/
dalek.py
executable file
·321 lines (290 loc) · 14.4 KB
/
dalek.py
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
from scipy.integrate import trapz
from scipy import polyfit, ndimage
import pdb
import os
import pickle
import copy
import numpy as np
import dalekSteps,launcherSteps,read,param,initialize,fitelem,plot,util
import pprint
import select
import sys
seedno=250819801106
def initLum(doPickle=True):
runDirs,params=launcherSteps.initLaunchLum()
lumModelGrid=read.modelGrid(runDirs,params)
basePath=os.path.split(runDirs[0])[0]
if doPickle:
pickle.dump(lumModelGrid,file(os.path.join(basePath,'lum0.pkl'),'w'))
return lumModelGrid,dalekSteps.getNextLumInterval(lumModelGrid)
def runLumCycle(doPickle=True,maxIter=None,meritThresh=0.01,samples=10,fitHist=None):
if fitHist==None: fitHist=param.fitHistory()
lumModelGrid,interval=initLum()
#fitHist.addHistItem([interval],[lumModelGrid],[None])
#print interval
i=1
while True:
runDirs,params=launcherSteps.launchLum(np.linspace(interval['interval'][0],interval['interval'][1],samples))
lumModelGrid=read.modelGrid(runDirs,params)
basePath=os.path.split(runDirs[0])[0]
if doPickle:
pickle.dump(lumModelGrid,file(os.path.join(basePath,'lum%d.pkl'%i),'w'))
i+=1
interval=dalekSteps.getNextLumInterval(lumModelGrid)
#fitHist.addHistItem([interval],[lumModelGrid],[None])
print "Current interval:",interval
if maxIter!=None:
if i>maxIter: break
#fitHist.write2pickle('lumCycle.pkl')
if interval['merit']<meritThresh: break
return interval
def initTriCycle(IGEElement,doPickle=True,samples=10):
fitHist=param.fitHistory()
lumInterval=runLumCycle(maxIter=5)
curParam=param.param()
curParam['lum']=lumInterval['suggestValue']
intervals={'luminterval':lumInterval['interval'],
'vphinterval':initialize.getVphBounds(),
#make sure that no other metal really shoots up
'igeinterval':initialize.getElementBounds(IGEElement,curParam.comp)}
#pdb.set_trace()
lumRange=np.linspace(intervals['luminterval'][0],intervals['luminterval'][1],num=samples)
vphRange=np.linspace(intervals['vphinterval'][0],intervals['vphinterval'][1],num=samples)
IGERange=np.linspace(intervals['igeinterval'][0],intervals['igeinterval'][1],num=samples)
lumMG,vphMG,IGEMG,curParamMG=launcherSteps.launchTriCycle(lumRange,vphRange,IGERange,IGEElement,initParam=curParam)
if doPickle:
pickle.dump(lumMG,file('lumMG0.pkl','w'))
pickle.dump(vphMG,file('vphMG0.pkl','w'))
pickle.dump(IGEMG,file('igeMG0.pkl','w'))
evalTriCycle(lumMG,vphMG,IGEMG,curParamMG,curParam,intervals,fitHist,IGEElement,mode='init')
return fitHist,curParam,intervals
def runTriCycle(param=None,samples=10,fitHist=None,doPickle=True):
IGEElement='Fe0'
fitHist,curParam,intervals=initTriCycle(IGEElement)
i=1
while True:
print "starting cycle %s"%i
lumRange=np.linspace(intervals['luminterval'][0],intervals['luminterval'][1],num=samples)
vphRange=np.linspace(intervals['vphinterval'][0],intervals['vphinterval'][1],num=samples)
# IGERange=np.linspace(intervals['igeinterval'][0],intervals['igeinterval'][1],num=samples)
IGERange=np.logspace(np.log10(intervals['igeinterval'][0]),np.log10(intervals['igeinterval'][1]),num=samples)
lumMG,vphMG,IGEMG,curParamMG=launcherSteps.launchTriCycle(lumRange,vphRange,IGERange,IGEElement,initParam=copy.deepcopy(curParam),procPath='tri%02d'%i)
print IGERange
print "EVALING Cycle %s"%i
print
done=evalTriCycle(lumMG,vphMG,IGEMG,curParamMG,curParam,intervals,fitHist,IGEElement)
#pdb.set_trace()
#if done: break
i+=1
pickle.dump(curParam,file('curTriParam.pkl','w'))
return curParam
def evalTriCycle(lumMG,vphMG,IGEMG,curParamMG,curParam,intervals,fitHist,IGEElement,mode='std'):
#IMPORTANT!!!!!!!! the function DOES NOT need curParamMG, only for adding it to the fithist object
#REMOVE AS SOON AS DEBUGGING for
pp=pprint.PrettyPrinter()
selectSteadyThresh=6
steadyThresh=4
pp.pprint(intervals)
pivots=[0,0,0]
lumBounds=initialize.getLumBounds()
vphBounds=initialize.getVphBounds()
IGEBounds=initialize.getElementBounds(IGEElement,curParam.comp)
lumInterval=dalekSteps.getNextLumInterval(lumMG)
vphInterval=dalekSteps.getNextVphInterval(vphMG)
IGEInterval=dalekSteps.getNextIGEInterval(IGEMG,IGEElement)
lastSuggestLum=np.mean(intervals['luminterval'])
lastSuggestVph=np.mean(intervals['vphinterval'])
lastSuggestIGE=np.mean(intervals['igeinterval'])
print "CurLum=%.4f CurVph=%.2f CurFe0=%.7f"%(curParam['lum'],curParam['vph'],curParam['Fe0'])
print "---------------------------------"
print "LUM:"
pp.pprint(lumInterval)
print
print "VPH:"
pp.pprint(vphInterval)
print
print "IGE:"
pp.pprint(IGEInterval)
#Updating the intervals if the solution is not steady
steadyIncFactor=2
selectSteadyIncFactor=2
curLum=curParam['lum']
closestLumID=np.argmin(lumMG['lum']-curLum)
#intervals=[lumInterval,vphInterval,IGEInterval]
merits=np.argsort([item[2] for item in intervals])
#pivots[merits[2]]+=3
#pivots[merits[1]]+=1
pivots[0]=3
pivots[1]=2
pivots[2]=1
#making sure that we are close in luminosity
print IGEMG['Fe0']
if not (abs((lumMG['lum'][closestLumID]-curLum))<0.05 and abs(lumInterval['merits'][closestLumID])<0.1):
print "lum doesnt allow other parameters atm"
print "closestLum %s closestMerit %s"%(lumMG['lum'][closestLumID]-curLum,lumInterval['merits'][closestLumID])
pivots[1]-3
pivots[2]-3
pivots[0]=10
#Checking if the interval is too small already
if lumInterval['mininterval']: pivots[0]-=5
if vphInterval['mininterval']: pivots[1]-=5
#IGE NOT WORKING properly
#pivots[2]=-10
print "PIVOTS %s"%pivots
#Checking w factors:
print 'vph w %s'%vphMG['w']
print 'ige w %s'%IGEMG['w']
if np.sum(np.abs(vphMG['w']-0.5)>0.1)>4: pivots[1]-=1.5
if np.sum(np.abs(IGEMG['w']-0.5)>0.1)>4: pivots[2]-=1
damping=True
#the factor does not do that much atm BEWARE
dampingFactor=2
if np.argmax(pivots)==0:
#param=copy.copy(lumMG.grid[lumInterval['bestFitID']].param)
if lumInterval['steady']>selectSteadyThresh:
if damping:
curParam['lum']=np.mean([lumInterval['suggestValue'],lastSuggestLum])
intervals.update({'luminterval':[curParam['lum']+i*dampingFactor*lumInterval['dev'] for i in [-1,1]]})
else:
curParam['lum']=lumInterval['suggestValue']
intervals.update({'luminterval':lumInterval['interval']})
else:
print "Not steady solution will increase last interval and try again"
dev=selectSteadyIncFactor*np.abs(intervals['luminterval'][0]-curParam['lum'])
intervals.update({'luminterval':[curParam['lum']-dev,curParam['lum']+dev]})
print "Chose Lum: updated for next run with suggestValue %s and Interval %s"%(curParam['lum'],intervals['luminterval'])
else:
if lumInterval['steady']<steadyThresh:
print "Updating the lum interval as the current solution is not steady"
dev=steadyIncFactor*np.abs(intervals['luminterval'][0]-curParam['lum'])
intervals.update({'luminterval':[curParam['lum']-dev,curParam['lum']+dev]})
if np.argmax(pivots)==1:
#param=copy.copy(vphMG.grid[vphInterval['bestFitID']].param)
if vphInterval['steady']>selectSteadyThresh:
if damping:
curParam['vph']=np.mean([vphInterval['suggestValue'],lastSuggestVph])
intervals.update({'vphinterval':[curParam['vph']+i*dampingFactor*vphInterval['dev'] for i in [-1,1]]})
else:
curParam['vph']=vphInterval['suggestValue']
intervals.update({'vphinterval':vphInterval['interval']})
else:
print "Not steady solution will increase last interval and try again"
dev=selectSteadyIncFactor*np.abs(intervals['vphinterval'][0]-curParam['vph'])
intervals.update({'vphinterval':[curParam['vph']-dev,curParam['vph']+dev]})
print "Chose Vph: updated for next run with suggestValue %s and Interval %s"%(curParam['vph'],intervals['vphinterval'])
else:
if vphInterval['steady']<steadyThresh:
print "Updating the vph interval as the current solution is not steady"
dev=steadyIncFactor*np.abs(intervals['vphinterval'][0]-curParam['vph'])
intervals.update({'vphinterval':[curParam['vph']-dev,curParam['vph']+dev]})
if np.argmax(pivots)==2:
if IGEInterval['steady']>selectSteadyThresh:
if damping:
curParam[IGEElement]=np.mean([IGEInterval['suggestValue'],lastSuggestIGE])
intervals.update({'igeinterval':[curParam[IGEElement]+i*dampingFactor*IGEInterval['dev'] for i in [-1,1]]})
else:
curParam[IGEElement]=IGEInterval['suggestValue']
intervals.update({'igeinterval':IGEInterval['interval']})
else:
print "Not steady solution will increase last interval and try again"
dev=selectSteadyIncFactor*np.abs(intervals['igeinterval'][0]-curParam[IGEElement])
intervals.update({'igeinterval':[curParam[IGEElement]-dev,curParam[IGEElement]+dev]})
print "Chose IGE: updated for next run with suggestValue %s and Interval %s"%(curParam[IGEElement],intervals['igeinterval'])
else:
if IGEInterval['steady']<steadyThresh:
print "Updating the IGE interval as the current solution is not steady"
dev=steadyIncFactor*np.abs(intervals['igeinterval'][0]-curParam[IGEElement])
intervals.update({'igeinterval':[curParam[IGEElement]-dev,curParam[IGEElement]+dev]})
#Checking for crossing Boundaries:
boundCheck=zip(lumBounds,intervals['luminterval'])
lumCheck=[np.max(boundCheck[0]),np.min(boundCheck[1])]
boundCheck=zip(vphBounds,intervals['vphinterval'])
vphCheck=[np.max(boundCheck[0]),np.min(boundCheck[1])]
boundCheck=zip(IGEBounds,intervals['igeinterval'])
igeCheck=[np.max(boundCheck[0]),np.min(boundCheck[1])]
intervals.update({'luminterval':lumCheck,'vphinterval':vphCheck,'igeinterval':igeCheck})
#Readjusting suggest values to be in the middle of the intervals:
curParam['lum']=np.mean(intervals['luminterval'])
curParam['vph']=np.mean(intervals['vphinterval'])
curParam[IGEElement]=np.mean(intervals['igeinterval'])
#Printing the intervals:
pp.pprint(intervals)
#Adding to fitHist
fitHist.addHistItem([lumInterval,vphInterval,IGEInterval],[lumMG,vphMG,IGEMG,curParamMG],pivots,intervals,reallyAdd=False,saveSingle=True)
#pdb.set_trace()
#Break condition
if np.argmax(pivots)>0:
# pdb.set_trace()
return True
else:
return False
def initElementCycle(initParam,elements,samples,scaling):
bounds=initialize.getElementsBounds(initParam.comp)
divIntervals={'Si':[5e-3,5],
'S':[1e-3,1],
'Ca':[0.5e-3,0.5],
'Mg':[0.5e-3,2],
'C':[0.5e-3,0.5],
'Fe0':[1e-3,1]}
intervals=copy.deepcopy(bounds)
initParam.lockIGE=False
initParam.lockIGEwNi=False
initParam.lockIGEwoNi=True
ranges={}
#intervalHistory={}
for element in intervals.keys():
if scaling=='log':
ranges[element]=np.logspace(*np.log10(intervals[element]),**dict(num=samples))
elif scaling=='lin':
ranges[element]=np.linspace(*intervals[element],**dict(num=samples))
else:
raise Exception('Unknown scaling option %s. Please use lin or log.'%scaling)
finished=False
for element in ['Si','S','Mg','C']:
if scaling=='log':
elementRange=np.logspace(*np.log10(intervals[element]),**dict(num=samples))
elif scaling=='lin':
elementRange=np.linspace(*intervals[element],**dict(num=samples))
elementMG=launcherSteps.launchElementCycle({element:elementRange},initParam=initParam)
plot.plotElementModelGrid(element,elementMG[0])
newInterval,finished=dalekSteps.getNextElementParams(element,elementMG[0],initParam,bounds[element])
if not finished: intervals[element]=newInterval
print "New Interval for %s: %s "%(element,intervals[element])
initParam[element]=np.mean(intervals[element])
bounds=initialize.getElementsBounds(initParam.comp)
intervals=dalekSteps.checkElementBounds(intervals,bounds)
#pdb.set_trace()
#allMG=launcherSteps.launchElementCycle(ranges,initParam=initParam)
#elementMG=dict(zip(ranges.keys(),allMG[:-1]))%debug
#curParamMG=allMG[-1]
return intervals,initParam
def runElementCycle(initParam,elements=['C','Ca','Mg','Si','S','Ti','Ni0','Fe0'],samples=10,scaling='lin'):
#initParam.autoRelAbund='O'
intervals,initParam=initElementCycle(initParam,elements,samples,scaling)
#print "Dumping initial run in initElemCycle.pkl"
#pickle.dump(elementMG,file('initElemCycle.pkl','w'))
bounds=initialize.getElementsBounds(initParam.comp)
intervalHistory={'Si':[],'S':[],'Mg':[],'C':[]}
i=1
while True:
converged=[]
for element in ['Si','S','Mg','C']:
if scaling=='log':
elementRange=np.logspace(*np.log10(intervals[element]),**dict(num=samples))
elif scaling=='lin':
elementRange=np.linspace(*intervals[element],**dict(num=samples))
elementMG=launcherSteps.launchElementCycle({element:elementRange},initParam=initParam)
plot.plotElementModelGrid(element,elementMG[0],suffix=str(i))
intervals[element],finished=dalekSteps.getNextElementParams(element,elementMG[0],initParam,bounds[element])
converged.append(finished)
intervalHistory[element].append(intervals[element])
print "New Interval for %s: %s "%(element,intervals[element])
initParam[element]=np.mean(intervals[element])
bounds=initialize.getElementsBounds(initParam.comp)
intervals=dalekSteps.checkElementBounds(intervals,bounds)
print "Convergence status %s"%converged
if all(converged): break
print "Looking for stop.txt in %s"%os.getcwd()
if os.path.exists('stop.txt'): break
i+=1
return intervalHistory