laydi/workflows/go_workflow.py

import gtk
import logger
from annotations import Annotations
from workflow import *
import plots
import dataset
#import geneontology
#import gostat
import rpy
from scipy import array,randn
import cPickle

class EinarsWorkflow (Workflow):

    def __init__(self, app):
        Workflow.__init__(self, app)
        self.name = 'Einar\'s Workflow'

        load = Stage('load', 'Load Data')
        load.add_function(Function('load', 'Load Microarrays'))
        load.add_function(CelFileImportFunction())
        load.add_function(TestDataFunction())
        load.add_function(DatasetLoadFunction())
        self.add_stage(load)

        preproc = Stage('preprocess', 'Preprocessing')
        preproc.add_function(Function('rma', 'RMA'))
        self.add_stage(preproc)

        go = Stage('go', 'Gene Ontology Data')
        go.add_function(LoadAnnotationsFunction())
        go.add_function(GODistanceFunction())
        self.add_stage(go)

        regression = Stage('regression', 'Regression')
        regression.add_function(Function('pls', 'PLS'))
        self.add_stage(regression)

        save = Stage('save', 'Save Data')
        save.add_function(DatasetSaveFunction())
        self.add_stage(save)
        
        logger.log('debug', '\tEinar\'s workflow is now active')

class LoadAnnotationsFunction(Function):

    def __init__(self):
        Function.__init__(self, 'load-go-ann', 'Load Annotations')
        self.annotations = None

    def load_file(self, filename):
        f = open(filename)
        self.annotations = Annotations('genes', 'go-terms')
        logger.log('notice', 'Loading annotation file: %s' % filename)

        for line in f.readlines():
            val = line.split(' \t')

            if len(val) > 1:
                val = [v.strip() for v in val]
                retval.add_annotations('genes', val[0], 
                                       'go-terms', set(val[1:]))
            
    def on_response(self, dialog, response):
        if response == gtk.RESPONSE_OK:
            logger.log('notice', 'Reading file: %s' % dialog.get_filename())
            self.load_file(dialog.get_filename())

    def run(self, data):
        btns = ('Open', gtk.RESPONSE_OK, \
                'Cancel', gtk.RESPONSE_CANCEL)
        dialog = gtk.FileChooserDialog('Open GO Annotation File',
                                       buttons=btns)
        dialog.connect('response', self.on_response)
        dialog.run()
        dialog.destroy()
        return [self.annotations]

class GODistanceFunction(Function):

    def __init__(self):
        Function.__init__(self, 'go_diatance', 'GO Distances')
        self.output = None

    def run(self, data):
        logger.log('debug', 'datatype: %s' % type(data))
        if not type(data) == Annotations:
            return None

        logger.log('debug', 'dimensions: %s' % data.dimensions)
        
        genes = data.get_ids('genes')
        gene_distances = array((len(genes), len(genes)))

        return gene_distances


class TestDataFunction(Function):  
    def __init__(self):
        Function.__init__(self, 'test_data', 'Generate Test Data')

    def run(self, data):
        logger.log('notice', 'Injecting foo test data')
        x = randn(20,30)
        axis_0 = ['rows',[]]
        axis_1 = ['cols',[]]
        X = dataset.Dataset(x,[axis_0,axis_1])
        return [X, plots.SinePlot(None)]


class DatasetLoadFunction(Function):
    """Loader for previously pickled Datasets."""
    def __init__(self):
        Function.__init__(self, 'load_data', 'Load Pickled Dataset')

    def run(self, data):
        chooser = gtk.FileChooserDialog(title="Select cel files...", parent=None,
                                        action=gtk.FILE_CHOOSER_ACTION_OPEN,
                                        buttons=(gtk.STOCK_CANCEL, gtk.RESPONSE_CANCEL,
                                                 gtk.STOCK_OPEN, gtk.RESPONSE_OK))
        pkl_filter = gtk.FileFilter()
        pkl_filter.set_name("Python pickled data files (*.pkl)")
        pkl_filter.add_pattern("*.[pP][kK][lL]")
        all_filter = gtk.FileFilter()
        all_filter.set_name("All Files (*.*)")
        all_filter.add_pattern("*")
        chooser.add_filter(pkl_filter)
        chooser.add_filter(all_filter)

        try:
            if chooser.run() == gtk.RESPONSE_OK:
                return [cPickle.load(open(chooser.get_filename()))]
        finally:
            chooser.destroy()


class DatasetSaveFunction(Function):
    """QND way to save data to file for later import to this program."""
    def __init__(self):
        Function.__init__(self, 'save_data', 'Save Pickled Dataset')

    def run(self, data):
        if not data:
            logger.log("notice", "No data to save.")
            return

        chooser = gtk.FileChooserDialog(title="Save pickled data...", parent=None,
                                        action=gtk.FILE_CHOOSER_ACTION_SAVE,
                                        buttons=(gtk.STOCK_CANCEL, gtk.RESPONSE_CANCEL,
                                                 gtk.STOCK_SAVE, gtk.RESPONSE_OK))
        pkl_filter = gtk.FileFilter()
        pkl_filter.set_name("Python pickled data files (*.pkl)")
        pkl_filter.add_pattern("*.[pP][kK][lL]")
        all_filter = gtk.FileFilter()
        all_filter.set_name("All Files (*.*)")
        all_filter.add_pattern("*")
        chooser.add_filter(pkl_filter)
        chooser.add_filter(all_filter)
        chooser.set_current_name(data.get_name() + ".pkl")

        try:
            if chooser.run() == gtk.RESPONSE_OK:
                cPickle.dump(data, open(chooser.get_filename(), "w"))
                logger.log("notice", "Saved data to %r." % chooser.get_filename())
        finally:
            chooser.destroy()
                

class CelFileImportFunction(Function):
    """Loads AffyMetrix .CEL-files into matrix."""
    def __init__(self):
        Function.__init__(self, 'cel_import', 'Import Affy')

    def run(self, data):
        chooser = gtk.FileChooserDialog(title="Select cel files...", parent=None,
                                        action=gtk.FILE_CHOOSER_ACTION_OPEN,
                                        buttons=(gtk.STOCK_CANCEL, gtk.RESPONSE_CANCEL,
                                                 gtk.STOCK_OPEN, gtk.RESPONSE_OK))
        chooser.set_select_multiple(True)
        cel_filter = gtk.FileFilter()
        cel_filter.set_name("Cel Files (*.cel)")
        cel_filter.add_pattern("*.[cC][eE][lL]")
        all_filter = gtk.FileFilter()
        all_filter.set_name("All Files (*.*)")
        all_filter.add_pattern("*")
        chooser.add_filter(cel_filter)
        chooser.add_filter(all_filter)

        if chooser.run() == gtk.RESPONSE_OK:
            logger.log('debug', "Selected files: %s" % ", ".join(chooser.get_filenames()))
            rpy.r.library("affy")
            # hack: we append ";1" to make sure no r-object is returned to python (faster)
            rpy.r('At.aBatch <- ReadAffy(filenames=c("%s"));1' % '", "'.join(chooser.get_filenames()))
            # we destroy it immediately to keep it from being on
            # screen while we do something with the files
            chooser.destroy()
            # also here we append ";1"
            rpy.r('At.eSet <- expresso(At.aBatch, bg.correct=F, summary.method="liwong", pmcorrect.method="pmonly", normalize.method="qspline");1')
            m = rpy.r('At.m <- exprs(At.eSet)')
            rownames = rpy.r('rownames(At.m)') 
            colnames = rpy.r('colnames(At.m)') 
            return [dataset.Dataset(m, (('ids', rownames), ('filename', colnames)))] 
        else:
            chooser.destroy()