Re: Emacs-orgmode Digest, Vol 30, Issue 39

Re: Emacs-orgmode Digest, Vol 30, Issue 39

[Stephan Schmitt]

Eric Schulte wrote:
> On Tuesday, August 26, at 14:36, Sebastian Rose wrote:
>  > The buffer is modified after C-M-g and emacs asks if I want to save it.
> 
> That is certainly something to figure out.
> 

This happens already if you press tab with the cursor 
located inside a table, so this is probably an org issue.

Eric, did you know that Carsten provided a function called 
org-table-to-lisp in org-6.06b?  I think with this you can 
refactor your code to be more modular and thus better 
maintainable and extensible.

Regards,
	Stephan

Re: Emacs-orgmode Digest, Vol 30, Issue 39

[Eric Schulte]

On Tuesday, August 26, at 19:57, Stephan Schmitt wrote:
 > Eric Schulte wrote:
 > > On Tuesday, August 26, at 14:36, Sebastian Rose wrote:
 > >  > The buffer is modified after C-M-g and emacs asks if I want to save it.
 > > 
 > > That is certainly something to figure out.
 > > 
 > 
 > This happens already if you press tab with the cursor 
 > located inside a table, so this is probably an org issue.

I had been using tab `org-cycle' in some cases to position the cursor
on selected columns, so that is probably the cause.

 > Eric, did you know that Carsten provided a function called 
 > org-table-to-lisp in org-6.06b?  I think with this you can 
 > refactor your code to be more modular and thus better 
 > maintainable and extensible.

I was not aware of that function.  It should certainly help, both in
cleaning up the code, and hopefully removing the reliance on
`org-cycle'.  Using `org-table-to-lisp' should also eliminate the 'Not
in table data field' error that Sebastian mentioned.  Thanks -- Eric

 > Regards,
 > 	Stephan

-- 
schulte

org-plot (generating graphs from org-mode)

[Eric Schulte]

On Tuesday, August 26, at 11:08, Eric Schulte wrote:
 > On Tuesday, August 26, at 19:57, Stephan Schmitt wrote:
 >  > Eric, did you know that Carsten provided a function called 
 >  > org-table-to-lisp in org-6.06b?  I think with this you can 
 >  > refactor your code to be more modular and thus better 
 >  > maintainable and extensible.

I have switched over to using `org-table-to-lisp' for all table
manipulation, and (at least in my environment), it seems to have made
the graphing functions faster, and much more reliable (no longer do I
have to run a function twice to generate a graph).

I believe this should fix most of the bugs mentioned previously.

Also, I have finished support for 3d graphing.  The updated version is
up on my git repo at

http://github.com/eschulte/org-plot/tree/master

or

git clone git://github.com/eschulte/org-plot.git

Of course if anyone has any comments, problems, or suggestions please
let me know.

Thanks -- Eric

-- 
schulte

Re: org-plot (generating graphs from org-mode)

[Sebastian Rose]

Hi Eric,


Great! This version fixes ALL the bugs I mentioned in my previous
email on emacs 23 !!!

I.e.

   - It always renders the plot after the first C-M-g
   - All three tables work

And it's much faster indeed!


Thank's, it's fantastic :-)


  - Sebastian


Eric Schulte wrote:
> On Tuesday, August 26, at 11:08, Eric Schulte wrote:
>  > On Tuesday, August 26, at 19:57, Stephan Schmitt wrote:
>  >  > Eric, did you know that Carsten provided a function called 
>  >  > org-table-to-lisp in org-6.06b?  I think with this you can 
>  >  > refactor your code to be more modular and thus better 
>  >  > maintainable and extensible.
> 
> I have switched over to using `org-table-to-lisp' for all table
> manipulation, and (at least in my environment), it seems to have made
> the graphing functions faster, and much more reliable (no longer do I
> have to run a function twice to generate a graph).
> 
> I believe this should fix most of the bugs mentioned previously.
> 
> Also, I have finished support for 3d graphing.  The updated version is
> up on my git repo at
> 
> http://github.com/eschulte/org-plot/tree/master
> 
> or
> 
> git clone git://github.com/eschulte/org-plot.git
> 
> Of course if anyone has any comments, problems, or suggestions please
> let me know.
> 
> Thanks -- Eric
> 

Re: org-plot (generating graphs from org-mode)

[Sebastian Rose]

Hi again,


Forgot to say: buffer-modified-p problem is gone too!!!


Woweeee!

Eric Schulte wrote:
> On Tuesday, August 26, at 11:08, Eric Schulte wrote:
>  > On Tuesday, August 26, at 19:57, Stephan Schmitt wrote:
>  >  > Eric, did you know that Carsten provided a function called 
>  >  > org-table-to-lisp in org-6.06b?  I think with this you can 
>  >  > refactor your code to be more modular and thus better 
>  >  > maintainable and extensible.
> 
> I have switched over to using `org-table-to-lisp' for all table
> manipulation, and (at least in my environment), it seems to have made
> the graphing functions faster, and much more reliable (no longer do I
> have to run a function twice to generate a graph).
> 
> I believe this should fix most of the bugs mentioned previously.
> 
> Also, I have finished support for 3d graphing.  The updated version is
> up on my git repo at
> 
> http://github.com/eschulte/org-plot/tree/master
> 
> or
> 
> git clone git://github.com/eschulte/org-plot.git
> 
> Of course if anyone has any comments, problems, or suggestions please
> let me know.
> 
> Thanks -- Eric
> 

Re: org-plot (generating graphs from org-mode)

[Eric Schulte]

On Wednesday, August 27, at 11:30, Sebastian Rose wrote:
 > Hi Eric,
 > 
 > 
 > Great! This version fixes ALL the bugs I mentioned in my previous
 > email on emacs 23 !!!
 > 

Wonderful!! thanks for testing this out.

-- 
schulte

Re: org-plot (generating graphs from org-mode)

[William Henney]

Hi Eric

On Tue, Aug 26, 2008 at 6:09 PM, Eric Schulte <schulte.eric@gmail.com> wrote:
> Of course if anyone has any comments, problems, or suggestions please
> let me know.

Org-plot is great! Thank you very much for writing this. I just found
a couple of small problems, as detailed below.

Cheers

Will

** Bugs in org-plot
*** [2008-08-28 Thu] Finding the table to plot
    =org-plot/goto-nearest-table= searches backward first, which means
    that if point is located in an options line for a following table,
    then it will find the previous table instead. This is usually
    undesired.

*** [2008-08-28 Thu] Columns with multi-word text
    Consider the following table:

#+PLOT: title:"Citas" ind:0 deps:(3) type:2d with:histograms set:"yrange [0:]"
    | Sede      | Max cites | H-index |
    |-----------+-----------+---------|
    | Chile     |    257.72 |   21.39 |
    | Leeds     |    165.77 |   19.68 |
    | São Paolo |     71.00 |   11.50 |
    | Stockholm |    134.19 |   14.33 |
    | Morelia   |    257.56 |   17.67 |

    In theory, this should plot the third column. However, the data
    point for São Paolo comes from the second column
    instead. Presumably this is because of the embedded space in "Sõo
    Paolo". If I replace it with a unicode non-breaking space,
    "São Paolo", then it works correctly.


-- 

 Dr William Henney, Centro de Radioastronomía y Astrofísica,
 Universidad Nacional Autónoma de México, Campus Morelia

Re: org-plot (generating graphs from org-mode)

[Eric Schulte]

Thanks for the feedback, much appreciated.  A new version of org-plot is
up on the git repo http://github.com/eschulte/org-plot/tree/master

detailed comments below...

"William Henney" <whenney@gmail.com> writes:
> *** [2008-08-28 Thu] Finding the table to plot
>     =org-plot/goto-nearest-table= searches backward first, which means
>     that if point is located in an options line for a following table,
>     then it will find the previous table instead. This is usually
>     undesired.

I changed this so it no longer looks backwards

> *** [2008-08-28 Thu] Columns with multi-word text
>     Consider the following table:
>
> #+PLOT: title:"Citas" ind:0 deps:(3) type:2d with:histograms set:"yrange [0:]"
>     | Sede      | Max cites | H-index |
>     |-----------+-----------+---------|
>     | Chile     |    257.72 |   21.39 |
>     | Leeds     |    165.77 |   19.68 |
>     | São Paolo |     71.00 |   11.50 |
>     | Stockholm |    134.19 |   14.33 |
>     | Morelia   |    257.56 |   17.67 |
>
>     In theory, this should plot the third column. However, the data
>     point for São Paolo comes from the second column
>     instead. Presumably this is because of the embedded space in "Sõo
>     Paolo". If I replace it with a unicode non-breaking space,
>     "São Paolo", then it works correctly.

you're right, I am now quoting non-text columns on export which takes
care of the problem you mentioned here.  Also, org-plot is now using to
gnuplot `xticlabel' command to use the ind columns values as labels in
the case that they are text.  (for an example plot your table from above
with the ind:1 option).  If you don't mind, I would like to add the
table you sent above to the example.org file in org-plot to demonstrate
this feature.

Thanks -- Eric

Re: org-plot (generating graphs from org-mode)

[Eric Schulte]

Thanks for the feedback, much appreciated.  A new version of org-plot is
up on the git repo http://github.com/eschulte/org-plot/tree/master

detailed comments below...

"William Henney" <whenney@gmail.com> writes:
> *** [2008-08-28 Thu] Finding the table to plot
>     =org-plot/goto-nearest-table= searches backward first, which means
>     that if point is located in an options line for a following table,
>     then it will find the previous table instead. This is usually
>     undesired.

I changed this so it no longer looks backwards

> *** [2008-08-28 Thu] Columns with multi-word text
>     Consider the following table:
>
> #+PLOT: title:"Citas" ind:0 deps:(3) type:2d with:histograms set:"yrange [0:]"
>     | Sede      | Max cites | H-index |
>     |-----------+-----------+---------|
>     | Chile     |    257.72 |   21.39 |
>     | Leeds     |    165.77 |   19.68 |
>     | São Paolo |     71.00 |   11.50 |
>     | Stockholm |    134.19 |   14.33 |
>     | Morelia   |    257.56 |   17.67 |
>
>     In theory, this should plot the third column. However, the data
>     point for São Paolo comes from the second column
>     instead. Presumably this is because of the embedded space in "Sõo
>     Paolo". If I replace it with a unicode non-breaking space,
>     "São Paolo", then it works correctly.

you're right, I am now quoting non-text columns on export which takes
care of the problem you mentioned here.  Also, org-plot is now using to
gnuplot `xticlabel' command to use the `ind' column's values as labels
in the case that they are text.  (to demonstrate, plot your table from
above with the ind:1 option).  If you don't mind, I would like to add
the table you sent above to the example.org file in org-plot to
demonstrate this feature.

Thanks -- Eric

Re: org-plot (generating graphs from org-mode)

[Seweryn Kokot]

Hi Eric,

I just want to ask if it is necessary to redefine 'debug' function in
org-plot.el? Because when I put the following lines into .emacs file:

(require 'org-plot)
(add-hook 'org-mode-hook
		  (lambda ()
			(local-set-key (kbd "\M-\C-g")
						   'org-plot/gnuplot)))

whenever I invoke an error, instead of getting *Backtrace* buffer I
get a strange message in the echo line.

-- 
regards,
Seweryn Kokot

Re: Re: org-plot (generating graphs from org-mode)

[Eric Schulte]

oh, not at all,

I was using it for convenience during my own debugging, and didn't
realize I was overwriting an existing function.  Please feel free to
remove the `debug' definition from org-plot.el (I will remove it from
the version in my git repo right now).

Thanks -- Eric

Re: org-plot (generating graphs from org-mode)

[William Henney]

Hi Eric

On Sun, Aug 31, 2008 at 1:14 PM, Eric Schulte <schulte.eric@gmail.com> wrote:
>
> Thanks for the feedback, much appreciated.  A new version of org-plot is
> up on the git repo http://github.com/eschulte/org-plot/tree/master
>

I can confirm that the new version fixes both the problems that I
reported. Furthermore, the new option to use a column of text labels
for the x-axis is wonderful. Thanks!

> If you don't mind, I would like to add the
> table you sent above to the example.org file in org-plot to demonstrate
> this feature.
>

Yes, by all means, go ahead.

I've run into a couple more minor problems that might be worth looking
into if you have the time.

**** Writing plot to file
     Two issues here:
     1. The suffix of the output file must exactly match the name of a
        gnuplot terminal type. Thus file:"foo.jpeg" works, but
        file:"foo.jpg" doesn't. Similarly, one needs to use
        ".postscript" instead of ".ps".
     2. Some terminals seem to work, but some don't (there is no error
        message and the ouput file is created, but it is empty):
	- Works :: jpeg, png, svg, postscript, mp
	- Fails :: pdf (OK, this seems to be the only one - weird)
**** Non-ascii characters in strings
     These tend not to work. However, the results vary according to
     the output terminal in use, so I suspect that this may be due to
     bugs in gnuplot. My version is
:	G N U P L O T:
:	Version 4.2 patchlevel 3
:	last modified Mar 2008
:	System: Darwin 9.4.0
     One strange thing is that, even if the .org file is utf8, the
     graphics file seems to get written as latin-1.

Cheers

Will

-- 

 Dr William Henney, Centro de Radioastronomía y Astrofísica,
 Universidad Nacional Autónoma de México, Campus Morelia