="none"'
+      stream.puts ' ];'
+    end
+    
+    def get_colour(processor_name) # :nodoc:
+      colour = @@processor_colours[processor_name]
+      if colour
+        colour
+      else 
+        'white'
+      end  
+    end
+    
+    # Returns true if the given name is a processor; false otherwise
+    def Dot.is_processor?(processor_name)
+      true if @@processor_colours[processor_name]
+    end
+    
+  end
+  
+end

data/lib/scufl/model.rb

@@ -0,0 +1,153 @@
+# This is the module containing the Scufl model implementation i.e. the model structure/definition and all its internals.
+
+module Scufl # :nodoc:
+  
+  # The model for a given Taverna 1 workflow.
+  class Model
+    # This returns a WorkflowDescription object.
+    attr_reader :description
+    
+    # Retrieve the list of processors specific to the workflow.
+    # Does not include those from nested workflows.
+    attr_reader :processors
+    
+    # Retrieve the list of datalinks specific to the workflow.
+    # Does not include those from nested workflows.
+    attr_reader :links
+    
+    # Retrieve the list of sources specific to the workflow.
+    # Does not include those from nested workflows.
+    attr_reader :sources
+    
+    # Retrieve the list of sinks specific to the workflow.
+    # Does not include those from nested workflows.
+    attr_reader :sinks
+    
+    # Retrieve the list of coordinations specific to the workflow.
+    # Does not include those from nested workflows.
+    attr_reader :coordinations
+    
+    # The list of any dependencies that have been found inside the workflow.
+    # Does not include those from nested workflows.
+    attr_accessor :dependencies
+    
+    # Creates an empty model for a Taverna 1 workflow.
+    def initialize
+      @description = WorkflowDescription.new
+      @processors = Array.new
+      @links = Array.new
+      @sources = Array.new
+      @sinks = Array.new
+      @coordinations = Array.new
+    end
+    
+    # Retrieve ALL the beanshell processors within the workflow.
+    def beanshells
+      return get_beanshells(self, [])
+    end
+    
+    private
+    
+    def get_beanshells(given_model, beans_collected) # :nodoc:
+      wf_procs = given_model.processors.select { |x| x.type == "workflow" }
+      wf_procs.each { |x| get_beanshells(x.model, beans_collected) }
+      
+      bean_procs = given_model.processors.select { |b| b.type == "beanshell" }
+      bean_procs.each { |a| beans_collected << a }
+      
+      return beans_collected
+    end
+  end
+  
+  
+  
+  # This is the (shim) object within the workflow.  This can be a beanshell,
+  # a webservice, a workflow, etc...
+  class Processor
+    # A string containing name of the processor.
+    attr_accessor :name 
+    
+    # A string containing the description of the processor if available.  
+    # Returns nil otherwise.
+    attr_accessor :description
+    
+    # A string for the type of processor, e.g. beanshell, workflow, webservice, etc...
+    attr_accessor :type 
+    
+    # For processors that have type == "workflow", model is the the workflow 
+    # definition.  For all other processor types, model is nil.
+    attr_accessor :model
+    
+    # This only has a value in beanshell processors.  This is the actual script
+    # embedded with the processor which does all the "work"
+    attr_accessor :script
+    
+    # This is a list of inputs that the processor can take in.
+    attr_accessor :inputs
+    
+    # This is a list of outputs that the processor can produce.
+    attr_accessor :outputs
+  end
+
+
+  
+  # This contains basic descriptive information about the workflow model.
+  class WorkflowDescription
+    # The author of the workflow.
+    attr_accessor :author
+    
+    # The name/title of the workflow.
+    attr_accessor :title
+    
+    # A small piece of descriptive text for the workflow.
+    attr_accessor :description
+  end
+
+  
+  
+  # This represents a connection between any of the following pair of entities:
+  # {processor -> processor}, {workflow -> workflow}, {workflow -> processor}, 
+  # and {processor -> workflow}.
+  class Link
+    # The name of the source (the starting point of the connection).
+    attr_accessor :source
+    
+    # The name of the sink (the endpoint of the connection).
+    attr_accessor :sink
+  end
+
+
+
+  # This is a representation of the 'Run after...' function in Taverna
+  # where the selected processor or workflow is set to run after another.
+  class Coordination
+    # The name of the processor/workflow which is to run first.
+    attr_accessor :controller
+    
+    # The name of the processor/workflow which is to run after the controller.
+    attr_accessor :target
+  end
+
+  
+  
+  # This is the start node of a Link.  Each source has a name and a port
+  # which is seperated by a colon; ":".
+  # This is represented as "source of a processor:port_name".
+  # A string that does not contain a colon can often be returned, signifiying
+  # a workflow source as opposed to that of a processor.
+  class Source
+  	attr_accessor :name, :description
+  end
+  
+  
+  
+  # This is the start node of a Link.  Each sink has a name and a port
+  # which is seperated by a colon; ":".
+  # This is represented as "sink of a processor:port_name".
+  # A string that does not contain a colon can often be returned, signifiying
+  # a workflow sink as opposed to that of a processor.
+  class Sink
+	  attr_accessor :name, :description
+  end  	
+  
+end

data/lib/scufl/parser.rb

@@ -0,0 +1,162 @@
+require "rexml/document"
+
+module Scufl
+  
+  class Parser
+    # Returns the model for the given t2flow_file.
+    # The method accepts objects of classes File and String only.
+    # ===Usage
+    #   foo = ... # stuff to initialize foo here
+    #   bar = Scufl::Parser.new.parse(foo)
+    def parse(scufl)
+      document = REXML::Document.new(scufl)
+      
+      root = document.root
+      raise "Doesn't appear to be a workflow!" if root.name != "scufl"
+      version = root.attribute('version').value
+      
+      create_model(root, version)
+    end
+    
+    def create_model(element, version) # :nodoc:
+      model = Model.new
+      
+      element.each_element('s:workflowdescription') { |description|  set_description(model, description, version)}
+      element.each_element('s:processor')           { |processor|    add_processor(model, processor, version)}
+      element.each_element('s:link')                { |link|         add_link(model, link, version)}
+      element.each_element('s:source')              { |source|       add_source(model, source, version)}
+      element.each_element('s:sink')                { |sink|         add_sink(model, sink, version)}
+      element.each_element('s:coordination')        { |coordination| add_coordination(model, coordination, version)}
+      
+      return model   
+    end
+    
+    def add_coordination(model, element, version) # :nodoc:
+      coordination = Coordination.new
+      
+      element.each_element('s:condition') do |condition|
+        condition.each_element('s:target') {|target| coordination.controller = target.text}
+      end
+      element.each_element('s:action') do |action|
+        action.each_element('s:target') {|target| coordination.target = target.text}
+      end
+      
+      model.coordinations.push coordination
+    end
+    
+    def add_link(model, element, version) # :nodoc:
+      link = Link.new
+      
+      if version == '0.1'
+        element.each_element('s:input') { |input| link.sink = input.text}
+        element.each_element('s:output') { |output| link.source = output.text}
+      else
+        source = element.attribute('source') 
+        link.source = source.value if source
+        
+        sink = element.attribute('sink')
+        link.sink = sink.value if sink
+      end
+      
+      model.links.push link
+    end
+    
+    def add_source(model, element, version) # :nodoc:
+      source = Source.new
+      
+      if version == '0.1'
+        name = element.text
+        source.name = name.value if name
+      else
+        name = element.attribute('name')
+        source.name = name.value if name
+        element.each_element('s:metadata') { |metadata|
+          metadata.each_element('s:description') {|description| source.description = description.text}			
+        }
+      end
+      
+      model.sources.push source
+    end
+    
+    def add_sink(model, element, version) # :nodoc:
+      sink = Sink.new
+      
+      if version == '0.1'
+        name = element.text
+        sink.name = name.value if name
+      else
+        name = element.attribute('name')
+        sink.name = name.value if name
+        element.each_element('s:metadata') { |metadata|
+          metadata.each_element('s:description') {|description| sink.description = description.text}			
+        }
+      end
+      
+      model.sinks.push sink
+    end
+    
+    def add_processor(model, element, version) # :nodoc:
+      processor = Processor.new
+      
+      name = element.attribute('name')
+      processor.name = name.value if name
+      
+      element.each_element() do |e|
+        case e.name
+          when 'description'
+            processor.description = e.text
+          when'beanshell'
+            processor.type = e.name
+            e.each_element do |bean|
+              case bean.name
+                when "scriptvalue"
+                  processor.script = bean.text
+                when "beanshellinputlist"
+                  bean.each_element do |input|
+                    if input.name == "beanshellinput"
+                      processor.inputs = [] if processor.inputs.nil?
+                      processor.inputs << input.text
+                    end # if
+                  end # bean.each_element
+                when "beanshelloutputlist"
+                  bean.each_element do |output|
+                    if output.name == "beanshelloutput"
+                      processor.outputs = [] if processor.outputs.nil?
+                      processor.outputs << output.text
+                    end # if
+                  end # bean.each_element
+                when "dependencies"
+                  bean.each_element do |dep|
+                    model.dependencies = [] if model.dependencies.nil?
+                    model.dependencies << dep.text unless dep.text =~ /^\s*$/
+                  end # bean.each_element
+                end # case bean.name
+              end # e.each_element
+            when 'mergemode'
+            when 'defaults'
+            when 'iterationstrategy'
+            else
+            if Dot.is_processor? e.name
+              processor.type = e.name
+              if processor.type == 'workflow'
+                e.each_element('s:scufl') {|e| processor.model = create_model(e, version)}
+              end
+            end
+          end
+        end
+      
+      model.processors.push processor
+      end
+    
+    def set_description(model, element, version) # :nodoc:
+      author = element.attribute('author')
+      title = element.attribute('title')
+      
+      model.description.author = author.value if author
+      model.description.title = title.value if title
+      model.description.description = element.text
+    end
+    
+  end
+  
+end