Warning, /eic-opticks/ana/tools.rst is written in an unsupported language. File is not indexed.
0001 Tools : bash, python, numpy, ipython, matplotlib
0002 ======================================================
0003
0004 Opticks uses bash functions extensively and
0005 analysis and debugging relies on:
0006
0007 * Python
0008 * IPython
0009 * NumPy http://www.numpy.org
0010 * MatPlotLib
0011
0012
0013 In general the best way to install common software such as
0014 these tools is with your systems package manager. For example:
0015
0016 * Linux: yum, apt-get
0017 * macOS: https://www.macports.org or homebrew
0018 * Windows: https://chocolatey.org
0019
0020 However there are other ways to install packages such as
0021 using **pip** for python packages.
0022
0023
0024 IPython Install on a mac with pip
0025 ----------------------------------
0026
0027 Installing ipython on a mac if you do not use macports.
0028
0029 ::
0030
0031 delta:tests blyth$ sudo pip install ipython==1.2.1
0032 Password:
0033 Downloading/unpacking ipython==1.2.1
0034 Downloading ipython-1.2.1.tar.gz (8.7MB): 8.7MB downloaded
0035 Running setup.py (path:/private/tmp/pip_build_root/ipython/setup.py) egg_info for package ipython
0036 running egg_info
0037
0038 Installing collected packages: ipython
0039 Running setup.py install for ipython
0040 running install
0041
0042 Installing iplogger script to /opt/local/Library/Frameworks/Python.framework/Versions/2.7/bin
0043 Installing iptest script to /opt/local/Library/Frameworks/Python.framework/Versions/2.7/bin
0044 Installing ipcluster script to /opt/local/Library/Frameworks/Python.framework/Versions/2.7/bin
0045 Installing ipython script to /opt/local/Library/Frameworks/Python.framework/Versions/2.7/bin
0046 Installing pycolor script to /opt/local/Library/Frameworks/Python.framework/Versions/2.7/bin
0047 Installing ipcontroller script to /opt/local/Library/Frameworks/Python.framework/Versions/2.7/bin
0048 Installing irunner script to /opt/local/Library/Frameworks/Python.framework/Versions/2.7/bin
0049 Installing ipengine script to /opt/local/Library/Frameworks/Python.framework/Versions/2.7/bin
0050 Successfully installed ipython
0051 Cleaning up...
0052 delta:tests blyth$
0053 delta:tests blyth$ which ipython
0054 /opt/local/bin/ipython
0055 delta:tests blyth$ ipython
0056 Python 2.7.11 (default, Dec 5 2015, 23:51:51)
0057 Type "copyright", "credits" or "license" for more information.
0058
0059 IPython 1.2.1 -- An enhanced Interactive Python.
0060 ? -> Introduction and overview of IPython's features.
0061 %quickref -> Quick reference.
0062 help -> Python's own help system.
0063 object? -> Details about 'object', use 'object??' for extra details.
0064
0065 In [1]: import numpy as np
0066
0067 In [2]: np.__version__
0068 Out[2]: '1.9.2'
0069
0070 In [3]: import matplotlib
0071
0072 In [4]: import matplotlib.pyplot as plt
0073
0074 In [5]: matplotlib.__version__
0075 Out[5]: '1.3.1'
0076
0077
0078
0079
0080
0081 Working with bash functions
0082 -----------------------------
0083
0084 Bash functions make you into a power user of the shell, by
0085 allowing you to operate at a higher level than individual commands
0086 without having to write separate scripts for everything.
0087
0088 .. code-block:: sh
0089
0090 delta:tests blyth$ alias t="typeset -f" # enables introspection of bash functions
0091 delta:tests blyth$ t opticks-
0092 opticks- ()
0093 {
0094 source $(opticks-source) && opticks-env $*
0095 }
0096 delta:tests blyth$ t tprism-
0097 tprism- ()
0098 {
0099 . $(opticks-home)/tests/tprism.bash && tprism-env $*
0100 }
0101 delta:tests blyth$ t tprism-- # not defined yet
0102 delta:tests blyth$ tprism- # running precursor defines functions including tprism-- tprism-usage and tprism-vi
0103 delta:tests blyth$ t tprism--
0104 tprism-- ()
0105 {
0106 local msg="=== $FUNCNAME :";
0107 local cmdline=$*;
0108 local pol=${1:-s};
0109 shift;
0110 local tag=$(tprism-tag $pol);
0111 echo $msg pol $pol tag $tag;
0112 local phifrac0=0.1667;
0113 local phifrac1=0.4167;
0114 local phifrac2=0.6667;
0115 local quadrant=$phifrac1,$phifrac2;
0116 local critical=0.4854,0.4855;
0117 local material=GlassSchottF2;
0118 local azimuth=$quadrant;
0119 local surfaceNormal=0,1,0;
0120 local torch_config=(type=invcylinder photons=500000 mode=${pol}pol,wavelengthComb polarization=$surfaceNormal frame=-1 transform=0.500,0.866,0.000,0.000,-0.866,0.500,0.000,0.000,0.000,0.000,1.000,0.000,-86.603,0.000,0.000,1.000 target=0,-500,0 source=0,0,0 radius=300 distance=25 zenithazimuth=0,1,$azimuth material=Vacuum wavelength=0);
0121 local test_config=(mode=BoxInBox analytic=1 shape=box parameters=-1,1,0,700 boundary=Rock//perfectAbsorbSurface/Vacuum shape=prism parameters=60,300,300,200 boundary=Vacuum///$material);
0122 op.sh $* --animtimemax 7 --timemax 7 --geocenter --eye 0,0,1 --test --testconfig "$(join _ ${test_config[@]})" --torch --torchconfig "$(join _ ${torch_config[@]})" --torchdbg --save --tag $tag --cat $(tprism-det)
0123 }
0124
0125 delta:tests blyth$ tprism- # pressing [tab] shows available functions beginning tprism-
0126 tprism- tprism-args tprism-det tprism-env tprism-py tprism-src tprism-test tprism-vi
0127 tprism-- tprism-cd tprism-dir tprism-pol tprism-source tprism-tag tprism-usage
0128 delta:tests blyth$ tprism-
0129
0130
0131
0132 A powerful way to rapidly develop bash functions is to use
0133 two terminal sessions, in the first edit the functions:
0134
0135 .. code-block:: sh
0136
0137 tprism-vi
0138
0139 In the second test the **tprism-foo** you are developing:
0140
0141 .. code-block:: sh
0142
0143 tprism-;tprism-foo # tprism- updates function definitions and the tprism-foo runs it
0144
0145
0146
0147 IPython : interactive python example
0148 ---------------------------------------
0149
0150 IPython enables you to inspect live python objects, so you can
0151 learn by discovery. In order to see the plots created by Opticks analysis
0152 scripts you will need to use IPython.
0153
0154 .. code-block:: py
0155
0156 delta:ana blyth$ ipython
0157 Python 2.7.11 (default, Dec 5 2015, 23:51:51)
0158 Type "copyright", "credits" or "license" for more information.
0159
0160 IPython 1.2.1 -- An enhanced Interactive Python.
0161 ? -> Introduction and overview of IPython's features.
0162 %quickref -> Quick reference.
0163 help -> Python's own help system.
0164 object? -> Details about 'object', use 'object??' for extra details.
0165
0166 IPython profile: g4opticks
0167
0168 In [1]: run tprism.py --tag 1
0169 tprism.py --tag 1
0170 INFO:__main__:sel prism/torch/ 1 : TO BT BT SA 20160716-1941 /tmp/blyth/opticks/evt/prism/fdomtorch/1.npy
0171 INFO:__main__:prism Prism(array([ 60., 300., 300., 0.]),Boundary Vacuum///GlassSchottF2 ) alpha 60.0
0172 ...
0173
0174
0175 IPython Tab Completion
0176 ~~~~~~~~~~~~~~~~~~~~~~~~
0177
0178 Discover available methods of an object by interactive exploration:
0179
0180 .. code-block:: py
0181
0182 In [2]: evt. # press [tab] to see the possibilities
0183 evt.RPOL evt.desc evt.histype evt.material_table evt.ph evt.rflgs_ evt.seqhis evt.td
0184 evt.RPOST evt.description evt.idom evt.mattype evt.polw evt.rpol_ evt.seqhis_or_not evt.tdii
0185 evt.RQWN_BINSCALE evt.det evt.incident_angle evt.msize evt.post evt.rpol_bins evt.seqmat evt.unique_wavelength
0186 evt.a_deviation_angle evt.deviation_angle evt.init_index evt.nrec evt.post_center_extent evt.rpost_ evt.seqs evt.valid
0187 evt.a_recside evt.dirw evt.init_metadata evt.ox evt.ps evt.rs evt.src evt.wl
0188 evt.a_side evt.fdom evt.init_photons evt.p0 evt.py evt.rsmry_ evt.stamp evt.x
0189 evt.all_history evt.flags evt.init_records evt.p_out evt.pyc evt.rsr evt.summary evt.y
0190 evt.all_material evt.flags_table evt.init_selection evt.path evt.rec evt.rx evt.t evt.z
0191 evt.brief evt.history evt.label evt.paths evt.recflags evt.rx_raw evt.tag evt.zrt_profile
0192 evt.c4 evt.history_table evt.material evt.pbins evt.recwavelength evt.selection evt.tbins
0193
0194
0195
0196 One question mark gives the documentation:
0197
0198 .. code-block:: py
0199
0200 In [2]: evt.deviation_angle?
0201 Type: instancemethod
0202 File: /Users/blyth/opticks/ana/evt.py
0203 Definition: evt.deviation_angle(self, side=None, incident=None)
0204 Docstring:
0205 Deviation angle for parallel squadrons of incident photons
0206 without assuming a bounce count
0207
0208 Two question marks gives the implementation:
0209
0210 .. code-block:: py
0211
0212 In [4]: evt.zrt_profile??
0213
0214 Type: instancemethod
0215 File: /Users/blyth/opticks/ana/evt.py
0216 Definition: evt.zrt_profile(self, n, pol=True)
0217 Source:
0218 def zrt_profile(self, n, pol=True):
0219 """
0220 :param n: number of bounce steps
0221 :return: min, max, mid triplets for z, r and t at n bounce steps
0222
0223 ::
0224
0225 In [7]: a_zrt
0226 Out[7]:
0227 array([[ 300. , 300. , 300. , 1.1748, 97.0913, 49.133 , 0.1001, 0.1001, 0.1001],
0228 [ 74.2698, 130.9977, 102.6337, 1.1748, 97.0913, 49.133 , 0.9357, 1.2165, 1.0761],
0229 [ 56.0045, 127.9946, 91.9996, 1.1748, 98.1444, 49.6596, 0.9503, 1.3053, 1.1278]])
0230
0231
0232 """
0233 slab = "z r t"
0234 if pol:
0235 slab += " lx ly lz"
0236
0237 labs = slab.split()
0238 nqwn = 3
0239 zrt = np.zeros((n,len(labs)*nqwn))
0240 tfmt = "%10.3f " * nqwn
0241 fmt = " ".join(["%s: %s " % (lab, tfmt) for lab in labs])
0242
0243 for i in range(n):
0244 p = self.rpost_(i)
0245 l = self.rpol_(i)
0246 lx = l[:,0]
0247 ly = l[:,1]
0248 lz = l[:,2]
0249
0250
0251
0252
0253
0254
0255
0256 NumPy
0257 -------
0258
0259 Opticks although mostly implemented in C++ uses the NPY serialization format
0260 for all geometry and event buffers allowing debugging and analysis
0261 to be done from the IPython using NumPy.
0262
0263 NPY serialized files used extension **.npy**, find some:
0264
0265 .. code-block:: sh
0266
0267 delta:~ blyth$ find /tmp/blyth/opticks/evt/lens -name 1.npy
0268 /tmp/blyth/opticks/evt/lens/fdomtorch/1.npy
0269 /tmp/blyth/opticks/evt/lens/idomtorch/1.npy
0270 /tmp/blyth/opticks/evt/lens/notorch/1.npy
0271 /tmp/blyth/opticks/evt/lens/oxtorch/1.npy
0272 /tmp/blyth/opticks/evt/lens/phtorch/1.npy
0273 /tmp/blyth/opticks/evt/lens/pstorch/1.npy
0274 /tmp/blyth/opticks/evt/lens/rstorch/1.npy
0275 /tmp/blyth/opticks/evt/lens/rxtorch/1.npy
0276
0277
0278 Determine maximum photon time **in one line**:
0279
0280 .. code-block:: sh
0281
0282 delta:~ blyth$ python -c "import numpy as np ; print np.load('/tmp/blyth/opticks/evt/lens/oxtorch/1.npy')[:,0,3].max() "
0283 11.2614
0284
0285 Do that more sedately with ipython:
0286
0287 .. code-block:: py
0288
0289 delta:~ blyth$ ipython
0290 Python 2.7.11 (default, Dec 5 2015, 23:51:51)
0291 Type "copyright", "credits" or "license" for more information.
0292
0293 IPython 1.2.1 -- An enhanced Interactive Python.
0294 ? -> Introduction and overview of IPython's features.
0295 %quickref -> Quick reference.
0296 help -> Python's own help system.
0297 object? -> Details about 'object', use 'object??' for extra details.
0298
0299 IPython profile: g4opticks
0300
0301 In [1]: ox = np.load("/tmp/blyth/opticks/evt/lens/oxtorch/1.npy")
0302
0303 In [2]: ox
0304 Out[2]:
0305 array([[[ 1.884, 0.202, 700. , 4.569],
0306 [ 0.098, 0.011, 0.995, 1. ],
0307 [ -0.107, 0.994, 0. , 630.652],
0308 [ 0. , 0. , 0. , 0. ]],
0309
0310 [[ 0.745, -1.288, 700. , 4.568],
0311 [ 0.05 , -0.087, 0.995, 1. ],
0312 [ 0.866, 0.501, 0. , 753.122],
0313 [ 0. , 0. , 0. , 0. ]],
0314
0315 [[ 290.308, -500.146, -700. , 6.057],
0316 [ 0.21 , -0.433, -0.877, 1. ],
0317 [ 0.501, 0.818, -0.283, 425.268],
0318 [ 0. , 0. , 0. , 0. ]],
0319
0320 ...,
0321 [[ 2.217, 0.106, 700. , 4.572],
0322 [ 0.078, 0.004, 0.997, 1. ],
0323 [ -0.048, 0.999, 0. , 472.495],
0324 [ 0. , 0. , 0. , 0. ]],
0325
0326 [[ -1.872, -0.614, 700. , 4.574],
0327 [ -0.055, -0.018, 0.998, 1. ],
0328 [ 0.312, -0.95 , 0. , 405.361],
0329 [ 0. , 0. , 0. , 0. ]],
0330
0331 [[ 2.042, -3.525, 700. , 4.571],
0332 [ 0.067, -0.116, 0.991, 1. ],
0333 [ 0.865, 0.501, 0. , 589.317],
0334 [ 0. , 0. , 0. , 0. ]]], dtype=float32)
0335
0336 In [3]: ox.shape
0337 Out[3]: (500000, 4, 4) # 0.5M photons with (4,4) floats each
0338
0339 In [4]: ox[0] # 1st photon
0340 Out[4]:
0341 array([[ 1.884, 0.202, 700. , 4.569],
0342 [ 0.098, 0.011, 0.995, 1. ],
0343 [ -0.107, 0.994, 0. , 630.652],
0344 [ 0. , 0. , 0. , 0. ]], dtype=float32)
0345
0346 In [5]: ox[-1] # last photon
0347 Out[5]:
0348 array([[ 2.042, -3.525, 700. , 4.571],
0349 [ 0.067, -0.116, 0.991, 1. ],
0350 [ 0.865, 0.501, 0. , 589.317],
0351 [ 0. , 0. , 0. , 0. ]], dtype=float32)
0352
0353 In [6]: ox[:,0,3] # "wildcard" first dimension with ":" and pick the other two, giving end times of the photons
0354 Out[6]: array([ 4.569, 4.568, 6.057, ..., 4.572, 4.574, 4.571], dtype=float32)
0355
0356 In [7]: ox[:,0,3].min() # earliest time
0357 Out[7]: 0.15406403
0358
0359 In [8]: ox[:,0,3].max() # latest time
0360 Out[8]: 11.261424
0361
0362
0363 See :doc:`../ana/evt` for a higher level way of loading event buffers.
0364
0365
0366
0367
0368
