Mailman 3 April 2010 - SciPy-User

How to find the intersections between two lines
by seasoul 10 Apr '10

10 Apr '10

This is my first post in the mailing list. I want to find the intersections between two lines. One line is an arbitrary straight line, the other is a fitting line created by a interpolation procedure. Below is a simple code: x = linspace(0,10,20) y = rand(20)*10 plt.plot(x,y,'o') f = interpolate.interp1d(x,y,kind=5) xx = linspace(0,10,100) yy = f(xx) plt.plot(xx,yy,'r-') y2 = x plt.plot(x,y2,'b-') def func(x): out = [x[1]-x[0]] out.append(x[1]-f(x[0])) return out fsolve(func,[1,1]) The fitting line is a polynomial curve, the straight line is y=x. This function can only find one of the intersection points,not all. I tried to use x[:,1] and x[:,2], and hope in the function func, the operation is element-by-element, but some errors came out. Any help is appreciated. The output figure: http://picasaweb.google.com/lh/photo/vJZRP6tAmmhpIh4J2BG0-g?feat=directlink

3 4

my own hist_equal (Vectorization gods required! ;)
by Michael Aye 09 Apr '10

09 Apr '10

Hi all, I am implementing my own histogram equalization, as the few lines of code that were mentioned before in this forum didn't really work for me and I wanted to really understand what is going on there as well anyway. So here is what I got so far, main effort was to wrap my brain around the double lookup table I needed to find the right values to replace and the new values, based on the transformed cdf. So, it seems to work fine, but the remaining loop of course disturbs me in terms of efficiency and I wondered if some of you vectorization gods could have a look if there's something possible. Especially when I have 16 bit images, there are a lot of different luminances I have to loop through, so it's not the fastest. Apart from that I think it can't be optimized further much (I think?) def hist_equal(data): # range is +2 to have the highest luminance to get into correct bin bins = numpy.arange(data.min(), data.max() + 2) # first the histogram of luminances h, bins = numpy.histogram(data, bins=bins) # now get the cdf cdf = h.cumsum() # now get the unique luminance values uniques = numpy.unique(data) # this to not look it up again everytime in the numpy array nPixel = data.size cdfmin = cdf.min() # don't change the incoming data nData = data.copy() for lum in uniques: # algorithm taken from wikipedia entry for histogram equalization # always scaling to 256 luminance values nData[numpy.where(data == lum)] = \ numpy.round((cdf[numpy.where(bins == lum)] - cdfmin) * 255 / nPixel) return nData I am thinking that this might even be useful to implement somewhere (unless it is already and I missed it??) Best regards, Michael

3 3

pythonxy scipy error win32
by bruce.labitt＠autoliv.com 09 Apr '10

09 Apr '10

Not sure how to diagnose this. Looking to the list for some guidance. Running scipy/numpy routines that use leastsq and svd. The code runs fine (to completion - and generates plots) under Ubuntu9.10 x86-64. I am also testing the routines using Pythonxy 2.6.2.0 bundle running on WinXP 32 bit. Installed plugins: CDT 6.0.1, cython 0.11.3, ETS 3.3.0, IPython 0.10.0, OpenGL 3.0.1a3, PIL 1.1.6, Photran 5.0.0, PyQt4 4.5.4, PyWavelets 0.1.6, Pydev 1.5.0, Qwt5 5.2.0, SciTE 2.01, WinMerge 2.12.4, console 2.0.144, itk 3.16, jinja2 2.2.1, matplotlib 0.99.1, mdp 2.5, mingw 4.4.0, nose 0.11.1, numexpr 1.3.1, numpy 1.3.0, opencv 1.1pre1a, pp 1.5.7, psyco 1.6, py2exe 0.6.9, pygments 1.1.1, pylint 0.18.1, pyreadline 1.5, py2win32 2.14, scipy 0.7.1, spyder 1.0.1, swig 1.3.40, vpython 5.13, vtk 5.4.2, winpdb 1.4.6, wxPython 2.8.10.1, xlutils 1.4.1, xy 1.1.0 I am running in an IPython (mlab) session. The routine runs a while, (outputting to the console) then the last message I get is: ** On entry to ZGELSS parameter number 12 had an illegal value Then the IPython session is terminated with no additional error messages. The IPython window merely reads: C:\Documents and Settings\me\.xy\startups> Google reveals that ZGELSS is an interface to LAPACK. If I run a Spyder session and execute from there, values are printed to the console then the Spyder application terminates and the entire Spyder Window vanishes. Is this a Scipy Win32 issue or Pythonxy Win32 issue, neither, both ??? In, other words, which list do I post to? TIA, Bruce ****************************** Neither the footer nor anything else in this E-mail is intended to or constitutes an electronic signature and/or legally binding agreement in the absence of an express statement or Autoliv policy and/or procedure to the contrary. This E-mail and any attachments hereto are Autoliv property and may contain legally privileged, confidential and/or proprietary information. The recipient of this E-mail is prohibited from distributing, copying, forwarding or in any way disseminating any material contained within this E-mail without prior written permission from the author. If you receive this E-mail in error, please immediately notify the author and delete this E-mail. Autoliv disclaims all responsibility and liability for the consequences of any person who fails to abide by the terms herein. ******************************

1 0

How to link against custom BLAS?
by John Reid 09 Apr '10

09 Apr '10

Perhaps I can put this another way. If I have a library libmyblasimplementation.so in directory /the/directory/where/I/store/my/libraries, what do I put in my site.cfg to get numpy and scipy to use it? I've been told the scipy build documentation isn't the best place to find this out. Also, can I share my site.cfg between numpy and scipy? Am I right in thinking that whatever I put in the numpy site.cfg is copied into the installation /usr/lib/python2.5/numpy/... somewhere and when I come to build scipy, these values are used? I'm on Ubuntu and OpenSuse. Thanks, John.

1 0

equivalent of ind2sub
by Anil C R 09 Apr '10

09 Apr '10

is there any function which is similar to matlab function ind2sub? Anil

5 5

Re: [SciPy-User] Interpolation of function on a line vs 2D interpolation
by igor Halperin 09 Apr '10

09 Apr '10

Hi Robert, On your questions: > Why would you use meshgrid? Can you show us the code you are trying? > It might be easier to see the problems that way. My bad. I glanced through examples in the cookbook and thought that interpolate.bisplev needs to pass matrices rather than 1D arrays, which is not the case, hence was my confusion. > No, not really. You will not get sensible results if you try that. You are right. Even though I do formally have a function defined on a line, this line does not have a structure, and thus I would not get sensible results in this way. I realized it after submitting the original question. My attempt to reformulate it as a line interpolation problem was due to my misinterpretation of the usage of function interpolate.bisplev per above. Looks like all I needed to do were these two standard lines of code: tck = interpolate.bisplrep(sValsOnReducedSet[:,0], sValsOnReducedSet[:,1], optValsOnReducedSet,s=0) optValsOnFullSet = interpolate.bisplev(sValsOnFullSet[:,0],sValsOnFullSet[:,1],tck) Thanks a lot for your help! Igor

1 0

re[SciPy-user] ading time series file using tsfromtxt
by Solomon M Negusse 08 Apr '10

08 Apr '10

Hello, I'm trying to read a time series of stream discharge using ts.genfromtxt. The file looks like: #agency_cd site_no datetime 02_00060_00003 02_00060_00003_cd #5s 15s 16d 14n 10s agency station 1940-07-01 1700 A agency station 1940-07-02 350 A I'm trying to read in the third 'date' column and the 4th 'data' column using: dateconverter = lambda y: datetime.date(year=int(y.split('-')[0]),month=int(y.split('-')[1]),day=int(y.split('-')[2])) flowts = ts.tsfromtxt('file', comments='#',datecols=2, usecols=(2,3),dateconverter=dateconverter,freq='D') This is returning an error Line #25309 (got 3 columns instead of 2). I'm fairly new with python and I'm hoping this will be a quick fix for the experienced users here. I appreciate any help I can get on this. Thanks, Solomon -- View this message in context: http://old.nabble.com/reading-time-series-file-using-tsfromtxt-tp28185028p2… Sent from the Scipy-User mailing list archive at Nabble.com.

1 0

Where can I get some help building scipy?
by John Reid 08 Apr '10

08 Apr '10

Is there a different mailing list for building/installing scipy?

2 2

Error in constants documentation?
by Florian Lindner 08 Apr '10

08 Apr '10

Hello, http://docs.scipy.org/doc/scipy-0.7.x/reference/constants.html chapter Force says: dyn one dyne in watts lbf one pound force in watts kgf one kilogram force in watts doesn't the watts actually have to newtons as the unit of force? Regards, Florian

10 14

Interpolation of function on a line vs 2D interpolation
by igor Halperin 08 Apr '10

08 Apr '10

Hi, I am a bit stuck with the following interpolation problem. I have to compute a function z = f(x,y) whose arguments are realizations of components of a two-dimensional random variable (X,Y), obtained by Monte Carlo. As each Monte Carlo simulation gives me a pair (x(t), y(t)) where t = 0,...,N-1 enumerates Monte Carlo runs, I can equivalently say that my function f is defined on a discretized line t = 0,..., N-1 in a 2D space. Now, as calculation of function f is costly but the function itself is smooth, I want to calculate it on a sparse subset of simulated 2D points (x(t),y(t)), and then interpolate for the rest of points. (my sparse set includes scenarios providing extreme values of x,y to avoid the need to extrapolate). Though this problem could be viewed as a 2D interpolation and solved using function interpolate.bisplrep, in practice I run out of memory with this method as the number N of 2D points is about 100,000, and I get memory error when trying to produce matrices of dimension 100,000x100,000 needed for interpolate.bisplrep using function meshgrid. However, in view of what is said above, the problem can alternatively be viewed as interpolation of a function defined on a line. I was unable to figure out how/whether I could use function interpolate.splprep to this end. Any help on this would be greatly appreciated. Igor

2 1