1 | """ |
---|
2 | Adds a linear fit plot to the chart |
---|
3 | """ |
---|
4 | import re |
---|
5 | import numpy |
---|
6 | from PyQt4 import QtGui |
---|
7 | from PyQt4 import QtCore |
---|
8 | |
---|
9 | #from sas.sasgui.transform import |
---|
10 | from sas.qtgui.GuiUtils import formatNumber |
---|
11 | from sas.sasgui.plottools import fittings |
---|
12 | from sas.sasgui.plottools import transform |
---|
13 | |
---|
14 | from sas.sasgui.plottools.LineModel import LineModel |
---|
15 | |
---|
16 | # Local UI |
---|
17 | from sas.qtgui.UI.LinearFitUI import Ui_LinearFitUI |
---|
18 | |
---|
19 | class LinearFit(QtGui.QDialog, Ui_LinearFitUI): |
---|
20 | def __init__(self, parent=None, |
---|
21 | data=None, |
---|
22 | max_range=(0.0, 0.0), |
---|
23 | fit_range=(0.0, 0.0), |
---|
24 | xlabel="", |
---|
25 | ylabel=""): |
---|
26 | super(LinearFit, self).__init__() |
---|
27 | |
---|
28 | self.setupUi(self) |
---|
29 | assert(isinstance(max_range, tuple)) |
---|
30 | assert(isinstance(fit_range, tuple)) |
---|
31 | |
---|
32 | self.data = data |
---|
33 | self.parent = parent |
---|
34 | |
---|
35 | self.max_range = max_range |
---|
36 | self.fit_range = fit_range |
---|
37 | self.xLabel = xlabel |
---|
38 | self.yLabel = ylabel |
---|
39 | |
---|
40 | self.x_is_log = self.xLabel == "log10(x)" |
---|
41 | self.y_is_log = self.yLabel == "log10(y)" |
---|
42 | |
---|
43 | self.txtFitRangeMin.setValidator(QtGui.QDoubleValidator()) |
---|
44 | self.txtFitRangeMax.setValidator(QtGui.QDoubleValidator()) |
---|
45 | |
---|
46 | # Default values in the line edits |
---|
47 | self.txtA.setText("1") |
---|
48 | self.txtB.setText("1") |
---|
49 | self.txtAerr.setText("0") |
---|
50 | self.txtBerr.setText("0") |
---|
51 | self.txtChi2.setText("0") |
---|
52 | |
---|
53 | # Initial ranges |
---|
54 | self.txtRangeMin.setText(str(max_range[0])) |
---|
55 | self.txtRangeMax.setText(str(max_range[1])) |
---|
56 | self.txtFitRangeMin.setText(str(fit_range[0])) |
---|
57 | self.txtFitRangeMax.setText(str(fit_range[1])) |
---|
58 | |
---|
59 | # cast xLabel into html |
---|
60 | label = re.sub(r'\^\((.)\)(.*)', r'<span style=" vertical-align:super;">\1</span>\2', |
---|
61 | str(self.xLabel).rstrip()) |
---|
62 | self.lblRange.setText('Fit range of ' + label) |
---|
63 | |
---|
64 | self.model = LineModel() |
---|
65 | # Display the fittings values |
---|
66 | self.default_A = self.model.getParam('A') |
---|
67 | self.default_B = self.model.getParam('B') |
---|
68 | self.cstA = fittings.Parameter(self.model, 'A', self.default_A) |
---|
69 | self.cstB = fittings.Parameter(self.model, 'B', self.default_B) |
---|
70 | self.transform = transform |
---|
71 | |
---|
72 | self.setFixedSize(self.minimumSizeHint()) |
---|
73 | |
---|
74 | # connect Fit button |
---|
75 | self.cmdFit.clicked.connect(self.fit) |
---|
76 | |
---|
77 | def setRangeLabel(self, label=""): |
---|
78 | """ |
---|
79 | Overwrite default fit range label to correspond to actual unit |
---|
80 | """ |
---|
81 | assert(isinstance(label, basestring)) |
---|
82 | self.lblRange.setText(label) |
---|
83 | |
---|
84 | def range(self): |
---|
85 | return (float(self.txtFitRangeMin.text()), float(self.txtFitRangeMax.text())) |
---|
86 | |
---|
87 | def fit(self, event): |
---|
88 | """ |
---|
89 | Performs the fit. Receive an event when clicking on |
---|
90 | the button Fit.Computes chisqr , |
---|
91 | A and B parameters of the best linear fit y=Ax +B |
---|
92 | Push a plottable to the caller |
---|
93 | """ |
---|
94 | tempx = [] |
---|
95 | tempy = [] |
---|
96 | tempdy = [] |
---|
97 | |
---|
98 | # Checks to assure data correctness |
---|
99 | if len(self.data.view.x) < 2: |
---|
100 | return |
---|
101 | if not self.checkFitValues(self.txtFitRangeMin): |
---|
102 | return |
---|
103 | |
---|
104 | self.xminFit, self.xmaxFit = self.range() |
---|
105 | |
---|
106 | xmin = self.xminFit |
---|
107 | xmax = self.xmaxFit |
---|
108 | xminView = xmin |
---|
109 | xmaxView = xmax |
---|
110 | |
---|
111 | # Set the qmin and qmax in the panel that matches the |
---|
112 | # transformed min and max |
---|
113 | #value_xmin = X_VAL_DICT[self.xLabel].floatTransform(xmin) |
---|
114 | #value_xmax = X_VAL_DICT[self.xLabel].floatTransform(xmax) |
---|
115 | |
---|
116 | value_xmin = self.floatInvTransform(xmin) |
---|
117 | value_xmax = self.floatInvTransform(xmax) |
---|
118 | self.txtRangeMin.setText(formatNumber(value_xmin)) |
---|
119 | self.txtRangeMax.setText(formatNumber(value_xmax)) |
---|
120 | |
---|
121 | tempx, tempy, tempdy = self.origData() |
---|
122 | |
---|
123 | # Find the fitting parameters |
---|
124 | self.cstA = fittings.Parameter(self.model, 'A', self.default_A) |
---|
125 | self.cstB = fittings.Parameter(self.model, 'B', self.default_B) |
---|
126 | tempdy = numpy.asarray(tempdy) |
---|
127 | tempdy[tempdy == 0] = 1 |
---|
128 | |
---|
129 | if self.x_is_log: |
---|
130 | xmin = numpy.log10(xmin) |
---|
131 | xmax = numpy.log10(xmax) |
---|
132 | |
---|
133 | chisqr, out, cov = fittings.sasfit(self.model, |
---|
134 | [self.cstA, self.cstB], |
---|
135 | tempx, tempy, tempdy, |
---|
136 | xmin, xmax) |
---|
137 | # Use chi2/dof |
---|
138 | if len(tempx) > 0: |
---|
139 | chisqr = chisqr / len(tempx) |
---|
140 | |
---|
141 | # Check that cov and out are iterable before displaying them |
---|
142 | errA = numpy.sqrt(cov[0][0]) if cov is not None else 0 |
---|
143 | errB = numpy.sqrt(cov[1][1]) if cov is not None else 0 |
---|
144 | cstA = out[0] if out is not None else 0.0 |
---|
145 | cstB = out[1] if out is not None else 0.0 |
---|
146 | |
---|
147 | # Reset model with the right values of A and B |
---|
148 | self.model.setParam('A', float(cstA)) |
---|
149 | self.model.setParam('B', float(cstB)) |
---|
150 | |
---|
151 | tempx = [] |
---|
152 | tempy = [] |
---|
153 | y_model = 0.0 |
---|
154 | |
---|
155 | # load tempy with the minimum transformation |
---|
156 | y_model = self.model.run(xmin) |
---|
157 | tempx.append(xminView) |
---|
158 | tempy.append(numpy.power(10, y_model) if self.y_is_log else y_model) |
---|
159 | |
---|
160 | # load tempy with the maximum transformation |
---|
161 | y_model = self.model.run(xmax) |
---|
162 | tempx.append(xmaxView) |
---|
163 | tempy.append(numpy.power(10, y_model) if self.y_is_log else y_model) |
---|
164 | |
---|
165 | # Set the fit parameter display when FitDialog is opened again |
---|
166 | self.Avalue = cstA |
---|
167 | self.Bvalue = cstB |
---|
168 | self.ErrAvalue = errA |
---|
169 | self.ErrBvalue = errB |
---|
170 | self.Chivalue = chisqr |
---|
171 | |
---|
172 | # Update the widget |
---|
173 | self.txtA.setText(formatNumber(self.Avalue)) |
---|
174 | self.txtAerr.setText(formatNumber(self.ErrAvalue)) |
---|
175 | self.txtB.setText(formatNumber(self.Bvalue)) |
---|
176 | self.txtBerr.setText(formatNumber(self.ErrBvalue)) |
---|
177 | self.txtChi2.setText(formatNumber(self.Chivalue)) |
---|
178 | |
---|
179 | #self.parent.updatePlot.emit((tempx, tempy)) |
---|
180 | self.parent.emit(QtCore.SIGNAL('updatePlot'), (tempx, tempy)) |
---|
181 | |
---|
182 | def origData(self): |
---|
183 | # Store the transformed values of view x, y and dy before the fit |
---|
184 | xmin_check = numpy.log10(self.xminFit) |
---|
185 | # Local shortcuts |
---|
186 | x = self.data.view.x |
---|
187 | y = self.data.view.y |
---|
188 | dy = self.data.view.dy |
---|
189 | |
---|
190 | if self.y_is_log: |
---|
191 | if self.x_is_log: |
---|
192 | tempy = [numpy.log10(y[i]) |
---|
193 | for i in range(len(x)) if x[i] >= xmin_check] |
---|
194 | tempdy = [transform.errToLogX(y[i], 0, dy[i], 0) |
---|
195 | for i in range(len(x)) if x[i] >= xmin_check] |
---|
196 | else: |
---|
197 | tempy = map(numpy.log10, y) |
---|
198 | tempdy = map(lambda t1,t2:transform.errToLogX(t1,0,t2,0),y,dy) |
---|
199 | else: |
---|
200 | tempy = y |
---|
201 | tempdy = dy |
---|
202 | |
---|
203 | if self.x_is_log: |
---|
204 | tempx = [numpy.log10(x) for x in self.data.view.x if x > xmin_check] |
---|
205 | else: |
---|
206 | tempx = x |
---|
207 | |
---|
208 | return tempx, tempy, tempdy |
---|
209 | |
---|
210 | def checkFitValues(self, item): |
---|
211 | """ |
---|
212 | Check the validity of input values |
---|
213 | """ |
---|
214 | flag = True |
---|
215 | value = item.text() |
---|
216 | p_white = item.palette() |
---|
217 | p_white.setColor(item.backgroundRole(), QtCore.Qt.white) |
---|
218 | p_pink = item.palette() |
---|
219 | p_pink.setColor(item.backgroundRole(), QtGui.QColor(255, 128, 128)) |
---|
220 | item.setAutoFillBackground(True) |
---|
221 | # Check for possible values entered |
---|
222 | if self.x_is_log: |
---|
223 | if float(value) > 0: |
---|
224 | item.setPalette(p_white) |
---|
225 | else: |
---|
226 | flag = False |
---|
227 | item.setPalette(p_pink) |
---|
228 | return flag |
---|
229 | |
---|
230 | def floatInvTransform(self, x): |
---|
231 | """ |
---|
232 | transform a float.It is used to determine the x.View min and x.View |
---|
233 | max for values not in x. Also used to properly calculate RgQmin, |
---|
234 | RgQmax and to update qmin and qmax in the linear range boxes on the |
---|
235 | panel. |
---|
236 | |
---|
237 | """ |
---|
238 | # TODO: refactor this. This is just a hack to make the |
---|
239 | # functionality work without rewritting the whole code |
---|
240 | # with good design (which really should be done...). |
---|
241 | if self.xLabel == "x": |
---|
242 | return x |
---|
243 | elif self.xLabel == "x^(2)": |
---|
244 | return numpy.sqrt(x) |
---|
245 | elif self.xLabel == "x^(4)": |
---|
246 | return numpy.sqrt(numpy.sqrt(x)) |
---|
247 | elif self.xLabel == "log10(x)": |
---|
248 | return numpy.power(10, x) |
---|
249 | elif self.xLabel == "ln(x)": |
---|
250 | return numpy.exp(x) |
---|
251 | elif self.xLabel == "log10(x^(4))": |
---|
252 | return numpy.sqrt(numpy.sqrt(numpy.power(10, x))) |
---|
253 | return x |
---|
254 | |
---|
255 | |
---|