class AnnealingSettingsTab(object):
'''This class describes the tab in the GUI where the user
can change setting that govern the monte carlo / annleaing
procedure. This also includes which information from the ccpn
analysis project is used and which information is
ignored. This includes:
* present sequential assignments
* tentative assignments
* amino acid type information
* whether to include untyped spin systems
* assignments to peak dimensions
ALso the chain can be selected here.
Furthermore the user can set the temperature
regime of the annealing, the amount of times the procedure
is repeated to obtain statistics. The fraction of peaks
that is left out in each run to diversify the results,
the treshhold score for amino acid typing and the treshhold
collabelling for a peak to be expected.
'''
def __init__(self, parent, frame):
'''Init. args: parent: the guiElement that this
tab is part of.
frame: the frame this part of the
GUI lives in.
'''
self.guiParent = parent
self.frame = frame
self.project = parent.project
self.nmrProject = parent.nmrProject
self.minIsoFrac = 0.1
self.leavePeaksOutFraction = 0.0
self.minTypeScore = 1.0
self.chain = None
self.amountOfRepeats = 10
self.amountOfSteps = 10000
self.acceptanceConstantList = [0.0, 0.01, 0.015, 0.022,
0.033, 0.050, 0.075, 0.113,
0.170, 0.256, 0.384, 0.576,
0.864, 1.297, 1.946, 2.919,
4.378, 6.568, 9.852, 14.77,
22.16, 33.25]
self.energyDataSets = [[]]
self.residues = []
self.body()
def body(self):
'''describes the body of this tab. It bascically consists
of some field to fill out for the user at the top and
a ScrolledGraph that shows the progess of the annealing
procedure a the bottom.
'''
frame = self.frame
# frame.expandGrid(13,0)
frame.expandGrid(15, 1)
row = 0
text = 'Calculate Assignment Suggestions'
command = self.runCalculations
self.startButton = Button(frame, command=command, text=text)
self.startButton.grid(row=row, column=0, sticky='nsew', columnspan=2)
row += 1
Label(frame, text='Amount of runs: ', grid=(row, 0))
tipText = 'The amount of times the whole optimization procedure is performed, each result is safed'
self.repeatEntry = IntEntry(frame, grid=(row, 1), width=7, text=10,
returnCallback=self.updateRepeatEntry,
tipText=tipText, sticky='nsew')
self.repeatEntry.bind('<Leave>', self.updateRepeatEntry, '+')
row += 1
Label(frame, text='Temperature regime: ', grid=(row, 0))
tipText = 'This list of numbers govern the temperature steps during the annealing, every number represents 1/(kb*t), where kb is the Boltzmann constant and t the temperature of one step.'
self.tempEntry = Entry(frame, text=map(str, self.acceptanceConstantList), width=64,
grid=(row, 1), isArray=True, returnCallback=self.updateAcceptanceConstantList,
tipText=tipText, sticky='nsew')
row += 1
Label(frame, text='Amount of attempts per temperature:', grid=(row, 0))
tipText = 'The amount of attempts to switch the position of two spinsystems in the sequence are performed for each temperature point'
self.NAStepEntry = IntEntry(frame, grid=(row, 1), width=7, text=10000,
returnCallback=self.updateStepEntry,
tipText=tipText, sticky='nsew')
self.NAStepEntry.bind('<Leave>', self.updateStepEntry, '+')
row += 1
Label(frame, text='Fraction of peaks to leave out:', grid=(row, 0))
tipText = 'In each run a fraction of the peaks can be left out of the optimization, thereby increasing the variability in the outcome and reducing false negatives. In each run this will be different randomly chosen sub-set of all peaks. 0.1 (10%) can be a good value.'
self.leaveOutPeaksEntry = FloatEntry(frame, grid=(row, 1), width=7, text=0.0,
returnCallback=self.updateLeavePeaksOutEntry,
tipText=tipText, sticky='nsew')
self.leaveOutPeaksEntry.bind(
'<Leave>', self.updateLeavePeaksOutEntry, '+')
row += 1
Label(frame, text='Minmal amino acid typing score:', grid=(row, 0))
tipText = 'If automatic amino acid typing is selected, a cut-off value has to set. Every amino acid type that scores higher than the cut-off is taken as a possible type. This is the same score as can be found under resonance --> spin systems --> predict type. Value should be between 0 and 100'
self.minTypeScoreEntry = FloatEntry(frame, grid=(row, 1), width=7, text=1.0,
returnCallback=self.updateMinTypeScoreEntry,
tipText=tipText, sticky='nsew')
self.minTypeScoreEntry.bind(
'<Leave>', self.updateMinTypeScoreEntry, '+')
row += 1
Label(frame, text='Minimal colabelling fraction:', grid=(row, 0))
tipText = 'The minimal amount of colabelling the different nuclei should have in order to still give rise to a peak.'
self.minLabelEntry = FloatEntry(frame, grid=(row, 1), width=7, text=0.1,
returnCallback=self.updateMinLabelEntry,
tipText=tipText, sticky='nsew')
self.minLabelEntry.bind('<Leave>', self.updateMinLabelEntry, '+')
row += 1
Label(frame, text='Use sequential assignments:', grid=(row, 0))
tipText = 'When this option is select the present sequential assignments will be kept in place'
self.useAssignmentsCheck = CheckButton(
frame, selected=True, tipText=tipText, grid=(row, 1))
row += 1
Label(frame, text='Use tentative assignments:', grid=(row, 0))
tipText = 'If a spin system has tentative assignments this can be used to narrow down the amount of possible sequential assignments.'
self.useTentativeCheck = CheckButton(
frame, selected=True, tipText=tipText, grid=(row, 1))
row += 1
Label(frame, text='Use amino acid types:', grid=(row, 0))
tipText = 'Use amino acid types of the spin systems. If this option is not checked the spin systems are re-typed, only resonance names and frequencies are used'
self.useTypeCheck = CheckButton(
frame, selected=True, tipText=tipText, grid=(row, 1))
row += 1
Label(frame, text='Include untyped spin systems:', grid=(row, 0))
tipText = 'Also include spin system that have no type information. Amino acid typing will be done on the fly.'
self.useAlsoUntypedSpinSystemsCheck = CheckButton(
frame, selected=True, tipText=tipText, grid=(row, 1))
row += 1
Label(frame, text='Use dimensional assignments:', grid=(row, 0))
tipText = 'If one or more dimensions of a peak is already assigned, assume that this assignment is the only option. If not the check the program will consider all possibilities for the assignment of the dimension.'
self.useDimensionalAssignmentsCheck = CheckButton(
frame, selected=True, tipText=tipText, grid=(row, 1))
row += 1
Label(frame, text='Chain:', grid=(row, 0))
self.molPulldown = PulldownList(
frame, callback=self.changeMolecule, grid=(row, 1))
self.updateChains()
row += 1
Label(frame, text='Residue ranges: ', grid=(row, 0))
tipText = 'Which residues should be included. Example: "10-35, 62-100, 130".'
self.residueRangeEntry = Entry(frame, text=None, width=64,
grid=(row, 1), isArray=True, returnCallback=self.updateResidueRanges,
tipText=tipText, sticky='nsew')
self.updateResidueRanges(fromChain=True)
row += 1
self.energyPlot = ScrolledGraph(frame, symbolSize=2, width=600,
height=200, title='Annealing',
xLabel='temperature step', yLabel='energy')
self.energyPlot.grid(row=row, column=0, columnspan=2, sticky='nsew')
def runCalculations(self):
'''Run all calculations. Also triggers the disabling of
some buttons and fields.
'''
self.startButton.disable()
self.disableIllegalButtonsAfterPrecalculations()
self.guiParent.connector.runAllCalculations()
self.startButton.configure(text='More runs')
self.startButton.enable()
def disableIllegalButtonsAfterPrecalculations(self):
'''Disable buttons and field the user can not alter
any longer after the model is set up and the
'pre-calculations' have finished.
This is done because this part of the calculation
should only be run once. All settings that would
be changed after this point will not have any influence.
'''
illegalButtons = [self.minTypeScoreEntry, self.minLabelEntry,
self.useAlsoUntypedSpinSystemsCheck, self.useAssignmentsCheck,
self.useTypeCheck, self.useDimensionalAssignmentsCheck,
self.useTentativeCheck]
for illegalButton in illegalButtons:
illegalButton.configure(state='disabled')
self.molPulldown.disable()
def getChainName(self, chain):
'''Get the name for a chain.
args: chain: ccpn analysis
chain object
returns: chain name
'''
return '%s:%s (%s)' % (chain.molSystem.code, chain.code, chain.molecule.molType)
def getChains(self):
'''Get all chains present in the project.
returns: list of ccpn analysis chain objects
'''
chains = []
if self.project:
for molSystem in self.project.sortedMolSystems():
for chain in molSystem.sortedChains():
if chain.residues:
chains.append(chain)
return chains
def updateChains(self, *opt):
'''Updates the list of chains if a new one is added
to or deleted from the project. Updates the
pull down list where a chain can be selected.
'''
index = 0
texts = []
chains = self.getChains()
chain = self.chain
if chains:
if chain not in chains:
chain = chains[0]
texts = [self.getChainName(c) for c in chains]
index = chains.index(chain)
else:
chain = None
self.molPulldown.setup(texts, chains, index)
if chain is not self.chain:
self.chain = chain
def changeMolecule(self, chain):
'''Select a molecular chain.'''
if chain is not self.chain:
self.chain = chain
self.updateResidueRanges(fromChain=True)
def updateStepEntry(self, event=None):
'''Update the value and entry that sets the amount of
steps per temperature point.
'''
value = self.NAStepEntry.get()
if value == self.amountOfSteps:
return
if value < 1:
self.NAStepEntry.set(1)
self.amountOfSteps = 1
else:
self.amountOfSteps = value
self.NAStepEntry.set(value)
def updateRepeatEntry(self, event=None):
'''Update the value and entry of that sets
the amount of times the whole annealing
procedure is repeated in order
to obtain statistics.
'''
value = self.repeatEntry.get()
if value == self.amountOfRepeats:
return
if value < 1:
self.repeatEntry.set(1)
self.amountOfRepeats = 1
else:
self.amountOfRepeats = value
self.repeatEntry.set(value)
def updateMinTypeScoreEntry(self, event=None):
'''Updates the value and the entry for the
treshhold value for amino acid typing.
'''
value = self.minTypeScoreEntry.get()
if value == self.minTypeScore:
return
if value < 0:
self.minTypeScoreEntry.set(0.0)
self.minTypeScore = 0.0
elif value > 100:
self.minTypeScoreEntry.set(100.0)
self.minTypeScore = 100.0
else:
self.minTypeScoreEntry.set(value)
self.minTypeScore = value
def updateMinLabelEntry(self, event=None):
'''Updates the minimum colabelling fraction
for which a peak is expected to be present
in the spectra.
'''
value = self.minLabelEntry.get()
if value == self.minIsoFrac:
return
if value < 0:
self.minIsoFrac = 0.0
self.minLabelEntry.set(0.0)
elif value > 1:
self.minIsoFrac = 1.0
self.minLabelEntry.set(1.0)
else:
self.minIsoFrac = value
self.minLabelEntry.set(value)
def updateLeavePeaksOutEntry(self, event=None):
'''Updates the value and entry of the fraction
of peaks that should be left out in each
run in order to diversify the results.
'''
value = self.leaveOutPeaksEntry.get()
if value == self.leavePeaksOutFraction:
return
if value < 0:
self.leavePeaksOutFraction = 0.0
self.leaveOutPeaksEntry.set(0.0)
elif value > 1:
self.leavePeaksOutFraction = 1.0
self.leaveOutPeaksEntry.set(1.0)
else:
self.leavePeaksOutFraction = value
self.leaveOutPeaksEntry.set(value)
def updateAcceptanceConstantList(self, event=None):
'''Updates the list with constants that are used
during the monte carlo procedure to decide whether
a changed is accepted or not.
'''
acList = self.tempEntry.get()
newList = []
for constant in acList:
try:
number = float(constant)
newList.append(number)
except ValueError:
string = constant + \
' in temperature constants is not a number.'
showWarning('Not A Number', string, parent=self.guiParent)
return False
self.acceptanceConstantList = newList
return True
def updateResidueRanges(self, event=None, fromChain=False):
self.residues = set()
subRanges = self.residueRangeEntry.get()
if not subRanges or fromChain:
self.residues = set(self.chain.residues)
residues = self.chain.sortedResidues()
text = '{}-{}'.format(residues[0].seqCode, residues[-1].seqCode)
self.residueRangeEntry.set(text=text)
return
for subRange in subRanges:
indeces = subRange.split('-')
start = int(indeces[0])
stop = int(indeces[-1]) + 1
for seqCode in range(start, stop):
residue = self.chain.findFirstResidue(seqCode=seqCode)
if not residue:
showWarning('Residue out of range.',
'There is no residue at position {}'.format(seqCode),
parent=self.guiParent)
self.residues = set()
return
self.residues.add(residue)
def addEnergyPoint(self, energy, time):
'''Adds a point to the graph that shows the progress
of the annealling procedure.
args: energy: the y-value
time: the x-value
'''
point = (time, energy * -1)
# This means one run has finished
if len(self.energyDataSets[-1]) / (len(self.acceptanceConstantList) + 1):
self.energyDataSets.append([point])
else:
self.energyDataSets[-1].append(point)
colors = colorSeries
Ncolors = len(colors)
NdataSets = len(self.energyDataSets)
colorList = (NdataSets / Ncolors) * colors + \
colors[:NdataSets % Ncolors]
self.energyPlot.update(dataSets=self.energyDataSets,
dataColors=colorList)
# Forcing the graph to draw, eventhough calculations
# are still running. Only do this with high numbers of
# steps, otherwise drawing takes longer than annealling.
if self.amountOfSteps >= 100000:
self.energyPlot.draw()
