Spyder and Eric5

[janwillem]

I have been using Eric5 for some while. I like it in particular
because of the simple way it offers for generating python code for the
PyQt4 slots and signals. It relieves one from the need to realy
understand slots, signals, events et cetra. However, I never could get
code completion to work properly. If I e.g. type sys. then I only get
the sys methods already used in my own source. In Spyder on the other
hand you get all sys methods listed which is for somebody with a
memory like mine very convenient. In Spyder you can start QT4Designer
but after that it stops; I cannot find more PyQt4 integration in
Spyder. So for now it seems Eric5 for the PyQt4 part and Spyder for
all other source editing. There must be a better solution. Any
suggestions welcome.
Janwillem
P.S. My backgound: 15 years Delphi-1 through 7 and since a few years
Python (on linux) mainly command line + matplotlib.
Placed on both Eric and Spyder lists

[Carlos Córdoba]

Hi Janwillem,

As you probably know, Spyder is more oriented to do scientific
computations than GUI programming. That doesn't mean your request is
valid, quite the contrary, it'd be quite handy to have in Spyder. It
means someone have to stand up and implement it.

The other way around is that Eric devs adopt our code to fill their
needs, since it's MIT/BSD licensed (unfortunately we can't do the same
because theirs is GPLed). For example, I developed a little library to
do import completions here:

http://code.google.com/p/spyderlib/source/browse/spyderlib/utils/module_completion.py

Cheers,
Carlos

El 13/10/11 03:39, janwillem escribió:

[Janwillem van Dijk]

Thanks, yes that is how I came across Spyder, somewhere in a numpy or
scipy discussion. However, as a lazy scientist I like to find my input
data with gui and alter/fine-tune parameters alter options etc. with
spin-, list- and check-boxes. The convenience of having the inputs and
the outputs in a grid and a mathplot figure together in an organised
window adds to that. That's where Qt4/PyQt4 comes in. Not having to
bother about the inner workings of the sender-signal-slot system would
leave more time for numerical math or for just being idle. Possibly I
should invest some more time in reading QtDesigner and PyQt4
documentation, understanding something can save a lot of time too :-).
Cheers, Janwillem

On 10/21/2011 04:02 AM, Carlos C�rdoba wrote:
> Hi Janwillem,
>
> As you probably know, Spyder is more oriented to do scientific
> computations than GUI programming. That doesn't mean your request is
> valid, quite the contrary, it'd be quite handy to have in Spyder. It
> means someone have to stand up and implement it.
>
> The other way around is that Eric devs adopt our code to fill their
> needs, since it's MIT/BSD licensed (unfortunately we can't do the same
> because theirs is GPLed). For example, I developed a little library to
> do import completions here:
>
> http://code.google.com/p/spyderlib/source/browse/spyderlib/utils/module_completion.py
>
>
> Cheers,
> Carlos
>

> El 13/10/11 03:39, janwillem escribi�:

[Carlos Córdoba]

Ok, I understand your position better now. There is people using Spyder
as a normal IDE to do all kinds of thing. I thought you were one of them :)

Enthought developed a library to create fancy "scientific" GUIs, i.e.
with plots which can be connected to input data and so on. It's called
TraitsUI. I never used it but perhaps it may serve you better.

Cheers,
Carlos

El 21/10/11 02:47, Janwillem van Dijk escribió:

> Thanks, yes that is how I came across Spyder, somewhere in a numpy or
> scipy discussion. However, as a lazy scientist I like to find my input
> data with gui and alter/fine-tune parameters alter options etc. with
> spin-, list- and check-boxes. The convenience of having the inputs and
> the outputs in a grid and a mathplot figure together in an organised
> window adds to that. That's where Qt4/PyQt4 comes in. Not having to
> bother about the inner workings of the sender-signal-slot system would
> leave more time for numerical math or for just being idle. Possibly I
> should invest some more time in reading QtDesigner and PyQt4
> documentation, understanding something can save a lot of time too :-).
> Cheers, Janwillem
>

> On 10/21/2011 04:02 AM, Carlos Córdoba wrote:
>> Hi Janwillem,
>>
>> As you probably know, Spyder is more oriented to do scientific
>> computations than GUI programming. That doesn't mean your request is
>> valid, quite the contrary, it'd be quite handy to have in Spyder. It
>> means someone have to stand up and implement it.
>>
>> The other way around is that Eric devs adopt our code to fill their
>> needs, since it's MIT/BSD licensed (unfortunately we can't do the
>> same because theirs is GPLed). For example, I developed a little
>> library to do import completions here:
>>
>> http://code.google.com/p/spyderlib/source/browse/spyderlib/utils/module_completion.py
>>
>>
>> Cheers,
>> Carlos
>>

>> El 13/10/11 03:39, janwillem escribió:

[pmav99]

http://code.enthought.com/projects/traits/docs/html/tutorials/traits_ui_scientific_app.html

A simple traits tutorial

[Steve]

I would also suggest Pierre's guidata/guiqwt libraries. I had been
hacking away in QtDesigner for weeks on an image manipulation GUI. I
had been coding my own routines for colormaps, image statistics, etc.
Guiqwt has these functions and more built in. I had a far more
advanced program coded in two days. Guidata provides the same
functionality as TraitsUI, but was much more intuitive/pythonic to me.
Guiqwt could be compared to Chacho, but when you add features like the
contrast tool, the image toolbars, and the cross-section tool, you may
never write a GUI any other way. I am an engineer first and a coder
second, and these libraries are a perfect fit.

[stefan]

OT, but @Steve, would you mind sharing what your doing with this image
manipulation GUI? I might have similar problems since I'm trying to
come up with a medical imaging processing toolkit that easily
integrates with our workflow...
Thanks, Stefan

[Steve]

Sure, I work with Optics, and we were evaluating a custom
multispectral camera, which of course came with very little in the way
of an interface, so I made one for it. The tools allowed me to easily
display all 16 images with the ability to autoscale based on the image
histogram, apply colormaps individually, and select a region of
interest for statistical analysis. The image stats tool returns its
bounds, so I could easily pull the same statistics across all images,
for bandwidth comparison. Check out the Sift tool demo that comes
with guiqwt. I am planning to use it as the basis for our own custom
tool for visible/IR image analysis. HTH.

[Janwillem van Dijk]

Hello Carlos, Steve, Stefan, Pierre,
Tanks for your contributions to my initial point. I browsed through some
TraitsUI docs (the traits_ui_scientific_app.html very well written by
the way) and thought this tip is among the best I ever got; I will try
that. But now the guidata/guiqwt looks also very good. I have no idea
why these did not come up when I looked for something better than the
Boa-constructor where I started with a few years ago because of its
simalarity to Delphi.
By the way one of my small things is incidentally also a camera control
but triggered by the computer using libgphoto2 of which the python
bindings (libgphoto2-python) seem to be broken with recent python
updates so have to re-do it and the above will help I guess.
Janwillem