[RQt-devel] scale of view for some primitive doesn't work

Michael Lawrence lawrence.michael at gene.com
Fri Nov 12 23:42:08 CET 2010 

On Fri, Nov 12, 2010 at 2:13 PM, Tengfei Yin <yintengfei at gmail.com> wrote:

>
>
> On Fri, Nov 12, 2010 at 3:27 PM, Michael Lawrence <
> lawrence.michael at gene.com> wrote:
>
>>
>>
>> On Fri, Nov 12, 2010 at 8:59 AM, Tengfei Yin <yintengfei at gmail.com>wrote:
>>
>>>
>>>
>>> On Fri, Nov 12, 2010 at 7:09 AM, Michael Lawrence <
>>> lawrence.michael at gene.com> wrote:
>>>
>>>>
>>>>
>>>> On Thu, Nov 11, 2010 at 9:35 PM, Tengfei Yin <yintengfei at gmail.com>wrote:
>>>>
>>>>>
>>>>>
>>>>> On Thu, Nov 11, 2010 at 10:27 PM, Michael Lawrence <
>>>>> lawrence.michael at gene.com> wrote:
>>>>>
>>>>>>
>>>>>>
>>>>>> On Thu, Nov 11, 2010 at 7:32 PM, Tengfei Yin <yintengfei at gmail.com>wrote:
>>>>>>
>>>>>>> So even when you use view$rotate(), the text direction won't
>>>>>>> change..is this also what we expected to see?
>>>>>>>
>>>>>>>
>>>>>> Yes. I can see how that may not always be desirable, but it would be
>>>>>> tough to work around it. That is, I cannot tease out the rotation from the
>>>>>> rest of the transformation. In the case of a word cloud, I think it would be
>>>>>> desirable, i.e., one is thinking of the text as glyphs that are independent
>>>>>> of the scales.
>>>>>>
>>>>>>
>>>>>>> and by the way, do you know any way to change to text size, the use
>>>>>>> case here is sth like "tag cloud", but I guess I can draw text as path to
>>>>>>> solve this problem.
>>>>>>>
>>>>>>>
>>>>>> Just change the font.. do you want some sort of vectorized "cex"
>>>>>> argument to qdrawText? Might not be that hard to add.
>>>>>>
>>>>>
>>>>> Yes, if you don't mind spend some time on it, it will be really nice to
>>>>> have a control over the text size, e.g. when I plot a sequence data, like
>>>>> ACTGGGC....I can resize one single character  to address the mis-alignment.
>>>>>
>>>>
>>>> You mean like a sequence logo plot, based on the counts of each base?
>>>> Sweet idea. The feature is committed, have fun.
>>>>
>>>
>>> Yes, I plan to do sth like this~the problem of this kind of graphic is
>>> that, no problem showing as a single plot, but when showing along a
>>> shot-read alighment, due to resale of all the text, even a single size
>>> change of one base will make it hard to align a single sequence with other
>>> "normal" text. I will work around this to see if I have some problem.
>>>
>>>
>> Text glyphs could be scaled differently in the horizontal and vertical.
>> Would this help with this issue? That is, you could stretch the letters
>> vertically.
>>
>
> Do you mean sth that is already in  qtpaint?
>

No it's not there yet but it would be easy to support it. Right now 'cex'
causes a symmetric scale; it does not need to be that way. I will add it.



> it's not in qdrawText(), and the qglyphText has no such a parameter, do you
> suggest me to create text as path and use qdrawPath to plot those stuff??
> and by the way, in the qt document,
> void QFont::setStretch ( int factor )
> only set width of character. How can I specify whether to stretch it
> vertically or horizontally? Or do you mean doing sth like transformation?
>
>
>>  Btw: any idea where is globalPos(), I try to use tooltip associated with
>>> locate() function to do this：when mouse over on exon, show the id, but I can
>>> only find pos, based on the documentation, it looks like I need to get
>>> globalPos..of course I have alternative workaround instead of using
>>> QToolTip.
>>>
>>>
>> I'm not sure I follow. Are you talking about globalPos() on the event
>> object? That is global screen position, not even relative to the window, so
>> you would need to be careful to adjust it.
>>
>
> Yes, I can not find globalPos() on event object?
>

What about screenPos()?


> and I think you are right, I guess I cannot use it.  I am thinking that
> maybe qdrawText will be faster than QToolTip, maybe tooltip is not used in
> this purpose to show the exon names, maybe it's designed to show widget name
> and function description.
>
>
Well the tooltip window has various features, like hypertext. The user could
click on something and launch a web browser to some external database, for
example.


> Thanks
>
>>
>> Thanks
>>>
>>>>
>>>>
>>>>>
>>>>>> Michael
>>>>>>
>>>>>>
>>>>>>>
>>>>>>> On Thu, Nov 11, 2010 at 6:14 PM, Tengfei Yin <yintengfei at gmail.com>wrote:
>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> On Thu, Nov 11, 2010 at 5:39 PM, Michael Lawrence <
>>>>>>>> lawrence.michael at gene.com> wrote:
>>>>>>>>
>>>>>>>>> This is by design. If this didn't work that way, then when the user
>>>>>>>>> zoomed into a chromosome, for example, the text labels would get really
>>>>>>>>> stretched out. Glyphs are the same way (they're essentially fancy
>>>>>>>>> characters). Zooming into a scatterplot should not make the glyphs larger.
>>>>>>>>> Circles (after much debate inside my head) behave the same as glyphs. This
>>>>>>>>> is because their radius, like in a bubble plot, is derived from something
>>>>>>>>> other than the X and Y axes.
>>>>>>>>>
>>>>>>>>> Shapes however, like rectangles, polygons and paths, are zoomed.
>>>>>>>>> You expect a mosaic plot rectangle or a map polygon to scale with the plot.
>>>>>>>>>
>>>>>>>>>
>>>>>>>> This make sense to me, in most cases, i guess people won't expect
>>>>>>>> the circle or text to be  zoomed in.
>>>>>>>>
>>>>>>>> Are you sure you really need the text to scale with the zoom?
>>>>>>>>> Deepayan and I decided a long time ago that zooming does not behave like a
>>>>>>>>> "magnifying glass." It rescales the plot.
>>>>>>>>>
>>>>>>>>>
>>>>>>>> No, you don't have to change that, it's convenient so far, I am just
>>>>>>>> curious why this doesn't work for text and circle, because in other rare
>>>>>>>> case, I will draw very small text as labels of axis for one circular track,
>>>>>>>> or if I draw points around the circular track, I am still expecting the
>>>>>>>> points will be zoomed in when I rescale the view. I guess I am confused the
>>>>>>>> "rescale of the view" with the one you defined with Deepayan, but it make
>>>>>>>> sense to me.
>>>>>>>>
>>>>>>>> There are some ways to get around this if you really want to:
>>>>>>>>> - Draw the text as a path.
>>>>>>>>> - Use the "render" method on the scene to render to a QPicture,
>>>>>>>>> rescale that and display it. This is how I imagined someone making a
>>>>>>>>> magnifying glass. QPicture is a vectorized representation, so no nasty
>>>>>>>>> scaling artifacts are introduced.
>>>>>>>>>
>>>>>>>>
>>>>>>>> Thanks, I think I need to do it in this way, if I really need them
>>>>>>>> to be rescaled.  I never try QPicture before, I will take a look at it.
>>>>>>>>
>>>>>>>>
>>>>>>>>> Michael
>>>>>>>>>
>>>>>>>>> On Thu, Nov 11, 2010 at 3:24 PM, Tengfei Yin <yintengfei at gmail.com
>>>>>>>>> > wrote:
>>>>>>>>>
>>>>>>>>>> Hi
>>>>>>>>>>
>>>>>>>>>> Here comes another question with text, please see the attached
>>>>>>>>>>  code below, only rect could be resized,  when I call v$scale()
>>>>>>>>>>
>>>>>>>>>> text, circle doesn't change. it happens to visnab, the circular
>>>>>>>>>> view, when I scale the view, every thing resize but the text, it's still in
>>>>>>>>>> the same size.
>>>>>>>>>>
>>>>>>>>>> library(qtpaint)
>>>>>>>>>> s <- qscene()
>>>>>>>>>> test_layer <- qlayer(s,function(layer,painter){
>>>>>>>>>>   qdrawRect(painter,30,30,70,70,stroke="blue")
>>>>>>>>>>   qdrawCircle(painter,50,50,20,stroke="red")
>>>>>>>>>>   qdrawText(painter,"TEST",50,50)
>>>>>>>>>> },limits=qrect(0,0,100,100))
>>>>>>>>>> v <- qplotView(s)
>>>>>>>>>> v$show()
>>>>>>>>>>
>>>>>>>>>> v$scale(1.5,1.5)
>>>>>>>>>>
>>>>>>>>>> Thanks
>>>>>>>>>>
>>>>>>>>>> Tengfei
>>>>>>>>>> --
>>>>>>>>>> Tengfei Yin
>>>>>>>>>> MCDB PhD student
>>>>>>>>>> 1620 Howe Hall, 2274,
>>>>>>>>>> Iowa State University
>>>>>>>>>> Ames, IA,50011-2274
>>>>>>>>>> Homepage: www.tengfei.name
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> _______________________________________________
>>>>>>>>>> Qtinterfaces-devel mailing list
>>>>>>>>>> Qtinterfaces-devel at lists.r-forge.r-project.org
>>>>>>>>>>
>>>>>>>>>> https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/qtinterfaces-devel
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> --
>>>>>>>> Tengfei Yin
>>>>>>>> MCDB PhD student
>>>>>>>> 1620 Howe Hall, 2274,
>>>>>>>> Iowa State University
>>>>>>>> Ames, IA,50011-2274
>>>>>>>> Homepage: www.tengfei.name
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> --
>>>>>>> Tengfei Yin
>>>>>>> MCDB PhD student
>>>>>>> 1620 Howe Hall, 2274,
>>>>>>> Iowa State University
>>>>>>> Ames, IA,50011-2274
>>>>>>> Homepage: www.tengfei.name
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> Qtinterfaces-devel mailing list
>>>>>>> Qtinterfaces-devel at lists.r-forge.r-project.org
>>>>>>>
>>>>>>> https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/qtinterfaces-devel
>>>>>>>
>>>>>>>
>>>>>>
>>>>>
>>>>>
>>>>> --
>>>>> Tengfei Yin
>>>>> MCDB PhD student
>>>>> 1620 Howe Hall, 2274,
>>>>> Iowa State University
>>>>> Ames, IA,50011-2274
>>>>> Homepage: www.tengfei.name
>>>>>
>>>>>
>>>>>
>>>>
>>>
>>>
>>> --
>>> Tengfei Yin
>>> MCDB PhD student
>>> 1620 Howe Hall, 2274,
>>> Iowa State University
>>> Ames, IA,50011-2274
>>> Homepage: www.tengfei.name
>>>
>>>
>>>
>>
>
>
> --
> Tengfei Yin
> MCDB PhD student
> 1620 Howe Hall, 2274,
> Iowa State University
> Ames, IA,50011-2274
> Homepage: www.tengfei.name
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.r-forge.r-project.org/pipermail/qtinterfaces-devel/attachments/20101112/1711f1fe/attachment.htm>

More information about the Qtinterfaces-devel mailing list