Repository navigation
Expand file tree
/
Copy pathAverageDataParser.py
More file actions
407 lines (385 loc) · 15.9 KB
/
Copy pathAverageDataParser.py
File metadata and controls
407 lines (385 loc) · 15.9 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
import numpy
import configparser
from math import sqrt, log
def stripLeadingDigits( word ):
for i in range( len( word ) ):
if word[i].isalpha():
return word[i:]
class AverageDataParser:
# C-tor, read inputs and calculate covariances:
def __init__( self, filename, llogNormal=False ):
self.__correlations= None
self.__filename= filename
self.__readInput( filename, llogNormal )
return
# Read inputs using ConfigParser:
def __readInput( self, filename, llogNormal ):
parser= configparser.ConfigParser( interpolation=None )
parser.read( filename )
self.__readData( parser )
self.__readRvalues( parser )
self.__readGlobals( parser )
self.__readCovariances( parser )
if llogNormal:
self.__transformLogNormal()
self.__makeCovariances()
return
def __transformLogNormal( self ):
herrors= {}
for key in self.__errors.keys():
errors= [ error/value for ( error, value ) in zip( self.__errors[key], self.__inputs ) ]
herrors[key]= errors
self.__errors= herrors
self.__inputs= [ log( value ) for value in self.__inputs ]
return
# Read "Data" section:
def __readData( self, parser ):
tuplelist= parser.items( "Data" )
herrors= {}
hcovopt= {}
grouplist= None
for item in tuplelist:
key= item[0]
value= item[1]
listvalue= value.split()
if key == "names":
names= listvalue
elif key == "values":
ldata= [ float(s) for s in listvalue ]
elif key == "groups":
grouplist= listvalue
else:
hcovopt[key]= listvalue.pop()
herrors[key]= [ float(s) for s in listvalue ]
for key in herrors.keys():
if "%" in hcovopt[key]:
for ierr in range( len(herrors[key]) ):
herrors[key][ierr]*= ldata[ierr]/100.0
if grouplist is None:
grouplist= [ "a" for name in names ]
self.__names= names
self.__inputs= ldata
self.__covopts= hcovopt
self.__errors= herrors
self.__groups= grouplist
groupset= set( grouplist )
ngroups= len( groupset )
groups= sorted( groupset )
groupmatrix= []
for g in grouplist:
groupmatrixrow= ngroups*[0]
index= groups.index( g )
groupmatrixrow[index]= 1
groupmatrix.append( groupmatrixrow )
self.__groupmatrix= groupmatrix
return
def __readRvalues( self, parser ):
hcovopt= self.__covopts
if sum( [ "R" in v for v in hcovopt.values() ] ):
hrvalues= {}
for key in hcovopt.keys():
if "R" in hcovopt[key]:
rvalue= parser.get( "Rvalues", key )
strippedKey= stripLeadingDigits( key )
hrvalues[strippedKey]= float( rvalue )
self.__hrvalues= hrvalues
else:
self.__hrvalues= None
return
def getRvalues( self ):
if self.__hrvalues is None:
return dict()
else:
return dict( self.__hrvalues )
def __readGlobals( self, parser ):
hglobals= {}
try:
tuplelist= parser.items( "Globals" )
except configparser.NoSectionError:
pass
else:
for item in tuplelist:
key= item[0]
value= item[1]
if key == "correlationfactor":
hglobals[key]= float( value )
self.__hglobals= hglobals
return
# Read "Covariances" section if it exists:
def __readCovariances( self, parser ):
hcovopt= self.__covopts
if sum( [ "c" in v or "m" in v for v in hcovopt.values() ] ):
hcovlists= {}
for key in hcovopt.keys():
if "c" in hcovopt[key] or "m" in hcovopt[key]:
covvalues= parser.get( "Covariances", key ).split()
if "c" in hcovopt[key]:
hcovlists[key]= [ float(s) for s in covvalues ]
elif "m" in hcovopt[key]:
hcovlists[key]= covvalues
self.__correlations= hcovlists
return
# Calculate covariances from inputs and keep as numpy matrices:
def __makeZeroMatrix( self ):
ndim= len( self.__inputs )
return numpy.matrix( numpy.zeros( shape=(ndim,ndim) ) )
def __makeMatrixFromList( self, lelements ):
m= numpy.matrix( lelements )
ndim= len( self.__inputs )
m.shape= ( ndim, ndim )
return m
def __calcCovariance( self, covoption, errors, iderr1, iderr2 ):
err1= errors[iderr1]
err2= errors[iderr2]
if "f" in covoption:
cov= err1*err2
elif "a" in covoption:
if iderr1 == iderr2:
cov= err1**2
else:
cov= -err1*err2
elif "p" in covoption:
cov= min( err1, err2 )**2
elif "u" in covoption:
if iderr1 == iderr2:
cov= err1**2
else:
cov= 0.0
return cov
def __makeCovariances( self ):
# The covariance matrices for each error source
hcov= {}
# The covariance matrices without fully correlated error components
# for each error source
hredcov= {}
systerrormatrix= {}
errorkeys= sorted( self.__errors.keys() )
covoptions= self.__covopts
for errorkey in errorkeys:
nerr= errorkeys.index( errorkey )
lcov= []
lredcov= []
errors= self.__errors[errorkey]
nerrors= len( errors )
covoption= self.__covopts[errorkey]
# Global options, all covariances according to
# same rule gpr, gp, p, f, a or u:
if "gpr" in covoption:
minrelerr= min( [ err/value for err, value in
zip( errors, self.__inputs ) if err > 0.0 ] )
systerrlist= []
for iderr1 in range( nerrors ):
for iderr2 in range( nerrors ):
if iderr1 == iderr2:
covelement= errors[iderr1]**2
redcovelement= errors[iderr1]**2 - (minrelerr*self.__inputs[iderr1])**2
redcovelement= max( redcovelement, 0.0 )
else:
covelement= minrelerr**2*self.__inputs[iderr1]*self.__inputs[iderr2]
redcovelement= 0.0
lcov.append( covelement )
lredcov.append( redcovelement )
systerrlist.append( minrelerr*self.__inputs[iderr1] )
systerrormatrix[nerr]= systerrlist
elif( "gp" in covoption ):
minerr= min( [ error for error in errors if error > 0.0 ] )
systerrlist= []
for iderr1 in range( nerrors ):
for iderr2 in range( nerrors ):
if iderr1 == iderr2:
covelement= errors[iderr1]**2
redcovelement= errors[iderr1]**2 - minerr**2
else:
covelement= minerr**2
redcovelement= 0.0
lcov.append( covelement )
lredcov.append( redcovelement )
systerrlist.append( minerr )
systerrormatrix[nerr]= systerrlist
# Direct calculation from "f", "p", "u" or "a":
elif( "f" in covoption or "p" in covoption or
"u" in covoption or "a" in covoption ):
lcov= [ self.__calcCovariance( covoption, errors, iderr1, iderr2 )
for iderr1 in range( nerrors )
for iderr2 in range( nerrors ) ]
if( "f" in covoption ):
systerrormatrix[nerr]= errors
lredcov= len(errors)**2*[0.0]
else:
lredcov= lcov
# Covariances from correlations and errors:
elif "c" in covoption:
corrlist= self.__correlations[errorkey]
err1err2= [ err1*err2 for err1 in errors for err2 in errors ]
lcov= [ corr*errprod
for corr, errprod in zip( corrlist, err1err2 ) ]
# "Onionisation":
if "o" in covoption:
for iderr1 in range( nerrors ):
for iderr2 in range( nerrors ):
if errors[iderr1] > 0.0 and errors[iderr2] > 0.0:
ierr= iderr1*nerrors+iderr2
lcov[ierr]= min( lcov[ierr],
min( errors[iderr1], errors[iderr2] )**2 )
lredcov= lcov
# Covariances from options:
elif "m" in covoption:
mcovopts= self.__correlations[errorkey]
iderr1err2= [ ( iderr1, iderr2 ) for iderr1 in range( nerrors )
for iderr2 in range( nerrors ) ]
lcov= [ self.__calcCovariance( mcovopt, errors, iderr[0], iderr[1] )
for mcovopt, iderr in zip( mcovopts, iderr1err2 ) ]
if( "f" in mcovopts and not "p" in mcovopts ):
systerrormatrix[nerr]= errors
lredcov= len(errors)**2*[0.0]
else:
lredcov= lcov
# Error in option:
else:
raise RuntimeError( "Option", covoption, "not recognised" )
# Final covariance matrix
m= self.__makeMatrixFromList( lcov )
# Apply correlation factor to final cov.matrix if present
if "correlationfactor" in self.__hglobals:
corrfac= self.__hglobals["correlationfactor"]
for i in range( m.shape[0] ):
for j in range( m.shape[1] ):
if i != j and m[i,i]*m[j,j] != 0.0:
m[i,j]*= m[i,j]/(sqrt(m[i,i]*m[j,j]))*corrfac
# Store final and reduced cov.matrices
redm= self.__makeMatrixFromList( lredcov )
hcov[errorkey]= m
hredcov[errorkey]= redm
# Build total full and reduced covariance matrices, reduced means
# all errors except fully correlated (see above)
totalcov= self.__makeZeroMatrix()
redcov= self.__makeZeroMatrix()
for errorkey in errorkeys:
redcov+= hredcov[errorkey]
for errorkey in errorkeys:
covoption= self.__covopts[errorkey]
if "fq" in covoption:
# Covariance matrix a la Neudecker et al and Bohm und Zech for
# not fully correlated errors, modify errors such that "f" treatment
# would produce the covariance matrix
gm= numpy.matrix( self.__groupmatrix )
inv= redcov.getI()
utvinvu= gm.getT()*inv*gm
utvinvuinv= utvinvu.getI()
wm= utvinvuinv*gm.getT()*inv
values= numpy.matrix( self.__inputs )
values.shape= ( len( self.__inputs ), 1 )
avg= wm*values
# Fully correlated cov.matrix elements
errors= numpy.matrix( self.__errors[errorkey] )
errors.shape= ( len( self.__inputs ), 1 )
relerrs= errors/values
avgerrs= numpy.multiply( gm*avg, relerrs )
cov= avgerrs*avgerrs.T
# Replace covariance matrix and systerrormatrix for this errorkey
hcov[errorkey]= cov
nerr= errorkeys.index( errorkey )
lerrors= [ avgerr.item() for avgerr in avgerrs ]
systerrormatrix[nerr]= lerrors
totalcov+= hcov[errorkey]
# Keep results as members:
self.__hcov= hcov
self.__hredcov= hredcov
self.__cov= totalcov
self.__redcov= redcov
self.__systerrormatrix= systerrormatrix
return
# Print inputs:
def printInputs( self, keys=None ):
if keys is None:
keys= sorted( self.__errors.keys() )
print( "\n AverageDataParser: input from", self.__filename )
print( "\n Variables:", end= " " )
for name in self.__names:
print( "{0:>10s}".format( name ), end=" " )
print( "Cov. opt.", end=" " )
if sum( [ "R" in opt for opt in self.__covopts.values() ] ):
print( "Error uncertainty" )
else:
print()
if len( set( self.__groups ) ) > 1:
print( "\n Groups:", end=" " )
for groupindex in self.__groups:
print( "{0:>10s}".format( groupindex ), end="" )
print()
print( "\n Values:", end=" " )
for value in self.__inputs:
print( "{0:10.4f}".format( value ), end=" " )
print()
for key in keys:
print( "{0:>10s}:".format( stripLeadingDigits( key ) ), end=" " )
for error in self.__errors[key]:
print( "{0:10.4f}".format( error ), end=" " )
print( self.__covopts[key], end=" " )
if "R" in self.__covopts[key]:
print( "{0:11.2f}".format( self.__hrvalues[stripLeadingDigits(key)] ) )
else:
print()
totalerrors= self.getTotalErrors()
print( "\n total:", end=" " )
for error in totalerrors:
print( "{0:10.4f}".format( error ), end=" " )
print()
if self.__correlations:
print( "\nCorrelations:" )
keys= sorted( self.__correlations.keys() )
for key in keys:
print( "\n{0:s}:".format( stripLeadingDigits( key ) ) )
covopt= self.__covopts[key]
correlations= self.__correlations[key]
n= int( sqrt( len(correlations) ) )
for i in range(n):
for j in range(n):
ii= i*n+j
if "c" in covopt:
print( "{0:6.3f}".format( correlations[ii] ),
end=" " )
else:
print( correlations[ii], end=" " )
print()
return
# Retrieve values:
def getFilename( self ):
return str( self.__filename )
def getNames( self ):
return list( self.__names )
def getValues( self ):
return list( self.__inputs )
def getErrors( self ):
return dict( self.__errors )
def getTotalErrors( self ):
totalerrors= len( self.__inputs )*[0]
for errors in self.__errors.values():
for ierr in range( len( errors ) ):
totalerrors[ierr]+= errors[ierr]**2
for ierr in range( len( totalerrors ) ):
totalerrors[ierr]= sqrt( totalerrors[ierr] )
return totalerrors
def getCovoption( self ):
return dict( self.__covopts )
def getCorrelations( self ):
if self.__correlations is None:
return None
else:
return dict( self.__correlations )
def getCovariances( self ):
return dict( self.__hcov )
def getTotalCovariance( self ):
return self.__cov.copy()
def getGroups( self ):
return list( self.__groups )
def getGroupMatrix( self ):
return list( self.__groupmatrix )
def getSysterrorMatrix( self ):
return dict( self.__systerrormatrix )
def getReducedCovariances( self ):
return dict( self.__hredcov )
def getTotalReducedCovariance( self ):
return self.__redcov.copy()
def getTotalReducedCovarianceAslist( self ):
return self.__redcov.tolist()