Mailman 3 March 2010 - SciPy-User

Scikits TimeSeries modifies names of features (dtypes
by Sergi Pons Freixes 26 Mar '10

26 Mar '10

I have this code, which simply creates a TimeSeries object from a datafile: data_labels = ["Temp", "Sal", "O2"] + self.species.keys() tmpdata = ts.tsfromtxt(join(dir,f), freq='D', delimiter='\t', \ usecols=range(len(data_labels)), datecols=0, \ names=data_labels, dateconverter=self.__dateconverter) where self.species.keys() is a list of strings, some of them with spaces, points, or parenthesis. A pair of examples: "Phaeocystis (colonies)", "Agr.fil. G. corsicum 200x". But, If I check the resulting tmpdata.dtypes.names (or simply tmpdata.dtypes), all the spaces has been changed to underscore, and the points and parenthesis removed. Examples: "Phaeocystis_colonies", "Agrfil_G_corsicum_200x". I've checked the documentation from timeseries and from record arrays of numpy, and cannot found the reason... Any hint abou it? Regards, Sergi -- Sergi Pons Freixes Ph.D. Student Marine Technology Unit http://www.utm.csic.es

2 3

Re: [SciPy-User] How to install scipy for opensuse 10.3?
by Joe Kington 26 Mar '10

26 Mar '10

Gah, meant to link to the base repo, but I managed to have the text show the base repo, and the link have the 64-bit repo... Here you go: http://download.opensuse.org/repositories/science:/ScientificLinux/SLE_10/ On Thu, Mar 25, 2010 at 2:00 PM, Joe Kington <jkington(a)wisc.edu> wrote: > It's rather dated (scipy 0.6) , but I think this repo definitely has scipy > and should be compatible with 10.3. > http://download.opensuse.org/repositories/science:/ScientificLinux/SLE_10/<http://download.opensuse.org/repositories/science:/ScientificLinux/SLE_10/x…> > > > <http://download.opensuse.org/repositories/science:/ScientificLinux/SLE_10/x…>Hope > that helps, > -Joe > > > On Thu, Mar 25, 2010 at 10:29 AM, Lam Waiman <gunbuster363(a)gmail.com>wrote: > >> But I cannot see scipy in that folder........ >> And I don't want to upgrade to 11.2 because I am not sure if my package in >> 10.3 still works.....I don't want to risk now.... >> >> 2010/3/25 Joe Kington <jkington(a)wisc.edu> >> >> I'm pretty certain scipy and numpy are in the repos for OpenSuse 10.3. >>> (or at least they were a few years ago... Is there some reason you can't >>> upgrade to 11.2? They're definitely in the repos for 11.0, 11.1, and 11.2) >>> >>> Do you have the science repos enabled? They may have dropped them for >>> 10.x, but you can always add the SEL_10 versions. They should be >>> compatible. http://download.opensuse.org/repositories/science/SLE_10/ >>> >>> Hope that helps, >>> -Joe >>> >>> On Thu, Mar 25, 2010 at 4:31 AM, Lam Waiman <gunbuster363(a)gmail.com>wrote: >>> >>>> Dear all, >>>> I want to install scipy for python usage. I see the >>>> installation isreally complex, can someone give on some advice on it? >>>> >>>> My system, opensuse 10.3 (i586) >>>> I am using python 2.6.4 >>>> >>>> when I see the doc, it said it require LAPACK to be installed, >>>> then I download lapack and see inside, it also require lots of things >>>> including fortran compiler >>>> and I don't know whether the gcc42-fortran compiler in my comp is ok. >>>> I try to dl the g77 compiler, which someone said I will need it for >>>> LAPACK >>>> then it g77 only works with a gcc compiler of version 2.8, but I am >>>> using a gcc compiler of version 4.2. >>>> >>>> SO COMPLEX! >>>> >>>> Looking forwards to your reply, thanks. >>>> >>>> Raymond >>>> >>>> _______________________________________________ >>>> SciPy-User mailing list >>>> SciPy-User(a)scipy.org >>>> http://mail.scipy.org/mailman/listinfo/scipy-user >>>> >>>> >>> >>> _______________________________________________ >>> SciPy-User mailing list >>> SciPy-User(a)scipy.org >>> http://mail.scipy.org/mailman/listinfo/scipy-user >>> >>> >> >> _______________________________________________ >> SciPy-User mailing list >> SciPy-User(a)scipy.org >> http://mail.scipy.org/mailman/listinfo/scipy-user >> >> > > _______________________________________________ > SciPy-User mailing list > SciPy-User(a)scipy.org > http://mail.scipy.org/mailman/listinfo/scipy-user > >

2 3

Re: [SciPy-User] milliseconds in matplotlib.dates?
by Tom Kuiper 25 Mar '10

25 Mar '10

> Date: Thu, 25 Mar 2010 13:17:01 -0500 > From: G?khan Sever <gokhansever(a)gmail.com> > Subject: Re: [SciPy-User] milliseconds in matplotlib.dates? > To: SciPy Users List <scipy-user(a)scipy.org> > Cc: Matplotlib Users <matplotlib-users(a)lists.sourceforge.net> > ... > Alternatively, you might use just floating-point version of your time values > and with a little adjustment (Ryan mentioned this in the recently, and will > forward there for other suggestions) millisecond resolutions should be > visible when zoomed in furthest. > > Something like: > > sci_fmt = plt.FormatStrFormatter("%.2f") > plt.gca().xaxis.set_major_formatter(sci_fmt) > Here's how I modified my code: sci_fmt = FormatStrFormatter("%.2f") fig = figure() top_axes = subplot(211) top_axes.xaxis.set_major_formatter(sci_fmt) #plot_date(times,kurts,fmt='b-') #plot_date(times,kurt_sm,fmt='r-',label=("Hamming 1-sec FWHM")) plot(times,kurts,'b-') plot(times,kurt_sm,'r-',label=("Hamming 1-sec FWHM")) I then restarted 'ipython -pylab' because I've noticed that date formatting problems seems to linger even after a code change. However, I see no change in behaviour. I can get six 1-sec ticks across my plot. If I expand the plot any more, the ticks disappear. Regards Tom

2 1

Re: [SciPy-User] milliseconds in matplotlib.dates?
by Tom Kuiper 25 Mar '10

25 Mar '10

scipy-user-request(a)scipy.org wrote: > Message: 1 > Date: Thu, 25 Mar 2010 13:17:01 -0500 > From: G?khan Sever <gokhansever(a)gmail.com> > Subject: Re: [SciPy-User] milliseconds in matplotlib.dates? > To: SciPy Users List <scipy-user(a)scipy.org> > Cc: Matplotlib Users <matplotlib-users(a)lists.sourceforge.net> > Message-ID: > <49d6b3501003251117v2e332a88t18bd52efc2d2cc4d(a)mail.gmail.com> > Content-Type: text/plain; charset="utf-8" > > On Wed, Mar 24, 2010 at 1:25 PM, Tom Kuiper <kuiper(a)jpl.nasa.gov> wrote: > > >> I'm using python 2.5.2 and matplotlib 1.1.0. >> I got mpl confused with pylab. In [2]: pylab.__version__ Out[2]: '1.1.0' > By the way, where did you get mpl v1.1.0? Mine says: > > In[19]: matplotlib.__version__ > Out[19]: '1.0.svn' > In [3]: print matplotlib.__version__ Out[3]: '0.98.1' Thanks for your help. Tom

1 0

How to install scipy for opensuse 10.3?
by Lam Waiman 25 Mar '10

25 Mar '10

Dear all, I want to install scipy for python usage. I see the installation isreally complex, can someone give on some advice on it? My system, opensuse 10.3 (i586) I am using python 2.6.4 when I see the doc, it said it require LAPACK to be installed, then I download lapack and see inside, it also require lots of things including fortran compiler and I don't know whether the gcc42-fortran compiler in my comp is ok. I try to dl the g77 compiler, which someone said I will need it for LAPACK then it g77 only works with a gcc compiler of version 2.8, but I am using a gcc compiler of version 4.2. SO COMPLEX! Looking forwards to your reply, thanks. Raymond

2 3

milliseconds in matplotlib.dates?
by Tom Kuiper 25 Mar '10

25 Mar '10

Greetings, everyone. I have some time tagged data spanning about three minutes with 4 ms resolution. If I plot these with plot_data and focus in on some event at high resolution I lose the tick marks. It seems that the minimum tick spacing is 1 sec. That appears consistent with strftime which does not provide for fractional seconds. I thought I saw this issue go by not long ago but I don't recall the solution, if any. I'm using python 2.5.2 and matplotlib 1.1.0. Thanks and regards, Tom

2 1

A newbie question on matrix multiplication
by Graziano Mirata 25 Mar '10

25 Mar '10

Hi to everybody, i am new here and above all I am new in python. After years of matlab programming I finally have switched to Python. Coming to question: i am trying to multiply a matrix transpose by itself but i've got an error on the shape. But it is impossible. Here there is the code: type(f) Out[213]: <type 'numpy.ndarray'> In [214]: type(ft) Out[214]: <type 'numpy.ndarray'> In [215]: f.size Out[215]: 12 In [216]: f.shape Out[216]: (4, 3) In [217]: ft = f.transpose() In [218]: ft.shape Out[218]: (3, 4) In [219]: ft*f --------------------------------------------------------------------------- ValueError Traceback (most recent call last) /Users/Graz/Documents/PhD researchs/Outsample Perf KT/PythonCodesOSPKT/<ipython console> in <module>() ValueError: shape mismatch: objects cannot be broadcast to a single shape Any idea where i am wrong? Thanx G

4 4

Hiring: Software Developer, Scientific Applications
by Amenity Applewhite 24 Mar '10

24 Mar '10

Enthought is hiring a Software Developer. See the description below, or on our website: http://www.enthought.com/company/sd-scientific-app.php Best, Amenity -- Amenity Applewhite Enthought, Inc. Scientific Computing Solutions www.enthought.com Software Developer The Software Developer at Enthought, Inc. participates in the development of scientific and technical applications involving GUIs, 2- D and 3-D graphics, workflow and pipeline architecture, and numerical algorithms. The position also involves some interaction with clients. Some travel may be required. We are interested both in experienced applicants as well as in recent graduates. Applicants should have a BS, MS, or PhD degree with a strong background in science and mathematics, as well as real experience developing quality software, either commercial or open source. More experienced applicants should also have demonstrated project management skills and the ability to lead a team of strong developers with highly technical backgrounds. Applicants will be measured against the following qualifications: • (Required) Bachelor's Degree in Computer Science or other scientific or engineering field with preferably an M.S. or Ph.D. degree. • (Required) Minimum 2 years of technical lead or development experience with 4 or more years preferred. • Ability to understand a problem domain and then conceive of and implement an intuitive user interface geared toward the scientist or engineer user. • Discipline, pride, and professionalism to write readable, documented, and unit-tested code that serves as an example to others who later study your work. • Strong work ethic and commitment to satisfying the customer. • Experience with Python, and a strong understanding of how to apply its capabilities to develop GUI frameworks, work flow frameworks, and elegant scientific applications. • Strong understanding of statistics, optimization, image processing, signal processing, or other technical area. • Experience with the following: • GUI frameworks such as NetBeans or Eclipse • wxPython, Qt • Low-level 2-D graphics APIs such as Quartz or GDI+ • 3-D graphics, preferably using VTK • Developing or working with plotting APIs • Experience using (and interest in contributing to) SciPy • numeric algorithms Enthought offers competitive salaries and the opportunity to work on varied and interesting technical projects. We are located in Austin, TX, consistently rated as one of the best places to live in the US. Benefits include health, dental, vision and a 401k plan. If you are interested in applying, submit a resume to jobs(a)enthought.com. Code samples and links to previous work are encouraged but not required.

1 0

Terminology question
by Ernest Adrogué 24 Mar '10

24 Mar '10

Hi, Sorry for this slightly off-topic question.. In a statistical model, what do you call the equation that defines the independent variable in terms of one or more dependent variables? Is it the "model equation"? The "regression function"? I have very good names for everything else, except for the bloody equation :( Ernest

4 4

Re: [SciPy-User] scipy.signal not installing (mac)
by Brian Larsen 24 Mar '10

24 Mar '10

David, I found part of the issue, I had some bogus env variables that I had to kill. The issue was the c was not compiling, but... I have a new error that has an odd name _aswfa_... Below is the error: Cheers, Brian In [4]: import scipy.signal --------------------------------------------------------------------------- ImportError Traceback (most recent call last) /Users/balarsen/<ipython console> in <module>() /Library/Python/2.6/site-packages/scipy/signal/__init__.py in <module>() 7 import sigtools 8 from waveforms import * ----> 9 from bsplines import * 10 from filter_design import * 11 from ltisys import * /Library/Python/2.6/site-packages/scipy/signal/bsplines.py in <module>() 2 ----> 3 import scipy.special 4 from numpy import logical_and, asarray, pi, zeros_like, \ 5 piecewise, array, arctan2, tan, zeros, arange, floor 6 from numpy.core.umath import sqrt, exp, greater, less, cos, add, sin, \ /Library/Python/2.6/site-packages/scipy/special/__init__.py in <module>() 6 #from special_version import special_version as __version__ 7 ----> 8 from basic import * 9 import specfun 10 import orthogonal /Library/Python/2.6/site-packages/scipy/special/basic.py in <module>() 6 7 from numpy import * ----> 8 from _cephes import * 9 import types 10 import specfun ImportError: dlopen(/Library/Python/2.6/site-packages/scipy/special/_cephes.so, 2): Symbol not found: _aswfa_ Referenced from: /Library/Python/2.6/site-packages/scipy/special/_cephes.so Expected in: flat namespace in /Library/Python/2.6/site-packages/scipy/special/_cephes.so -- Brian A. Larsen Space Science and Applications Group ISR-1, MS D466 Los Alamos National Laboratory Los Alamos, NM 87545 USA (For overnight add: SM-30, Bikini Atoll Road) Phone: TBD FAX: 505-665-7395 email: balarsen(a)lanl.gov Correspondence / Technical data or Software Publicly Available

1 0