[aa36f96] | 1 | |
---|
| 2 | |
---|
[792db7d5] | 3 | """ |
---|
[aa36f96] | 4 | ScipyFitting module contains FitArrange , ScipyFit, |
---|
| 5 | Parameter classes.All listed classes work together to perform a |
---|
| 6 | simple fit with scipy optimizer. |
---|
[792db7d5] | 7 | """ |
---|
[61cb28d] | 8 | |
---|
[88b5e83] | 9 | import numpy |
---|
[7705306] | 10 | from scipy import optimize |
---|
| 11 | |
---|
[342d9197] | 12 | from AbstractFitEngine import FitEngine, SansAssembly,FitAbort |
---|
[61cb28d] | 13 | |
---|
[e0072082] | 14 | class fitresult(object): |
---|
[48882d1] | 15 | """ |
---|
[aa36f96] | 16 | Storing fit result |
---|
[48882d1] | 17 | """ |
---|
[e0072082] | 18 | def __init__(self, model=None, paramList=None): |
---|
| 19 | self.calls = None |
---|
| 20 | self.fitness = None |
---|
| 21 | self.chisqr = None |
---|
| 22 | self.pvec = None |
---|
| 23 | self.cov = None |
---|
| 24 | self.info = None |
---|
| 25 | self.mesg = None |
---|
| 26 | self.success = None |
---|
| 27 | self.stderr = None |
---|
| 28 | self.parameters = None |
---|
| 29 | self.model = model |
---|
| 30 | self.paramList = paramList |
---|
| 31 | |
---|
| 32 | def set_model(self, model): |
---|
[aa36f96] | 33 | """ |
---|
| 34 | """ |
---|
[e0072082] | 35 | self.model = model |
---|
| 36 | |
---|
[90c9cdf] | 37 | def set_fitness(self, fitness): |
---|
[aa36f96] | 38 | """ |
---|
| 39 | """ |
---|
[90c9cdf] | 40 | self.fitness = fitness |
---|
| 41 | |
---|
[e0072082] | 42 | def __str__(self): |
---|
[aa36f96] | 43 | """ |
---|
| 44 | """ |
---|
[e0072082] | 45 | if self.pvec == None and self.model is None and self.paramList is None: |
---|
| 46 | return "No results" |
---|
| 47 | n = len(self.model.parameterset) |
---|
| 48 | |
---|
| 49 | result_param = zip(xrange(n), self.model.parameterset) |
---|
| 50 | L = ["P%-3d %s......|.....%s"%(p[0], p[1], p[1].value) for p in result_param if p[1].name in self.paramList ] |
---|
| 51 | L.append("=== goodness of fit: %s"%(str(self.fitness))) |
---|
| 52 | return "\n".join(L) |
---|
[48882d1] | 53 | |
---|
[e0072082] | 54 | def print_summary(self): |
---|
[aa36f96] | 55 | """ |
---|
| 56 | """ |
---|
[e0072082] | 57 | print self |
---|
[88b5e83] | 58 | |
---|
[4c718654] | 59 | class ScipyFit(FitEngine): |
---|
[7705306] | 60 | """ |
---|
[aa36f96] | 61 | ScipyFit performs the Fit.This class can be used as follow: |
---|
| 62 | #Do the fit SCIPY |
---|
| 63 | create an engine: engine = ScipyFit() |
---|
| 64 | Use data must be of type plottable |
---|
| 65 | Use a sans model |
---|
| 66 | |
---|
| 67 | Add data with a dictionnary of FitArrangeDict where Uid is a key and data |
---|
| 68 | is saved in FitArrange object. |
---|
| 69 | engine.set_data(data,Uid) |
---|
| 70 | |
---|
| 71 | Set model parameter "M1"= model.name add {model.parameter.name:value}. |
---|
| 72 | |
---|
| 73 | :note: Set_param() if used must always preceded set_model() |
---|
| 74 | for the fit to be performed.In case of Scipyfit set_param is called in |
---|
| 75 | fit () automatically. |
---|
| 76 | |
---|
| 77 | engine.set_param( model,"M1", {'A':2,'B':4}) |
---|
| 78 | |
---|
| 79 | Add model with a dictionnary of FitArrangeDict{} where Uid is a key and model |
---|
| 80 | is save in FitArrange object. |
---|
| 81 | engine.set_model(model,Uid) |
---|
| 82 | |
---|
| 83 | engine.fit return chisqr,[model.parameter 1,2,..],[[err1....][..err2...]] |
---|
| 84 | chisqr1, out1, cov1=engine.fit({model.parameter.name:value},qmin,qmax) |
---|
[7705306] | 85 | """ |
---|
[792db7d5] | 86 | def __init__(self): |
---|
| 87 | """ |
---|
[aa36f96] | 88 | Creates a dictionary (self.fitArrangeDict={})of FitArrange elements |
---|
| 89 | with Uid as keys |
---|
[792db7d5] | 90 | """ |
---|
[393f0f3] | 91 | self.fitArrangeDict={} |
---|
| 92 | self.paramList=[] |
---|
[d9dc518] | 93 | #def fit(self, *args, **kw): |
---|
| 94 | # return profile(self._fit, *args, **kw) |
---|
[393f0f3] | 95 | |
---|
[e0072082] | 96 | def fit(self, q=None, handler=None, curr_thread=None): |
---|
[aa36f96] | 97 | """ |
---|
| 98 | """ |
---|
[393f0f3] | 99 | fitproblem=[] |
---|
| 100 | for id ,fproblem in self.fitArrangeDict.iteritems(): |
---|
| 101 | if fproblem.get_to_fit()==1: |
---|
| 102 | fitproblem.append(fproblem) |
---|
| 103 | if len(fitproblem)>1 : |
---|
[e0072082] | 104 | msg = "Scipy can't fit more than a single fit problem at a time." |
---|
| 105 | raise RuntimeError, msg |
---|
[a9e04aa] | 106 | return |
---|
[393f0f3] | 107 | elif len(fitproblem)==0 : |
---|
[a9e04aa] | 108 | raise RuntimeError, "No Assembly scheduled for Scipy fitting." |
---|
| 109 | return |
---|
| 110 | |
---|
[393f0f3] | 111 | listdata=[] |
---|
| 112 | model = fitproblem[0].get_model() |
---|
| 113 | listdata = fitproblem[0].get_data() |
---|
[792db7d5] | 114 | # Concatenate dList set (contains one or more data)before fitting |
---|
[e0072082] | 115 | data = listdata |
---|
[393f0f3] | 116 | self.curr_thread= curr_thread |
---|
[e0072082] | 117 | result = fitresult(model=model, paramList=self.paramList) |
---|
| 118 | if handler is not None: |
---|
| 119 | handler.set_result(result=result) |
---|
[fd6b789] | 120 | #try: |
---|
[e0072082] | 121 | functor = SansAssembly(self.paramList, model, data, handler=handler, |
---|
| 122 | fitresult=result,curr_thread= self.curr_thread) |
---|
| 123 | |
---|
| 124 | |
---|
| 125 | out, cov_x, info, mesg, success = optimize.leastsq(functor, |
---|
| 126 | model.getParams(self.paramList), |
---|
| 127 | full_output=1, warning=True) |
---|
[fd6b789] | 128 | |
---|
| 129 | chisqr = functor.chisq(out) |
---|
[e71440c] | 130 | |
---|
[fd6b789] | 131 | if cov_x is not None and numpy.isfinite(cov_x).all(): |
---|
| 132 | stderr = numpy.sqrt(numpy.diag(cov_x)) |
---|
| 133 | else: |
---|
[e0072082] | 134 | stderr = None |
---|
[fd6b789] | 135 | if not (numpy.isnan(out).any()) or ( cov_x !=None) : |
---|
| 136 | result.fitness = chisqr |
---|
| 137 | result.stderr = stderr |
---|
| 138 | result.pvec = out |
---|
| 139 | result.success = success |
---|
[59d8b56] | 140 | #print result |
---|
[e0072082] | 141 | if q is not None: |
---|
| 142 | #print "went here" |
---|
[fd6b789] | 143 | q.put(result) |
---|
[e0072082] | 144 | #print "get q scipy fit enfine",q.get() |
---|
[fd6b789] | 145 | return q |
---|
| 146 | return result |
---|
| 147 | else: |
---|
[393f0f3] | 148 | raise ValueError, "SVD did not converge"+str(success) |
---|
[e0072082] | 149 | |
---|
[d8a2e31] | 150 | |
---|
| 151 | |
---|
[9c648c7] | 152 | def profile(fn, *args, **kw): |
---|
| 153 | import cProfile, pstats, os |
---|
| 154 | global call_result |
---|
| 155 | def call(): |
---|
| 156 | global call_result |
---|
| 157 | call_result = fn(*args, **kw) |
---|
| 158 | cProfile.runctx('call()', dict(call=call), {}, 'profile.out') |
---|
| 159 | stats = pstats.Stats('profile.out') |
---|
| 160 | #stats.sort_stats('time') |
---|
| 161 | stats.sort_stats('calls') |
---|
| 162 | stats.print_stats() |
---|
| 163 | os.unlink('profile.out') |
---|
| 164 | return call_result |
---|
| 165 | |
---|
[48882d1] | 166 | |
---|