laydi/system/workflow.py

import gtk
import sys
import os
import inspect
from system import logger

def _workflow_classes(modname):
    """Returns a list of all subclasses of Workflow in a given module"""
    workflow_classes = []

    __import__('workflows.%s' % modname)
    module = sys.modules['workflows.%s' % modname]

    d = module.__dict__
    for wf in d.values():
        try:
            if issubclass(wf, Workflow):
                workflow_classes.append(wf)
        except TypeError, e:
            pass
    return workflow_classes
        
def workflow_list():
    """Returns a ListStore containing all new workflows"""
    retval = []

    # List all .py files that can contain workflow classes
    wf_path = sys.modules['workflows'].__path__
    wf_files = []

    for dir in wf_path:
        for fn in os.listdir(dir):
            if fn.endswith('.py') and ('#' not in fn):
                wf_files.append(fn[:-3])

    # Try to load each file and look for Workflow derived classes
    for fn in wf_files:
        try:
            for wf in _workflow_classes(fn):
                retval.append(wf)
        except Exception, e:
            logger.log('warning', 'Cannot load workflow: %s' % fn)
            logger.log('warning', e)
        
    return retval

class Workflow:
    """Defines a workflow that contains a set of analysis stages.

    A Workflow is a set of analysis stages for a certain type of analysis.
    Each stage contains some possible operations to do accomplish that 
    task.
    """

    name = "Workflow"
    ident = None
    description = "Workflow Description"

    def __init__(self, app):
        self.stages = []
        self.stages_by_id = {}
        self.app = app

    def add_stage(self, stage):
        self.stages.append(stage)
        self.stages_by_id[stage.id] = stage

    def print_tree(self):
        print self.name
        for stage in self.stages:
            print '   %s' % stage.name
            for fun in stage.functions:
                print '        %s' % fun.name

    def add_project(self,project):
        if project == None:
            logger.log('notice','Proejct is empty')
        logger.log('notice','Project added in : %s' %self.name)
        self.project = project

class EmptyWorkflow(Workflow):
    name = 'Empty Workflow'

    def __init__(self, app):
        Workflow.__init__(self, None)
    
class Stage:
    """A stage is a part of the data analysis process.

    Each stage contains a set of functions that can be used to
    accomplish the task. A typical early stage is 'preprocessing', which
    can be done in several ways, each represented by a function.
    """

    def __init__(self, id, name):
        self.id = id
        self.name = name
        self.functions = []
        self.functions_by_id = {}

    def add_function(self, fun):
        self.functions.append(fun)
        self.functions_by_id[fun.id] = fun


class Function:
    """A Function object encapsulates a function on a data set.

    Each Function instance encapsulates some function that can be applied
    to one or more types of data.
    """

    def __init__(self, id, name):
        self.id = id
        self.name = name

    # ,ust return a Validation object
    def validate_input(input):
        return Validation(True,"Validation Not Implemented")

    def run(self):
        pass

class Validation:
    def __init__(self,result, reason):
        self.succeeded = result
        self.reason = reason


class WorkflowView (gtk.VBox):
   
    def __init__(self, wf):
        gtk.VBox.__init__(self)
        self.workflow = wf
        self.setup_workflow(wf)

    def setup_workflow(self, wf):
        # Add stage in the process
        for stage in wf.stages:
            exp = gtk.Expander(stage.name)
            btn_box = gtk.VBox()
            btn_box.show()
            exp.add(btn_box)

            # Add functions in each stage 
            for fun in stage.functions:
                btn = gtk.Button(fun.name)
                btn.connect('clicked',
                            lambda button, f=fun : self.run_function(f))

                btn_box.add(btn)
                btn.show()
            
            exp.show()
            self.pack_start(exp, expand=False, fill=False)

    def remove_workflow(self):
        for c in self.get_children():
            c.hide()
            self.remove(c)

    def set_workflow(self, workflow):
        self.workflow = workflow
        self.remove_workflow()
        self.setup_workflow(workflow)

    def run_function(self, function):
        logger.log('debug', 'Starting function: %s' % function.name)
        project = self.workflow.app.project
        parent_data = project.current_data

        validation  = function.validate_input()

        if not validation.succeeded:
            logger.log('warning','Invalid Inputdata: ' + str(reason))
            return

        args, varargs, varkw, defaults = inspect.getargspec(function.run)

        # first argument is 'self' and second should be the selection
        # and we don't care about those...
        args.remove('self')
        if "selection" in args:
            pass_selection = True
            args.remove('selection')
        else:
            pass_selection = False

        if varargs and len(parent_data) < len(args):
            logger.log('warning', "Function requires minimum %d datasets selected." % len(args))
            return
        elif not varargs and args and len(args) != len(parent_data):
            # functions requiring datasets have to have the right number
            logger.log('warning', "Function requires %d datasets, but only %d selected." % (len(args), len(parent_data)))
            return

        if not args:
            # we allow functions requiring no data to be run even if a
            # dataset is is selected
            data = []
        else:
            data = parent_data
            
        if pass_selection:
            # if the function has a 'selection' argument, we pass in
            # the selection
            new_data = function.run(selection=project.get_selection(), *data)
        else:
            new_data = function.run(*data)
            
        if new_data != None:
            project.add_data(parent_data, new_data, function.name)
        
        logger.log('debug', 'Function ended: %s' % function.name)