forked from Alkistis/class_IDE
-
Notifications
You must be signed in to change notification settings - Fork 0
/
CPU
executable file
·427 lines (391 loc) · 15.7 KB
/
CPU
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
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
###################################################################
#
# CPU, a CLASS Plotting Utility
# v1.3
# Benjamin Audren, 07.11.2011
#
# This is a small python program aimed to gain time when comparing two
# spectra, i.e. from CAMB and CLASS, or a non-linear spectrum to a
# linear one. It is designed to be used in a command line fashion, not
# being restricted to your CLASS directory, though it recognized mainly
# CLASS output format. Far from perfect, or complete, it could use any
# suggestion for enhancing it, just to avoid losing time on useless
# matters for others. Be warned that, when comparing with other
# format, the following is assumed: there are no empty line (especially
# at the end of file). Gnuplot comment lines (starting with a # are
# allowed). This issue will cause a non-very descriptive error in CPU,
# any suggestion for testing it is welcome. Example of use: To
# superimpose two different spectra and see their global shape : python
# CPU output/lcdm_z2_pk.dat output/lncdm_z2_pk.dat To precisely compare
# the non-linear contribution of one-loop w.r.t. to TRG method: python
# CPU -i output/lcdm_1l_z1_pk.dat output/lcdm_1l_z1_pk_nl_density.dat
# output/lcdm_trg_pk_nl_density.dat The documentation available with
# --help should cover any question.
###################################################################
from scipy import *
from scipy import interpolate
import numpy as np
import os,sys,string,io,subprocess
import argparse
# parser for input arguments
parser = argparse.ArgumentParser(description='CPU, a CLASS Plotting Utility, specify wether you want to superimpose, divide or interpolate different files, and behold!',
epilog='''A standard usage would be, for instance :
python CPU output/test_1l_pk.dat output/test_1l_pk_nl_density.dat -i 0
python CPU output/wmap_cl.dat output/planck_cl.dat''',
formatter_class=argparse.RawDescriptionHelpFormatter)
parser.add_argument('files', metavar='Files',
type=str, nargs='*',
help='the relative path of the desired file to plot\nfrom the root directory of CLASS')
parser.add_argument('-d, --divide',
dest='merging',action='store_const',const='blend_against',default='blend_together',
help='divide the spectra from different files. The k-values must be strictly identical (default: plot every graph on the same plot)')
parser.add_argument('-i, --interpolate',metavar='index',
nargs='?',dest='interp',const=True,default=False,
help='interpolate the spectra on each set of k-values if different. Use to compare two (or more) graphs with different k-values. If nothing more is specified, it will plot both index in gnuplot. Otherwise, just add -i 0 to plot only the first one.')
parser.add_argument('-t', metavar='pk, cl_lin,etc...',
help='specify the file type (whether pk or cl (if not specified, cl==cl_lin, other choices are cl_log and cl_ll for log linear), only needed if your file name does not already contain one of these keywords...')
parser.add_argument('-colnum',metavar='2, 3, ...',
help='specify the column you want to plot. By default, column is 2: you are plotting TT spectrum')
parser.add_argument('-x, --x11',metavar='gnuplot terminal',
dest='term',action='store_const',const='x11',default='aqua',
help='if your version of gnuplot does not support the aqua terminal, it will pick instead the x11 one, for crappier displays')
parser.add_argument('-p, --print',
dest='printfile',action='store_true',default=False,
help='print the graph directly in a .png file')
parser.add_argument('-r, --repeat',
dest='repeat',action='store_true',default=False,
help='repeat the step for all redshifts with same base name, whatever the desired action')
parser.add_argument('-c, --cleaning',metavar='path',
nargs='?',dest='cleaning',const=True,default=False,
help='remove all .plt files in the current directory (if none specified) : keep your folders clean !')
# Remove all .plt, interp and divided files generated in output/
# By default, will clean current directory.
def clean(path):
pattern1,pattern2,pattern3='.plt','_interp.dat','_divided.dat'
path+="/"
print 'Cleaning {0} directory from .plt files ...'.format(path)
i=0
for each in os.listdir(path):
if ( (each.rfind(pattern1)!=-1) or (each.rfind(pattern2)!=-1) or (each.rfind(pattern3)!=-1) ):
name= "{0}{1}".format(path,each)
print ' deleting {0}'.format(name)
if os.remove(name)!= None:
break
i+=1
if i==0:
print ' ==> Already as clean as possible'
else:
print ' ==> Done'
# Errors ###########################
def error_format():
print " Hum... have you really provided a .dat file ?"
print " We apologise for the inconvenience, but CPU is exiting now"
exit()
def error_type():
print " Spectrum type unrecognized or unsupported yet, sorry!"
print " Please try the -t command in addition to your previous call"
print " We apologise for the inconvenience"
exit()
def error_number_of_files():
print " You specified a wrong number of files for the operation you demanded, maybe you do not want to divide only one file, for instance ?"
print " We apologise for the inconvenience"
exit()
####################################
# Subfunction called to write on the gnuplot script file the proper file names.
def printstring(names):
print_file=names[0].rstrip(".dat")
print_file+=".ps"
print_string="set terminal po enhanced color\nset output '{0}'\nreplot".format(print_file)
print '{0} has been generated'.format(print_file)
return print_string
# Subfunction called to write the most of the gnuplot script file according to options.
def headers_plot_file(spectrum_type,names,plot_line,term):
tmp=open(names[0],"r")
# printing option
if ((args.printfile is True) and (plot_line is not False)):
print_string=printstring(names)
else:
print_string=""
# for interpolation: index choice option
if plot_line is True:
if args.interp is not True:
if args.interp is not False:
index_string=" index {0} ".format(args.interp)
else:
index_string=" "
else:
index_string=" "
# if args colnum not specified, put it to 2,
if args.colnum is None:
args.colnum=2
if spectrum_type=='cl_lin':
if plot_line is True:
plot_string="plot "
for name in names:
plot_string+="'{0}'".format(name)+"{0}u 1:{1} w l,".format(index_string,args.colnum)
plot_string=plot_string.rstrip(",")
plot_string+="\n"
else:
plot_string=''
return "set terminal {0} enhanced\n".format(term)+"set xlabel 'l'\nset ylabel 'l(l+1)C_l / 2{/Symbol p}'\nset xr [2:]\nset format y '%.0t*10^{%T}'\nset key right\nset title 'CLASS output'\n"+plot_string+print_string
elif spectrum_type=='cl_log':
for line in tmp:
if line.rfind('multipoles')!=-1:
lmax= line.split(None)[-1]
break
if plot_line is True:
plot_string="plot "
for name in names:
plot_string+="'{0}'".format(name)+"{0}u 1:{1} w l,".format(index_string,args.colnum)
plot_string=plot_string.rstrip(",")
plot_string+="\n"
else:
plot_string=''
return "set terminal {0} enhanced\n".format(term)+"set logscale x\nset xr [2:{0}]\n".format(lmax)+"set xlabel 'l'\nset format y '%.0t*10^{%T}'\n"+"set ylabel 'l(l+1)C_l / 2{/Symbol p}'\nset key left\nset title 'CLASS output'\n"+plot_string+print_string
elif spectrum_type=='cl_ll':
for line in tmp:
if line.rfind('multipoles')!=-1:
lmax= line.split()[-1]
break
if plot_line is True:
plot_string="plot "
for name in names:
plot_string+="'{0}'".format(name)+"{0}u (sqrt($1)):{1} w l,".format(index_string,args.colnum)
plot_string=plot_string.rstrip(",")
plot_string=plot_string+"\n"
else:
plot_string=''
return "set terminal {0} enhanced\n".format(term)+"set xlabel 'l'\nset ylabel 'l(l+1)C_l / 2{/Symbol p}'\nset key right\nset title 'CLASS output'\nset xtics ('2' (2),'100' (100), '500' (500), '1000' (1000), '1500' (1500), '2000' (2000),'2500' (2500))"+"\nset format y '%.0t*10^{%T}'\n"+"set xr [2:{0}]\n".format(lmax)+plot_string+print_string
elif spectrum_type=='pk':
if plot_line is True:
plot_string="plot "
for name in names:
plot_string+="'{0}'".format(name)+"{0}w l,".format(index_string)
plot_string=plot_string.rstrip(",")
plot_string+="\n"
else:
plot_string=""
for line in tmp:
if line.rfind('redshift')!=-1:
z= line.split("=")[1]
z= z.rstrip("\n")
break
else:
z= 'I have no idea'
return "set terminal {0} enhanced\n".format(term)+"set logscale\nset xlabel 'k (h/Mpc)'\nset ylabel 'P_k (Mpc/h)^3'\nset key right\nset title 'Power spectrum at z={0}'\n".format(z)+plot_string+print_string
else:
return None
tmp.close()
# Print all asked files one next to the other (default operation if files with different k values)
def blend_together(names,spectrum_type,term):
gnuplot_file=names[0].replace(".dat",".plt")
if gnuplot_file==names[0]:
error_format()
print ' creating {0}'.format(gnuplot_file)
plotfile = open(gnuplot_file, "w")
plotfile.write(headers_plot_file(spectrum_type, names,True,term))
plotfile.close()
_plot(gnuplot_file)
# Print all asked files with respect to the same k (or anything) values, all y-data being divided by the y data of the first file.
# Only valid if the k values are exactly the same (values and number of values).
def blend_against(names,spectrum_type,term):
gnuplot_file=names[0].replace(".dat",".plt")
data_file=names[0].replace(".dat","_divided.dat")
if gnuplot_file==names[0]:
error_format()
print ' creating {0} and {1}'.format(gnuplot_file,data_file)
string=['' for rows in range(10000)]
imax=0
datafile=open(data_file,"w")
for name in names:
i=0
tmp=open(name,"r")
for line in tmp:
if ((line.find('#')==-1) and (line.rfind('000000')==-1)):
string[i]=string[i]+line.rstrip("\n")+"\t"
if i>imax:
imax=i
i+=1
tmp.close()
for i in range(imax):
datafile.write(string[i]+"\n")
plotfile = open(gnuplot_file,"w")
plotfile.write(headers_plot_file(spectrum_type,names,False,term))
plot_string='unset logscale\nplot '
x_base=1
field_base=2
field=2
size=len(names)
for i in range(size):
field+=2
plot_string+="'{0}' u {1}:(${2}/${3}) w l,".format(data_file,x_base,field,field_base)
plot_string=plot_string.rstrip(',')
plot_string+="\n"
plotfile.write(plot_string)
plotfile.close()
datafile.close()
_plot(gnuplot_file)
# If k values are different, an interpolation is done and outputs a data file. For a two files case:
# File 1 contains ( k1 | P1(k1) ), File 2 ( k2 | P2(k2) ). The data file created will contain:
# k1 | P1(k1) | P2(k1)_interp
#(blank space for gnuplot using index 0, etc)
# k2 | P1(k2)_interp | P2(k2)
def blend_against_interp(names,spectrum_type,term):
gnuplot_file=names[0].replace(".dat",".plt")
data_file=names[0].replace(".dat","_interp.dat")
if gnuplot_file==names[0]:
error_format()
print ' creating {0} and {1}'.format(gnuplot_file,data_file)
plotfile = open(gnuplot_file,"w")
l=0
jmax=[0 for col in range(10)]
lmax=0
kmax=10000000
kmin=0
spam = np.array([[[0 for col in range(10)] for row in range(10000)] for depth in range(2)],dtype=float)
for name in names:
currentfile = open(name,"r")
print ' reading {0}..'.format(name)
j=0
for line in currentfile:
if (line.find('#')==-1):
line=line.split()
spam[0,j,l]=float(line[0])
spam[1,j,l]=float(line[1])
j+=1
jmax[l]=j
if spam[0,jmax[l]-1,l]<=kmax:
kmax=spam[0,jmax[l]-1,l]
if spam[0,0,l]>=kmin:
kmin=spam[0,0,l]
l+=1
currentfile.close()
lmax=l
lower_bound=[0 for col in range(10)]
upper_bound=[0 for col in range(10)]
#determining the upper and lower bound for each file, to only do interpolation
for l in range (lmax):
for j in range (jmax[l]):
if ((spam[0,j,l]<kmin) and (spam[0,j+1,l]>=kmin)):
lower_bound[l]=j+1
if (spam[0,j,l]<=kmax):
upper_bound[l]=j
print ' -> done'
#creating the new data file
curves=[np.array for row in range(lmax)]
interpolated=[np.array for row in range(lmax)]
for l in range (lmax):
x=spam[0,lower_bound[l]:upper_bound[l],l]
y=spam[1,lower_bound[l]:upper_bound[l],l]
curves[l]=interpolate.splrep(x,y)
datafile = open(data_file,"w")
for l in range (lmax):
for ll in range(lmax):
if ll==l:
interpolated[ll]=spam[1,lower_bound[l]:upper_bound[l],l]
else:
x2=spam[0,lower_bound[l]:upper_bound[l],l]
interpolated[ll]=interpolate.splev(x2,curves[ll],der=0)
for j in range(upper_bound[l]-lower_bound[l]-2):
data_string=''
for ll in range(lmax):
data_string=data_string+str((interpolated[ll])[j+1])+"\t"
datafile.write(str(spam[0,j+lower_bound[l]+1,l])+"\t"+data_string+"\n")
datafile.write("\n\n")
datafile.close()
# if index chosen
if args.interp is not True:
index_string_interp=" index {0}".format(args.interp)
else:
index_string_interp=""
plotfile = open(gnuplot_file,"w")
plotfile.write(headers_plot_file(spectrum_type,names,False,term))
plotfile.write("unset logscale \n")
if args.t in ['cl_log', 'pk']:
plotfile.write("set logscale x\n")
plotfile.write("set xr [{0}:{1}]\n".format(kmin,kmax))
#plotfile.write("set yr [-0.001:0.005]\n")
plot_string="plot "
for l in range(lmax-1):
plot_string+="'{0}' {1} u 1:(${2}/$2-1) w l,".format(data_file,index_string_interp,l+3)
plot_string=plot_string.rstrip(",")
plot_string=plot_string+"\n"
plotfile.write(plot_string)
if (args.printfile is True):
plotfile.write(printstring(names))
plotfile.close()
_plot(gnuplot_file)
# Launch session of gnuplot with generated gnuplot script file
def _plot(gnuplot_file):
os.system("gnuplot -persist '{0}'".format(gnuplot_file))
#######################################################
################## MAIN PART ##########################
#######################################################
print '~~~ CPU, a CLASS Plotting Utility ~~~'
args = parser.parse_args()
# check if the user want to clean its directory first
if args.cleaning is not False:
if args.cleaning is not True:
clean(args.cleaning)
else:
clean(os.getcwd())
exit()
# if there are no argument in the input, print usage
if len(args.files)==0:
parser.print_usage()
exit()
# if the first file name contains cl or pk, infer the type of desired spectrum
if ((args.files[0].rfind('cl')!=-1) and (args.t is None)):
spectrum_type='cl_lin'
args.t = 'cl_lin'
elif args.files[0].rfind('pk')!=-1:
spectrum_type='pk'
args.t = 'pk'
elif args.t is not None:
spectrum_type=args.t
else:
error_type()
# repeater
temp_path=args.files[0].split("/")
if len(temp_path)> 1:
path=temp_path[-2]
else:
path=os.getcwd()
list_file=['' for i in range(len(args.files)*10)]
if (temp_path[-1].rfind("z")!=-1 and args.repeat is True):
root_name=temp_path[-1].split("z")[-2]
for any in range (0,len(args.files)):
extension_name=args.files[any].split("/")[-1].split("_")[-1]
i=0
for each in os.listdir(path):
if (("z" in each) and (each.split("z")[-2]==root_name) and (each.split("_")[-1]==extension_name)):
if len(temp_path)>1:
list_file[any+len(args.files)*i]=temp_path[-2]+"/"+each
else:
list_file[any+len(args.files)*i]=each
i+=1
if args.repeat is True:
repeat_len=i-1
else:
repeat_len=0
else:
for any in range(0,len(args.files)):
list_file[any]=args.files[any]
repeat_len=0
# actual computation
for i in range(0,repeat_len+1):
local_files=['' for k in range(len(args.files))]
for j in range(0,len(args.files)):
local_files[j]=list_file[j+len(args.files)*i]
if args.interp is False:
if args.merging=='blend_together':
blend_together(local_files,spectrum_type,args.term)
else:
if len(local_files)<2:
error_number_of_files()
else:
blend_against(local_files,spectrum_type,args.term)
else:
print '**interpolating (please wait)'
blend_against_interp(local_files,spectrum_type,args.term)
exit()