cll35 / mplh5canvas Goto Github PK

What steps will reproduce the problem?
1. Run examples/monitor_plot.py
2. Go to local webpage in webbrowser (http://xxx.xxx.xxx.xxx:9000)

What is the expected output? What do you see instead?
I expected to see the list of sensor, as defined in the monitor_plot.py code, 
together with the plot canvas. Instead, I see the default page (thumbs at the 
top, canvas below, etc.). The plot has the correct title ("No sensors 
selected"), but is empty otherwise.

What version of the product are you using? On what operating system?
SVN r37, tested with with Safari 5.1 and Chrome 21 on Mac OS X.

What steps will reproduce the problem?
1. Run anyone of the demos (e.g. size_demo.py)
2. Use Firefox (4 or earlier versions) to view the page

The main figure frame shows up, but no plot is seen.

The backend works with Safari and Chrome.

This is with

mplh5canvas-trunk-py2.6.egg
Python 2.6.4 (r264:75821M, Oct 27 2009, 19:48:32)
matplotlib.__version__ = 1.0.1

Nice work and very usefull

It seems that the maximum number of graphs is 4, and it is impossible to select 
the desired graphs amont those in the thumbnail strip (the thumbnail strip 
display more than 4 graphe)

Do you confirm, or it is my configuration (Chrome last version, python 2.6)

Sincerely yours,

Loïc

Hi 

This is not a bug but I did not find a mailing list for suggestions or general 
questions.

I have been working with matplotlib for quite some time. I have done some 
modifications to the tk and gtk backends mainly to include tabbed figure 
manager and style management (colors, line thickness).
https://github.com/fariza/MPL-Experimental-Backend

I have a small project where it would be perfect to use the mplh5canvas for 
plotting automated test results on a webpage and be able to interact with those 
graphics.

The only problem is the websocket limitation (scarce suport from popular 
browsers).

The question is: How difficult would be to provide an alternative to websockets 
using ajax?
I have been looking at your source code but I got lost pretty soon.
If you could point me in the right direction I can give it a try.

Thanks and sorry for the not a bug issue.
Federico

If I run the subplot_demo.py example, and try and change the view in the top 
subplot, the view changes in the bottom panel instead, and vice-versa. Attached 
is a movie of the bug.

I am using mplh5canvas svn r10, matplotlib svn r8544, and numpy svn r8480, as 
well as Safari Version 5.0 (6533.16).

python interactive_plot.py

Bring the browser page and click the plot or next button doesn't register 
anything on screen.

This is Google Chrome 10.0.648.127

python
Python 2.7 (r27:82500) 

matplotlib.__version__
 '1.1.0svn'

uname -a
Linux 2.6.35.11-83.fc14.x86_64 #1 SMP Mon Feb 7 07:06:44 UTC 2011 x86_64 
GNU/Linux

* What steps will reproduce the problem?
1. easy_install mplh5canvas

* What is the expected output? What do you see instead?
$ easy_install mplh5canvas
Searching for mplh5canvas
Reading http://pypi.python.org/simple/mplh5canvas/
Reading http://code.google.com/p/mplh5canvas/
Best match: mplh5canvas 0.5
Downloading 
http://pypi.python.org/packages/source/m/mplh5canvas/mplh5canvas-0.5.tar.gz#md5=
de1a64e81a635d589e85dc33a229e4f7
Processing mplh5canvas-0.5.tar.gz
Running mplh5canvas-0.5/setup.py -q bdist_egg --dist-dir 
/tmp/easy_install-W4IGPw/mplh5canvas-0.5/egg-dist-tmp-_p3XCG
error: SandboxViolation: os.open('/home/***/.matplotlib/tmpWUu1ke', 131266, 
384) {}

The package setup script has attempted to modify files on your system
that are not within the EasyInstall build area, and has been aborted.

This package cannot be safely installed by EasyInstall, and may not
support alternate installation locations even if you run its setup
script by hand.  Please inform the package's author and the EasyInstall
maintainers to find out if a fix or workaround is available.

* What version of the product are you using? On what operating system?
Linux ubuntu 2.6.38-10-generic #46-Ubuntu SMP Tue Jun 28 15:05:41 UTC 2011 i686 
i686 i386 GNU/Linux

* Please provide any additional information below.

The imshow_plot.py fails when run:

valinor:examples tom$ python imshow_plot.py 
/Users/tom/Library/Python/2.6/site-packages/mplh5canvas-trunk-py2.6.egg/mplh5can
vas/handshake/handshake.py:41: DeprecationWarning: the md5 module is 
deprecated; use hashlib instead
  from md5 import md5
Web server active. Browse to http://131.142.156.6:8080 to view plots.
Traceback (most recent call last):
  File "imshow_plot.py", line 17, in <module>
    show(block=False)
  File "/Users/tom/Library/Python/2.6/site-packages/mplh5canvas-trunk-py2.6.egg/mplh5canvas/backend_h5canvas.py", line 518, in show
    Gcf.get_active().canvas.draw()
  File "/Users/tom/Library/Python/2.6/site-packages/mplh5canvas-trunk-py2.6.egg/mplh5canvas/backend_h5canvas.py", line 727, in draw
    self.figure.draw(renderer)
  File "/Users/tom/Library/Python/2.6/site-packages/matplotlib/artist.py", line 55, in draw_wrapper
    draw(artist, renderer, *args, **kwargs)
  File "/Users/tom/Library/Python/2.6/site-packages/matplotlib/figure.py", line 798, in draw
    func(*args)
  File "/Users/tom/Library/Python/2.6/site-packages/matplotlib/artist.py", line 55, in draw_wrapper
    draw(artist, renderer, *args, **kwargs)
  File "/Users/tom/Library/Python/2.6/site-packages/matplotlib/axes.py", line 1934, in draw
    a.draw(renderer)
  File "/Users/tom/Library/Python/2.6/site-packages/matplotlib/artist.py", line 55, in draw_wrapper
    draw(artist, renderer, *args, **kwargs)
  File "/Users/tom/Library/Python/2.6/site-packages/matplotlib/image.py", line 352, in draw
    renderer.draw_image(gc, l, b, im)
  File "/Users/tom/Library/Python/2.6/site-packages/mplh5canvas-trunk-py2.6.egg/mplh5canvas/backend_h5canvas.py", line 362, in draw_image
    h,w = im.get_size_out()
AttributeError: 'numpy.float64' object has no attribute 'get_size_out'

I am using mplh5canvas svn r10, matplotlib svn r8544, and numpy svn r8480

RendererH5Canvas.points_to_pixels got passed a string instead of a number by a 
recent version of matplotlib. casting it explicitly seemed to fix the issue.

Running python single_plot.py, then accessing the plot page in Chromium, I get 
an exception that readline was given 2 arguments but expected just 1.

The file interface specifies that readline() should accept an optional 'size' 
argument, which is actually used by BaseHTTPServer.  

I'm using python 2.7, in a fresh install of Kubuntu Natty.

The fix is just to modify memorizingfile.py:68 to add size: 

    def readline(self,size=None):
        """Override file.readline and memorize the line read."""

        line = self._file.readline(size=size)
        if line and len(self._memorized_lines) < self._max_memorized_lines:
            self._memorized_lines.append(line)
        return line

What steps will reproduce the problem?
1. Run one of the example plots (eg anim_plot.py)
2. Embed the produced plot in an iframe, eg. with:
<html>
    <head>
     Embedded
    </head>
    <body>  
                An mplh5canvas graph embedded in an iframe!
        <iframe src="http://127.0.0.1:9000" width=100% height=100%></iframe>
    </body>
</html>

3. Navigate to the url of the containing page (127.0.0.1:80 in this case)

What is the expected output? What do you see instead?
It should graph immediately.

What version of the product are you using? On what operating system?
mplh5canvas v0.6. Running ubuntu 11.10 with apache as the webserver

Please provide any additional information below.
It would be great to have a way to display the graphs in a page with other 
elements. Perhaps I'm missing something obvious?

matplotlib 1.0.0 changes the API for backends' draw_image functions. A patch is 
attached.

Running an animated plot over a period of time (5-10 minutes) crashes the 
Chrome browser.

What version of the product are you using? On what operating system?
0.6. Happens on Windows and Linux

None of the example scripts (for instance, single_plot.py) work for me unless I 
use version 0.7.6 or older of mod_pywebsocket. With version 0.7.7 or later of 
mod_pywebsocket, I get a traceback that looks like this:

[2013-05-14 22:01:43,653] [WARNING] mplh5canvas.simple_server: 
'_StandaloneRequest' object has no attribute 'protocol'
[2013-05-14 22:01:43,654] [WARNING] mplh5canvas.simple_server: Traceback (most 
recent call last):
  File "/opt/local/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/mplh5canvas/simple_server.py", line 421, in parse_request
    strict=self._options.strict)
  File "/opt/local/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/mod_pywebsocket/handshake/__init__.py", line 90, in do_handshake
    handshaker.do_handshake()
  File "/opt/local/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/mod_pywebsocket/handshake/hybi.py", line 137, in do_handshake
    check_request_line(self._request)
  File "/opt/local/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/mod_pywebsocket/handshake/_base.py", line 178, in check_request_line
    if request.protocol != 'HTTP/1.1':
AttributeError: '_StandaloneRequest' object has no attribute 'protocol'

It looks like mplh5canvas.simple_server is behind in some API changes in 
mod_pywebsocket.

This is on Mac OS 10.8.3 (Mountain Lion) and Python 2.7.3 installed from 
MacPorts. In fact, I came across this bug while preparing MacPorts packages for 
mplh5canvas and its dependencies mod_pywebsocket and netifaces.

What steps will reproduce the problem?
1. Create a button and an event listener, called when button is clicked
2. click the button
3. check if your event-listener gets called

What is the expected output? What do you see instead?
button_click_event should be triggered when button is clicked, but is only 
triggered after button is clicked and then another click is done (on button or 
something else).

What version of the product are you using? On what operating system?
Current svn version (rev 32)

Please provide any additional information below.
The fix is trivial:
in handle_click(...) in backend_h5canvas.py the button_press and button_release 
events get called in the wrong order (release, press), so only after a second 
click a (release),press->release,(press) cycle triggers the "clicked" event. If 
I change the order and call first button_press_event and then 
button_release_event the events are processed as expected: A button click 
triggers my event directly.

So the handle_click function should be changed to:

def handle_click(self, x, y, button):
  self.button_press_event(float(x), float(y), int(button))
  self.button_release_event(float(x),float(y),int(button))


Thanks for this very useful piece of software!

  Christoph

What steps will reproduce the problem?
1. In an interactive interpreter, import matplotlib, set mplh5canvas backend
2. import pyplot
3. Go to the alleged management interface address

What is the expected output? What do you see instead?
I expect to see some HTML page.  Instead I see the source of the interface 
page.  It's not rendered as HTML.

What version of the product are you using? On what operating system?
This is on Firefox 17 on Win7.

My more general question here is: is this just because the html5 backend 
doesn't work on Firefox, or is it something else?  Are there any plans to 
support Firefox?  Is this meant to be an HTML5 backend, or a Chrome-web-browser 
backend?  The backend looks like a useful tool, but not supporting Firefox is a 
dealbreaker for me.

Not a big deal, but the extension of the "mplh5canvas.r7.24jun10.tar.bz2" in 
the download area is not corresponding to the actual compression.  

$ file -i mplh5canvas.r7.24jun10.tar.bz2 
mplh5canvas.r7.24jun10.tar.bz2: application/x-gzip; charset=binary

What steps will reproduce the problem?
$> python setup.py build 
The HTML 5 Canvas Backend requires matplotlib 0.99.1.1 or newer. Your version 
(1.0.0) appears older than this. Unable to continue...

What is the expected output? What do you see instead?
The build should continue, instead of being aborted.

What version of the product are you using? On what operating system?
SVN r7 (june 24, 2010); Mac OS X

Please provide any additional information below.
Shouldn't version checking be replaced by feature detection? That seems to be 
the general recommendation for Python module dependencies.
In addition, mpl 0.99.0 would pass the check, but isn't in agreement with 
apparent requirement of > 0.99.1.1