- cctbxbb - phenix-online.org

Flex array slice assignment work-alike
by James Stroud 15 Apr '13

15 Apr '13

Hello All, It seems that flex arrays do not support slice assignment: py> from scitbx.array_family import flex py> ary = flex.float(range(5)) py> list(ary) [0.0, 1.0, 2.0, 3.0, 4.0] py> ary[:3] = flex.float([41, 42, 43]) --------------------------------------------------------------------------- ArgumentError Traceback (most recent call last) /usr/local/cctbx/cctbx_sources/mmtbx/dynamics/<ipython-input-13-5ce3d648bc6f> in <module>() ----> 1 ary[:3] = flex.float([41, 42, 43]) ArgumentError: Python argument types in float.__setitem__(float, slice, float) did not match C++ signature: __setitem__(scitbx::af::versa<float, scitbx::af::flex_grid<scitbx::af::small<long, 10ul> > > {lvalue}, scitbx::af::small<long, 10ul>, float) __setitem__(scitbx::af::versa<float, scitbx::af::flex_grid<scitbx::af::small<long, 10ul> > > {lvalue}, long, float) Slice assignment is common in python: py> alist = range(5) py> alist[:3] = [41, 42, 43] py> alist [41, 42, 43, 3, 4] Is there a flex array work-alike for this operation? I've found set_selected(): py> sel_ary = flex.bool([True, True, True, False, False]) py> ary.set_selected(sel_ary, flex.float([41, 42, 43])) <scitbx_array_family_flex_ext.float at 0x101df1158> py> list(ary) [41.0, 42.0, 43.0, 3.0, 4.0] But set_selected() is cumbersome for slice assignment because (1) you have to build a flex selection array somehow and (2) you have to put the new values into a flex array of the appropriate type. Is set_selected() the best way? If so, is there something in the cctbx libraries that makes building "slicing" selection arrays more efficient than say a list comprehension? sel_ary = flex.bool([True if i < 3 else False for i in xrange(5)]) James

7 16

cif parser & ccp4 mon_lib incompatibility
by James Stroud 15 Apr '13

15 Apr '13

Hi All, It's a fresh install of the latest-greatest cctbx and of ccp4. Any ideas? % cctbx.python ./tst_sa.py CifParserError: /Applications/ccp4-6.3.0/lib/data/monomers/list/mon_lib_list.cif(line 1) : error 4 : Unexpected token, at offset -1 near global : unexpected input... expected one of : Actually dude, we didn't seem to be expecting anything here, or at least I could not work out what I was expecting, like so many of us these days! Here's the top of the cif file: global _lib_name mon_lib _lib_version 5.36 _lib_update 03/05/12 --------------------------------- # # --- LIST OF MONOMERS --- # data_comp_list loop_ _chem_comp.id _chem_comp.three_letter_code _chem_comp.name _chem_comp.group James

7 24

Re: [cctbxbb] [ccp4bb] CCP4 Update victim of own success
by James Stroud 11 Apr '13

11 Apr '13

On Apr 11, 2013, at 6:17 AM, <eugene.krissinel(a)stfc.ac.uk> <eugene.krissinel(a)stfc.ac.uk> wrote: > Sorry that this was unclear. We assume that updater is used primarily from ccp4i, where nothing changed (and why it should be used from command line at all ?:)). Scriptability. On Apr 11, 2013, at 10:34 AM, Jim Pflugrath wrote: > I think James gets to 'fight' like in the old game of rogue by pressing the h, j, k, l keys on his keyboard (not a detachable one either). While Eugene gets to use a modern game controller or a Wii. > > Ooops, game is already over and James [Holton] has lost. But he wrote the game. James

1 1

bin_centers() and d_star_sq
by James Stroud 09 Apr '13

09 Apr '13

Hello All, I'm genuinely perplexed about what factors should go into choosing an appropriate d_star_sq to calculate bin_centers() in cctbx.miller.binner. This decision may be a nuance because it doesn't seem change the output significantly (except for low resolution bins). But if it doesn't matter significantly, then why have the parameter in the first place? James py> binner = miller.setup_binner(n_bins=10) py> binner.show_summary() unused: - 81.7405 [ 0/0 ] bin 1: 81.7405 - 6.1396 [2341/2341] bin 2: 6.1396 - 4.8734 [2213/2213] bin 3: 4.8734 - 4.2574 [2185/2185] bin 4: 4.2574 - 3.8681 [2168/2168] bin 5: 3.8681 - 3.5909 [2177/2177] bin 6: 3.5909 - 3.3792 [2156/2156] bin 7: 3.3792 - 3.2099 [2137/2137] bin 8: 3.2099 - 3.0702 [2144/2144] bin 9: 3.0702 - 2.9520 [2133/2133] bin 10: 2.9520 - 2.8501 [2135/2135] unused: 2.8501 - [ 0/0 ] py> centers = binner.bin_centers(1) py> list(1 / centers) [11.421375552155686, 5.4336997182888345, 4.544594791289144, 4.053431451445772, 3.7243527058119903, 3.481806175995981, 3.292371389128289, 3.138504394416836, 3.0099330507138475, 2.9001646196929975] py> centers = binner.bin_centers(4) py> list(1 / centers**(1 / 4)) [7.301226942489155, 5.330776416864701, 4.514065417212759, 4.039592121777967, 3.716672320747147, 3.4770032890247498, 3.28912114749567, 3.13617782414047, 3.0081962210988116, 2.8988250671976883]

1 1

match_indices.paired_miller_indices() ungraceful overshoot
by James Stroud 08 Apr '13

08 Apr '13

Hello all, I accidentally gave paired_miller_indices() a value > 1 for i_array, and got a segfault. The problem looks to be in match_indices.cpp, line 94: result.push_back(miller_indices_[i_array][pairs_[i][i_array]]); Maybe there should be a test that raises an exception on the python side, if possible. James

3 3

flex and numpy caveats
by James Stroud 06 Apr '13

06 Apr '13

Hello All, I've noticed some incompatibilities between numpy and cctbx, which is understandable. However, the incompatibilities manifest either as nonsense errors or, worse, not at all. Following are a couple of examples I have found. 1. The first example should really throw an exception, because silent failures like this can be catastrophic: py> from numpy import random py> r = random.randint(2, size=10) py> r array([1, 0, 1, 1, 0, 1, 0, 1, 1, 1]) py> list(flex.bool(r)) [True, False, False, False, False, False, False, False, False, False] This example is clearly due to incorrect assumptions about the internal representations of numpy ndarrays. Much better behavior can be seen when using a sequence of python `int` objects: py> list(flex.bool(range(10))) --------------------------------------------------------------------------- TypeError Traceback (most recent call last) /Users/jstroud/Unison/Code/radialx/testdata/<ipython-input-754-dce96b02979e> in <module>() ----> 1 list(flex.bool(range(10))) TypeError: No registered converter was able to produce a C++ rvalue of type bool from this Python object of type int 2. Although not as potentially catastrophic as the first, the second example should (in a perfect world) either work or throw a more meaningful exception: py> flex.int(range(10))[r[0]] --------------------------------------------------------------------------- TypeError Traceback (most recent call last) /Users/jstroud/Unison/Code/radialx/testdata/<ipython-input-753-f14587f0ceeb> in <module>() ----> 1 flex.int(range(10))[r[0]] TypeError: 'numpy.int64' object is not iterable This example is obviously a result of type checking for a python `int`. The usual python approach of duck-type checking would avoid this problem: def __getitem__(self, i): try: i = int(i) except: pass return do_whatever_with_i(i) James

2 9

Wilson Plot
by James Stroud 01 Apr '13

01 Apr '13

Hello All, The docsting for cctbx.miller.array.wilson_plot() does not seem to agree with its function (see below). Am I misinterpreting something? James py> miller.wilson_plot.__doc__ '<data^2>' py> miller.wilson_plot() (0.6109909654097065+4.386371170677923j) py> flex.mean(miller.data()) (0.6109909654097065+4.386371170677923j) py> flex.mean(miller.data() * flex.conj(miller.data())) (366421.87112270086+0j) py> flex.mean(miller.data() * miller.data()) (70775.39818541733-1856.4852414296488j)

2 1

Miller array by index
by James Stroud 28 Mar '13

28 Mar '13

Hello All, I ask this question mostly out of curiosity, but what is the most direct way to access the value for a Miller index? The best I can do is the method below, but it seems rather indirect. James py> miller <cctbx.miller.array at 0x107bfa290> py> idx = (0, 0, 2) py> miller.select_indices([idx]).data()[0] (-26338.974329482127+0j)

2 1

Re: [cctbxbb] [ccp4bb] refinement protein structure
by James Stroud 27 Mar '13

27 Mar '13

On Mar 27, 2013, at 12:32 PM, Anastassis Perrakis wrote: > These said, I am rather easily irritated these days, so I will not comment on the irritable character of the email. ... > we will soon start calling disordered loops > positionally challenged polypeptide segments (*). You might want to embrace the power of positive thinking. Where you see "positionally challenged", an optimist would see "entropically gifted". James

1 0

Rotate & translate a structure
by James Stroud 26 Mar '13

26 Mar '13

Hello all, I have a rotation + translation matrix. Let's say py> mtx array([[ 0.88302225, 0.11697778, 0.45451948, 5. ], [ 0.11697778, 0.88302225, -0.45451948, 10. ], [ -0.45451948, 0.45451948, 0.76604444, 20. ], [ 0. , 0. , 0. , 1. ]]) My usual way of transforming a set of coordinates with a 4x4 matrix is to use numpy: py> xyz_1 array([[ 5., 10., 15., 1.], [ 20., 15., 10., 1.]]) py> numpy.inner(xyz_1, mtx) array([[ 17.40268126, 12.59731919, 33.76326397, 1. ], [ 28.9603065 , 21.03969455, 25.38784698, 1. ]]) What is the equivalent in cctbx? Is there a way to avoid converting coordinates back and fourth between numpy arrays with atoms().extract_xyz() & .set_xyz() (as might be used from a iotbx_pdb_hierarchy_ext.root). Thank you for any help. James

3 2