jzaremba/guiqwt

Automatically exported from code.google.com/p/guiqwt

guiqwt's Issues

LoadItemsTool fails in interactive mode

What steps will reproduce the problem?
1. Open example tests/loadsavecanvas.py
2. Click on "Load items" button

What is the expected output? What do you see instead?
I expect to get an open file dialog, instead the following error is raised:

Building items and add them to plotting canvas
Traceback (most recent call last):
  File "C:\Python27\lib\site-packages\guiqwt\tools.py", line 1239, in activate
    self.activate_command(plot, checked)
  File "C:\Python27\lib\site-packages\guiqwt\tools.py", line 1704, in activate_command
    filename = self.get_filename(plot)
  File "C:\Python27\lib\site-packages\guiqwt\tools.py", line 1667, in get_filename
    self.directory, self.formats)
AttributeError: 'LoadItemsTool' object has no attribute 'directory'

What version of the product are you using? On what operating system?
guiqwt v2.1.5 on Win 7 x64 with Python v2.7.2 (32 bit)

I looked at the source code and found that although LoadItemsTool is derived 
from OpenFileTool it doesn't call the super class' constructor. Hence the 
missing attribute is not created.
Moving the declaration from the constructor the class root fixes the problem:

class OpenFileTool(CommandTool):
    directory = "" # Define it here to fix issue with LoadItemsTool

    def __init__(self, manager, formats='*.*', toolbar_id=DefaultToolbarID):
        CommandTool.__init__(self, manager, _("Open..."),
                             get_std_icon("DialogOpenButton", 16),
                             toolbar_id=toolbar_id)
        self.formats = formats

Original issue reported on code.google.com by [email protected] on 22 Aug 2012 at 9:20

Event gives wrong x-coordiante (from google groups)

What steps will reproduce the problem?

1.in the minimal example attached I get a wrong x value, the y value is right. 
If I position the mouse in the upper left corner of the plot area, the 
event tells me y=0 what is right, but x>0. I have to move the mouse 
to the left of the y-axis tick markers to get x = 0. 


What is the expected output? What do you see instead?

The correct coordinates.

What version of the product are you using? On what operating system?
Python 2.6.5, guiqwt '2.1.4', windows xp 32 bit

Original issue reported on code.google.com by [email protected] on 4 Oct 2011 at 2:46

Attachments:

• minimal_example.py

Multi-peak fitting

It would be nice to change the guifit() function so that it is convenient for 
multi-peak fitting. If I want to fit multiple peaks, e.g. 3 Gaussians, I would 
do it like that (adapted from the example in the docs):

def fit(x, params):
y0, A0, xc0, w0, A1, xc1, w1, A2, xc2, w2 = params
return y0 + A0*np.exp(-(x-xc0)**2/(2*w0**2)) + A1*np.exp(-(x-xc1)**2/(2*w1**2)) 
+ A2*np.exp(-(x-xc2)**2/(2*w2**2))

y0 = FitParam("Offset", 0., -10., 10.)
A0 = FitParam("Amplitude 0", 1., -10., 10.)
xc0 = FitParam("location 0", 2.5, 2., 5.)
w0 = FitParam("width 0", 1., -30., 30.)
A1 = FitParam("Amplitude 1", 1., -10., 10.)
xc1 = FitParam("location 1", 2.5, 2., 5.)
w1 = FitParam("width 1", 1., -30., 30.)
A2 = FitParam("Amplitude 2", 1., -10., 10.)
xc2 = FitParam("location 2", 2.5, 2., 5.)
w2 = FitParam("width 2", 1., -30., 30.)
params = [y0, A0, xc0, w0, A1, xc1, w1, A2, xc2, w2]
values = guifit(x, y, fit, params, xlabel="x", ylabel="Signal(x)")

Using guifit, the overall function, that was fitted, is now plotted. However, 
it would be nice to visualise the contributions of the individual peaks. So 
that you define fit=[fit1,fit2,fit2] with fit1=y0 + 
A0*np.exp(-(x-xc0)**2/(2*w0**2)) and the subfits are plotted additionally to 
the overall fit.

What version of the product are you using? On what operating system?

Version 2.1.6 under Linux

Original issue reported on code.google.com by [email protected] on 26 Nov 2012 at 4:57

PlotManager.create_action has fewer kwargs than qthelpers.create_action

PlotManager.create_action calls qthelpers.create_action, but it lacks some of 
its keyword arguments.

Specifically, it lacks the "checkable", "context" and "enabled" keywords.

I miss the checkable one since I want to create actions that are checkable 
without having to connect to the "toggled" signal.

Suggestion: unless there is some reason against it, I would simply implement 
PlotManager.create_action() as follows:

def create_action(self, *a, **kw): 
   return create_action(self.main, *a, **kw)

Original issue reported on code.google.com by [email protected] on 9 Jan 2013 at 8:59

Provide a PyQwt package for python 2.7

What steps will reproduce the problem?
1. install enthought python distribution
2. install guiqwt
3. try to import quiqwt

What is the expected output? What do you see instead?
i expected the import to work by it's missing pyqwt

What version of the product are you using? On what operating system?
win7 enthought python 2.7.2

Please provide any additional information below.
i can get pyqt running but not pyqwt. I'm unable to find a package or compile 
pyqwt for python 2.7. Can these be included in the installed if not already 
present?

Original issue reported on code.google.com by [email protected] on 31 Aug 2012 at 4:29

changleog of 2.2.0 refer to 2.1.6

Hello, just a small inconsistency

the changelog of 2.2.0 speak about 2.1.6 :)

cheers


Frederic

Original issue reported on code.google.com by [email protected] on 9 Sep 2012 at 8:42

guiqwt-tests missing manpage and .desktop file

idem guidata [1]

[1] http://code.google.com/p/guidata/issues/detail?id=1

See you

Original issue reported on code.google.com by [email protected] on 4 Jan 2011 at 9:31

Attachments:

• guiqwt-tests.desktop

patch - convert setup.py to use distribute/setuptools

* Scripts are auto generated and lunch with no console on win32
* dependencies are auto installed (also depends on Spyder/guidata fixes)

Original issue reported on code.google.com by grizzly.nyo on 6 Dec 2012 at 8:40

Attachments:

• 589.patch

ImageItem should accept int64 and uint64 dtypes

If I do ImageItem.set_data(foo) where foo is a numpy array with dtype=int64, I 
get the following exception:

TypeError: src data type must be one of the following: double, float, uint32, 
int32, uint16, int16, uint8, int8, bool


It would be nice if int64 and uint64 are also accepted.

Original issue reported on code.google.com by [email protected] on 11 Feb 2013 at 10:53

french docstring

Just a minor thing
The docstring of class CommandTool (in guiqwt/tool.py) is not translated to 
English

Original issue reported on code.google.com by [email protected] on 28 Apr 2011 at 4:14

test_line.py makes illegal call to _line_test

What steps will reproduce the problem?
1. either run the entire test suite, or
2. run test_line.py

What is the expected output? 
something other than:

In [1]: run test_line.py
0 , 0 -> 9 , 9
------------------------------------------------------------
Traceback (most recent call last):
  File "test_line.py", line 27, in <module>
    test_line(0,0, 9,9)
  File "test_line.py", line 23, in test_line
    line(x0, y0, x1, y1, imin, imax)
TypeError: _vert_line() takes exactly 7 arguments (6 given)


What version of the product are you using? On what operating system?
$ python -c 'import guiqwt; print(guiqwt.__version__)'
2.1.1


Please provide any additional information below.
I tried to unravel the call in pcolor.cpp:
PyObject *py_vert_line(PyObject *self, PyObject *args)
...
if (!PyArg_ParseTuple(args, "ddddiOO:_vert_line", 
&x0,&y0,&x1,&y1,&xmax,&p_min,&p_max)) {return NULL;}

but somehow failed to uncover what I would have to correct - sorry.

Original issue reported on code.google.com by [email protected] on 20 Apr 2011 at 10:05

Add mechanism to rangeselections for displaying data while moving the range borders

Currently no (accurate) feedback is given when the borders of a range tool are 
moved. I think it could be useful to display the current position and/or 
additional data dynamically while the borders are moved (similar like when 
pressing alt to get the cursor coordinates). This would make it possible to 
select a specific value for the borders with the mouse or display a range 
dependent computation without clobbering up the rest of the graph while the 
borders are not moved.


As a picture says more than a thousand words, I included a picture of how I 
imagined it could look like.

Original issue reported on code.google.com by [email protected] on 18 Aug 2011 at 4:23

Attachments:

• range_tool_improvement.png

Baseplot captures mouseDoubleClickEvent for whole plot area instead of just the axis area

Since Baseplot now implements mouseDoubleClickEvent, a reimplementation at a 
higher level is ignored (it never sees the event) so you can't implement custom 
actions (for instance, double-clicking the plot auto-zooms).

What steps will reproduce the problem?
1. Use the plot.py example from tests
2. Subclass CurveDialog to a local copy (say CurveDialogLocal) and implement a 
mouseDoubleClickEvent() there.
3. Verify that double-clicking on the plot area does not trigger the event in 
CurveDialogLocal.

This occurs because the mouseDoubleClickEvent in BasePlot doesn't ignore 
double-clicks that are on the canvas area (it just accepts them and does 
nothing).  Modifying BasePlot's mouseDoubleClickEvent to detect when the 
d-click is on the canvas and calling event.ignore() will pass the event up and 
allow higher level widgets to respond.  Suggested fix:

    def mouseDoubleClickEvent(self, event):
        """Reimplement QWidget method"""
        if self.canvas().geometry().contains(event.pos()): #NEW
            event.ignore()                                 #NEW
            return                                         #NEW
        for axis_id in self.AXIS_IDS:
            widget = self.axisWidget(axis_id)
            if widget.geometry().contains(event.pos()):
                self.edit_axis_parameters(axis_id)


What version of the product are you using? On what operating system?
Using guiqwt 2.2.1-1 (installed via PythonXY) on Windows7 (and Scientific Linux 
6.2 - where I have the older version installed that works since mouseDCEvent 
isn't implemented there).

Please provide any additional information below.

Original issue reported on code.google.com by [email protected] on 6 Feb 2013 at 5:04

guiqwt.fit.guifit: standard fit boundary excludes rightmost value

What steps will reproduce the problem?

The standard boundary generated automatically for the fitting excludes the 
rightmost value (The leftmost value is included in the fit). Manually setting 
the boundary conditions can be used as a work around.   
Example code:

import numpy as np

from guiqwt.fit import FitParam, guifit

def test():
    x = np.arange(1, 10)
    y = np.arange(1, 10)
    # setting the rightmost value to 11 instead of 10 => slope > 1
    y[-1] = 11

    def fit(x, params):
        a, b = params
        return (a*x)+b

    a = FitParam("Slope", 1., 0., 2.)
    b = FitParam("Offset", 0.01, -1., 1.)
    params = [a, b]
    values = guifit(x, y, fit, params)

if __name__ == "__main__":
    test()

What is the expected output? What do you see instead?

I expect a slope >1 and a non-zero offset auf using the autofit function. 
Instead the slope is 1 with zero offset. Once the upper boundary is increased 
manually, everything works as expected.


What version of the product are you using? On what operating system?

guiqwt 2.0, debian testing, spyder right from the repository (2.08+)

Original issue reported on code.google.com by [email protected] on 9 Mar 2011 at 6:48

The values display od markers has a problem with very small values (< 1e-6)

What steps will reproduce the problem?
1. create a plot
2. fill it with data with values smaller than 1e-6
3. display a marker (free or bound to items)

What is the expected output? What do you see instead?
I would like to see the value at the marker but 0.000000 is displayed. (e.g. 
see attached screen shot)

What version of the product are you using? On what operating system?
guiqwt 2.2.1-1 (from debian sid)
guidata 1.3.0-1.1 (from Debian sid)


Please provide any additional information below.
I propose to switch from %6f to %6g for the marker string format as this 
ensures that small values are displayed correctly when a log axis and small 
values are displayed.

Original issue reported on code.google.com by [email protected] on 26 Apr 2011 at 2:19

Attachments:

• guiqwt_marker_bug.png

Circular reference preventing CurvePlot object from being garbage collected

See here:
http://code.google.com/p/spyderlib/issues/detail?id=1021

Would you be kind enough to take a look at it, Ludovic?

Original issue reported on code.google.com by pierre.raybaut on 24 May 2012 at 12:32

Some minor issues about provided CHM documentation

Hi,

I think that it is great you provide ready made CHM output from sphinx, and 
would further suggest:

1. store CHM file in ZIP archive and provide that as a download link, because 
direct CHM download causes CHM file to be blocked, and user needs to unblock it 
before being able to see contents (I guess not many would know about this)

2. "sphinxdoc.css" assigns margins for HTML output which are not needed for CHM

I addressed both issues, and supply file in attachment.


Cheers

Original issue reported on code.google.com by [email protected] on 12 Mar 2013 at 5:20

Attachments:

• guiqwtdoc.zip

guiqwt won't build on MacOSX

Ran setup with Python 2.7.2 and guidata 1.4.0
See attached files for build and run output.

Original issue reported on code.google.com by [email protected] on 9 Sep 2011 at 2:19

Attachments:

• guiqwt_build.txt
• guiqwt-tests.txt

numpy.nan support request

What steps will reproduce the problem?
1. Trying to display a numpy.ndarray (dtype = float32) containing one or more 
numpy.nan (not a number) in an ImageDialog.

What is the expected output? What do you see instead?
It would be nice to see the image, with np.nan points set to an arbitrary value 
(img.min(), img.max(), ...), just like with matplotlib, instead of a blue 
image. Surfing the image in the ImageDialog still shows the right value for the 
points though.

What version of the product are you using? On what operating system?
guiqwt 2.0.8.1 on Gnu/Linux x86_64, python 2.7.1, qt 4.7.1, pyqt 4.8.3.

Please provide any additional information below.

Original issue reported on code.google.com by [email protected] on 20 Feb 2011 at 8:40

PDF output has wrong aspect

What steps will reproduce the problem?
1. program "simple_widows.py" displays brain.png with correct aspect ratio
2. "Print" command (icon) in "simple_widows.py" creates a PDF with wrong aspect 
ratio

What is the expected output? What do you see instead?
I expect the same aspect ratio as in the image window 

What version of the product are you using? On what operating system?
Python 2.6.6 Qt 4.6.3 PyQt 4.7.3 GUIqwt 2.2.0 GUIdata 1.5.0 on Linux 
2.6.32-5-amd64

Please provide any additional information below.
The aspect ratio seems to change between the display in "simple_windows" and 
the PDF output. The ration changes with the ratio of the page hight and width.

Original issue reported on code.google.com by [email protected] on 30 Jan 2013 at 1:44

Attachments:

• print.pdf
• brain.png

guiqwt does not build on amd64

Hello when I try to build guiqwt on an amd64 debian system it failes with

> /usr/bin/gfortran -Wall -shared
> build/temp.linux-x86_64-2.5/build/src.linux-x86_64-2.5/guiqwt/_extmodule.o
> build/temp.linux-x86_64-2.5/build/src.linux-x86_64-2.5/fortranobject.o
> build/temp.linux-x86_64-2.5/src/histogram.o -lgfortran -o
> build/lib.linux-x86_64-2.5/guiqwt/_ext.so
> /usr/bin/ld: build/temp.linux-x86_64-2.5/src/histogram.o: relocation
> R_X86_64_PC32 against undefined symbol `lround@@GLIBC_2.2.5' can not be used
> when making a shared object; recompile with -fPIC
> /usr/bin/ld: final link failed: Bad value

reading the documentation I have this

       long int lround(double x);

      All functions shown above:
           _XOPEN_SOURCE >= 600 || _ISOC99_SOURCE ||
           _POSIX_C_SOURCE >= 200112L;
           or cc -std=c99

maybe there is a problem with the build flags, or a -lm link missing ?

See you

Original issue reported on code.google.com by [email protected] on 4 Jan 2011 at 9:28

Rename tests/qtdesigner.pyw to tests/qtdesigner.py

This will avoid a bug on Linux platforms when running the GUI-based test 
launcher.

Original issue reported on code.google.com by pierre.raybaut on 17 Jul 2011 at 12:50

no able to build windows executable

I have Python XY with latest code.

http://packages.python.org/guiqwt/disthelpers.html#example

Building this example yields to:


>python setup.py py2exe

C:\Python26\lib\site-packages\py2exe\build_exe.py:16: DeprecationWarning: the 
sets module is deprecated
  import sets
Removing 'build' and 'dist' directories
Traceback (most recent call last):
  File "setup.py", line 21, in <module>
    DATA_FILES = create_vs2008_data_files()
  File "C:\Python26\lib\site-packages\guidata\disthelpers.py", line 146, in create_vs2008_data_files
    assembly = ElementTree.fromstring(file(manifest).read())
IOError: [Errno 2] No such file or directory: 
'C:\\Python26\\Microsoft.VC90.CRT.manifest'

Original issue reported on code.google.com by [email protected] on 9 Mar 2011 at 9:45

Legend with restricted items

What steps will reproduce the problem?
1. Use builder to create plots
2. Use builder to add legend restricted to only some of the plots
3. - Error - TypeError: include_item() takes exactly 2 arguments (1 given)

What is the expected output? What do you see instead?
You should get the legend restricted to the selected items

What version of the product are you using? On what operating system?
guiqwt 2.0.8.1
Windows 

Please provide any additional information below.
in order to solve this problem replace in guiqwt.label.SelectedLegendBoxItem 
class the entry:
**
def include_item(self, item):
        return LegendBoxItem.include_item(self) and item in self.itemlist
**
with
**
def include_item(self, item):
        return LegendBoxItem.include_item(self, item) and item in self.itemlist
**

Original issue reported on code.google.com by [email protected] on 28 Mar 2011 at 3:16

Curve plot not shown because underlying data was deleted

What steps will reproduce the problem?
1.
    plot(make.curve(np.arange(10), np.arange(10))) works
2.
    y,x = np.histogram(areas)
    plot(make.curve(x[:-1],y))
crashes with the error: "underlying c/c++ object has been deleted"

I'm on guiqwt 2.2.1 (latest release), cvs won't compile for me, different bug 
report

Original issue reported on code.google.com by [email protected] on 18 Mar 2013 at 5:55

Request about guiqwt.pyplot.savefig

Hi Pierre , hi all

We have a request about guiqwt.pyplot.savefig  which has less possibilities than
its matplotlib conterpart.

In guiqwt, the only possible result of savefig is an image file, 
either png or pdf.

In matplotlib, we can save the result as a string buffer as in
....
f=cStringIO.StringIO()
matplotlib.pyplot.savefig(f,'png')
....

This saving in a meomry buffer is not possible with guiqwt.pyplot.savefig

We have the request it could be possible, as in matplotlib.

Thanks in adance for eventually consider this enhancement.

Claude

Original issue reported on code.google.com by [email protected] on 3 Nov 2011 at 12:20

Item List tools looses names of curves

What steps will reproduce the problem?
1. curvewidget with added curves which have titles
2. Item List tool will show the titles 
3. by enable/disable the titles will eventually disappear
4. setting titles by item list tool will not disappear  

What is the expected output? What do you see instead?

titles in item list disappear by enabling/disabling curves 

What version of the product are you using? On what operating system?

guiqwt-2.2.1-py2.7.egg-info

Please provide any additional information below.

Original issue reported on code.google.com by [email protected] on 22 Mar 2013 at 9:49

sift examples : 'Marker' object has no attribute 'setSpacing'

What steps will reproduce the problem?
1. launch SIFT examples (/usr/bin/sift)


What is the expected output? What do you see instead?
Testing guiqwt with demonstration example.


What version of the product are you using? On what operating system?
OS : Centos 6.3
Python : 2.6.6
PyQt4 : 4.6.2 (installation from rpm)
PyQwt : 5.2.0 (installation from rpm)
guidata : 1.4.1 (installation from sources)
numpy : 1.4.1 (installation from rpm)
scipy : 0.7.2 (installation from rpm)
PIL : 1.1.7 (installation from sources)

Please provide any additional information below.
Traceback (most recent call last):
  File "/usr/bin/sift", line 3, in <module>
    sift.run()
  File "/usr/lib/python2.6/site-packages/guiqwt/tests/sift.py", line 1271, in run
    window = MainWindow()
  File "/usr/lib/python2.6/site-packages/guiqwt/tests/sift.py", line 1108, in __init__
    curveplot_toolbar)
  File "/usr/lib/python2.6/site-packages/guiqwt/tests/sift.py", line 1031, in __init__
    self.plotwidget = plotwidgetclass()
  File "/usr/lib/python2.6/site-packages/guiqwt/plot.py", line 725, in __init__
    section, show_itemlist, gridparam)
  File "/usr/lib/python2.6/site-packages/guiqwt/plot.py", line 700, in __init__
    section=section, gridparam=gridparam)
  File "/usr/lib/python2.6/site-packages/guiqwt/curve.py", line 1293, in __init__
    self.cross_marker = Marker()
  File "/usr/lib/python2.6/site-packages/guiqwt/shapes.py", line 288, in __init__
    self.markerparam.update_marker(self)
  File "/usr/lib/python2.6/site-packages/guiqwt/styles.py", line 1362, in update_marker
    obj.setSpacing(self.spacing)
AttributeError: 'Marker' object has no attribute 'setSpacing'

Original issue reported on code.google.com by [email protected] on 28 Jul 2012 at 5:37

cannot build on Windows 64 bit

What steps will reproduce the problem?

* Environment
1. Windows Server 2008 Std 64 bit 
2. MSVC2010 Professional
3. Python 2.6.6 64 bit

* build command
> python setup.py install



What is the expected output? What do you see instead?

* error message

(...omitted...)

histogram.o : error LNK2019: unresolved external symbol "lround" referenced in 
function "hist2d_"
build\lib.win-amd64-2.6\guiqwt\_ext.pyd : fatal error LNK1120: 1 unresolved 
external
error: Command "C:\Program Files (x86)\Microsoft Visual Studio 
10.0\VC\BIN\amd64\link.exe /DLL /nologo /INCREMENTAL:NO 
/LIBPATH:d:\mingw-w64\lib\gcc\x
86_64-w64-mingw32\4.5.4 /LIBPATH:d:\mingw-w64\x86_64-w64-mingw32\lib 
/LIBPATH:C:\Python26\libs /LIBPATH:C:\Python26\PCbuild\amd64 
/LIBPATH:C:\Python26
\libs /LIBPATH:C:\Python26\PCbuild\amd64 /EXPORT:init_ext 
build\temp.win-amd64-2.6\Release\build\src.win-amd64-2.6\guiqwt\_extmodule.obj 
build\temp.wi
n-amd64-2.6\Release\build\src.win-amd64-2.6\fortranobject.obj 
build\temp.win-amd64-2.6\Release\src\histogram.o 
/OUT:build\lib.win-amd64-2.6\guiqwt\_ex
t.pyd 
/IMPLIB:build\temp.win-amd64-2.6\Release\build\src.win-amd64-2.6\guiqwt\_ext.lib
 /MANIFESTFILE:build\temp.win-amd64-2.6\Release\build\src.win-am
d64-2.6\guiqwt\_ext.pyd.manifest" failed with exit status 1120




What version of the product are you using? On what operating system?
guiqwt-2.1.4



Please provide any additional information below.

Original issue reported on code.google.com by [email protected] on 20 Jul 2011 at 1:38

Add spectrogram/raster plot type

The image plot type is good enough for doing static displays of 3D curves but 
doesn't perform well for plotting in realtime as I believe the Qwt Spectrogram 
plot is.

Original issue reported on code.google.com by [email protected] on 14 Feb 2012 at 7:50

Problemn when changing the contrast of an image

Hello, I got this error when playing with a tiff image

Traceback (most recent call last):
  File "/usr/lib/python2.7/dist-packages/guiqwt/histogram.py", line 364, in set_full_range
    imin, imax = item.get_lut_range_full()
  File "/usr/lib/python2.7/dist-packages/guiqwt/image.py", line 488, in get_lut_range_full
    return _nanmin(self.data), _nanmax(self.data)
  File "/usr/lib/python2.7/dist-packages/guiqwt/image.py", line 190, in _nanmin
    return np.nanmin(data)
  File "/usr/lib/pymodules/python2.7/numpy/lib/function_base.py", line 1507, in nanmin
    return np.fmin.reduce(a.flat)
TypeError: cannot reduce on a scalar

if you want the image just tell me

Original issue reported on code.google.com by [email protected] on 18 Mar 2013 at 2:29

sift: image data not accessible in sift namespace of console

What steps will reproduce the problem?
1. start sift (python $LOCALDISTPACKAGES/guiqwt/tests/sift.py)
2. open an image or create new image
3. attempt to access it in the console: 
>>> len(sift.i)
0

What is the expected output? What do you see instead?
A new ImageParam object should turn up there, instead the sift.i list is empty. 
What's more, if you create or open signals they are visible as sift.s *and* 
sift.i


What version of the product are you using? On what operating system?
sift: 0.2.2
guiqwt: 2.1.0

Please provide any additional information below.
suggested bugfix in class SiftProxy works for me and restores expected 
behaviour:
diff /usr/local/lib/python2.6/dist-packages/guiqwt/tests/sift.py sift.py 
976c976
<         self.i = self.win.signalft.objects

---
>         self.i = self.win.imageft.objects

Original issue reported on code.google.com by [email protected] on 11 Apr 2011 at 10:43

does not build on few architectures due to hard linked flag "-msse2"

Hello, I attached a patch sent by a Debian user to help fix a few FTBFS (fail 
to build from sources) on different architectures.

I found a problem with the sparc kfreebsd architecture with this patch, but I 
do not know if this is a bug in pyton or numpy.

Maybe it could be interesting to check also for sse3 on architectures suporting 
it :).

See you

Frederic

Original issue reported on code.google.com by [email protected] on 20 Feb 2011 at 8:06

Attachments:

• sse2.patch

How to dinamically ad plots after a toolbar has been created ?

Hi, I figured that you can't dynamically add or remove items from plots once a 
toolbar has been added.

for example the next code fails to add 1000 sinusoid waveforms to the plot:

from guiqwt.plot import CurveDialog
fig = CuveDialog()
p = fig.get_plot()

t=linspace(0,1,31)
a=[]

for k in range(1000):
    a.append(0.8*(rand()+0.5)*sin(2*pi*t+0.7*rand()))

curves=[]
for k in a:
    curves.append(make.mcurve(t,k))

for k in curves:
    p.add_item(k)

p.replot()


Is there any workaround to this problem. Specially when your are creating 
applications that need to dynamically use the existing axes to plot new data.

Thanks

Original issue reported on code.google.com by [email protected] on 23 Jun 2012 at 3:10

fix FTBFS on kfreebsd due to Werror

Hello, I rebuilt guiqwt on the debian infrastructure, and now I got a failed to 
build on kfreebsd arch.

this is due to a change in the compilation toolchain

/usr/include/sys/syslimits.h:38:2: error: #warning "No user-serviceable parts 
inside." [-Werror=cpp]
cc1plus: all warnings being treated as errors
error: Command "g++ -pthread -fno-strict-aliasing -DNDEBUG -g -fwrapv -O2 -Wall 
-fPIC -I/usr/lib/pymodules/python2.6/numpy/core/include 
-I/usr/include/python2.6 -c src/pcolor.cpp -o 
build/temp.gnukfreebsd-8.1-1-686-smp-i686-2.6/src/pcolor.o -Wall -Werror" 
failed with exit status 1In file included from 
/usr/include/i386-kfreebsd-gnu/bits/local_lim.h:20:0,

                 from /usr/include/i386-kfreebsd-gnu/bits/posix1_lim.h:157,
                 from /usr/include/limits.h:145,
                 from /usr/lib/gcc/i486-kfreebsd-gnu/4.6/include-fixed/limits.h:169,
                 from /usr/lib/gcc/i486-kfreebsd-gnu/4.6/include-fixed/syslimits.h:7,
                 from /usr/lib/gcc/i486-kfreebsd-gnu/4.6/include-fixed/limits.h:34,
                 from /usr/include/python2.6/Python.h:19,
                 from src/pcolor.cpp:8:
/usr/include/sys/syslimits.h:38:2: error: #warning "No user-serviceable parts 
inside." [-Werror=cpp]
cc1plus: all warnings being treated as errors
dh_auto_build: python2.6 setup.py build --force returned exit code 1

It is better to avoir using Werror as it can cause some troubles on a few 
architectures for no good reasons.

thanks

Frederic

Original issue reported on code.google.com by [email protected] on 26 Feb 2012 at 1:57

Attachments:

• 0001-fix-forwarded-upstream-for-FTBFS-on-kfreebsd.patch

Add new way to enter a mask

Hello, I put here some remarks of one of our user.

- It would be nice if free form could be used to define a mask. (at least 
polygone)

- It would be nice to define a mask using an histogram, inside min, max or 
outside min, max.

Original issue reported on code.google.com by [email protected] on 18 Mar 2013 at 3:45

Build fails "lipo: can't open input file: /var/tmp//ccz8lan2.out (No such file or directory)" osx 10.7

What steps will reproduce the problem?
1. install pyqt from homebrew, other dependencies from source/setup.py
2. download guiqwt, run sudo python setup.py install
3.

What is the expected output? What do you see instead?
Build stops with 

lipo: can't open input file: /var/tmp//ccz8lan2.out (No such file or directory)


What version of the product are you using? On what operating system?
2.1.4 (all dependencies current). Mac OSX 10.7 Lion

Original issue reported on code.google.com by [email protected] on 18 Oct 2011 at 12:04

Attachments:

• guiqwt_build.txt

guiqwt 2.1.4 needs guidata 1.3.2 which is not compatible with PyQt 4.4

The new versions of guiqwt (2.1.4) and guidata (1.3.2) introduce many changes 
to make them PyQt4.4 compatible.

But the following issue prevents me from installing on a system with PyQt4.4:
http://code.google.com/p/guidata/issues/detail?id=14

Therefore it is not yet possible to install any (recent) version of guiqwt on 
machines with PyQt4.4 without patching it.

Original issue reported on code.google.com by [email protected] on 26 Jul 2011 at 1:32

RuntimeError: Failed to make contiguous array of PyArray_DOUBLE

What steps will reproduce the problem?
1. from terminal: ipython testplot.py

What is the expected output? What do you see instead?
expected output should be comparable to running testplot.py with importing 
matplotlib instead of guiqwt. the output is a plot of the 10 data points

after running testplot.py:
RuntimeError                              Traceback (most recent call last)
/usr/local/Cellar/python/2.7.3/Frameworks/Python.framework/Versions/2.7/lib/pyth
on2.7/site-packages/IPython/utils/py3compat.pyc in execfile(fname, *where)
    176             else:
    177                 filename = fname
--> 178             __builtin__.execfile(filename, *where)

/Users/Jonah/programs/testplot.py in <module>()
      8 ydata = [23,34,45,56,44,37,23,18,12,6]
      9 
---> 10 plt.plot(xdata,ydata)
     11 plt.show()

/usr/local/Cellar/python/2.7.3/Frameworks/Python.framework/Versions/2.7/lib/pyth
on2.7/site-packages/guiqwt/pyplot.pyc in plot(*args, **kwargs)
    474     """
    475     axe = gca()
--> 476     curves = make.mcurve(*args, **kwargs)
    477     if not isinstance(curves, list):
    478         curves = [curves]

/usr/local/Cellar/python/2.7.3/Frameworks/Python.framework/Versions/2.7/lib/pyth
on2.7/site-packages/guiqwt/builder.pyc in mcurve(self, *args, **kwargs)
    307                 param.label = make_title(basename, CURVE_COUNT)
    308             update_style_attr(stylei, param)
--> 309             curves.append(self.pcurve(x, yi, param, **kwargs))
    310         if len(curves) == 1:
    311             return curves[0]

/usr/local/Cellar/python/2.7.3/Frameworks/Python.framework/Versions/2.7/lib/pyth
on2.7/site-packages/guiqwt/builder.pyc in pcurve(self, x, y, param, xaxis, 
yaxis)
    322         """
    323         curve = CurveItem(param)
--> 324         curve.set_data(x, y)
    325         curve.update_params()
    326         self.__set_curve_axes(curve, xaxis, yaxis)

/usr/local/Cellar/python/2.7.3/Frameworks/Python.framework/Versions/2.7/lib/pyth
on2.7/site-packages/guiqwt/curve.pyc in set_data(self, x, y)
    449         self._x = np.array(x, copy=False)
    450         self._y = np.array(y, copy=False)
--> 451         self.setData(self._x, self._y)
    452 
    453     def is_empty(self):

RuntimeError: Failed to make contiguous array of PyArray_DOUBLE


What version of the product are you using? On what operating system?
OS X 10.8
guiqwt: 2.1.6
gt: 4.8.2
qwt: 6.0.1
python: 2.7.3
pyqwt: 5.2.0
pyqt 4.9.4

Please provide any additional information below.
error only occurs when using guiqwt.pyplot as the graphing engine but not when 
using matplotlib

Original issue reported on code.google.com by [email protected] on 8 Aug 2012 at 5:51

Attachments:

• testplot.py

How to customize a shape produced by a tool like FreeFormTool

Hi guys !
First of all, thanks for the job, it's quite nice !

I am building an app in which I need to select clusters of points. For that 
purpose I saw the FreeFormTool and it is able to build one polygon for me. The 
problem is that I want to create several polygons each one with a different 
color/marker. Is there ant way to specify in advance the properties of a newly 
created polygon ? Once a Polygon is created, how can I get a handle to interact 
with it, i.e. obtain the coordinates ?

Thanks

Original issue reported on code.google.com by [email protected] on 19 Jun 2012 at 6:44

Source code repository online

Hello,
I would like to import the sourcecode of this module into Launchpad.net in 
order to be able create Ubuntu packages of the module automatically.

Please load the guiqwt source code up to google code.

Thanks in advance.

Original issue reported on code.google.com by [email protected] on 20 Feb 2011 at 6:22

Saving plot (picture)

What steps will reproduce the problem?
1. Create a curve plot with the standard tools
2. Try to save the plot writing only the name, without the type (.png, .pdf).

What is the expected output? What do you see instead?

I expect the program save the picture as a "default" option (.png).

What version of the product are you using? On what operating system?

2.1.5

Please provide any additional information below.

A way to fix the problem can be in the attached file.

Original issue reported on code.google.com by [email protected] on 19 Sep 2011 at 11:54

Attachments:

• file_name.py

Is it possible to display an image in log scale

With scientific images, usually we have a great dynamic so it is very 
convenient to plot them in logscale.

Indeed I can transfor the image with something like this

numpy.log(1 + img + img.min())

but , I want when I read the stat of the image, data based on the real values, 
not the transformed one.

Or with the shift button to read the real z valu of the image not the log of 
this value.

thanks

Frederic

Original issue reported on code.google.com by [email protected] on 18 Mar 2013 at 3:17

set_manager/plot_id

Hi,

while trying to synchronize several plot axes, I came across what looks like a 
bug in the call to set_manager in plot.py. If I understand the logic in plot.py 
right, the attached fix seems to be needed to enable synchronized axes.

Cheers,

Onno

Original issue reported on code.google.com by [email protected] on 13 Oct 2012 at 3:07

Attachments:

• set_manager_plot_id.diff

Double click on axis to change its range (min/max)

See this thread:
https://groups.google.com/group/guidata_guiqwt/browse_thread/thread/f2b11459058a
35fa

For example, the idea would be to double click on any part of the X-axis to set 
the Xmin/Xmax values. (This is a simplified version of the feature suggested in 
cited thread above).

Original issue reported on code.google.com by pierre.raybaut on 5 Nov 2012 at 8:56

guiqwt unable to import pyplot

What steps will reproduce the problem?
1. import guiqwt.pyplot

What is the expected output? What do you see instead?

In [4]: import guiqwt.pyplot
Module 'guiqwt.image': missing fortran or C extension
try running :python setup.py build_ext --inplace -c mingw32
------------------------------------------------------------
Traceback (most recent call last):
  File "<ipython console>", line 1, in <module>
  File "C:\Python26\lib\site-packages\guiqwt\pyplot.py", line 111, in <module>
    from guiqwt.plot import PlotManager
  File "C:\Python26\lib\site-packages\guiqwt\plot.py", line 120, in <module>
    from guiqwt.image import ImagePlot
  File "C:\Python26\lib\site-packages\guiqwt\image.py", line 146, in <module>
    from guiqwt._ext import hist2d
ImportError: DLL load failed: The specified module could not be found.



What version of the product are you using? On what operating system?
guiqwt 2.0.8.1 running on windows 7 64 bit

Please provide any additional information below.
I have pythonxy 2.6.5.6 installed

Original issue reported on code.google.com by [email protected] on 10 Jan 2011 at 10:58

Add "Coding guidelines" to ensure PyQt v4.4+ compatibility

Following this thread:
http://groups.google.com/group/guidata_guiqwt/t/bc194cb081f15e6f

We should add a text file in the source repository on "coding guidelines for 
guiqwt developers" to handle PyQt v4.4+ compatibility issues. Besides, this 
will certainly be useful when the PyQt API #1/2 compatibility issues will be 
handled.

Original issue reported on code.google.com by pierre.raybaut on 21 May 2011 at 5:04

bug in XYImageItem for big values in x

Hi, 

I posted this to the list some days ago, and I got no answer, so I am 
re-posting here as a bug (I should have done it in the first place, I guess... 
sorry)

What steps will reproduce the problem?
Run the following code:

+++++++++++++++++++++++++++++++++++++
def xyimagebug(offset):
    from guiqwt.plot import ImageDialog
    from guiqwt.builder import make
    import numpy
    import guidata
    app = guidata.qapplication()
    data = numpy.random.rand(100,100)
    x = numpy.arange(100)+offset 
    y = numpy.arange(100)
    image = make.xyimage(x,y,data=data)
    win = ImageDialog()
    plot = win.get_plot()
    plot.add_item(image)
    plot.select_item(image) #this helps in seeing where the image should be
    win.exec_()

xyimagebug(1e9) #offset=1e9 makes it fail, but offset=1e4 is ok
+++++++++++++++++++++++++++++++++++

What is the expected output?
If called as "xyimagebug(1e4)", everything works as expected (see ok.png 
attached). You can pan and zoom in/out without problems...

What do you see instead?
but if launched as "xyimagebug(1e9)", you will see that either:
a) the image does not appear at all (see bad2.png)
b) a wrong image appears (see bad1.png)
c) the image appears/disapears when panning or zooming in/out

What version of the product are you using? On what operating system?

Bug seen in guiqwt 2.1.3 and 2.1.4beta on Ubuntu 10.10 and SuSe 11.1


Please provide any additional information below.

I tried using ints instead of doubles for "data" and "x", but the problem 
still occurs
Also, I did a similar plot using QwtPlotSpectrogram from PyQwt and it worked, 
so my guess is that the problem is not caused by PyQwt (but I cannot be sure 
since I don't know all the details on how guiqwt calls PyQwt)

Original issue reported on code.google.com by [email protected] on 29 Jun 2011 at 8:36

Attachments:

• ok.png
• bad1.png
• bad2.png

Add "Auto range" checkable entry to the plot widget context menu

See this thread:
https://groups.google.com/group/guidata_guiqwt/browse_thread/thread/f2b11459058a
35fa

Original issue reported on code.google.com by pierre.raybaut on 5 Nov 2012 at 8:57

ImagePlot ignores lock_aspect_ratio when zooming (and possibly under other circumstances)

What steps will reproduce the problem?
1. Run image.py from the tests.run()
2. Verify that Apect Ratio is set to locked (right-click on image etc)
3. Zoom to an area of aspect ratio different from the one you're currently 
viewing, e.g. 'very' rectangular

What is the expected output? What do you see instead?
It should zoom in maybe taking the longer of the two sides of the zoom 
rectangle as an indication for the desired zoom factor *but* maintain the 
aspect ratio. Instead, it will zoom to the area asked for ignoring the current 
aspect ratio. The behaviour is deterministic but not what I expect - maybe it's 
me? One can go back to the desired aspect ratio by right clicking, Aspect 
Ratio/Edit.../OK and pressing OK. 

What version of the product are you using? On what operating system?
guiqwt __version__ = '2.0.8.1' on ubuntu


Please provide any additional information below.
Bizarrely, even on first startup the aspect ratio has been fiddle with and 
comes up as 0.972 (current) rather than 1.

Original issue reported on code.google.com by [email protected] on 7 Apr 2011 at 10:11