def validateAndReturnErrors(cls, choices):
'''
Should validate the selected input parameters. If the parameters are not
valid, an error text explaining the problem should be returned. The GUI
then shows this text to the user (if not empty) and greys out the
execute button (even if the text is empty). If all parameters are valid,
the method should return None, which enables the execute button.
'''
errorStr = cls._checkTrack(choices,
trackChoiceIndex='track',
genomeChoiceIndex='genome')
if errorStr:
return errorStr
if choices.track and not choices.attr:
return 'You have chosen a track with no attributes (columns) supported for splitting. ' \
'Attributes that do not support splitting are: ' + ', '.join(cls.UNSUPPORTED_ATTRS)
geSource = etm.getGESourceFromGalaxyOrVirtualTN(
choices.track, choices.genome)
trackFormat = TrackFormat.createInstanceFromGeSource(geSource)
if trackFormat.isDense():
return 'The track format of the selected track file is: %s' % trackFormat.getFormatName() +\
' This tool only supports track types Points, Segments, or variations of these.'
def _getBasicTrackFormat(choices, tnChoiceIndex=1, genomeChoiceIndex=0):
genome = GeneralGuiTool._getGenomeChoice(choices, genomeChoiceIndex)[0]
tn = GeneralGuiTool._getTrackChoice(choices, tnChoiceIndex)[0]
from quick.application.GalaxyInterface import GalaxyInterface
from gold.description.TrackInfo import TrackInfo
from quick.application.ExternalTrackManager import ExternalTrackManager
from gold.track.TrackFormat import TrackFormat
if ExternalTrackManager.isGalaxyTrack(tn):
geSource = ExternalTrackManager.getGESourceFromGalaxyOrVirtualTN(
tn, genome)
try:
tf = GeneralGuiTool._convertToBasicTrackFormat(
TrackFormat.createInstanceFromGeSource(
geSource).getFormatName())
except Warning:
return genome, tn, ''
else:
if GalaxyInterface.isNmerTrackName(genome, tn):
tfName = 'Points'
else:
tfName = TrackInfo(genome, tn).trackFormatName
tf = GeneralGuiTool._convertToBasicTrackFormat(tfName)
return genome, tn, tf
def getOptionsBox3(prevChoices):
if prevChoices[-1]:
return ('__hidden__', prevChoices[-1])
elif prevChoices[0] and prevChoices[1]:
from quick.application.ExternalTrackManager import ExternalTrackManager
geSource = ExternalTrackManager.getGESourceFromGalaxyOrVirtualTN(
prevChoices[-2].split(':'), prevChoices[0])
catSet = set([ge.val for ge in geSource if ge.val])
return ('__hidden__', repr(catSet))
def _getRawTrackView(self, region, borderHandling, allowOverlaps):
assert region.start == 0 and region.end == 1
from collections import OrderedDict
from gold.track.CommonMemmapFunctions import findEmptyVal
from gold.track.TrackView import TrackView
import numpy as np
geSource = ExternalTrackManager.getGESourceFromGalaxyOrVirtualTN(self.trackName, region.genome)
prefixList = geSource.getPrefixList()
valDataType = geSource.getValDataType()
valDim = geSource.getValDim()
weightDataType = geSource.getEdgeWeightDataType()
weightDim = geSource.getEdgeWeightDim()
startList, endList, valList, strandList, idList, edgesList, weightsList = [None]*7
extraLists=OrderedDict()
tf = TrackFormat.createInstanceFromPrefixList(prefixList, valDataType, valDim, \
weightDataType, weightDim)
if allowOverlaps and (tf.isDense() or geSource.hasNoOverlappingElements()):
raise IncompatibleTracksError(prettyPrintTrackName(self.trackName) + ' with format: '\
+ str(tf) + ' does not satisfy ' + str(self._trackFormatReq))
denseAndInterval = tf.isDense() and tf.isInterval()
numEls = 2 if denseAndInterval else 1
if valDataType == 'S':
valDataType = 'S2'
if weightDataType == 'S':
weightDataType = 'S2'
for prefix in prefixList:
if prefix == 'start':
startList = np.array([-1], dtype='int32')
elif prefix == 'end':
if denseAndInterval:
endList = np.array([0, 1], dtype='int32')
else:
endList = np.array([0], dtype='int32')
elif prefix == 'val':
valList = np.array([findEmptyVal(valDataType)] * valDim * numEls, \
dtype=valDataType).reshape((numEls, valDim) if valDim > 1 else numEls)
elif prefix == 'strand':
strandList = np.array([1] * numEls, dtype='int8')
elif prefix == 'id':
idList = np.array([''] * numEls, dtype='S1')
elif prefix == 'edges':
edgesList = np.array([['']] * numEls, dtype='S1')
elif prefix == 'weights':
weightsList = np.array([[[findEmptyVal(weightDataType)]]] * weightDim * numEls, \
dtype=weightDataType).reshape((numEls, 1, weightDim) if weightDim > 1 else (numEls, 1))
else:
extraLists[prefix] = np.array([''] * numEls, dtype='S1')
return TrackView(region, startList, endList, valList, strandList, idList, edgesList, weightsList, borderHandling, allowOverlaps, extraLists)
def _getRawTrackView(self, region, borderHandling, allowOverlaps):
assert len(region) == 1
from collections import OrderedDict
from gold.track.CommonMemmapFunctions import findEmptyVal
from gold.track.TrackView import TrackView
import numpy as np
geSource = ExternalTrackManager.getGESourceFromGalaxyOrVirtualTN(self.trackName, region.genome)
prefixList = geSource.getPrefixList()
valDataType = geSource.getValDataType()
valDim = geSource.getValDim()
weightDataType = geSource.getEdgeWeightDataType()
weightDim = geSource.getEdgeWeightDim()
startList, endList, valList, strandList, idList, edgesList, weightsList = [None]*7
extraLists=OrderedDict()
tf = TrackFormat.createInstanceFromPrefixList(prefixList, valDataType, valDim, \
weightDataType, weightDim)
if allowOverlaps and (tf.isDense() or geSource.hasNoOverlappingElements()):
raise IncompatibleTracksError(prettyPrintTrackName(self.trackName) + ' with format: '\
+ str(tf) + ' does not satisfy ' + str(self._trackFormatReq))
denseAndInterval = tf.isDense() and tf.isInterval()
numEls = 2 if denseAndInterval else 1
if valDataType == 'S':
valDataType = 'S2'
if weightDataType == 'S':
weightDataType = 'S2'
for prefix in prefixList:
if prefix == 'start':
startList = np.array([-1], dtype='int32')
elif prefix == 'end':
if denseAndInterval:
endList = np.array([0, 1], dtype='int32')
else:
endList = np.array([0], dtype='int32')
elif prefix == 'val':
valList = np.array([findEmptyVal(valDataType)] * valDim * numEls, \
dtype=valDataType).reshape((numEls, valDim) if valDim > 1 else numEls)
elif prefix == 'strand':
strandList = np.array([1] * numEls, dtype='int8')
elif prefix == 'id':
idList = np.array([''] * numEls, dtype='S1')
elif prefix == 'edges':
edgesList = np.array([['']] * numEls, dtype='S1')
elif prefix == 'weights':
weightsList = np.array([[[findEmptyVal(weightDataType)]]] * weightDim * numEls, \
dtype=weightDataType).reshape((numEls, 1, weightDim) if weightDim > 1 else (numEls, 1))
else:
extraLists[prefix] = np.array([''] * numEls, dtype='S1')
return TrackView(region, startList, endList, valList, strandList, idList, edgesList, weightsList, borderHandling, allowOverlaps, extraLists)
def getExtraHistElements(cls, choices):
if choices.genome and choices.catTrack:
from quick.webtools.GeneralGuiTool import HistElement
genome = choices.genome
outputType = choices.outputType
catTrack = choices.catTrack.split(':')
geSource = ExternalTrackManager.getGESourceFromGalaxyOrVirtualTN(
catTrack, genome)
return [HistElement(ge.val, outputType) for ge in geSource
] + [HistElement('GSuite from categorical', 'gsuite')]
def getOptionsBoxFormat(prevChoices):
if prevChoices.track:
geSource = etm.getGESourceFromGalaxyOrVirtualTN(
prevChoices.track, prevChoices.genome)
tf = TrackFormat.createInstanceFromGeSource(geSource)
matchingComposers = findMatchingFileFormatComposers(tf)
conversions = [geSource.getFileFormatName() + \
' (no conversion, track type: %s)' % tf.getFormatName()]
conversions += ['%s -> %s (track type: %s)' % (geSource.getFileFormatName(), \
composerInfo.fileFormatName, composerInfo.trackFormatName) \
for composerInfo in matchingComposers \
if geSource.getFileFormatName() != composerInfo.fileFormatName]
return conversions
def execute(cls, choices, galaxyFn=None, username=''):
'''
Is called when execute-button is pushed by web-user. Should print
output as HTML to standard out, which will be directed to a results page
in Galaxy history. If getOutputFormat is anything else than HTML, the
output should be written to the file with path galaxyFn. If needed,
StaticFile can be used to get a path where additional files can be put
(e.g. generated image files). choices is a list of selections made by
web-user in each options box.
'''
genome = choices.genome
trackChoice = choices.history if choices.source == 'history' else choices.track
trackName = trackChoice.split(':')
if ExternalTrackManager.isGalaxyTrack(trackName):
geSource = ExternalTrackManager.getGESourceFromGalaxyOrVirtualTN(
trackName, genome)
if not geSource.hasOrigFile(): #hbfunction
stdTrackName = ExternalTrackManager.getStdTrackNameFromGalaxyTN(
trackName)
ExternalTrackManager.renameExistingStdTrackIfNeeded(
genome, stdTrackName)
geSource = FullTrackGenomeElementSource(genome,
stdTrackName,
allowOverlaps=False)
else:
try:
geSource = FullTrackGenomeElementSource(genome,
trackName,
allowOverlaps=True)
for ge in geSource:
break
except:
geSource = FullTrackGenomeElementSource(genome,
trackName,
allowOverlaps=False)
threshold = float(choices.threshold)
rule = RULE_DICT[choices.rule]
composerCls = getComposerClsFromFileFormatName(choices.format)
geModifier = GEValueThresholder(geSource, genome, threshold, rule)
if choices.merge == 'Yes':
if not geModifier.isSorted():
geModifier = GenomeElementSorter(geModifier)
geModifier = GEAdjacencyClusterer_Segment(geModifier)
composerCls(geModifier).composeToFile(galaxyFn, ignoreEmpty=True)
def execute(cls, choices, galaxyFn=None, username=''):
from quick.application.ExternalTrackManager import ExternalTrackManager
histFn = ExternalTrackManager.extractFnFromGalaxyTN(choices[1])
chosenCatsList = [k for k, v in choices[-1].items() if v]
geSource = ExternalTrackManager.getGESourceFromGalaxyOrVirtualTN(
choices[1].split(':'), choices[0])
utfil = open(galaxyFn, 'w')
for ge in geSource:
if ge.val in chosenCatsList:
print >> utfil, '\t'.join(
[ge.chr, str(ge.start),
str(ge.end), ge.val])
utfil.close()
def getOptionsBoxStatistics(cls, prevChoices):
if prevChoices.showStatistics == cls.SHOW_STATISTICS_CHOICE_YES:
geSource = etm.getGESourceFromGalaxyOrVirtualTN(
prevChoices.track, prevChoices.genome)
valDict = defaultdict(int)
numEls = 0
for ge in geSource:
valDict[str(getattr(ge, prevChoices.attr))] += 1
numEls += 1
return [['Statistic', 'Result'], \
['Number of elements in track', str(numEls)],
['Number of unique attribute values', str(len(valDict))],
['Attribute(s) least represented in track (number of elements)', \
'|'.join(findKeysWithMinVal(valDict)) + ' (%s)' % min(valDict.values())],
['Attribute(s) most represented in track (number of elements)', \
'|'.join(findKeysWithMaxVal(valDict)) + ' (%s)' % max(valDict.values())]]
def _getValueTypeName(choices, tnChoiceIndex=1, genomeChoiceIndex=0):
genome = GeneralGuiTool._getGenomeChoice(choices, genomeChoiceIndex)[0]
tn = GeneralGuiTool._getTrackChoice(choices, tnChoiceIndex)[0]
from quick.application.GalaxyInterface import GalaxyInterface
from gold.description.TrackInfo import TrackInfo
from quick.application.ExternalTrackManager import ExternalTrackManager
from gold.track.TrackFormat import TrackFormat
if ExternalTrackManager.isGalaxyTrack(tn):
geSource = ExternalTrackManager.getGESourceFromGalaxyOrVirtualTN(
tn, genome)
valTypeName = TrackFormat.createInstanceFromGeSource(
geSource).getValTypeName()
else:
if GalaxyInterface.isNmerTrackName(genome, tn):
valTypeName = ''
else:
valTypeName = TrackInfo(genome, tn).markType
return valTypeName.lower()
def execute(cls, choices, galaxyFn=None, username=''):
'''
Is called when execute-button is pushed by web-user. Should print
output as HTML to standard out, which will be directed to a results page
in Galaxy history. If getOutputFormat is anything else than HTML, the
output should be written to the file with path galaxyFn. If needed,
StaticFile can be used to get a path where additional files can be put
(e.g. generated image files). choices is a list of selections made by
web-user in each options box.
'''
genome = choices.genome
geSource = etm.getGESourceFromGalaxyOrVirtualTN(choices.track, genome)
#hiddenStorageFn = cls.extraGalaxyFn[cls.HISTORY_HIDDEN_TRACK_STORAGE]
hiddenStorageFn = galaxyFn
composerCls = cls._getComposerCls(choices)
valAttr = choices.attr
gSuite = createGalaxyGSuiteBySplittingInputFileOnAttribute\
(hiddenStorageFn, geSource, genome, composerCls, valAttr)
GSuiteComposer.composeToFile(gSuite, galaxyFn)
def execute(choices, galaxyFn=None, username=''):
utfil = open(galaxyFn, 'w')
genome = choices[0]
trackName = choices[1].split(':')
intervals = sorted([float(v.strip()) for v in choices[2].split(',')])
geSource = ExternalTrackManager.getGESourceFromGalaxyOrVirtualTN(
trackName, genome)
for ge in geSource:
value = ge.val
counter = 0
while counter < len(intervals) and intervals[counter] <= value:
counter += 1
if counter == len(intervals):
category = str(intervals[-1]) + '-inf'
else:
category = str(intervals[counter - 1]) + '-' + str(
intervals[counter]) if counter > 0 else '0.0-' + str(
intervals[0])
print >> utfil, '%s\t%i\t%i\t%s' % (ge.chr, ge.start, ge.end,
category)
def execute(cls, choices, galaxyFn=None, username=''):
'''
Is called when execute-button is pushed by web-user. Should print
output as HTML to standard out, which will be directed to a results page
in Galaxy history. If getOutputFormat is anything else than HTML, the
output should be written to the file with path galaxyFn. If needed,
StaticFile can be used to get a path where additional files can be put
(e.g. generated image files). choices is a list of selections made by
web-user in each options box.
'''
categoryToGenomeElementListDict = defaultdict(list)
genome = choices.genome
outputType = choices.outputType
catTrack = choices.catTrack.split(':')
geSource = ExternalTrackManager.getGESourceFromGalaxyOrVirtualTN(
catTrack, genome)
for ge in geSource:
categoryToGenomeElementListDict[ge.val].append(ge)
for category, genomeElementList in categoryToGenomeElementListDict.iteritems(
):
geSourceWrapper = ListGESourceWrapper(geSource, genomeElementList)
composer = cls.getComposer(geSourceWrapper, outputType)
# staticFile = GalaxyRunSpecificFile(catTrack + [category, outputType], galaxyFn)
composer.composeToFile(cls.extraGalaxyFn[category])
outGSuite = GSuite()
for category, galaxyFileName in OrderedDict([
(x, cls.extraGalaxyFn[x])
for x in categoryToGenomeElementListDict.keys()
]).iteritems():
uri = GalaxyGSuiteTrack.generateURI(galaxyFn=galaxyFileName,
suffix=outputType)
outGSuite.addTrack(GSuiteTrack(uri, title=category, genome=genome))
GSuiteComposer.composeToFile(
outGSuite, cls.extraGalaxyFn['GSuite from categorical'])
print 'Execution done!'
def validateAndReturnErrors(choices):
'''
Should validate the selected input parameters. If the parameters are not
valid, an error text explaining the problem should be returned. The GUI
then shows this text to the user (if not empty) and greys out the
execute button (even if the text is empty). If all parameters are valid,
the method should return None, which enables the execute button.
'''
if not choices.genome:
return 'Please select genome'
if not choices.catTrack:
return 'Please select categorical track from history'
geSource = ExternalTrackManager.getGESourceFromGalaxyOrVirtualTN(
choices.catTrack.split(':'), choices.genome)
trackFormat = TrackFormat.createInstanceFromGeSource(geSource)
if trackFormat.getValTypeName() != 'Category':
return 'Please select <b>categorical</b> track from history, current is of type ' + trackFormat.getValTypeName(
)
def getOptionsBoxAttr(cls, prevChoices):
if prevChoices.track:
geSource = etm.getGESourceFromGalaxyOrVirtualTN(
prevChoices.track, prevChoices.genome)
return [prefix for prefix in geSource.getPrefixList() \
if prefix not in cls.UNSUPPORTED_ATTRS ]