Meeting...?

Development matplotlib-devel
1

It's probably a good time to schedule another matplotlib Google Hangout.

Is this Thursday at 1500 UTC (10 am ET) too short notice for the usual candidates?

I know there was discussion of getting Michiel de Hoon on today (which I just saw, unfortunately). Is there another time in the future that works for you, Michiel?

Mike

···

--
                    _

\/|o _|_ _. _ | | \.__ __|__|_|_ _ _ ._ _
>>(_| |(_|(/_| |_/|(_)(/_|_ |_|_)(_)(_)| | |

2

I can do Thursday.

Tom

···

On Mon, Jan 13, 2014 at 10:36 AM, Michael Droettboom <mdroe@...31...> wrote:

It's probably a good time to schedule another matplotlib Google Hangout.

Is this Thursday at 1500 UTC (10 am ET) too short notice for the usual
candidates?

I know there was discussion of getting Michiel de Hoon on today (which I
just saw, unfortunately). Is there another time in the future that
works for you, Michiel?

Mike

--
                    _
>\/|o _|_ _. _ | | \.__ __|__|_|_ _ _ ._ _
> >>(_| |(_|(/_| |_/|(_)(/_|_ |_|_)(_)(_)| | |

http://www.droettboom.com

------------------------------------------------------------------------------
CenturyLink Cloud: The Leader in Enterprise Cloud Services.
Learn Why More Businesses Are Choosing CenturyLink Cloud For
Critical Workloads, Development Environments & Everything In Between.
Get a Quote or Start a Free Trial Today.
http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk
_______________________________________________
Matplotlib-devel mailing list
Matplotlib-devel@lists.sourceforge.net
matplotlib-devel List Signup and Options

--
Thomas Caswell
tcaswell@...149...

3

I can join this Thursday if we start with the discussion on timers.
If we can start 1 hour earlier (14:00 UTC, 9 am ET, 23:00 in Japan) that would be even better.
-Michiel.

···

--------------------------------------------
On Mon, 1/13/14, Michael Droettboom <mdroe@...31...> wrote:

Subject: [matplotlib-devel] Meeting...?
To: "matplotlib-devel@lists.sourceforge.net" <matplotlib-devel@...958...eforge.net>
Date: Monday, January 13, 2014, 11:36 AM

It's probably a good time to schedule
another matplotlib Google Hangout.

Is this Thursday at 1500 UTC (10 am ET) too short notice for
the usual
candidates?

I know there was discussion of getting Michiel de Hoon on
today (which I
just saw, unfortunately). Is there another time in the
future that
works for you, Michiel?

Mike

--

 \_

>\/|o _|_ _. _ | | \.__ __|__|_|_ _
_ ._ _
> >>(_| |(_|(/_| |_/|(_)(/_|_ |_|_)(_)(_)| | |

http://www.droettboom.com

------------------------------------------------------------------------------
CenturyLink Cloud: The Leader in Enterprise Cloud Services.
Learn Why More Businesses Are Choosing CenturyLink Cloud
For
Critical Workloads, Development Environments &
Everything In Between.
Get a Quote or Start a Free Trial Today.
http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk
_______________________________________________
Matplotlib-devel mailing list
Matplotlib-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/matplotlib-devel

4

I'm fine with starting the meeting an hour early. How about others?

Mike

···

On 01/14/2014 04:57 AM, Michiel de Hoon wrote:

I can join this Thursday if we start with the discussion on timers.
If we can start 1 hour earlier (14:00 UTC, 9 am ET, 23:00 in Japan) that would be even better.
-Michiel.

--------------------------------------------
On Mon, 1/13/14, Michael Droettboom <mdroe@...31...> wrote:

  Subject: [matplotlib-devel] Meeting...?
  To: "matplotlib-devel@lists.sourceforge.net" <matplotlib-devel@lists.sourceforge.net>
  Date: Monday, January 13, 2014, 11:36 AM
    It's probably a good time to schedule
  another matplotlib Google Hangout.
    Is this Thursday at 1500 UTC (10 am ET) too short notice for
  the usual
  candidates?
    I know there was discussion of getting Michiel de Hoon on
  today (which I
  just saw, unfortunately). Is there another time in the
  future that
  works for you, Michiel?
    Mike
    --
                       _
  >\/|o _|_ _. _ | | \.__ __|__|_|_ _
  _ ._ _
  > >>(_| |(_|(/_| |_/|(_)(/_|_ |_|_)(_)(_)| | |
    http://www.droettboom.com
  
  CenturyLink Cloud: The Leader in Enterprise Cloud Services.
  Learn Why More Businesses Are Choosing CenturyLink Cloud
  For
  Critical Workloads, Development Environments &
  Everything In Between.
  Get a Quote or Start a Free Trial Today.
  http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk
  _______________________________________________
  Matplotlib-devel mailing list
  Matplotlib-devel@lists.sourceforge.net
  matplotlib-devel List Signup and Options
  
--
                    _
\/|o _|_ _. _ | | \.__ __|__|_|_ _ _ ._ _
>>(_| |(_|(/_| |_/|(_)(/_|_ |_|_)(_)(_)| | |

5

That would actually work a little bit better for me… I just have to remember to get into work a little bit earlier.

···

Ben

On Tue, Jan 14, 2014 at 11:36 AM, Michael Droettboom <mdroe@…31…> wrote:

I’m fine with starting the meeting an hour early. How about others?

Mike

On 01/14/2014 04:57 AM, Michiel de Hoon wrote:

I can join this Thursday if we start with the discussion on timers.

If we can start 1 hour earlier (14:00 UTC, 9 am ET, 23:00 in Japan) that would be even better.

-Michiel.

On Mon, 1/13/14, Michael Droettboom <mdroe@…31…> wrote:

Subject: [matplotlib-devel] Meeting…?

To: “matplotlib-devel@lists.sourceforge.net” <matplotlib-devel@…1148…t>

Date: Monday, January 13, 2014, 11:36 AM

It’s probably a good time to schedule

another matplotlib Google Hangout.

Is this Thursday at 1500 UTC (10 am ET) too short notice for

the usual

candidates?

I know there was discussion of getting Michiel de Hoon on

today (which I

just saw, unfortunately). Is there another time in the

future that

works for you, Michiel?

Mike

–

  _

|/|o | . _ | | ._ ||| _

_ ._ _

| ||(| |(|(/| |/|()(/|_ ||)()()| | |

http://www.droettboom.com

CenturyLink Cloud: The Leader in Enterprise Cloud Services.

Learn Why More Businesses Are Choosing CenturyLink Cloud

For

Critical Workloads, Development Environments &

Everything In Between.

Get a Quote or Start a Free Trial Today.

http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk

Matplotlib-devel mailing list

Matplotlib-devel@lists.sourceforge.net

https://lists.sourceforge.net/lists/listinfo/matplotlib-devel

–

                _

|/|o | . _ | | ._ ||| _ _ ._ _

| ||(| |(|(/| |/|()(/|_ ||)()()| | |

http://www.droettboom.com

CenturyLink Cloud: The Leader in Enterprise Cloud Services.

Learn Why More Businesses Are Choosing CenturyLink Cloud For

Critical Workloads, Development Environments & Everything In Between.

Get a Quote or Start a Free Trial Today.

http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk

Matplotlib-devel mailing list

Matplotlib-devel@lists.sourceforge.net

https://lists.sourceforge.net/lists/listinfo/matplotlib-devel

6

I’ll probably not be able to swing 6am on the west coast, but other folks are more important for this call, I think :slight_smile:
Jake

···

On Tue, Jan 14, 2014 at 8:51 AM, Benjamin Root <ben.root@…553…> wrote:

That would actually work a little bit better for me… I just have to remember to get into work a little bit earlier.

CenturyLink Cloud: The Leader in Enterprise Cloud Services.

Learn Why More Businesses Are Choosing CenturyLink Cloud For

Critical Workloads, Development Environments & Everything In Between.

Get a Quote or Start a Free Trial Today.

http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk

Matplotlib-devel mailing list

Matplotlib-devel@lists.sourceforge.net

https://lists.sourceforge.net/lists/listinfo/matplotlib-devel

Ben

On Tue, Jan 14, 2014 at 11:36 AM, Michael Droettboom <mdroe@…31…> wrote:

I’m fine with starting the meeting an hour early. How about others?

Mike

On 01/14/2014 04:57 AM, Michiel de Hoon wrote:

I can join this Thursday if we start with the discussion on timers.

If we can start 1 hour earlier (14:00 UTC, 9 am ET, 23:00 in Japan) that would be even better.

-Michiel.

On Mon, 1/13/14, Michael Droettboom <mdroe@…31…> wrote:

Subject: [matplotlib-devel] Meeting…?

To: “matplotlib-devel@lists.sourceforge.net” matplotlib-devel@lists.sourceforge.net

Date: Monday, January 13, 2014, 11:36 AM

It’s probably a good time to schedule

another matplotlib Google Hangout.

Is this Thursday at 1500 UTC (10 am ET) too short notice for

the usual

candidates?

I know there was discussion of getting Michiel de Hoon on

today (which I

just saw, unfortunately). Is there another time in the

future that

works for you, Michiel?

Mike

–

  _

|/|o | . _ | | ._ ||| _

_ ._ _

| ||(| |(|(/| |/|()(/|_ ||)()()| | |

http://www.droettboom.com

CenturyLink Cloud: The Leader in Enterprise Cloud Services.

Learn Why More Businesses Are Choosing CenturyLink Cloud

For

Critical Workloads, Development Environments &

Everything In Between.

Get a Quote or Start a Free Trial Today.

http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk

Matplotlib-devel mailing list

Matplotlib-devel@lists.sourceforge.net

https://lists.sourceforge.net/lists/listinfo/matplotlib-devel

–

                _

|/|o | . _ | | ._ ||| _ _ ._ _

| ||(| |(|(/| |/|()(/|_ ||)()()| | |

http://www.droettboom.com

CenturyLink Cloud: The Leader in Enterprise Cloud Services.

Learn Why More Businesses Are Choosing CenturyLink Cloud For

Critical Workloads, Development Environments & Everything In Between.

Get a Quote or Start a Free Trial Today.

http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk

Matplotlib-devel mailing list

Matplotlib-devel@lists.sourceforge.net

https://lists.sourceforge.net/lists/listinfo/matplotlib-devel

7

Jake: I'd definitely like to get you into one of these calls at some point. If you're able to pop in late, that would still be great -- or we can save that for another date. Trying to get Japan, three NA timezones and the UK all together is challenging :wink:

In any event, with Thomas, Ben, Michiel and myself confirmed, I think that's enough to go ahead, and hopefully others who have yet to respond can join as well.

Mike

···

On 01/14/2014 11:57 AM, Jacob Vanderplas wrote:

I'll probably not be able to swing 6am on the west coast, but other folks are more important for this call, I think :slight_smile:
   Jake

On Tue, Jan 14, 2014 at 8:51 AM, Benjamin Root <ben.root@...553... > <mailto:ben.root@…553…>> wrote:

    That would actually work a little bit better for me... I just have
    to remember to get into work a little bit earlier.

    Ben

    On Tue, Jan 14, 2014 at 11:36 AM, Michael Droettboom > <mdroe@...31... <mailto:mdroe@…31…>> wrote:

        I'm fine with starting the meeting an hour early. How about
        others?

        Mike

        On 01/14/2014 04:57 AM, Michiel de Hoon wrote:
        > I can join this Thursday if we start with the discussion on
        timers.
        > If we can start 1 hour earlier (14:00 UTC, 9 am ET, 23:00 in
        Japan) that would be even better.
        > -Michiel.
        >
        > --------------------------------------------
        > On Mon, 1/13/14, Michael Droettboom <mdroe@...31... > <mailto:mdroe@…31…>> wrote:
        >
        > Subject: [matplotlib-devel] Meeting...?
        > To: "matplotlib-devel@lists.sourceforge.net
        <mailto:matplotlib-devel@lists.sourceforge.net>"
        <matplotlib-devel@lists.sourceforge.net
        <mailto:matplotlib-devel@lists.sourceforge.net>>
        > Date: Monday, January 13, 2014, 11:36 AM
        >
        > It's probably a good time to schedule
        > another matplotlib Google Hangout.
        >
        > Is this Thursday at 1500 UTC (10 am ET) too short notice for
        > the usual
        > candidates?
        >
        > I know there was discussion of getting Michiel de Hoon on
        > today (which I
        > just saw, unfortunately). Is there another time in the
        > future that
        > works for you, Michiel?
        >
        > Mike
        >
        > --
        >
        > _
        > >\/|o _|_ _. _ | | \.__ __|__|_|_ _
        > _ ._ _
        > > >>(_| |(_|(/_| |_/|(_)(/_|_ |_|_)(_)(_)| | |
        >
        > http://www.droettboom.com
        >
        ------------------------------------------------------------------------------
        > CenturyLink Cloud: The Leader in Enterprise Cloud Services.
        > Learn Why More Businesses Are Choosing CenturyLink Cloud
        > For
        > Critical Workloads, Development Environments &
        > Everything In Between.
        > Get a Quote or Start a Free Trial Today.
        >
        http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk
        > _______________________________________________
        > Matplotlib-devel mailing list
        > Matplotlib-devel@lists.sourceforge.net
        <mailto:Matplotlib-devel@lists.sourceforge.net>
        > matplotlib-devel List Signup and Options
        >

        --
                            _
        >\/|o _|_ _. _ | | \.__ __|__|_|_ _ _ ._ _
        > >>(_| |(_|(/_| |_/|(_)(/_|_ |_|_)(_)(_)| | |

        http://www.droettboom.com

        ------------------------------------------------------------------------------
        CenturyLink Cloud: The Leader in Enterprise Cloud Services.
        Learn Why More Businesses Are Choosing CenturyLink Cloud For
        Critical Workloads, Development Environments & Everything In
        Between.
        Get a Quote or Start a Free Trial Today.
        http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk
        _______________________________________________
        Matplotlib-devel mailing list
        Matplotlib-devel@lists.sourceforge.net
        <mailto:Matplotlib-devel@lists.sourceforge.net>
        matplotlib-devel List Signup and Options

    ------------------------------------------------------------------------------
    CenturyLink Cloud: The Leader in Enterprise Cloud Services.
    Learn Why More Businesses Are Choosing CenturyLink Cloud For
    Critical Workloads, Development Environments & Everything In Between.
    Get a Quote or Start a Free Trial Today.
    http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk
    _______________________________________________
    Matplotlib-devel mailing list
    Matplotlib-devel@lists.sourceforge.net
    <mailto:Matplotlib-devel@lists.sourceforge.net>
    matplotlib-devel List Signup and Options

--
                    _

\/|o _|_ _. _ | | \.__ __|__|_|_ _ _ ._ _
>>(_| |(_|(/_| |_/|(_)(/_|_ |_|_)(_)(_)| | |

8

Thanks - we’ll make it happen at some point.

Perhaps I can give the seed for a discussion: the stuff I’ve been doing with mpld3 is a lot of fun, but it’s fundamentally limited by the fact that I have to dig around the internals of the figure object to find the relevant information to construct a plot representation. I may be able to do the same thing by creating a backend, but the problem is that the draw() methods of most objects call the renderer with no reference to whether the points lie in the data space or figure space: that is, paths and points are usually specified in figure/pixel coordinates or some transformed version thereof, which makes it near impossible to construct interactive representations absent Python kernel callbacks.

What I’d love to see is some enhancement of the backend framework where there are some extra flags and information passed to the renderer: i.e. for each draw command, we need to know whether the drawn object should be linked to static figure coordinates or to dynamic axes/data coordinates.

I’ve been in touch with Cyrille Rossant from the vispy team, Chris Beaumont from the Glue team, and Matt Sundwuist from the plotly team, all of whom asked if there might be a way to use what I’ve done with mpld3 to enable matplotlib to export into their own front-end format. I didn’t start mpld3 with that sort of extensibility in mind, but I’m starting to invest some time thinking about how to design that.

With the current matplotlib package, I think there are two ways to accomplish it: one is to create a general backend-like interface based on the figure introspection that mpld3 currently uses. The artist elements in each figure contain enough information to be able to infer whether the elements should move & zoom with the axes or not. The problem is, a lot of elements (like legends, axes aspects, etc.) are not fully established until the draw() command is called, so there are a few ugly hacks required to make it happen.

The other option is to use an even uglier hack, and wrap the current backend framework with an object that somehow links back into the figure and infers from the draw_*() commands whether the path/point/marker/etc. should be drawn in static figure coordinates or in dynamic axes coordinates. I’ve started a simple prototype backend translator which has a renderer class that uses inspect back-trace the stack and accomplish this: It’s really ugly, and I’m not particularly proud about it, but I think it’s the current best way to accomplish the desired behavior.

Ugly hacks aside, I think all of this points to a general desire for a new type of backend-like hook that can export dynamic plot elements in data coordinates, and static plot elements in figure coordinates. An enhancement in that direction could pave the way for a lot of interesting interactive front-ends to matplotlib figures.

Anyway - if any of you have suggestions or responses to this, I’d love to hear them! Thanks,

Jake

···

On Tue, Jan 14, 2014 at 9:11 AM, Michael Droettboom <mdroe@…31…> wrote:

  Jake: I'd definitely like to get you

into one of these calls at some point. If you’re able to pop in
late, that would still be great – or we can save that for another
date. Trying to get Japan, three NA timezones and the UK all
together is challenging :wink:

  In any event, with Thomas, Ben, Michiel and myself confirmed, I

think that’s enough to go ahead, and hopefully others who have yet
to respond can join as well.

  Mike




  On 01/14/2014 11:57 AM, Jacob Vanderplas wrote:

    I'll probably not be able to swing 6am on the west

coast, but other folks are more important for this call, I think
:slight_smile:
Jake

                   -- _ |\/|o _|_ _. _ | | \.__ __|__|_|_ _ _ ._ _ | ||(_| |(_|(/_| |_/|(_)(/_|_ |_|_)(_)(_)| | | [http://www.droettboom.com](http://www.droettboom.com)

      On Tue, Jan 14, 2014 at 8:51 AM,

Benjamin Root <ben.root@…553…>
wrote:

          That would actually work a little bit better

for me… I just have to remember to get into work a
little bit earlier.

        CenturyLink Cloud: The Leader in Enterprise Cloud Services.

        Learn Why More Businesses Are Choosing CenturyLink Cloud For

        Critical Workloads, Development Environments &

Everything In Between.

        Get a Quote or Start a Free Trial Today.

        [http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk](http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk)

        _______________________________________________

        Matplotlib-devel mailing list

        Matplotlib-devel@lists.sourceforge.net

        [https://lists.sourceforge.net/lists/listinfo/matplotlib-devel](https://lists.sourceforge.net/lists/listinfo/matplotlib-devel)

Ben

                    On Tue, Jan 14, 2014 at

11:36 AM, Michael Droettboom <mdroe@…31…>
wrote:

                      I'm fine with starting the meeting an hour

early. How about others?

                      Mike


                          On 01/14/2014 04:57 AM, Michiel de Hoon

wrote:

                          > I can join this Thursday if we start

with the discussion on timers.

                          > If we can start 1 hour earlier (14:00

UTC, 9 am ET, 23:00 in Japan) that would
be even better.

                          > -Michiel.

                          >

                          >

                          >

                          >

                          > On Mon, 1/13/14, Michael Droettboom

<mdroe@…31… >
wrote:

                          >

                          >   Subject: [matplotlib-devel]

Meeting…?

                          >   To: "matplotlib-devel@...758...urceforge.net                              "

matplotlib-devel@lists.sourceforge.net

                          >   Date: Monday, January 13, 2014,

11:36 AM

                          >

                          >   It's probably a good time to

schedule

                          >   another matplotlib Google Hangout.

                          >

                          >   Is this Thursday at 1500 UTC (10 am

ET) too short notice for

                          >   the usual

                          >   candidates?

                          >

                          >   I know there was discussion of

getting Michiel de Hoon on

                          >   today (which I

                          >   just saw, unfortunately).  Is there

another time in the

                          >   future that

                          >   works for you, Michiel?

                          >

                          >   Mike

                          >

                          >   --

                          >

                          >       _

                          >   |\/|o _|_  _. _ | | \.__  __|__|_|_

_

                          >   _ ._ _

                          >   |  ||(_| |(_|(/_| |_/|(_)(/_|_

||)()()| | |

                          >

                          >   [http://www.droettboom.com](http://www.droettboom.com)

                          >

                          >

                          >  

                          >   CenturyLink Cloud: The Leader in

Enterprise Cloud Services.

                          >   Learn Why More Businesses Are

Choosing CenturyLink Cloud

                          >   For

                          >   Critical Workloads, Development

Environments &

                          >   Everything In Between.

                          >   Get a Quote or Start a Free Trial

Today.

                          >   [http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk](http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk)


                          >  

                          >   Matplotlib-devel mailing list

                          >   Matplotlib-devel@...925...net

                          >   [https://lists.sourceforge.net/lists/listinfo/matplotlib-devel](https://lists.sourceforge.net/lists/listinfo/matplotlib-devel)

                          >





                          --

                                              _

                          |\/|o _|_  _. _ | | \.__  __|__|_|_  _  _

._ _

                          |  ||(_| |(_|(/_| |_/|(_)(/_|_

||)()()| | |

                          [http://www.droettboom.com](http://www.droettboom.com)

                          CenturyLink Cloud: The Leader in

Enterprise Cloud Services.

                          Learn Why More Businesses Are Choosing

CenturyLink Cloud For

                          Critical Workloads, Development

Environments & Everything In Between.

                          Get a Quote or Start a Free Trial Today.

                          [http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk](http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk)

                          Matplotlib-devel mailing list

                          Matplotlib-devel@lists.sourceforge.net

                          [https://lists.sourceforge.net/lists/listinfo/matplotlib-devel](https://lists.sourceforge.net/lists/listinfo/matplotlib-devel)
9

A *very* big +1. I'm really excited about doing IPython widgets these days, and this would make a matplotlib IPython widget pretty exciting.

Thanks,

Jason

···

On 1/14/14, 1:30 PM, Jacob Vanderplas wrote:

Ugly hacks aside, I think all of this points to a general desire for a
new type of backend-like hook that can export dynamic plot elements in
data coordinates, and static plot elements in figure coordinates. An
enhancement in that direction could pave the way for a lot of
interesting interactive front-ends to matplotlib figures.

10

I'm sure it's obvious, but the backend API was never really designed with this in mind. It was designed for "dumb" things like screens and printers that know about simple primitives like lines and text, but not much else. From that, I don't think it makes much sense to change or even extend that API, as the current targets of the backends are not going to suddenly get any smarter. You also have to remember that the drawing commands that matplotlib emits to a backend are essentially a stream -- they don't have the necessary structure (in all cases) to write out to something like d3. You really want tree traversal, IMHO, not a stream of commands.

For a long time, I've felt what is needed is a standard way to navigate over the tree that makes up the figure -- something like a DOM. So you would start at the root (the "figure") and drill down through all of the objects, getting their properties such as location, data and drawing styles and convert into some other format as you go. Your code wouldn't have to care about what coordinate space the data is in -- it would have access to the raw data and any custom transformations the user specified etc., and it could handle accordingly. There's two major pieces that make this hard right now. One is that each class has some arbitrary set of "children", and there's no uniform way to iterate over them. If each Artist had a children method that returns a list of its children, you could write a "Visitor" pattern [1] class over the tree that would let you write out your output format.

The second problem, as you point out, is that some things aren't really calculated until draw time. In some cases, we could just be doing more in the constructor. In other cases, we can't calculate the final locations of things until we know things like font size etc., but hopefully those aren't the cases a tool to write d3 would care about. I could also imagine doing drawing in two phases, if necessary, one to just finalize things, and the other to actually write to a "dumb" backend. A d3 converter would only do the first phase, and then iterate over the tree to generate its output.

[1] See Python's ast module or docutils for great examples of this pattern. docutils is a particular good analogy because it converts text in one complex logical structure (reStructuredText) to n different other logical structures (including HTML and LaTeX etc.)

I hope all of the above makes sense...

Mike

···

On 01/14/2014 01:30 PM, Jacob Vanderplas wrote:

Thanks - we'll make it happen at some point.

Perhaps I can give the seed for a discussion: the stuff I've been doing with mpld3 is a lot of fun, but it's fundamentally limited by the fact that I have to dig around the internals of the figure object to find the relevant information to construct a plot representation. I may be able to do the same thing by creating a backend, but the problem is that the draw() methods of most objects call the renderer with no reference to whether the points lie in the data space or figure space: that is, paths and points are usually specified in figure/pixel coordinates or some transformed version thereof, which makes it near impossible to construct interactive representations absent Python kernel callbacks.

What I'd love to see is some enhancement of the backend framework where there are some extra flags and information passed to the renderer: i.e. for each draw command, we need to know whether the drawn object should be linked to static figure coordinates or to dynamic axes/data coordinates.

I've been in touch with Cyrille Rossant from the vispy team, Chris Beaumont from the Glue team, and Matt Sundwuist from the plotly team, all of whom asked if there might be a way to use what I've done with mpld3 to enable matplotlib to export into their own front-end format. I didn't start mpld3 with that sort of extensibility in mind, but I'm starting to invest some time thinking about how to design that.

With the current matplotlib package, I think there are two ways to accomplish it: one is to create a general backend-like interface based on the figure introspection that mpld3 currently uses. The artist elements in each figure contain enough information to be able to infer whether the elements should move & zoom with the axes or not. The problem is, a lot of elements (like legends, axes aspects, etc.) are not fully established until the draw() command is called, so there are a few ugly hacks required to make it happen.

The other option is to use an even uglier hack, and wrap the current backend framework with an object that somehow links back into the figure and infers from the draw_*() commands whether the path/point/marker/etc. should be drawn in static figure coordinates or in dynamic axes coordinates. I've started a simple prototype backend translator which has a renderer class that uses ``inspect`` back-trace the stack and accomplish this: It's really ugly, and I'm not particularly proud about it, but I think it's the current best way to accomplish the desired behavior.

Ugly hacks aside, I think all of this points to a general desire for a new type of backend-like hook that can export dynamic plot elements in data coordinates, and static plot elements in figure coordinates. An enhancement in that direction could pave the way for a lot of interesting interactive front-ends to matplotlib figures.

Anyway - if any of you have suggestions or responses to this, I'd love to hear them! Thanks,
  Jake

On Tue, Jan 14, 2014 at 9:11 AM, Michael Droettboom <mdroe@...31... > <mailto:mdroe@…31…>> wrote:

    Jake: I'd definitely like to get you into one of these calls at
    some point. If you're able to pop in late, that would still be
    great -- or we can save that for another date. Trying to get
    Japan, three NA timezones and the UK all together is challenging :wink:

    In any event, with Thomas, Ben, Michiel and myself confirmed, I
    think that's enough to go ahead, and hopefully others who have yet
    to respond can join as well.

    Mike

    On 01/14/2014 11:57 AM, Jacob Vanderplas wrote:

    I'll probably not be able to swing 6am on the west coast, but
    other folks are more important for this call, I think :slight_smile:
       Jake

    On Tue, Jan 14, 2014 at 8:51 AM, Benjamin Root <ben.root@...553... >> <mailto:ben.root@…553…>> wrote:

        That would actually work a little bit better for me... I just
        have to remember to get into work a little bit earlier.

        Ben

        On Tue, Jan 14, 2014 at 11:36 AM, Michael Droettboom >> <mdroe@...31... <mailto:mdroe@…31…>> wrote:

            I'm fine with starting the meeting an hour early. How
            about others?

            Mike

            On 01/14/2014 04:57 AM, Michiel de Hoon wrote:
            > I can join this Thursday if we start with the
            discussion on timers.
            > If we can start 1 hour earlier (14:00 UTC, 9 am ET,
            23:00 in Japan) that would be even better.
            > -Michiel.
            >
            > --------------------------------------------
            > On Mon, 1/13/14, Michael Droettboom <mdroe@...31... >> <mailto:mdroe@…31…>> wrote:
            >
            > Subject: [matplotlib-devel] Meeting...?
            > To: "matplotlib-devel@lists.sourceforge.net
            <mailto:matplotlib-devel@lists.sourceforge.net>"
            <matplotlib-devel@lists.sourceforge.net
            <mailto:matplotlib-devel@lists.sourceforge.net>>
            > Date: Monday, January 13, 2014, 11:36 AM
            >
            > It's probably a good time to schedule
            > another matplotlib Google Hangout.
            >
            > Is this Thursday at 1500 UTC (10 am ET) too short
            notice for
            > the usual
            > candidates?
            >
            > I know there was discussion of getting Michiel de Hoon on
            > today (which I
            > just saw, unfortunately). Is there another time in the
            > future that
            > works for you, Michiel?
            >
            > Mike
            >
            > --
            >
            > _
            > >\/|o _|_ _. _ | | \.__ __|__|_|_ _
            > _ ._ _
            > > >>(_| |(_|(/_| |_/|(_)(/_|_ |_|_)(_)(_)| | |
            >
            > http://www.droettboom.com
            >
            ------------------------------------------------------------------------------
            > CenturyLink Cloud: The Leader in Enterprise Cloud
            Services.
            > Learn Why More Businesses Are Choosing CenturyLink Cloud
            > For
            > Critical Workloads, Development Environments &
            > Everything In Between.
            > Get a Quote or Start a Free Trial Today.
            >
            http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk
            > _______________________________________________
            > Matplotlib-devel mailing list
            > Matplotlib-devel@lists.sourceforge.net
            <mailto:Matplotlib-devel@lists.sourceforge.net>
            >
            matplotlib-devel List Signup and Options
            >

            --
                                _
            >\/|o _|_ _. _ | | \.__ __|__|_|_ _ _ ._ _
            > >>(_| |(_|(/_| |_/|(_)(/_|_ |_|_)(_)(_)| | |

            http://www.droettboom.com

            ------------------------------------------------------------------------------
            CenturyLink Cloud: The Leader in Enterprise Cloud Services.
            Learn Why More Businesses Are Choosing CenturyLink Cloud For
            Critical Workloads, Development Environments & Everything
            In Between.
            Get a Quote or Start a Free Trial Today.
            http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk
            _______________________________________________
            Matplotlib-devel mailing list
            Matplotlib-devel@lists.sourceforge.net
            <mailto:Matplotlib-devel@lists.sourceforge.net>
            matplotlib-devel List Signup and Options

        ------------------------------------------------------------------------------
        CenturyLink Cloud: The Leader in Enterprise Cloud Services.
        Learn Why More Businesses Are Choosing CenturyLink Cloud For
        Critical Workloads, Development Environments & Everything In
        Between.
        Get a Quote or Start a Free Trial Today.
        http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk
        _______________________________________________
        Matplotlib-devel mailing list
        Matplotlib-devel@lists.sourceforge.net
        <mailto:Matplotlib-devel@lists.sourceforge.net>
        matplotlib-devel List Signup and Options

    -- _
    >\/|o _|_ _. _ | | \.__ __|__|_|_ _ _ ._ _
    > >>(_| |(_|(/_| |_/|(_)(/_|_ |_|_)(_)(_)| | |

    http://www.droettboom.com

--
                    _

\/|o _|_ _. _ | | \.__ __|__|_|_ _ _ ._ _
>>(_| |(_|(/_| |_/|(_)(/_|_ |_|_)(_)(_)| | |

11

I hope all of the above makes sense...

Definitely makes sense: what I've built-up in mpld3 is essentially
something that mimics this sort of visitor pattern, though it misses some
things because of the draw-time difficulties you mention.
I think a two-stage draw() would be a _very_ helpful restructure.
Currently, I'm forced to achieve this result by writing a png to a
throwaway byte-stream...
   Jake

···

On Tue, Jan 14, 2014 at 12:04 PM, Michael Droettboom <mdroe@...31...>wrote:

Mike

On 01/14/2014 01:30 PM, Jacob Vanderplas wrote:

Thanks - we'll make it happen at some point.

Perhaps I can give the seed for a discussion: the stuff I've been doing
with mpld3 is a lot of fun, but it's fundamentally limited by the fact that
I have to dig around the internals of the figure object to find the
relevant information to construct a plot representation. I may be able to
do the same thing by creating a backend, but the problem is that the draw()
methods of most objects call the renderer with no reference to whether the
points lie in the data space or figure space: that is, paths and points are
usually specified in figure/pixel coordinates or some transformed version
thereof, which makes it near impossible to construct interactive
representations absent Python kernel callbacks.

What I'd love to see is some enhancement of the backend framework where
there are some extra flags and information passed to the renderer: i.e. for
each draw command, we need to know whether the drawn object should be
linked to static figure coordinates or to dynamic axes/data coordinates.

I've been in touch with Cyrille Rossant from the vispy team, Chris
Beaumont from the Glue team, and Matt Sundwuist from the plotly team, all
of whom asked if there might be a way to use what I've done with mpld3 to
enable matplotlib to export into their own front-end format. I didn't
start mpld3 with that sort of extensibility in mind, but I'm starting to
invest some time thinking about how to design that.

With the current matplotlib package, I think there are two ways to
accomplish it: one is to create a general backend-like interface based on
the figure introspection that mpld3 currently uses. The artist elements in
each figure contain enough information to be able to infer whether the
elements should move & zoom with the axes or not. The problem is, a lot of
elements (like legends, axes aspects, etc.) are not fully established until
the draw() command is called, so there are a few ugly hacks required to
make it happen.

The other option is to use an even uglier hack, and wrap the current
backend framework with an object that somehow links back into the figure
and infers from the draw_*() commands whether the path/point/marker/etc.
should be drawn in static figure coordinates or in dynamic axes
coordinates. I've started a simple prototype backend translator which has a
renderer class that uses ``inspect`` back-trace the stack and accomplish
this: It's really ugly, and I'm not particularly proud about it, but I
think it's the current best way to accomplish the desired behavior.

Ugly hacks aside, I think all of this points to a general desire for a
new type of backend-like hook that can export dynamic plot elements in data
coordinates, and static plot elements in figure coordinates. An
enhancement in that direction could pave the way for a lot of interesting
interactive front-ends to matplotlib figures.

Anyway - if any of you have suggestions or responses to this, I'd love
to hear them! Thanks,
   Jake

On Tue, Jan 14, 2014 at 9:11 AM, Michael Droettboom <mdroe@...31...>wrote:

Jake: I'd definitely like to get you into one of these calls at some
point. If you're able to pop in late, that would still be great -- or we
can save that for another date. Trying to get Japan, three NA timezones
and the UK all together is challenging :wink:

In any event, with Thomas, Ben, Michiel and myself confirmed, I think
that's enough to go ahead, and hopefully others who have yet to respond can
join as well.

Mike

On 01/14/2014 11:57 AM, Jacob Vanderplas wrote:

I'll probably not be able to swing 6am on the west coast, but other folks
are more important for this call, I think :slight_smile:
   Jake

On Tue, Jan 14, 2014 at 8:51 AM, Benjamin Root <ben.root@...553...> wrote:

That would actually work a little bit better for me... I just have to
remember to get into work a little bit earlier.

Ben

On Tue, Jan 14, 2014 at 11:36 AM, Michael Droettboom <mdroe@...31...>wrote:

I'm fine with starting the meeting an hour early. How about others?

Mike

On 01/14/2014 04:57 AM, Michiel de Hoon wrote:
> I can join this Thursday if we start with the discussion on timers.
> If we can start 1 hour earlier (14:00 UTC, 9 am ET, 23:00 in Japan)
that would be even better.
> -Michiel.
>
>
>
> --------------------------------------------
> On Mon, 1/13/14, Michael Droettboom <mdroe@...31...> wrote:
>
> Subject: [matplotlib-devel] Meeting...?
> To: "matplotlib-devel@lists.sourceforge.net" <
matplotlib-devel@lists.sourceforge.net>
> Date: Monday, January 13, 2014, 11:36 AM
>
> It's probably a good time to schedule
> another matplotlib Google Hangout.
>
> Is this Thursday at 1500 UTC (10 am ET) too short notice for
> the usual
> candidates?
>
> I know there was discussion of getting Michiel de Hoon on
> today (which I
> just saw, unfortunately). Is there another time in the
> future that
> works for you, Michiel?
>
> Mike
>
> --
>
> _
> >\/|o _|_ _. _ | | \.__ __|__|_|_ _
> _ ._ _
> > >>(_| |(_|(/_| |_/|(_)(/_|_ |_|_)(_)(_)| | |
>
> http://www.droettboom.com
>
>
>
------------------------------------------------------------------------------
> CenturyLink Cloud: The Leader in Enterprise Cloud Services.
> Learn Why More Businesses Are Choosing CenturyLink Cloud
> For
> Critical Workloads, Development Environments &
> Everything In Between.
> Get a Quote or Start a Free Trial Today.
>
http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk
> _______________________________________________
> Matplotlib-devel mailing list
> Matplotlib-devel@lists.sourceforge.net
> matplotlib-devel List Signup and Options
>

--
                    _
>\/|o _|_ _. _ | | \.__ __|__|_|_ _ _ ._ _
> >>(_| |(_|(/_| |_/|(_)(/_|_ |_|_)(_)(_)| | |

http://www.droettboom.com

------------------------------------------------------------------------------
CenturyLink Cloud: The Leader in Enterprise Cloud Services.
Learn Why More Businesses Are Choosing CenturyLink Cloud For
Critical Workloads, Development Environments & Everything In Between.
Get a Quote or Start a Free Trial Today.

http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk
_______________________________________________
Matplotlib-devel mailing list
Matplotlib-devel@lists.sourceforge.net
matplotlib-devel List Signup and Options

------------------------------------------------------------------------------
CenturyLink Cloud: The Leader in Enterprise Cloud Services.
Learn Why More Businesses Are Choosing CenturyLink Cloud For
Critical Workloads, Development Environments & Everything In Between.
Get a Quote or Start a Free Trial Today.

http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk
_______________________________________________
Matplotlib-devel mailing list
Matplotlib-devel@lists.sourceforge.net
matplotlib-devel List Signup and Options

--
                   _
>\/|o _|_ _. _ | | \.__ __|__|_|_ _ _ ._ _
> >>(_| |(_|(/_| |_/|(_)(/_|_ |_|_)(_)(_)| | |
http://www.droettboom.com

--
                   _
>\/|o _|_ _. _ | | \.__ __|__|_|_ _ _ ._ _
> >>(_| |(_|(/_| |_/|(_)(/_|_ |_|_)(_)(_)| | |
http://www.droettboom.com

12

What I'd love to see is some enhancement of the backend framework where
there are some extra flags and information passed to the renderer: i.e. for
each draw command, we need to know whether the drawn object should be linked
to static figure coordinates or to dynamic axes/data coordinates.

+1. It would also be nice to have the possibility to bypass dynamic
data transformations performed by matplotlib. A potential GPU-based
backend would be the most efficient if it was granted the right to
perform all data transformations by itself (for example when panning
and zooming).

Cyrille

13

I have another long-simmering feature request along these lines: if Matplotlib were to evolve a formal DOM-like figure structure like mentioned above, it would be cool if this structure retained more semantic information about the visualization itself. By this, I mean that many high-level commands like hist, scatter, etc. spawn a bunch of low level artists like rectangles and circles. After these methods exit, it’s hard/impossible to introspect a Figure and diagnose that it, for example, is a histogram and not a bar chart.

Retaining a better high-level description of a plot (which probably amounts to creating more compound artist types) would make it easier to build tools like mpld3 and other cool things that involve runtime editing or optimization of tree-like data structures.

chris

···

On Tue, Jan 14, 2014 at 3:34 PM, Jacob Vanderplas <jakevdp@…1179…> wrote:

CenturyLink Cloud: The Leader in Enterprise Cloud Services.

Learn Why More Businesses Are Choosing CenturyLink Cloud For

Critical Workloads, Development Environments & Everything In Between.

Get a Quote or Start a Free Trial Today.

http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk

Matplotlib-devel mailing list

Matplotlib-devel@lists.sourceforge.net

https://lists.sourceforge.net/lists/listinfo/matplotlib-devel

–

Chris Beaumont
Graduate Student
Institute for Astronomy
University of Hawaii at Manoa
2680 Woodlawn Drive

Honolulu, HI 96822
www.ifa.hawaii.edu/~beaumont

On Tue, Jan 14, 2014 at 12:04 PM, Michael Droettboom <mdroe@…31…> wrote:

  I hope all of the above makes sense...

Definitely makes sense: what I’ve built-up in mpld3 is essentially something that mimics this sort of visitor pattern, though it misses some things because of the draw-time difficulties you mention.

I think a two-stage draw() would be a very helpful restructure. Currently, I’m forced to achieve this result by writing a png to a throwaway byte-stream…

Jake

  Mike




  On 01/14/2014 01:30 PM, Jacob Vanderplas wrote:

Thanks - we’ll make it happen at some point.

          Perhaps I can give the seed for a discussion: the stuff

I’ve been doing with mpld3 is a lot of fun, but it’s
fundamentally limited by the fact that I have to dig
around the internals of the figure object to find the
relevant information to construct a plot representation.
I may be able to do the same thing by creating a backend,
but the problem is that the draw() methods of most objects
call the renderer with no reference to whether the points
lie in the data space or figure space: that is, paths and
points are usually specified in figure/pixel coordinates
or some transformed version thereof, which makes it near
impossible to construct interactive representations absent
Python kernel callbacks.

          What I'd love to see is some enhancement of the backend

framework where there are some extra flags and information
passed to the renderer: i.e. for each draw command, we
need to know whether the drawn object should be linked to
static figure coordinates or to dynamic axes/data
coordinates.

          I've been in touch with Cyrille Rossant from the vispy

team, Chris Beaumont from the Glue team, and Matt
Sundwuist from the plotly team, all of whom asked if there
might be a way to use what I’ve done with mpld3 to enable
matplotlib to export into their own front-end format. I
didn’t start mpld3 with that sort of extensibility in
mind, but I’m starting to invest some time thinking about
how to design that.

          With the current matplotlib package, I think there are

two ways to accomplish it: one is to create a general
backend-like interface based on the figure introspection
that mpld3 currently uses. The artist elements in each
figure contain enough information to be able to infer
whether the elements should move & zoom with the axes
or not. The problem is, a lot of elements (like legends,
axes aspects, etc.) are not fully established until the
draw() command is called, so there are a few ugly hacks
required to make it happen.

          The other option is to use an even uglier hack, and

wrap the current backend framework with an object that
somehow links back into the figure and infers from the
draw_*() commands whether the path/point/marker/etc.
should be drawn in static figure coordinates or in dynamic
axes coordinates. I’ve started a simple prototype backend
translator which has a renderer class that uses
inspect back-trace the stack and accomplish this: It’s
really ugly, and I’m not particularly proud about it, but
I think it’s the current best way to accomplish the
desired behavior.

          Ugly hacks aside, I think all of this points to a

general desire for a new type of backend-like hook that
can export dynamic plot elements in data coordinates, and
static plot elements in figure coordinates. An
enhancement in that direction could pave the way for a lot
of interesting interactive front-ends to matplotlib
figures.

          Anyway - if any of you have suggestions or responses to

this, I’d love to hear them! Thanks,

Jake

                   -- _ |\/|o _|_ _. _ | | \.__ __|__|_|_ _ _ ._ _ | ||(_| |(_|(/_| |_/|(_)(/_|_ |_|_)(_)(_)| | | [http://www.droettboom.com](http://www.droettboom.com)

      On Tue, Jan 14, 2014 at 9:11 AM,

Michael Droettboom <mdroe@…31…> wrote:

            Jake: I'd definitely like to get you into one of

these calls at some point. If you’re able to pop in
late, that would still be great – or we can save that
for another date. Trying to get Japan, three NA
timezones and the UK all together is challenging :wink:

            In any event, with Thomas, Ben, Michiel and myself

confirmed, I think that’s enough to go ahead, and
hopefully others who have yet to respond can join as
well.

            Mike




                On 01/14/2014 11:57 AM, Jacob Vanderplas wrote:

                  I'll probably not be able to swing

6am on the west coast, but other folks are more
important for this call, I think :slight_smile:
Jake

                   -- _ |\/|o _|_ _. _ | | \.__ __|__|_|_ _ _ ._ _ | ||(_| |(_|(/_| |_/|(_)(/_|_ |_|_)(_)(_)| | | [http://www.droettboom.com](http://www.droettboom.com)

                    On Tue, Jan 14, 2014 at

8:51 AM, Benjamin Root <ben.root@…553…>
wrote:

                        That would actually work a

little bit better for me… I just have to
remember to get into work a little bit
earlier.

                      CenturyLink Cloud: The Leader in Enterprise

Cloud Services.

                      Learn Why More Businesses Are Choosing

CenturyLink Cloud For

                      Critical Workloads, Development Environments

& Everything In Between.

                      Get a Quote or Start a Free Trial Today.

                      [http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk](http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk)

                      Matplotlib-devel mailing list

                      Matplotlib-devel@lists.sourceforge.net

                      [https://lists.sourceforge.net/lists/listinfo/matplotlib-devel](https://lists.sourceforge.net/lists/listinfo/matplotlib-devel)

Ben

                                  On Tue, Jan

14, 2014 at 11:36 AM, Michael
Droettboom <mdroe@…31…>
wrote:

                                    I'm

fine with starting the meeting
an hour early. How about
others?

                                    Mike


                                        On 01/14/2014 04:57 AM,

Michiel de Hoon wrote:

                                        > I can join this

Thursday if we start with
the discussion on timers.

                                        > If we can start 1 hour

earlier (14:00 UTC, 9 am ET,
23:00 in Japan) that would
be even better.

                                        > -Michiel.

                                        >

                                        >

                                        >

                                        >

                                        > On Mon, 1/13/14,

Michael Droettboom <mdroe@…31…
>
wrote:

                                        >

                                        >   Subject:

[matplotlib-devel]
Meeting…?

                                        >   To: "matplotlib-devel@lists.sourceforge.net                                            "

<matplotlib-devel@…1052…rceforge.net>

                                        >   Date: Monday, January

13, 2014, 11:36 AM

                                        >

                                        >   It's probably a good

time to schedule

                                        >   another matplotlib

Google Hangout.

                                        >

                                        >   Is this Thursday at

1500 UTC (10 am ET) too
short notice for

                                        >   the usual

                                        >   candidates?

                                        >

                                        >   I know there was

discussion of getting
Michiel de Hoon on

                                        >   today (which I

                                        >   just saw,

unfortunately). Is there
another time in the

                                        >   future that

                                        >   works for you,

Michiel?

                                        >

                                        >   Mike

                                        >

                                        >   --

                                        >

                                        >       _

                                        >   |\/|o _|_  _. _ | |

.__ ||| _

                                        >   _ ._ _

                                        >   |  ||(_| |(_|(/_|

|/|()(/| ||)()()| |
|

                                        >

                                        >   [http://www.droettboom.com](http://www.droettboom.com)

                                        >

                                        >

                                        >  

                                        >   CenturyLink Cloud:

The Leader in Enterprise
Cloud Services.

                                        >   Learn Why More

Businesses Are Choosing
CenturyLink Cloud

                                        >   For

                                        >   Critical Workloads,

Development Environments
&

                                        >   Everything In

Between.

                                        >   Get a Quote or Start

a Free Trial Today.

                                        >   [http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk](http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk)



                                        >  

                                        >   Matplotlib-devel

mailing list

                                        >   Matplotlib-devel@...1034...s.sourceforge.net

                                        >   [https://lists.sourceforge.net/lists/listinfo/matplotlib-devel](https://lists.sourceforge.net/lists/listinfo/matplotlib-devel)

                                        >





                                        --

                                                            _

                                        |\/|o _|_  _. _ | | \.__

||| _ _ ._ _

                                        |  ||(_| |(_|(/_|

|/|()(/| ||)()()| |
|

                                        [http://www.droettboom.com](http://www.droettboom.com)

                                        CenturyLink Cloud: The

Leader in Enterprise Cloud
Services.

                                        Learn Why More Businesses

Are Choosing CenturyLink
Cloud For

                                        Critical Workloads,

Development Environments
& Everything In Between.

                                        Get a Quote or Start a Free

Trial Today.

                                        [http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk](http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk)

                                        Matplotlib-devel mailing

list

                                        Matplotlib-devel@...1035...orge.net

                                        [https://lists.sourceforge.net/lists/listinfo/matplotlib-devel](https://lists.sourceforge.net/lists/listinfo/matplotlib-devel)
14

In the specific case of data transformations, this is already the case. The backends are sent the raw data + a transformation. This allows file formats to write out untransformed data and have the renderer of that format do the transformation. Unfortunately, all of the backends end up doing the transformation in software anyway, since all of them scale the line width along with the path data, which is never what you want for plotting. (The exception to this is the SVG spec, which now contains a flag for this, but that spec is so new we can't use it because it isn't widely implemented).

Mike

···

On 01/14/2014 03:43 PM, Cyrille Rossant wrote:

What I'd love to see is some enhancement of the backend framework where
there are some extra flags and information passed to the renderer: i.e. for
each draw command, we need to know whether the drawn object should be linked
to static figure coordinates or to dynamic axes/data coordinates.

+1. It would also be nice to have the possibility to bypass dynamic
data transformations performed by matplotlib. A potential GPU-based
backend would be the most efficient if it was granted the right to
perform all data transformations by itself (for example when panning
and zooming).

--
                    _

\/|o _|_ _. _ | | \.__ __|__|_|_ _ _ ._ _
>>(_| |(_|(/_| |_/|(_)(/_|_ |_|_)(_)(_)| | |

15

I agree, this would be very nice to have, but is (obviously) a much
bigger step from what we currently have in matplotlib. Along those
lines, if the “frontend” (i.e. the pylab interface and even the OO
interface) basically generate a tree, then you could serialize that
tree (here XML would be a great fit, don’t knock it), and even
non-Python based tools could transform it into something else.
Agreed.
No argument against this at all from me – but knowing how much work
this would be, the obstacle there is just finding the time to do
it. It would be a significant rewrite…
Mike

···

On 01/14/2014 03:49 PM, Chris Beaumont
wrote:

    I have another long-simmering feature request along

these lines: if Matplotlib were to evolve a formal DOM-like
figure structure like mentioned above, it would be cool if this
structure retained more semantic information about the
visualization itself. By this, I mean that many high-level
commands like hist, scatter, etc. spawn a bunch of low level
artists like rectangles and circles. After these methods exit,
it’s hard/impossible to introspect a Figure and diagnose that
it, for example, is a histogram and not a bar chart.

      Retaining a better high-level description of a plot (which

probably amounts to creating more compound artist types) would
make it easier to build tools like mpld3 and other cool things
that involve runtime editing or optimization of tree-like data
structures.

chris

      On Tue, Jan 14, 2014 at 3:34 PM, Jacob

Vanderplas <jakevdp@…1179…>
wrote:

                On Tue, Jan 14, 2014 at 12:04 PM,

Michael Droettboom <mdroe@…31…>
wrote:

                      I hope all of the above makes sense...

                Definitely makes sense: what I've built-up in

mpld3 is essentially something that mimics this sort
of visitor pattern, though it misses some things
because of the draw-time difficulties you mention.

                I think a two-stage draw() would be a _very_

helpful restructure. Currently, I’m forced to
achieve this result by writing a png to a throwaway
byte-stream…

Jake

                        Mike




                            On 01/14/2014 01:30 PM, Jacob Vanderplas

wrote:

                              Thanks - we'll make it

happen at some point.

                                    Perhaps I can give the seed

for a discussion: the stuff I’ve
been doing with mpld3 is a lot
of fun, but it’s fundamentally
limited by the fact that I have
to dig around the internals of
the figure object to find the
relevant information to
construct a plot representation.
I may be able to do the same
thing by creating a backend, but
the problem is that the draw()
methods of most objects call the
renderer with no reference to
whether the points lie in the
data space or figure space: that
is, paths and points are usually
specified in figure/pixel
coordinates or some transformed
version thereof, which makes it
near impossible to construct
interactive representations
absent Python kernel callbacks.

                                    What I'd love to see is some

enhancement of the backend
framework where there are some
extra flags and information
passed to the renderer: i.e. for
each draw command, we need to
know whether the drawn object
should be linked to static
figure coordinates or to dynamic
axes/data coordinates.

                                    I've been in touch with Cyrille

Rossant from the vispy team,
Chris Beaumont from the Glue
team, and Matt Sundwuist from
the plotly team, all of whom
asked if there might be a way to
use what I’ve done with mpld3 to
enable matplotlib to export into
their own front-end format. I
didn’t start mpld3 with that
sort of extensibility in mind,
but I’m starting to invest some
time thinking about how to
design that.

                                    With the current matplotlib

package, I think there are two
ways to accomplish it: one is to
create a general backend-like
interface based on the figure
introspection that mpld3
currently uses. The artist
elements in each figure contain
enough information to be able to
infer whether the elements
should move & zoom with the
axes or not. The problem is, a
lot of elements (like legends,
axes aspects, etc.) are not
fully established until the
draw() command is called, so
there are a few ugly hacks
required to make it happen.

                                    The other option is to use an

even uglier hack, and wrap the
current backend framework with
an object that somehow links
back into the figure and infers
from the draw_*() commands
whether the
path/point/marker/etc. should be
drawn in static figure
coordinates or in dynamic axes
coordinates. I’ve started a
simple prototype backend
translator which has a renderer
class that uses inspect
back-trace the stack and
accomplish this: It’s really
ugly, and I’m not particularly
proud about it, but I think it’s
the current best way to
accomplish the desired behavior.

                                    Ugly hacks aside, I think all

of this points to a general
desire for a new type of
backend-like hook that can
export dynamic plot elements in
data coordinates, and static
plot elements in figure
coordinates. An enhancement in
that direction could pave the
way for a lot of interesting
interactive front-ends to
matplotlib figures.

                                    Anyway - if any of you have

suggestions or responses to
this, I’d love to hear them!
Thanks,

Jake

                                On Tue, Jan

14, 2014 at 9:11 AM, Michael
Droettboom <mdroe@…31…>
wrote:

                                      Jake: I'd definitely like

to get you into one of these
calls at some point. If
you’re able to pop in late,
that would still be great –
or we can save that for
another date. Trying to get
Japan, three NA timezones and
the UK all together is
challenging :wink:

                                      In any event, with Thomas,

Ben, Michiel and myself
confirmed, I think that’s
enough to go ahead, and
hopefully others who have yet
to respond can join as well.

                                      Mike




                                          On 01/14/2014 11:57 AM,

Jacob Vanderplas wrote:

                                            I'll

probably not be able to
swing 6am on the west
coast, but other folks
are more important for
this call, I think :slight_smile:
Jake

                                              On

Tue, Jan 14, 2014 at
8:51 AM, Benjamin Root
<ben.root@…553…>
wrote:

                                                  That

would actually
work a little bit
better for me… I
just have to
remember to get
into work a little
bit earlier.

Ben

                                                      On

Tue, Jan 14,
2014 at 11:36
AM, Michael
Droettboom <mdroe@…31…>
wrote:

                                                      I'm fine with

starting the
meeting an
hour early.
How about
others?

                                                      Mike


                                                      On 01/14/2014

04:57 AM,
Michiel de
Hoon wrote:

                                                      > I can

join this
Thursday if we
start with the
discussion on
timers.

                                                      > If we can

start 1 hour
earlier (14:00
UTC, 9 am ET,
23:00 in
Japan) that
would be even
better.

                                                      > -Michiel.

                                                      >

                                                      >

                                                      >

                                                      >

                                                      > On Mon,

1/13/14,
Michael
Droettboom
<mdroe@…31… >
wrote:

                                                      >

                                                      >

Subject:
[matplotlib-devel]
Meeting…?

                                                      >   To: "matplotlib-devel@lists.sourceforge.net" <matplotlib-devel@lists.sourceforge.net>

                                                      >   Date:

Monday,
January 13,
2014, 11:36 AM

                                                      >

                                                      >   It's

probably a
good time to
schedule

                                                      >   another

matplotlib
Google
Hangout.

                                                      >

                                                      >   Is this

Thursday at
1500 UTC (10
am ET) too
short notice
for

                                                      >   the

usual

                                                      >

candidates?

                                                      >

                                                      >   I know

there was
discussion of
getting
Michiel de
Hoon on

                                                      >   today

(which I

                                                      >   just

saw,
unfortunately).
Is there
another time
in the

                                                      >   future

that

                                                      >   works

for you,
Michiel?

                                                      >

                                                      >   Mike

                                                      >

                                                      >   --

                                                      >

                                                      >       _

                                                      >   |\/|o

| . _ | |
._
||| _

                                                      >   _ ._ _

                                                      >   |

||(|
|(|(/|
|/|()(/|_
||)()()| |
|

                                                      >

                                                      >   [http://www.droettboom.com](http://www.droettboom.com)

                                                      >

                                                      >

                                                      >  

                                                      >

CenturyLink
Cloud: The
Leader in
Enterprise
Cloud
Services.

                                                      >   Learn

Why More
Businesses Are
Choosing
CenturyLink
Cloud

                                                      >   For

                                                      >

Critical
Workloads,
Development
Environments
&

                                                      >

Everything In
Between.

                                                      >   Get a

Quote or Start
a Free Trial
Today.

                                                      >   [http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk](http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk)

                                                      >  

                                                      >

Matplotlib-devel
mailing list

                                                      >   Matplotlib-devel@lists.sourceforge.net

                                                      >   [https://lists.sourceforge.net/lists/listinfo/matplotlib-devel](https://lists.sourceforge.net/lists/listinfo/matplotlib-devel)

                                                      >





                                                      --

                                                                   
  _

                                                      |\/|o _|_  _.

_ | | .__
||| _
_ ._ _

                                                      |  ||(_|

|(|(/|
|/|()(/|
||)()()| |
|

                                                      [http://www.droettboom.com](http://www.droettboom.com)

                                                      CenturyLink

Cloud: The
Leader in
Enterprise
Cloud
Services.

                                                      Learn Why More

Businesses Are
Choosing
CenturyLink
Cloud For

                                                      Critical

Workloads,
Development
Environments
&
Everything In
Between.

                                                      Get a Quote or

Start a Free
Trial Today.

                                                      [http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk](http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk)

                                                      Matplotlib-devel

mailing list

                                                      Matplotlib-devel@lists.sourceforge.net

                                                      [https://lists.sourceforge.net/lists/listinfo/matplotlib-devel](https://lists.sourceforge.net/lists/listinfo/matplotlib-devel)

                                                CenturyLink Cloud:

The Leader in
Enterprise Cloud
Services.

                                                Learn Why More

Businesses Are
Choosing CenturyLink
Cloud For

                                                Critical Workloads,

Development
Environments &
Everything In
Between.

                                                Get a Quote or Start

a Free Trial Today.

                                                [http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk](http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk)

                                                Matplotlib-devel

mailing list

                                                Matplotlib-devel@lists.sourceforge.net

                                                [https://lists.sourceforge.net/lists/listinfo/matplotlib-devel](https://lists.sourceforge.net/lists/listinfo/matplotlib-devel)

                   -- _ |\/|o _|_ _. _ | | \.__ __|__|_|_ _ _ ._ _ | ||(_| |(_|(/_| |_/|(_)(/_|_ |_|_)(_)(_)| | | [http://www.droettboom.com](http://www.droettboom.com)

                   -- _ |\/|o _|_ _. _ | | \.__ __|__|_|_ _ _ ._ _ | ||(_| |(_|(/_| |_/|(_)(/_|_ |_|_)(_)(_)| | | [http://www.droettboom.com](http://www.droettboom.com)

        CenturyLink Cloud: The Leader in Enterprise Cloud Services.

        Learn Why More Businesses Are Choosing CenturyLink Cloud For

        Critical Workloads, Development Environments &

Everything In Between.

        Get a Quote or Start a Free Trial Today.

        [http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk](http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk)

        _______________________________________________

        Matplotlib-devel mailing list

        Matplotlib-devel@lists.sourceforge.net

        [https://lists.sourceforge.net/lists/listinfo/matplotlib-devel](https://lists.sourceforge.net/lists/listinfo/matplotlib-devel)

–

    ************************************

    Chris Beaumont

    Graduate Student

    Institute for Astronomy

    University of Hawaii at Manoa

    2680 Woodlawn Drive

    Honolulu, HI 96822

    [www.ifa.hawaii.edu/~beaumont](http://www.ifa.hawaii.edu/%7Ebeaumont)

    ************************************

                   -- _ |\/|o _|_ _. _ | | \.__ __|__|_|_ _ _ ._ _ | ||(_| |(_|(/_| |_/|(_)(/_|_ |_|_)(_)(_)| | |

http://www.droettboom.com

16

Is there room for a lurker in the hangout? If so, is there a link for joining the hangout? (I tried using the hangout calendar link from awhile back, but I couldn’t get my google calendar to recognize it.)

Cheers,

-Tony

···

On Wed, Jan 15, 2014 at 9:57 AM, Michael Droettboom <mdroe@…31…> wrote:

  On 01/14/2014 03:49 PM, Chris Beaumont

wrote:

    I have another long-simmering feature request along

these lines: if Matplotlib were to evolve a formal DOM-like
figure structure like mentioned above, it would be cool if this
structure retained more semantic information about the
visualization itself. By this, I mean that many high-level
commands like hist, scatter, etc. spawn a bunch of low level
artists like rectangles and circles. After these methods exit,
it’s hard/impossible to introspect a Figure and diagnose that
it, for example, is a histogram and not a bar chart.

I agree, this would be very nice to have, but is (obviously) a much

bigger step from what we currently have in matplotlib. Along those
lines, if the “frontend” (i.e. the pylab interface and even the OO
interface) basically generate a tree, then you could serialize that
tree (here XML would be a great fit, don’t knock it), and even
non-Python based tools could transform it into something else.

      Retaining a better high-level description of a plot (which

probably amounts to creating more compound artist types) would
make it easier to build tools like mpld3 and other cool things
that involve runtime editing or optimization of tree-like data
structures.

Agreed.

No argument against this at all from me -- but knowing how much work

this would be, the obstacle there is just finding the time to do
it. It would be a significant rewrite…

Mike

chris

                   -- _ |\/|o _|_ _. _ | | \.__ __|__|_|_ _ _ ._ _ | ||(_| |(_|(/_| |_/|(_)(/_|_ |_|_)(_)(_)| | | [http://www.droettboom.com](http://www.droettboom.com)

CenturyLink Cloud: The Leader in Enterprise Cloud Services.

Learn Why More Businesses Are Choosing CenturyLink Cloud For

Critical Workloads, Development Environments & Everything In Between.

Get a Quote or Start a Free Trial Today.

http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk

Matplotlib-devel mailing list

Matplotlib-devel@lists.sourceforge.net

https://lists.sourceforge.net/lists/listinfo/matplotlib-devel

      On Tue, Jan 14, 2014 at 3:34 PM, Jacob

Vanderplas <jakevdp@…1179…>
wrote:

        CenturyLink Cloud: The Leader in Enterprise Cloud Services.

        Learn Why More Businesses Are Choosing CenturyLink Cloud For

        Critical Workloads, Development Environments &

Everything In Between.

        Get a Quote or Start a Free Trial Today.

        [http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk](http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk)

        _______________________________________________

        Matplotlib-devel mailing list

        Matplotlib-devel@lists.sourceforge.net

        [https://lists.sourceforge.net/lists/listinfo/matplotlib-devel](https://lists.sourceforge.net/lists/listinfo/matplotlib-devel)

–

    ************************************

    Chris Beaumont

    Graduate Student

    Institute for Astronomy

    University of Hawaii at Manoa

    2680 Woodlawn Drive

    Honolulu, HI 96822

    [www.ifa.hawaii.edu/~beaumont](http://www.ifa.hawaii.edu/%7Ebeaumont)

    ************************************

                On Tue, Jan 14, 2014 at 12:04 PM,

Michael Droettboom <mdroe@…31…>
wrote:

                      I hope all of the above makes sense...

                Definitely makes sense: what I've built-up in

mpld3 is essentially something that mimics this sort
of visitor pattern, though it misses some things
because of the draw-time difficulties you mention.

                I think a two-stage draw() would be a _very_

helpful restructure. Currently, I’m forced to
achieve this result by writing a png to a throwaway
byte-stream…

Jake

                        Mike




                            On 01/14/2014 01:30 PM, Jacob Vanderplas

wrote:

                              Thanks - we'll make it

happen at some point.

                                    Perhaps I can give the seed

for a discussion: the stuff I’ve
been doing with mpld3 is a lot
of fun, but it’s fundamentally
limited by the fact that I have
to dig around the internals of
the figure object to find the
relevant information to
construct a plot representation.
I may be able to do the same
thing by creating a backend, but
the problem is that the draw()
methods of most objects call the
renderer with no reference to
whether the points lie in the
data space or figure space: that
is, paths and points are usually
specified in figure/pixel
coordinates or some transformed
version thereof, which makes it
near impossible to construct
interactive representations
absent Python kernel callbacks.

                                    What I'd love to see is some

enhancement of the backend
framework where there are some
extra flags and information
passed to the renderer: i.e. for
each draw command, we need to
know whether the drawn object
should be linked to static
figure coordinates or to dynamic
axes/data coordinates.

                                    I've been in touch with Cyrille

Rossant from the vispy team,
Chris Beaumont from the Glue
team, and Matt Sundwuist from
the plotly team, all of whom
asked if there might be a way to
use what I’ve done with mpld3 to
enable matplotlib to export into
their own front-end format. I
didn’t start mpld3 with that
sort of extensibility in mind,
but I’m starting to invest some
time thinking about how to
design that.

                                    With the current matplotlib

package, I think there are two
ways to accomplish it: one is to
create a general backend-like
interface based on the figure
introspection that mpld3
currently uses. The artist
elements in each figure contain
enough information to be able to
infer whether the elements
should move & zoom with the
axes or not. The problem is, a
lot of elements (like legends,
axes aspects, etc.) are not
fully established until the
draw() command is called, so
there are a few ugly hacks
required to make it happen.

                                    The other option is to use an

even uglier hack, and wrap the
current backend framework with
an object that somehow links
back into the figure and infers
from the draw_*() commands
whether the
path/point/marker/etc. should be
drawn in static figure
coordinates or in dynamic axes
coordinates. I’ve started a
simple prototype backend
translator which has a renderer
class that uses inspect
back-trace the stack and
accomplish this: It’s really
ugly, and I’m not particularly
proud about it, but I think it’s
the current best way to
accomplish the desired behavior.

                                    Ugly hacks aside, I think all

of this points to a general
desire for a new type of
backend-like hook that can
export dynamic plot elements in
data coordinates, and static
plot elements in figure
coordinates. An enhancement in
that direction could pave the
way for a lot of interesting
interactive front-ends to
matplotlib figures.

                                    Anyway - if any of you have

suggestions or responses to
this, I’d love to hear them!
Thanks,

Jake

                   -- _ |\/|o _|_ _. _ | | \.__ __|__|_|_ _ _ ._ _ | ||(_| |(_|(/_| |_/|(_)(/_|_ |_|_)(_)(_)| | | [http://www.droettboom.com](http://www.droettboom.com)

                                On Tue, Jan

14, 2014 at 9:11 AM, Michael
Droettboom <mdroe@…31…>
wrote:

                                      Jake: I'd definitely like

to get you into one of these
calls at some point. If
you’re able to pop in late,
that would still be great –
or we can save that for
another date. Trying to get
Japan, three NA timezones and
the UK all together is
challenging :wink:

                                      In any event, with Thomas,

Ben, Michiel and myself
confirmed, I think that’s
enough to go ahead, and
hopefully others who have yet
to respond can join as well.

                                      Mike




                                          On 01/14/2014 11:57 AM,

Jacob Vanderplas wrote:

                                            I'll

probably not be able to
swing 6am on the west
coast, but other folks
are more important for
this call, I think :slight_smile:
Jake

                   -- _ |\/|o _|_ _. _ | | \.__ __|__|_|_ _ _ ._ _ | ||(_| |(_|(/_| |_/|(_)(/_|_ |_|_)(_)(_)| | | [http://www.droettboom.com](http://www.droettboom.com)

                                              On

Tue, Jan 14, 2014 at
8:51 AM, Benjamin Root
<ben.root@…553…>
wrote:

                                                  That

would actually
work a little bit
better for me… I
just have to
remember to get
into work a little
bit earlier.

                                                CenturyLink Cloud:

The Leader in
Enterprise Cloud
Services.

                                                Learn Why More

Businesses Are
Choosing CenturyLink
Cloud For

                                                Critical Workloads,

Development
Environments &
Everything In
Between.

                                                Get a Quote or Start

a Free Trial Today.

                                                [http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk](http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk)

                                                Matplotlib-devel

mailing list

                                                Matplotlib-devel@...923....sourceforge.net

                                                [https://lists.sourceforge.net/lists/listinfo/matplotlib-devel](https://lists.sourceforge.net/lists/listinfo/matplotlib-devel)

Ben

                                                      On

Tue, Jan 14,
2014 at 11:36
AM, Michael
Droettboom <mdroe@…55…31…>
wrote:

                                                      I'm fine with

starting the
meeting an
hour early.
How about
others?

                                                      Mike


                                                      On 01/14/2014

04:57 AM,
Michiel de
Hoon wrote:

                                                      > I can

join this
Thursday if we
start with the
discussion on
timers.

                                                      > If we can

start 1 hour
earlier (14:00
UTC, 9 am ET,
23:00 in
Japan) that
would be even
better.

                                                      > -Michiel.

                                                      >

                                                      >

                                                      >

                                                      >

                                                      > On Mon,

1/13/14,
Michael
Droettboom
<mdroe@…31… >
wrote:

                                                      >

                                                      >

Subject:
[matplotlib-devel]
Meeting…?

                                                      >   To: "matplotlib-devel@lists.sourceforge.net" <matplotlib-devel@...101...ists.sourceforge.net>



                                                      >   Date:

Monday,
January 13,
2014, 11:36 AM

                                                      >

                                                      >   It's

probably a
good time to
schedule

                                                      >   another

matplotlib
Google
Hangout.

                                                      >

                                                      >   Is this

Thursday at
1500 UTC (10
am ET) too
short notice
for

                                                      >   the

usual

                                                      >

candidates?

                                                      >

                                                      >   I know

there was
discussion of
getting
Michiel de
Hoon on

                                                      >   today

(which I

                                                      >   just

saw,
unfortunately).
Is there
another time
in the

                                                      >   future

that

                                                      >   works

for you,
Michiel?

                                                      >

                                                      >   Mike

                                                      >

                                                      >   --

                                                      >

                                                      >       _

                                                      >   |\/|o

| . _ | |
._
||| _

                                                      >   _ ._ _

                                                      >   |

||(|
|(|(/|
|/|()(/|_
||)()()| |
|

                                                      >

                                                      >   [http://www.droettboom.com](http://www.droettboom.com)

                                                      >

                                                      >

                                                      >  

                                                      >

CenturyLink
Cloud: The
Leader in
Enterprise
Cloud
Services.

                                                      >   Learn

Why More
Businesses Are
Choosing
CenturyLink
Cloud

                                                      >   For

                                                      >

Critical
Workloads,
Development
Environments
&

                                                      >

Everything In
Between.

                                                      >   Get a

Quote or Start
a Free Trial
Today.

                                                      >   [http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk](http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk)



                                                      >  

                                                      >

Matplotlib-devel
mailing list

                                                      >   Matplotlib-devel@lists.sourceforge.net

                                                      >   [https://lists.sourceforge.net/lists/listinfo/matplotlib-devel](https://lists.sourceforge.net/lists/listinfo/matplotlib-devel)



                                                      >





                                                      --

                                                                   
  _

                                                      |\/|o _|_  _.

_ | | .__
||| _
_ ._ _

                                                      |  ||(_|

|(|(/|
|/|()(/|
||)()()| |
|

                                                      [http://www.droettboom.com](http://www.droettboom.com)

                                                      CenturyLink

Cloud: The
Leader in
Enterprise
Cloud
Services.

                                                      Learn Why More

Businesses Are
Choosing
CenturyLink
Cloud For

                                                      Critical

Workloads,
Development
Environments
&
Everything In
Between.

                                                      Get a Quote or

Start a Free
Trial Today.

                                                      [http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk](http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk)

                                                      Matplotlib-devel

mailing list

                                                      Matplotlib-devel@lists.sourceforge.net

                                                      [https://lists.sourceforge.net/lists/listinfo/matplotlib-devel](https://lists.sourceforge.net/lists/listinfo/matplotlib-devel)
17

Yes, there's room.

This is the link that Google gives me for the calendar:

https://www.google.com/calendar/feeds/79hk8jhvlks8jn8ds4ri1e6q4g%40group.calendar.google.com/public/basic

Let me know whether that works...

Mike

···

On 01/15/2014 11:42 PM, Tony Yu wrote:

Is there room for a lurker in the hangout? If so, is there a link for joining the hangout? (I tried using the hangout calendar link from awhile back, but I couldn't get my google calendar to recognize it.)

Cheers,
-Tony

On Wed, Jan 15, 2014 at 9:57 AM, Michael Droettboom <mdroe@...31... > <mailto:mdroe@…31…>> wrote:

    On 01/14/2014 03:49 PM, Chris Beaumont wrote:

    I have another long-simmering feature request along these lines:
    if Matplotlib were to evolve a formal DOM-like figure structure
    like mentioned above, it would be cool if this structure retained
    more semantic information about the visualization itself. By
    this, I mean that many high-level commands like hist, scatter,
    etc. spawn a bunch of low level artists like rectangles and
    circles. After these methods exit, it's hard/impossible to
    introspect a Figure and diagnose that it, for example, is a
    histogram and not a bar chart.

    I agree, this would be very nice to have, but is (obviously) a
    much bigger step from what we currently have in matplotlib. Along
    those lines, if the "frontend" (i.e. the pylab interface and even
    the OO interface) basically generate a tree, then you could
    serialize that tree (here XML would be a great fit, don't knock
    it), and even non-Python based tools could transform it into
    something else.

    Retaining a better high-level description of a plot (which
    probably amounts to creating more compound artist types) would
    make it easier to build tools like mpld3 and other cool things
    that involve runtime editing or optimization of tree-like data
    structures.

    Agreed.

    No argument against this at all from me -- but knowing how much
    work this would be, the obstacle there is just finding the time to
    do it. It would be a significant rewrite...

    Mike

    chris

    On Tue, Jan 14, 2014 at 3:34 PM, Jacob Vanderplas >> <jakevdp@...1179... <mailto:jakevdp@…1179…>> wrote:

        On Tue, Jan 14, 2014 at 12:04 PM, Michael Droettboom >> <mdroe@...31... <mailto:mdroe@…31…>> wrote:

            I hope all of the above makes sense...

        Definitely makes sense: what I've built-up in mpld3 is
        essentially something that mimics this sort of visitor
        pattern, though it misses some things because of the
        draw-time difficulties you mention.
        I think a two-stage draw() would be a _very_ helpful
        restructure. Currently, I'm forced to achieve this result by
        writing a png to a throwaway byte-stream...
           Jake

            Mike

            On 01/14/2014 01:30 PM, Jacob Vanderplas wrote:

            Thanks - we'll make it happen at some point.

            Perhaps I can give the seed for a discussion: the stuff
            I've been doing with mpld3 is a lot of fun, but it's
            fundamentally limited by the fact that I have to dig
            around the internals of the figure object to find the
            relevant information to construct a plot representation.
             I may be able to do the same thing by creating a
            backend, but the problem is that the draw() methods of
            most objects call the renderer with no reference to
            whether the points lie in the data space or figure
            space: that is, paths and points are usually specified
            in figure/pixel coordinates or some transformed version
            thereof, which makes it near impossible to construct
            interactive representations absent Python kernel callbacks.

            What I'd love to see is some enhancement of the backend
            framework where there are some extra flags and
            information passed to the renderer: i.e. for each draw
            command, we need to know whether the drawn object should
            be linked to static figure coordinates or to dynamic
            axes/data coordinates.

            I've been in touch with Cyrille Rossant from the vispy
            team, Chris Beaumont from the Glue team, and Matt
            Sundwuist from the plotly team, all of whom asked if
            there might be a way to use what I've done with mpld3 to
            enable matplotlib to export into their own front-end
            format. I didn't start mpld3 with that sort of
            extensibility in mind, but I'm starting to invest some
            time thinking about how to design that.

            With the current matplotlib package, I think there are
            two ways to accomplish it: one is to create a general
            backend-like interface based on the figure introspection
            that mpld3 currently uses. The artist elements in each
            figure contain enough information to be able to infer
            whether the elements should move & zoom with the axes or
            not. The problem is, a lot of elements (like legends,
            axes aspects, etc.) are not fully established until the
            draw() command is called, so there are a few ugly hacks
            required to make it happen.

            The other option is to use an even uglier hack, and wrap
            the current backend framework with an object that
            somehow links back into the figure and infers from the
            draw_*() commands whether the path/point/marker/etc.
            should be drawn in static figure coordinates or in
            dynamic axes coordinates. I've started a simple
            prototype backend translator which has a renderer class
            that uses ``inspect`` back-trace the stack and
            accomplish this: It's really ugly, and I'm not
            particularly proud about it, but I think it's the
            current best way to accomplish the desired behavior.

            Ugly hacks aside, I think all of this points to a
            general desire for a new type of backend-like hook that
            can export dynamic plot elements in data coordinates,
            and static plot elements in figure coordinates. An
            enhancement in that direction could pave the way for a
            lot of interesting interactive front-ends to matplotlib
            figures.

            Anyway - if any of you have suggestions or responses to
            this, I'd love to hear them! Thanks,
              Jake

            On Tue, Jan 14, 2014 at 9:11 AM, Michael Droettboom >>> <mdroe@...31... <mailto:mdroe@…31…>> wrote:

                Jake: I'd definitely like to get you into one of
                these calls at some point. If you're able to pop in
                late, that would still be great -- or we can save
                that for another date. Trying to get Japan, three
                NA timezones and the UK all together is challenging :wink:

                In any event, with Thomas, Ben, Michiel and myself
                confirmed, I think that's enough to go ahead, and
                hopefully others who have yet to respond can join as
                well.

                Mike

                On 01/14/2014 11:57 AM, Jacob Vanderplas wrote:

                I'll probably not be able to swing 6am on the west
                coast, but other folks are more important for this
                call, I think :slight_smile:
                   Jake

                On Tue, Jan 14, 2014 at 8:51 AM, Benjamin Root >>>> <ben.root@...553... <mailto:ben.root@…553…>> wrote:

                    That would actually work a little bit better
                    for me... I just have to remember to get into
                    work a little bit earlier.

                    Ben

                    On Tue, Jan 14, 2014 at 11:36 AM, Michael >>>> Droettboom <mdroe@...31... >>>> <mailto:mdroe@…31…>> wrote:

                        I'm fine with starting the meeting an hour
                        early. How about others?

                        Mike

                        On 01/14/2014 04:57 AM, Michiel de Hoon wrote:
                        > I can join this Thursday if we start with
                        the discussion on timers.
                        > If we can start 1 hour earlier (14:00
                        UTC, 9 am ET, 23:00 in Japan) that would be
                        even better.
                        > -Michiel.
                        >
                        > --------------------------------------------
                        > On Mon, 1/13/14, Michael Droettboom >>>> <mdroe@...31... <mailto:mdroe@…31…>> >>>> wrote:
                        >
                        > Subject: [matplotlib-devel] Meeting...?
                        > To:
                        "matplotlib-devel@lists.sourceforge.net
                        <mailto:matplotlib-devel@lists.sourceforge.net>"
                        <matplotlib-devel@lists.sourceforge.net
                        <mailto:matplotlib-devel@lists.sourceforge.net>>
                        > Date: Monday, January 13, 2014, 11:36 AM
                        >
                        > It's probably a good time to schedule
                        > another matplotlib Google Hangout.
                        >
                        > Is this Thursday at 1500 UTC (10 am ET)
                        too short notice for
                        > the usual
                        > candidates?
                        >
                        > I know there was discussion of getting
                        Michiel de Hoon on
                        > today (which I
                        > just saw, unfortunately). Is there
                        another time in the
                        > future that
                        > works for you, Michiel?
                        >
                        > Mike
                        >
                        > --
                        >
                        > _
                        > >\/|o _|_ _. _ | | \.__ __|__|_|_ _
                        > _ ._ _
                        > > >>(_| |(_|(/_| |_/|(_)(/_|_
                        >_|_)(_)(_)| | |
                        >
                        > http://www.droettboom.com
                        >
                        ------------------------------------------------------------------------------
                        > CenturyLink Cloud: The Leader in
                        Enterprise Cloud Services.
                        > Learn Why More Businesses Are Choosing
                        CenturyLink Cloud
                        > For
                        > Critical Workloads, Development
                        Environments &
                        > Everything In Between.
                        > Get a Quote or Start a Free Trial Today.
                        >
                        http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk
                        >
                        _______________________________________________
                        > Matplotlib-devel mailing list
                        > Matplotlib-devel@lists.sourceforge.net
                        <mailto:Matplotlib-devel@lists.sourceforge.net>
                        >
                        matplotlib-devel List Signup and Options
                        >

                        --
                              _
                        >\/|o _|_ _. _ | | \.__ __|__|_|_ _ _ ._ _
                        > >>(_| |(_|(/_| |_/|(_)(/_|_ |_|_)(_)(_)| | |

                        http://www.droettboom.com

                        ------------------------------------------------------------------------------
                        CenturyLink Cloud: The Leader in Enterprise
                        Cloud Services.
                        Learn Why More Businesses Are Choosing
                        CenturyLink Cloud For
                        Critical Workloads, Development
                        Environments & Everything In Between.
                        Get a Quote or Start a Free Trial Today.
                        http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk
                        _______________________________________________
                        Matplotlib-devel mailing list
                        Matplotlib-devel@lists.sourceforge.net
                        <mailto:Matplotlib-devel@lists.sourceforge.net>
                        matplotlib-devel List Signup and Options

                    ------------------------------------------------------------------------------
                    CenturyLink Cloud: The Leader in Enterprise
                    Cloud Services.
                    Learn Why More Businesses Are Choosing
                    CenturyLink Cloud For
                    Critical Workloads, Development Environments &
                    Everything In Between.
                    Get a Quote or Start a Free Trial Today.
                    http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk
                    _______________________________________________
                    Matplotlib-devel mailing list
                    Matplotlib-devel@lists.sourceforge.net
                    <mailto:Matplotlib-devel@lists.sourceforge.net>
                    matplotlib-devel List Signup and Options

                -- _
                >\/|o _|_ _. _ | | \.__ __|__|_|_ _ _ ._ _
                > >>(_| |(_|(/_| |_/|(_)(/_|_ |_|_)(_)(_)| | |

                http://www.droettboom.com

            -- _
            >\/|o _|_ _. _ | | \.__ __|__|_|_ _ _ ._ _
            > >>(_| |(_|(/_| |_/|(_)(/_|_ |_|_)(_)(_)| | |

            http://www.droettboom.com

        ------------------------------------------------------------------------------
        CenturyLink Cloud: The Leader in Enterprise Cloud Services.
        Learn Why More Businesses Are Choosing CenturyLink Cloud For
        Critical Workloads, Development Environments & Everything In
        Between.
        Get a Quote or Start a Free Trial Today.
        http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk
        _______________________________________________
        Matplotlib-devel mailing list
        Matplotlib-devel@lists.sourceforge.net
        <mailto:Matplotlib-devel@lists.sourceforge.net>
        matplotlib-devel List Signup and Options

    -- ************************************
    Chris Beaumont
    Graduate Student
    Institute for Astronomy
    University of Hawaii at Manoa
    2680 Woodlawn Drive
    Honolulu, HI 96822
    www.ifa.hawaii.edu/~beaumont <http://www.ifa.hawaii.edu/~beaumont&gt;
    ************************************

    -- _
    >\/|o _|_ _. _ | | \.__ __|__|_|_ _ _ ._ _
    > >>(_| |(_|(/_| |_/|(_)(/_|_ |_|_)(_)(_)| | |

    http://www.droettboom.com

    ------------------------------------------------------------------------------
    CenturyLink Cloud: The Leader in Enterprise Cloud Services.
    Learn Why More Businesses Are Choosing CenturyLink Cloud For
    Critical Workloads, Development Environments & Everything In Between.
    Get a Quote or Start a Free Trial Today.
    http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk
    _______________________________________________
    Matplotlib-devel mailing list
    Matplotlib-devel@lists.sourceforge.net
    <mailto:Matplotlib-devel@lists.sourceforge.net>
    matplotlib-devel List Signup and Options

--
                    _

\/|o _|_ _. _ | | \.__ __|__|_|_ _ _ ._ _
>>(_| |(_|(/_| |_/|(_)(/_|_ |_|_)(_)(_)| | |

18

My bad: I expected clicking the link to add the calendar to my GCal
account. I didn't realize I had to copy the address to the "Add by URL"
dialog under "Other Calendars". So primitive :stuck_out_tongue:

Thanks for letting me lurk during the meeting,
-Tony

···

On Thu, Jan 16, 2014 at 7:30 AM, Michael Droettboom <mdroe@...31...> wrote:

Yes, there's room.

This is the link that Google gives me for the calendar:

https://www.google.com/calendar/feeds/79hk8jhvlks8jn8ds4ri1e6q4g%40group.calendar.google.com/public/basic

Let me know whether that works...

Mike

19

So that’s how you do it?! That has been driving me nuts!

···

On Thu, Jan 16, 2014 at 9:35 PM, Tony Yu <tsyu80@…149…> wrote:

On Thu, Jan 16, 2014 at 7:30 AM, Michael Droettboom <mdroe@…31…> wrote:

Yes, there’s room.

  This is the link that Google gives me for the calendar:

https://www.google.com/calendar/feeds/79hk8jhvlks8jn8ds4ri1e6q4g%40group.calendar.google.com/public/basic

  Let me know whether that works...



  Mike

My bad: I expected clicking the link to add the calendar to my GCal account. I didn’t realize I had to copy the address to the “Add by URL” dialog under “Other Calendars”. So primitive :stuck_out_tongue:

Thanks for letting me lurk during the meeting,

-Tony