| 1 | import _asap |
|---|
| 2 | |
|---|
| 3 | class fitter: |
|---|
| 4 | """ |
|---|
| 5 | The fitting class for ASAP. |
|---|
| 6 | """ |
|---|
| 7 | def _verbose(self, *args): |
|---|
| 8 | """ |
|---|
| 9 | Set stdout output. |
|---|
| 10 | """ |
|---|
| 11 | if type(args[0]) is bool: |
|---|
| 12 | self._vb = args[0] |
|---|
| 13 | return |
|---|
| 14 | elif len(args) == 0: |
|---|
| 15 | return self._vb |
|---|
| 16 | |
|---|
| 17 | def __init__(self): |
|---|
| 18 | """ |
|---|
| 19 | Create a fitter object. No state is set. |
|---|
| 20 | """ |
|---|
| 21 | self.fitter = _asap.fitter() |
|---|
| 22 | self.x = None |
|---|
| 23 | self.y = None |
|---|
| 24 | self.mask = None |
|---|
| 25 | self.fitfunc = None |
|---|
| 26 | self.fitted = False |
|---|
| 27 | self.data = None |
|---|
| 28 | self._p = None |
|---|
| 29 | self._vb = True |
|---|
| 30 | |
|---|
| 31 | def set_data(self, xdat, ydat, mask=None): |
|---|
| 32 | """ |
|---|
| 33 | Set the absissa and ordinate for the fit. Also set the mask |
|---|
| 34 | indicationg valid points. |
|---|
| 35 | This can be used for data vectors retrieved from a scantable. |
|---|
| 36 | For scantable fitting use 'fitter.set_scan(scan, mask)'. |
|---|
| 37 | Parameters: |
|---|
| 38 | xdat: the abcissa values |
|---|
| 39 | ydat: the ordinate values |
|---|
| 40 | mask: an optional mask |
|---|
| 41 | |
|---|
| 42 | """ |
|---|
| 43 | self.fitted = False |
|---|
| 44 | self.x = xdat |
|---|
| 45 | self.y = ydat |
|---|
| 46 | if mask == None: |
|---|
| 47 | from numarray import ones |
|---|
| 48 | self.mask = ones(len(xdat)) |
|---|
| 49 | else: |
|---|
| 50 | self.mask = mask |
|---|
| 51 | return |
|---|
| 52 | |
|---|
| 53 | def set_scan(self, thescan=None, mask=None): |
|---|
| 54 | """ |
|---|
| 55 | Set the 'data' (a scantable) of the fitter. |
|---|
| 56 | Parameters: |
|---|
| 57 | thescan: a scantable |
|---|
| 58 | mask: a msk retireved from the scantable |
|---|
| 59 | """ |
|---|
| 60 | if not thescan: |
|---|
| 61 | print "Please give a correct scan" |
|---|
| 62 | self.fitted = False |
|---|
| 63 | self.data = thescan |
|---|
| 64 | if mask is None: |
|---|
| 65 | from numarray import ones |
|---|
| 66 | self.mask = ones(self.data.nchan()) |
|---|
| 67 | else: |
|---|
| 68 | self.mask = mask |
|---|
| 69 | return |
|---|
| 70 | |
|---|
| 71 | def set_function(self, **kwargs): |
|---|
| 72 | """ |
|---|
| 73 | Set the function to be fit. |
|---|
| 74 | Parameters: |
|---|
| 75 | poly: use a polynomial of the order given |
|---|
| 76 | gauss: fit the number of gaussian specified |
|---|
| 77 | Example: |
|---|
| 78 | fitter.set_function(gauss=2) # will fit two gaussians |
|---|
| 79 | fitter.set_function(poly=3) # will fit a 3rd order polynomial |
|---|
| 80 | """ |
|---|
| 81 | #default poly order 0 |
|---|
| 82 | self.fitfunc = 'poly' |
|---|
| 83 | n=0 |
|---|
| 84 | if kwargs.has_key('poly'): |
|---|
| 85 | self.fitfunc = 'poly' |
|---|
| 86 | n = kwargs.get('poly') |
|---|
| 87 | elif kwargs.has_key('gauss'): |
|---|
| 88 | n = kwargs.get('gauss') |
|---|
| 89 | self.fitfunc = 'gauss' |
|---|
| 90 | |
|---|
| 91 | self.fitter.setexpression(self.fitfunc,n) |
|---|
| 92 | return |
|---|
| 93 | |
|---|
| 94 | def fit(self): |
|---|
| 95 | """ |
|---|
| 96 | Execute the actual fitting process. All the state has to be set. |
|---|
| 97 | Parameters: |
|---|
| 98 | none |
|---|
| 99 | Example: |
|---|
| 100 | s= scantable('myscan.asap') |
|---|
| 101 | f = fitter() |
|---|
| 102 | f.set_scan(s) |
|---|
| 103 | f.set_function(poly=0) |
|---|
| 104 | f.fit() |
|---|
| 105 | """ |
|---|
| 106 | if ((self.x is None or self.y is None) and self.data is None) \ |
|---|
| 107 | or self.fitfunc is None: |
|---|
| 108 | print "Fitter not yet initialised. Please set data & fit function" |
|---|
| 109 | return |
|---|
| 110 | else: |
|---|
| 111 | if self.data is not None: |
|---|
| 112 | self.x = self.data.getabcissa() |
|---|
| 113 | self.y = self.data.getspectrum() |
|---|
| 114 | print "Fitting:" |
|---|
| 115 | vb = self.data._verbose |
|---|
| 116 | self.data._verbose(True) |
|---|
| 117 | s = self.data.get_selection() |
|---|
| 118 | self.data._verbose(vb) |
|---|
| 119 | |
|---|
| 120 | self.fitter.setdata(self.x,self.y,self.mask) |
|---|
| 121 | if self.fitfunc == 'gauss': |
|---|
| 122 | ps = self.fitter.getparameters() |
|---|
| 123 | if len(ps) == 0: |
|---|
| 124 | self.fitter.estimate() |
|---|
| 125 | self.fitter.fit() |
|---|
| 126 | self.fitted = True |
|---|
| 127 | return |
|---|
| 128 | |
|---|
| 129 | def set_parameters(self, params, fixed=None): |
|---|
| 130 | self.fitter.setparameters(params) |
|---|
| 131 | if fixed is not None: |
|---|
| 132 | self.fitter.setfixedparameters(fixed) |
|---|
| 133 | return |
|---|
| 134 | |
|---|
| 135 | def get_parameters(self): |
|---|
| 136 | """ |
|---|
| 137 | Return the fit paramters. |
|---|
| 138 | |
|---|
| 139 | """ |
|---|
| 140 | if not self.fitted: |
|---|
| 141 | print "Not yet fitted." |
|---|
| 142 | pars = list(self.fitter.getparameters()) |
|---|
| 143 | fixed = list(self.fitter.getfixedparameters()) |
|---|
| 144 | if self._vb: |
|---|
| 145 | print self._format_pars(pars) |
|---|
| 146 | return pars,fixed |
|---|
| 147 | |
|---|
| 148 | def _format_pars(self, pars): |
|---|
| 149 | out = '' |
|---|
| 150 | if self.fitfunc == 'poly': |
|---|
| 151 | c = 0 |
|---|
| 152 | for i in pars: |
|---|
| 153 | out += ' p%d = %3.3f, ' % (c,i) |
|---|
| 154 | c+=1 |
|---|
| 155 | elif self.fitfunc == 'gauss': |
|---|
| 156 | i = 0 |
|---|
| 157 | c = 0 |
|---|
| 158 | unit = '' |
|---|
| 159 | if self.data: |
|---|
| 160 | unit = self.data.get_unit() |
|---|
| 161 | while i < len(pars): |
|---|
| 162 | out += ' %d: peak = %3.3f , centre = %3.3f %s, FWHM = %3.3f %s \n' % (c,pars[i],pars[i+1],unit,pars[i+2],unit) |
|---|
| 163 | c+=1 |
|---|
| 164 | i+=3 |
|---|
| 165 | return out |
|---|
| 166 | |
|---|
| 167 | def get_estimate(self): |
|---|
| 168 | """ |
|---|
| 169 | Return the paramter estimates (for non-linear functions). |
|---|
| 170 | """ |
|---|
| 171 | pars = self.fitter.getestimate() |
|---|
| 172 | if self._vb: |
|---|
| 173 | print self._format_pars(pars) |
|---|
| 174 | return pars |
|---|
| 175 | |
|---|
| 176 | |
|---|
| 177 | def get_residual(self): |
|---|
| 178 | """ |
|---|
| 179 | Return the residual of the fit. |
|---|
| 180 | """ |
|---|
| 181 | if not self.fitted: |
|---|
| 182 | print "Not yet fitted." |
|---|
| 183 | return self.fitter.getresidual() |
|---|
| 184 | |
|---|
| 185 | def get_chi2(self): |
|---|
| 186 | """ |
|---|
| 187 | Return chi^2. |
|---|
| 188 | """ |
|---|
| 189 | |
|---|
| 190 | if not self.fitted: |
|---|
| 191 | print "Not yet fitted." |
|---|
| 192 | ch2 = self.fitter.getchi2() |
|---|
| 193 | if self._vb: |
|---|
| 194 | print 'Chi^2 = %3.3f' % (ch2) |
|---|
| 195 | return ch2 |
|---|
| 196 | |
|---|
| 197 | def get_fit(self): |
|---|
| 198 | """ |
|---|
| 199 | Return the fitted ordinate values. |
|---|
| 200 | """ |
|---|
| 201 | if not self.fitted: |
|---|
| 202 | print "Not yet fitted." |
|---|
| 203 | return self.fitter.getfit() |
|---|
| 204 | |
|---|
| 205 | def commit(self): |
|---|
| 206 | """ |
|---|
| 207 | Return a new scan where teh fits have been commited. |
|---|
| 208 | """ |
|---|
| 209 | if not self.fitted: |
|---|
| 210 | print "Not yet fitted." |
|---|
| 211 | if self.data is not scantable: |
|---|
| 212 | print "Only works with scantables" |
|---|
| 213 | return |
|---|
| 214 | scan = self.data.copy() |
|---|
| 215 | scan.setspectrum(self.fitter.getresidual()) |
|---|
| 216 | |
|---|
| 217 | def plot(self, residual=False): |
|---|
| 218 | """ |
|---|
| 219 | Plot the last fit. |
|---|
| 220 | Parameters: |
|---|
| 221 | residual: an optional parameter indicating if the residual |
|---|
| 222 | should be plotted (default 'False') |
|---|
| 223 | """ |
|---|
| 224 | if not self.fitted: |
|---|
| 225 | return |
|---|
| 226 | if not self._p: |
|---|
| 227 | from asap.asaplot import ASAPlot |
|---|
| 228 | self._p = ASAPlot() |
|---|
| 229 | self._p.clear() |
|---|
| 230 | tlab = 'Spectrum' |
|---|
| 231 | xlab = 'Abcissa' |
|---|
| 232 | if self.data: |
|---|
| 233 | tlab = self.data._getsourcename(0) |
|---|
| 234 | xlab = self.data.getabcissalabel(0) |
|---|
| 235 | ylab = r'Flux' |
|---|
| 236 | m = self.data.getmask(0) |
|---|
| 237 | self._p.set_line(colour='blue',label='Spectrum') |
|---|
| 238 | self._p.plot(self.x, self.y, m) |
|---|
| 239 | if residual: |
|---|
| 240 | self._p.set_line(colour='green',label='Residual') |
|---|
| 241 | self._p.plot(self.x, self.get_residual(), m) |
|---|
| 242 | self._p.set_line(colour='red',label='Fit') |
|---|
| 243 | self._p.plot(self.x, self.get_fit(), m) |
|---|
| 244 | |
|---|
| 245 | self._p.set_axes('xlabel',xlab) |
|---|
| 246 | self._p.set_axes('ylabel',ylab) |
|---|
| 247 | self._p.set_axes('title',tlab) |
|---|
| 248 | self._p.release() |
|---|
| 249 | |
|---|
| 250 | |
|---|
| 251 | def auto_fit(self): |
|---|
| 252 | """ |
|---|
| 253 | Return a scan where the function is applied to all Beams/IFs/Pols. |
|---|
| 254 | |
|---|
| 255 | """ |
|---|
| 256 | from asap import scantable |
|---|
| 257 | if not isinstance(self.data,scantable) : |
|---|
| 258 | print "Only works with scantables" |
|---|
| 259 | return |
|---|
| 260 | scan = self.data.copy() |
|---|
| 261 | vb = scan._verbose |
|---|
| 262 | scan._verbose(False) |
|---|
| 263 | sel = scan.get_selection() |
|---|
| 264 | for i in range(scan.nbeam()): |
|---|
| 265 | scan.setbeam(i) |
|---|
| 266 | for j in range(scan.nif()): |
|---|
| 267 | scan.setif(j) |
|---|
| 268 | for k in range(scan.npol()): |
|---|
| 269 | scan.setpol(k) |
|---|
| 270 | self.x = scan.getabcissa() |
|---|
| 271 | self.y = scan.getspectrum() |
|---|
| 272 | self.data = None |
|---|
| 273 | self.fit() |
|---|
| 274 | if self._vb: |
|---|
| 275 | print "Fitting:" |
|---|
| 276 | print 'Beam[%d], IF[%d], Pol[%d]' % (i,j,k) |
|---|
| 277 | x = self.get_parameters() |
|---|
| 278 | scan.setspectrum(self.fitter.getresidual()) |
|---|
| 279 | scan.set_selection(sel[0],sel[1],sel[2]) |
|---|
| 280 | scan._verbose(vb) |
|---|
| 281 | return scan |
|---|
| 282 | |
|---|
