1 #+TITLE: org-babel --- facilitating communication between programming languages and people
2 #+SEQ_TODO: PROPOSED TODO STARTED | DONE DEFERRED REJECTED
3 #+OPTIONS: H:3 num:nil toc:2 \n:nil @:t ::t |:t ^:t -:t f:t *:t TeX:t LaTeX:t skip:nil d:(HIDE) tags:not-in-toc
4 #+STARTUP: oddeven hideblocks
8 <p>executable source code blocks in org-mode</p>
12 <img src="images/tower-of-babel.png" alt="images/tower-of-babel.png" />
13 <div id="attr">from <a href="http://www.flickr.com/photos/23379658@N05/" title=""><b>Martijn Streefkerk</b></a></div>
18 Org-Babel makes source-code blocks in Org-Mode executable and allows
19 data to pass seamlessly between different programming languages,
20 Org-Mode constructs (tables, file links, example text) and interactive
24 - The [[* Introduction][Introduction]] :: provides a brief overview of the design and use
25 of Org-Babel including tutorials and examples.
26 - In [[* Getting started][Getting Started]] :: find instructions for installing org-babel
27 into your emacs configuration.
28 - The [[* Tasks][Tasks]] :: section contains current and past tasks roughly ordered
29 by TODO state, then importance or date-completed. This would be
30 a good place to suggest ideas for development.
31 - The [[* Bugs][Bugs]] :: section contains bug reports.
32 - The [[* Tests][Tests]] :: section consists of a large table which can be
33 evaluated to run Org-Babel's functional test suite. This
34 provides a good overview of the current functionality with
35 pointers to example source blocks.
36 - The [[* Sandbox][Sandbox]] :: demonstrates much of the early/basic functionality
37 through commented source-code blocks.
39 Also see the [[file:library-of-babel.org][Library of Babel]], an extensible collection of ready-made
40 and easily-shortcut-callable source-code blocks for handling common
45 Org-Babel enables *communication* between programming languages and
49 - communication between programs :: Data passes seamlessly between
50 different programming languages, Org-Mode constructs (tables,
51 file links, example text) and interactive comint buffers.
52 - communication between people :: Data and calculations are embedded
53 in the same document as notes explanations and reports.
55 ** communication between programs
57 Org-Mode supports embedded blocks of source code (in any language)
58 inside of Org documents. Org-Babel allows these blocks of code to be
59 executed from within Org-Mode with natural handling of their inputs
63 with both scalar, file, and table output
65 *** reading information from tables
67 *** reading information from other source blocks (disk usage in your home directory)
69 This will work for Linux and Mac users, not so sure about shell
70 commands for windows users.
72 To run place the cursor on the =#+begin_src= line of the source block
73 labeled directory-pie and press =\C-c\C-c=.
75 #+srcname: directories
76 #+begin_src bash :results replace
77 cd ~ && du -sc * |grep -v total
80 #+resname: directories
82 | 11882808 | "Documents" |
83 | 8210024 | "Downloads" |
84 | 879800 | "Library" |
87 | 5307664 | "Pictures" |
96 #+srcname: directory-pie
97 #+begin_src R :var dirs = directories :session R-pie-example
98 pie(dirs[,1], labels = dirs[,2])
106 *** operations in/on tables
108 #+tblname: grades-table
109 | student | grade | letter |
110 |---------+-------+--------|
117 #+TBLFM: $2='(sbe random-score-generator)::$3='(sbe assign-grade (score $2))
119 #+srcname: assign-grade
120 #+begin_src ruby :var score=99
131 #+srcname: random-score-generator
136 #+srcname: show-distribution
137 #+begin_src R :var grades=grades-table :session *R*
145 ** communication between people
146 Quick overview of Org-Mode's exportation abilities, with links to the
147 online Org-Mode documentation, a focus on source-code blocks, and the
148 exportation options provided by Org-Babel.
150 *** Interactive tutorial
151 This would demonstrate applicability to Reproducible Research, and
152 Literate Programming.
154 *** Tests embedded in documentation
155 org-babels own functional tests are contained in a large org-mode
156 table, allowing the test suite to be run be evaluation of the table
157 and the results to be collected in the same table.
159 *** Emacs initialization files stored in Org-Mode buffers
160 Using `org-babel-tangle' it is possible to embed your Emacs
161 initialization into org-mode files. This allows for folding,
162 note-taking, todo's etc... embedded with the source-code of your Emacs
163 initialization, and through org-mode's publishing features aids in
164 sharing your customizations with others.
166 It may be worthwhile to create a fork of Phil Hagelberg's
167 [[http://github.com/technomancy/emacs-starter-kit/tree/master][emacs-starter-kit]] which uses literate org-mode files for all of the
168 actual elisp customization. These org-mode files could then be
169 exported to html and used to populate the repositories wiki on [[http://github.com/][github]].
174 *** code evaluation (comint buffer sessions and external processes)
175 There are two main ways to evaluate source blocks with org-babel.
177 - external :: By default (if the =:session= header argument is not
178 present) all source code blocks are evaluated in
179 external processes. In these cases an external process
180 is used to evaluate the source-code blocks.
181 - session :: Session based evaluation uses persistent sessions in
182 comint buffers. Sessions can be used across multiple
183 source blocks setting and accessing variables in the
186 Evaluating source blocks in sessions also allows for
187 interaction with the code. To jump to the session of a
188 source block use the `org-babel-pop-to-session' command
189 or press =M-[down]= while inside of a source code block.
190 When called with a prefix argument
191 `org-babel-pop-to-session' will evaluate all header
192 arguments before jumping to the source-code block.
194 *** results (values and outputs)
195 Either the *value* or the *output* of source code blocks can be
196 collected after evaluation.
198 - value :: The default way to collect results from a source-code block
199 is to return the value of the last statement in the block.
200 This can be thought of as the return value of the block.
201 In this case any printed output of the block is ignored.
202 This can be though of a similar to a "functional" value of
204 - output :: Another way of generating results from a source-code block
205 is to collect the output generated by the execution of the
206 block. In this case all printed output is collected
207 throughout the execution of the block. This can be
208 thought of as similar to a "script" style of evaluation.
212 Add the following lines to your .emacs, replacing the path as
213 appropriate. A good place to check that things are up and running
214 would then be [[#sandbox][the sandbox]].
215 #+begin_src emacs-lisp
216 (add-to-list 'load-path "/path/to/org-babel/lisp")
217 (require 'org-babel-init)
222 ** TODO source-name visible in LaTeX and html exports
223 Maybe this should be done in backend specific manners.
225 The listings package may provide for naming a source-code block...
227 ** STARTED new results types (org, html, latex)
228 Thanks to Tom Short for this recommendation.
230 - raw or org :: in which case the results are implemented raw, unquoted
231 into the org-mode file. This would also handle links as
233 - html :: the results are inserted inside of a #+BEGIN_HTML block
234 - latex :: the results are inserted inside of a #+BEGIN_LATEX block
237 : #+begin_src R :session *R* :results org
238 : cat("***** This is a table\n")
239 : cat("| 1 | 2 | 3 |\n")
240 : cat("[[http://google.com][Google it here]]\n"
244 : ***** This is a table
246 [[http://google.com][: Google it here]]
248 We actually might want to remove the =#+resname= line if the results
249 type is org-mode, not sure... Either way I don't think there is a
250 good way to capture/remove org type results.
254 Added a =raw= results header argument, which will insert the results
255 of a source-code block into an org buffer un-escaped. Also, if the
256 results look like a table, then the table will be aligned.
258 #+srcname: raw-table-demonstration
259 #+begin_src ruby :results output raw
260 puts "| root | square |"
263 puts "| #{n} | #{n*n} |"
280 ** PROPOSED raise elisp error when source-blocks return errors
281 Not sure how/if this would work, but it may be desirable.
283 ** PROPOSED allow `anonymous' function block with function call args?
284 My question here is simply whether we're going to allow
285 #+begin_src python(arg=ref)
289 but with preference given to
290 #+srcname blockname(arg=ref)
291 ** PROPOSED allow :result as synonym for :results?
292 ** PROPOSED allow 'output mode to return stdout as value?
293 Maybe we should allow this. In fact, if block x is called
294 with :results output, and it references blocks y and z, then
295 shouldn't the output of x contain a concatenation of the outputs of
296 y and z, together with x's own output? That would raise the
297 question of what happens if y is defined with :results output and z
298 with :results value. I guess z's (possibly vector/tabular) output
299 would be inside a literal example block containing the whole lot.
300 ** PROPOSED optional timestamp for output
301 Add option to place an (inactive) timestamp at the #+resname, to
302 record when that output was generated.
304 *** source code block timestamps (optional addition)
305 [Eric] If we did this would we then want to place a timestamp on the
306 source-code block, so that we would know if the results are
307 current or out of date? This would have the effect of caching the
308 results of calculations and then only re-running if the
309 source-code has changed. For the caching to work we would need to
310 check not only the timestamp on a source-code block, but also the
311 timestamps of any tables or source-code blocks referenced by the
312 original source-code block.
314 [Dan] I do remember getting frustrated by Sweave always having to
315 re-do everything, so this could be desirable, as long as it's easy
316 to over-ride of course. I'm not sure it should be the default
317 behaviour unless we are very confident that it works well.
319 **** maintaining source-code block timestamps
320 It may make sense to add a hook to `org-edit-special' which could
321 update the source-code blocks timestamp. If the user edits the
322 contents of a source-code block directly I can think of no
323 efficient way of maintaining the timestamp.
324 ** TODO make tangle files read-only?
325 With a file-local variable setting, yea that makes sense. Maybe
326 the header should reference the related org-mode file.
328 ** TODO support for working with =*Org Edit Src Example*= buffers [4/6]
329 *** STARTED Patch against org source.
330 I've worked on several related changes to source code edit buffer
331 behaviour in the org core. My current patch (below) does the
332 following. Detailed explanation / working notes are below.
333 - C-x s offers to save edit buffers
334 - C-x C-c offers to save edit buffers
335 - C-x k warns that you're killing an edit buffer
336 - If you do kill an edit buffer, the overlay in the parent buffer is removed
337 - Edit buffers are named *Org Src <orgbuf>[<lang>]*, where
338 <orgbuf> is the name of the org-mode buffer containing this
339 source code block, and lang is the language major mode. The
340 latter might be unnecessary?
343 diff --git a/lisp/org-src.el b/lisp/org-src.el
344 index 2083c77..2be21e6 100644
345 --- a/lisp/org-src.el
346 +++ b/lisp/org-src.el
347 @@ -113,7 +113,7 @@ but which mess up the display of a snippet in Org exported files.")
349 (defvar org-src-mode-map (make-sparse-keymap))
350 (define-key org-src-mode-map "\C-c'" 'org-edit-src-exit)
351 -(define-key org-src-mode-map "\C-x\C-s" 'org-edit-src-save)
352 +;; (define-key org-src-mode-map "\C-x\C-s" 'org-edit-src-save)
353 (defvar org-edit-src-force-single-line nil)
354 (defvar org-edit-src-from-org-mode nil)
355 (defvar org-edit-src-picture nil)
356 @@ -168,7 +168,8 @@ the edited version."
357 (if (boundp 'org-edit-src-overlay)
358 (org-delete-overlay org-edit-src-overlay)))
359 (kill-buffer buffer))
360 - (setq buffer (generate-new-buffer "*Org Edit Src Example*"))
361 + (setq buffer (generate-new-buffer
362 + (concat "*Org Src " (file-name-nondirectory buffer-file-name) "[" lang "]*")))
363 (setq ovl (org-make-overlay beg end))
364 (org-overlay-put ovl 'face 'secondary-selection)
365 (org-overlay-put ovl 'edit-buffer buffer)
366 @@ -186,8 +187,7 @@ the edited version."
367 '(display nil invisible nil intangible nil))
368 (org-do-remove-indentation)
369 (let ((org-inhibit-startup t))
373 (set (make-local-variable 'org-edit-src-force-single-line) single)
374 (set (make-local-variable 'org-edit-src-from-org-mode) org-mode-p)
376 @@ -201,6 +201,7 @@ the edited version."
377 (org-set-local 'org-edit-src-end-marker end)
378 (org-set-local 'org-edit-src-overlay ovl)
379 (org-set-local 'org-edit-src-nindent nindent)
381 (and org-edit-src-persistent-message
382 (org-set-local 'header-line-format msg)))
384 @@ -400,12 +401,13 @@ the language, a switch telling of the content should be in a single line."
385 (defun org-edit-src-exit ()
386 "Exit special edit and protect problematic lines."
388 - (unless (string-match "\\`*Org Edit " (buffer-name (current-buffer)))
389 - (error "This is not an sub-editing buffer, something is wrong..."))
390 + (unless org-edit-src-from-org-mode
391 + (error "This is not a sub-editing buffer, something is wrong..."))
392 (let ((beg org-edit-src-beg-marker)
393 (end org-edit-src-end-marker)
394 (ovl org-edit-src-overlay)
395 (buffer (current-buffer))
396 + (buffer-file-name nil)
397 (nindent org-edit-src-nindent)
399 (untabify (point-min) (point-max))
400 @@ -464,6 +466,17 @@ the language, a switch telling of the content should be in a single line."
401 (goto-char (min p (point-max)))
402 (message (or msg ""))))
404 +(defun org-src-mode-configure-buffer ()
405 + (setq buffer-offer-save t)
406 + (setq buffer-file-name
407 + (concat (buffer-file-name (marker-buffer org-edit-src-beg-marker))
408 + "[" (buffer-name) "]"))
409 + (setq write-contents-functions '(org-edit-src-save))
410 + (org-add-hook 'kill-buffer-hook
411 + '(lambda () (org-delete-overlay org-edit-src-overlay)) nil 'local))
413 +(org-add-hook 'org-src-mode-hook 'org-src-mode-configure-buffer)
417 ;; arch-tag: 6a1fc84f-dec7-47be-a416-64be56bea5d8
421 **** Detailed working notes to go with that patch
422 ***** Recap of current org-src-mode
424 If you use C-c ' to work on code in a begin_source block, the code
425 buffer is put in minor mode org-src-mode, which features the
426 following two useful key-bindings:
428 | C-x s | org-edit-src-save | save the code in the source code block in the parent org file |
429 | C-c ' | org-edit-src-exit | return to the parent org file with new code |
431 Furthermore, while the edit buffer is alive, the originating code
432 block is subject to a special overlay which links to the edit
433 buffer when you click on it.
435 This is all excellent, and I use it daily, but I think there's
436 still a couple of improvements that we should make.
439 C-x k kills the buffer without questions; the overlay remains, but
440 now links to a deleted buffer.
441 ***** Proposed bug II
442 C-x C-c kills a modified edit buffer silently, without offering to
443 save your work. I have lost work like that a number of times
445 ***** Proposed bug III
446 C-x s does not offer to save a modified edit buffer
447 ***** Notes on solution
448 ****** write-contents-functions
449 A good start seems to be to use org-src-mode-hook to add
450 org-edit-src-save to the write-contents-functions list. This
451 means that when it comes to saving, org-edit-src-save will be
452 called and no subsequent attempt will be made to save the buffer
453 in the normal way. (This should obviate the remapping of C-x C-s
454 to org-edit-src-save in org-src.el)
455 ****** buffer-offer-save
456 We also want to set this to t.
458 ****** Where does this get us?
460 - C-x s still does *not* offer to save the edit buffer. That's
461 because buffer-file-name is nil.
463 - C-x C-c does ask us whether we want to save the
464 edit buffer. However, since buffer-file-name is nil it asks us
465 for a file name. The check in org-edit-src-exit throws an error
466 unless the buffer is named '* Org Edit '...
468 - C-x k kills the buffer silently, leaving a broken overlay
469 link. If buffer-file-name were set, it would have warned that
470 the buffer was modified.
472 ****** buffer-file-name
473 So, that all suggests that we need to set buffer-file-name, even
474 though we don't really want to associate this buffer with a file
475 in the normal way. As for the file name, my current suggestion
476 is parent-org-filename[edit-buffer-name].
478 [I had to move the (org-src-mode) call to the end of
479 org-edit-src-code to make sure that the required variables were
480 defined when the hook was called.]
482 ****** And so where are we now?
483 - C-x s *does* offer to save the edit buffer, but in saving
484 produces a warning that the edit buffer is modified.
485 - C-x k now gives a warning that the edit buffer is modified
487 - C-x C-c is working as desired, except that again we get
488 warnings that the edit buffer is modified, once when we save,
489 and again just before exiting emacs.
490 - And C-c ' now issues a warning that the edit buffer is
491 modified when we leave it, which we don't want.
492 ****** So, we need to get rid of the buffer modification warnings.
493 I've made buffer-file-name nil inside the let binding in
496 - C-x s behaves as desired, except that as was already the case,
497 the edit buffer is always considered modified, and so repeated
498 invocations keep saving it.
499 - As was already the case, C-x k always gives a warning that the
500 edit buffer has been modified.
501 - C-x C-c is as desired (offers to save the edit buffer) except
502 that it warns of the modified buffer just before exiting.
503 - C-c ' is as it should be (silent)
505 We've got the desired behaviour, at the cost of being forced to
506 assign a buffer-file-name to the edit buffer. The consequence is
507 that the edit buffer is considered to always be modified, since
508 a file of that name is never actually written to (doesn't even
509 exist). I couldn't see a way to trick emacs into believing that
510 the buffer was unmodified since last save. But in any case, I
511 think there's an argument that these modifications warnings are
512 a good thing, because one should not leave active edit buffers
513 around: you should always have exited with C-c ' first.
515 *** DONE name edit buffer according to #+srcname (and language?)
516 See above patch agains org.
517 *** DONE optionally evaluate header references when we switch to =*Org Edit Src*= buffer
518 That seems to imply that the header references need to be evaluated
519 and transformed into the target language object when we hit C-c ' to
520 enter the *Org Edit Src* buffer [DED]
522 Good point, I heartily agree that this should be supported [Eric]
524 (or at least before the first time we attempt to evaluate code in that
525 buffer -- I suppose there might be an argument for lazy evaluation, in
526 case someone hits C-c ' but is "just looking" and not actually
527 evaluating anything.) Of course if evaluating the reference is
528 computationally intensive then the user might have to wait before they
529 get the *Org Edit Src* buffer. [DED]
531 I fear that it may be hard to anticipate when the references will be
532 needed, some major-modes do on-the-fly evaluation while the buffer is
533 being edited. I think that we should either do this before the buffer
534 is opened or not at all, specifically I think we should resolve
535 references if the user calls C-c ' with a prefix argument. Does that
536 sound reasonable? [Eric]
540 [Dan] So now that we have org-src-mode and org-src-mode-hook, I guess
541 org-babel should do this by using the hook to make sure that, when C-c
542 C-' is issued on a source block, any references are resolved and
543 assignments are made in the appropriate session.
545 #+tblname: my-little-table
549 #+srcname: resolve-vars-on-edit
550 #+begin_src ruby :var table=my-little-table :results silent :session test
551 table.size.times.do |n|
557 *** TODO set buffer-local-process variables appropriately [DED]
558 I think something like this would be great. You've probably
559 already thought of this, but just to note it down: it would be really
560 nice if org-babel's notion of a buffer's 'session/process' played
561 nicely with ESS's notion of the buffer's session/process. ESS keeps
562 the current process name for a buffer in a buffer-local variable
563 ess-local-process-name. So one thing we will probably want to do is
564 make sure that the *Org Edit Src Example* buffer sets that variable
567 I had not thought of that, but I agree whole heartedly. [Eric]
569 Once this is done every variable should be able to dump regions into
570 their inferior-process buffer using major-mode functions.
571 *** REJECTED send code to inferior process
572 Another thought on this topic: I think we will want users to send
573 chunks of code to the interpreter from within the *Org Edit Src*
574 buffer, and I think that's what you have in mind already. In ESS that
575 is done using the ess-eval-* functions. [DED]
577 I think we can leave this up to the major-mode in the source code
578 buffer, as almost every source-code major mode will have functions for
579 doing things like sending regions to the inferior process. If
580 anything we might need to set the value of the buffer local inferior
581 process variable. [Eric]
583 *** DONE some possible requests/proposed changes for Carsten [4/4]
584 While I remember, some possible requests/proposed changes for Carsten
585 come to mind in that regard:
587 **** DONE Remap C-x C-s to save the source to the org buffer?
588 I've done this personally and I find it essential. I'm using
589 #+begin_src emacs-lisp
590 (defun org-edit-src-save ()
591 "Update the parent org buffer with the edited source code, save
592 the parent org-buffer, and return to the source code edit
601 (define-key org-exit-edit-mode-map "\C-x\C-s" 'org-edit-src-save)
605 I think this is great, but I think it should be implemented in the
608 **** DEFERRED Rename buffer and minor mode?
609 Something shorter than *Org Edit Src Example* for the buffer
610 name. org-babel is bringing org's source code interaction to a
611 level of maturity where the 'example' is no longer
612 appropriate. And if further keybindings are going to be added to
613 the minor mode then maybe org-edit-src-mode is a better name than
616 Maybe we should name the buffer with a combination of the source
617 code and the session. I think that makes sense.
619 [ES] Are you also suggesting a new org-edit-src minor mode?
620 [DED] org-exit-edit-mode is a minor mode that already exists:
622 Minor mode installing a single key binding, "C-c '" to exit special edit.
624 org-edit-src-save now has a binding in that mode, so I guess all
625 I'm saying at this stage is that it's a bit of a misnomer. But
626 perhaps we will also have more functionality to add to that minor
627 mode, making it even more of a misnomer. Perhaps something like
628 org-src-mode would be better.
629 **** DONE Changed minor mode name and added hooks
631 **** DONE a hook called when the src edit buffer is created
632 This should be implemented in the org-mode core
633 ** TODO resolve references to other org buffers/files
634 This would allow source blocks to call upon tables, source-blocks,
635 and results in other org buffers/files.
638 - [[file:lisp/org-babel-ref.el::TODO%20allow%20searching%20for%20names%20in%20other%20buffers][org-babel-ref.el:searching-in-other-buffers]]
639 - [[file:lisp/org-babel.el::defun%20org-babel%20find%20named%20result%20name][org-babel.el#org-babel-find-named-result]]
640 ** TODO resolve references to other non-org files
641 - tabular data in .csv, .tsv etc format
642 - files of interpreted code: anything stopping us giving such files
643 similar status to a source code block?
644 - Would be nice to allow org and non-org files to be remote
645 ** TODO figure out how to handle errors during evaluation
646 I expect it will be hard to do this properly, but ultimately it
647 would be nice to be able to specify somewhere to receive STDERR,
648 and to be warned if it is non-empty.
650 Probably simpler in non-session evaluation than session? At least
651 the mechanism will be different I guess.
653 R has a try function, with error handling, along the lines of
654 python. I bet ruby does too. Maybe more of an issue for functional
655 style; in my proposed scripting style the error just gets dumped to
656 the org buffer and the user is thus alerted.
657 ** STARTED figure out how to handle graphic output
659 This is listed under [[* graphical output][graphical output]] in out objectives.
661 This should take advantage of the =:results file= option, and
662 languages which almost always produce graphical output should set
663 =:results file= to true by default (this is currently done for the
664 gnuplot and ditaa languages). That would handle placing these results
665 in the buffer. Then if there is a combination of =silent= and =file=
666 =:results= headers we could drop the results to a temp buffer and pop
669 Display of file results is addressed in the [[* =\C-c \C-o= to open results of source block][open-results-task]].
671 *** TODO R graphics to screen means session evaluation
672 If R graphical output is going to screen then evaluation must be
673 in a session, otherwise the graphics will disappear as soon as the
676 *** Adding to a discussion started in email
677 I'm not deeply wedded to these ideas, just noting them down. I'm
678 probably just thinking of R and haven't really thought about how
679 this fits with the other graphics-generating languages.
681 > I used the approach below to get graphical file output
682 > today, which is one idea at least. Maybe it could be linked up with
683 > your :results file variable. (Or do we need a :results image for R?)
686 I don't think we need a special image results variable, but I may be
687 missing what the code below accomplishes. Would the task I added about
688 adding org-open-at-point functionality to source code blocks take care
691 Dan: I'm not sure. I think the ability for a script to generate both
692 text and graphical output might be a natural expectation, at least for
698 > #+srcname: cohort-scatter-plots-2d(org_babel_graphical_output_file="cohort-scatter-plots-2d.png")
700 > if(exists("org_babel_output_file"))
701 > png(filename=org_babel_graphical_output_file, width=1000, height=1000)
702 > ## plotting code in here
703 > if(exists("org_babel_graphical_output_file")) dev.off()
706 Dan: Yes, the results :file option is nice for dealing with graphical
707 output, and that could well be enough. Something based on the scheme
708 above would have a couple of points in its favour:
709 1. It's easy to switch between output going to on-screen graphics and
710 output going to file: Output will go to screen unless a string variable
711 with a standard name (e.g. ""org_babel_graphical_output_file"")
712 exists in which case it will go to the file indicated by the value
714 2. The block can return a result / script output, as well as produce
717 In interactive use we might want to allow the user to choose between
718 screen and file output. In non-interactive use such as export, it
719 would be file output (subject to the :exports directives).
720 ** TODO Finalise behaviour regarding vector/scalar output
721 *** DONE Stop spaces causing vector output
722 This simple example of multilingual chaining produces vector output if
723 there are spaces in the message and scalar otherwise.
727 #+srcname: msg-from-R(msg=msg-from-python)
729 paste(msg, "und R", sep=" ")
733 : org-babel speaks elisp y python und R
735 #+srcname: msg-from-python(msg=msg-from-elisp)
740 #+srcname: msg-from-elisp(msg="org-babel speaks")
741 #+begin_src emacs-lisp
742 (concat msg " elisp")
744 ** STARTED share org-babel [1/6]
745 how should we share org-babel?
746 *** DONE post to org-mode
747 *** TODO post to ess mailing list
748 *** TODO create a org-babel page on worg
749 *** TODO create a short screencast demonstrating org-babel in action
750 *** PROPOSED a peer-reviewed publication?
752 The following notes are biased towards statistics-oriented
753 journals because ESS and Sweave are written by people associated
754 with / in statistics departments. But I am sure there are suitable
755 journals out there for an article on using org mode for
756 reproducible research (and literate programming etc).
758 Clearly, we would invite Carsten to be involved with this.
760 ESS is described in a peer-reviewed journal article:
761 Emacs Speaks Statistics: A Multiplatform, Multipackage Development Environment for Statistical Analysis [Abstract]
762 Journal of Computational & Graphical Statistics 13(1), 247-261
763 Rossini, A.J, Heiberger, R.M., Sparapani, R.A., Maechler, M., Hornik, K. (2004)
764 [[http://www.amstat.org/publications/jcgs.cfm][Journal of Computational and Graphical Statistics]]
766 Also [[http://www.amstat.org/publications/jss.cfm][Journal of Statistical Software]] Established in 1996, the
767 Journal of Statistical Software publishes articles, book reviews,
768 code snippets, and software reviews. The contents are freely
769 available online. For both articles and code snippets, the source
770 code is published along with the paper.
774 Friedrich Leisch and Anthony J. Rossini. Reproducible statistical
775 research. Chance, 16(2):46-50, 2003. [ bib ]
779 Friedrich Leisch. Sweave: Dynamic generation of statistical reports
780 using literate data analysis. In Wolfgang Härdle and Bernd Rönz,
781 editors, Compstat 2002 - Proceedings in Computational Statistics,
782 pages 575-580. Physica Verlag, Heidelberg, 2002. ISBN 3-7908-1517-9.
786 We could also look at the Journals publishing these [[http://www.reproducibleresearch.net/index.php/RR_links#Articles_about_RR_.28chronologically.29][Reproducible
789 *** PROPOSED an article in [[http://journal.r-project.org/][The R Journal]]
790 This looks good. It seems that their main topic to software tools for
791 use by R programmers, and Org-babel is certainly that.
793 *** existing similar tools
794 try to collect pointers to similar tools
796 Reproducible Research
797 - [[http://en.wikipedia.org/wiki/Sweave][Sweave]]
800 - [[http://www.cs.tufts.edu/~nr/noweb/][Noweb]]
801 - [[http://www-cs-faculty.stanford.edu/~knuth/cweb.html][Cweb]]
802 - [[http://www.lri.fr/~filliatr/ocamlweb/][OCamlWeb]]
804 Meta Functional Programming
807 Programmable Spreadsheet
811 we need to think up some good examples
813 **** interactive tutorials
814 This could be a place to use [[* org-babel assertions][org-babel assertions]].
816 for example the first step of a tutorial could assert that the version
817 of the software-package (or whatever) is equal to some value, then
818 source-code blocks could be used with confidence (and executed
819 directly from) the rest of the tutorial.
821 **** answering a text-book question w/code example
822 org-babel is an ideal environment enabling both the development and
823 demonstrationg of the code snippets required as answers to many
826 **** something using tables
827 maybe something along the lines of calculations from collected grades
830 Maybe something like the following which outputs sizes of directories
831 under the home directory, and then instead of the trivial =emacs-lisp=
832 block we could use an R block to create a nice pie chart of the
836 #+begin_src bash :results replace
840 #+begin_src emacs-lisp :var sizes=sizes :results replace
843 ** TODO command line execution
844 Allow source code blocks to be called form the command line. This
845 will be easy using the =sbe= function in [[file:lisp/org-babel-table.el][org-babel-table.el]].
847 This will rely upon [[* resolve references to other buffers][resolve references to other buffers]].
849 ** TODO inline source code blocks [3/5]
850 Like the =\R{ code }= blocks
852 not sure what the format should be, maybe just something simple
853 like =src_lang[]{}= where lang is the name of the source code
854 language to be evaluated, =[]= is optional and contains any header
855 arguments and ={}= contains the code.
857 (see [[* (sandbox) inline source blocks][the-sandbox]])
859 *** DONE evaluation with \C-c\C-c
860 Putting aside the header argument issue for now we can just run these
861 with the following default header arguments
862 - =:results= :: silent
863 - =:exports= :: results
865 *** DONE inline exportation
866 Need to add an interblock hook (or some such) through org-exp-blocks
867 *** DONE header arguments
868 We should make it possible to use header arguments.
870 *** TODO fontification
871 we should color these blocks differently
873 *** TODO refine html exportation
874 should use a span class, and should show original source in tool-tip
875 ** TODO LoB: re-implement plotting and analysis functions from org-R
876 I'll do this soon, now that we things are a bit more settled and we
877 have column names in R.
878 ** PROPOSED Creating presentations
879 The [[mairix:t:@@9854.1246500519@gamaville.dokosmarshall.org][recent thread]] containing posts by Nick Dokos and Sebastian
880 Vaubán on exporting to beamer looked very interesting, but I
881 haven't had time to try it out yet. I would really like it if,
882 eventually, we can generate a presentation (with graphics generated
883 by code blocks) from the same org file that contains all the notes
884 and code etc. I just wanted that to be on record in this document;
885 I don't have anything more profound to say about it at the moment,
886 and I'm not sure to what extent it is an org-babel issue.
887 ** PROPOSED conversion between org-babel and noweb (e.g. .Rnw) format
888 I haven't thought about this properly. Just noting it down. What
889 Sweave uses is called "R noweb" (.Rnw).
891 I found a good description of noweb in the following article (see
892 the [[http://www.cs.tufts.edu/~nr/pubs/lpsimp.pdf][pdf]]).
894 I think there are two parts to noweb, the construction of
895 documentation and the extraction of source-code (with notangle).
897 *documentation*: org-mode handles all of our documentation needs in
898 a manner that I believe is superior to noweb.
900 *source extraction* At this point I don't see anyone writing large
901 applications with 100% of the source code contained in org-babel
902 files, rather I see org-babel files containing things like
903 - notes with active code chunks
904 - interactive tutorials
905 - requirements documents with code running test suites
906 - and of course experimental reports with the code to run the
907 experiment, and perform analysis
909 Basically I think the scope of the programs written in org-babel
910 (at least initially) will be small enough that it wont require the
911 addition of a tangle type program to extract all of the source code
912 into a running application.
914 On the other hand, since we already have named blocks of source
915 code which reference other blocks on which they rely, this
916 shouldn't be too hard to implement either on our own, or possibly
917 relying on something like noweb/notangle.
919 ** PROPOSED support for passing paths to files between source blocks
920 Maybe this should be it's own result type (in addition to scalars and
921 vectors). The reason being that some source-code blocks (for example
922 ditaa or anything that results in the creation of a file) may want to
923 pass a file path back to org-mode which could then be inserted into
924 the org-mode buffer as a link to the file...
926 This would allow for display of images upon export providing
927 functionality similar to =org-exp-blocks= only in a more general
929 ** DEFERRED Support rownames and other org babel table features?
931 The full org table features are detailed in the manual [[http://orgmode.org/manual/Advanced-features.html#Advanced-features][here]].
934 Perhaps add a :rownames header arg. This would be an integer
935 (usually 1) which would have the effect of post-processing all the
936 variables created in the R session in the following way: if the
937 integer is j, set the row names to the contents of column j and
938 delete column j. Perhaps it is artificial to allow this integer to
939 take any value other than 1. The default would be nil which would
940 mean no such behaviour.
942 Actually I don't know about that. If multiple variables are passed
943 in, it's not appropriate to alter them all in the same way. The
944 rownames specification would normally refer to just one of the
945 variables. For now maybe just say this has to be done in R. E.g.
947 #+TBLNAME: sample-sizes
948 | collection | size | exclude | include | exclude2 | include2 |
949 |-----------------+------+---------+---------+----------+----------|
950 | 58C | 2936 | 8 | 2928 | 256 | 2680 |
951 | MS | 5852 | 771 | 5081 | 771 | 5081 |
952 | NBS | 2929 | 64 | 2865 | 402 | 2527 |
953 | POBI | 2717 | 1 | 2716 | 1 | 2716 |
954 | 58C+MS+NBS+POBI | | | 13590 | | 13004 |
955 #+TBLFM: @2$4=@2$2 - @2$3::@2$6=@2$2 - @2$5::@3$4=@3$2-@3$3::@3$6=@3$2 - @3$5::@4$4=@4$2 - @4$3::@4$6=@4$2 - @4$5::@5$4=@5$2-@5$3::@5$6=@5$2 - @5$5::@6$4=vsum(@2$4..@5$4)::@6$6=vsum(@2$6..@5$6)
957 #+srcname: make-size-table(size=sample-sizes)
959 rownames(size) <- size[,1]
965 [I don't think it's as problematic as this makes out]
966 This is non-trivial, but may be worth doing, in particular to
967 develop a nice framework for sending data to/from R.
969 In R, indexing vector elements, and rows and columns, using
970 strings rather than integers is an important part of the
972 - elements of a vector may have names
973 - matrices and data.frames may have "column names" and "row names"
974 which can be used for indexing
975 - In a data frame, row names *must* be unique
983 > mat <- matrix(1:4, nrow=2, ncol=2, dimnames=list(c("r1","r2"), c("c1","c2")))
988 > # The names are separate from the data: they do not interfere with operations on the data
995 > df <- data.frame(var1=1:26, var2=26:1, row.names=letters)
997 [1] 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1
1003 So it's tempting to try to provide support for this in org-babel. For example
1004 - allow R to refer to columns of a :var reference by their names
1005 - When appropriate, results from R appear in the org buffer with "named
1008 However none (?) of the other languages we are currently supporting
1009 really have a native matrix type, let alone "column names" or "row
1010 names". Names are used in e.g. python and perl to refer to entries
1013 It currently seems to me that support for this in org-babel would
1014 require setting rules about when org tables are considered to have
1015 named columns/fields, and ensuring that (a) languages with a notion
1016 of named columns/fields use them appropriately and (b) languages
1017 with no such notion do not treat then as data.
1019 - Org allows something that *looks* like column names to be separated
1021 - Org also allows a row to *function* as column names when special
1022 markers are placed in the first column. An hline is unnecessary
1023 (indeed hlines are purely cosmetic in org [correct?]
1024 - Org does not have a notion of "row names" [correct?]
1026 The full org table functionality exeplified [[http://orgmode.org/manual/Advanced-features.html#Advanced-features][here]] has features that
1027 we would not support in e.g. R (like names for the row below).
1029 **** Initial statement: allow tables with hline to be passed as args into R
1030 This doesn't seem to work at the moment (example below). It would
1031 also be nice to have a natural way for the column names of the org
1032 table to become the column names of the R data frame, and to have
1033 the option to specify that the first column is to be used as row
1034 names in R (these must be unique). But this might require a bit of
1039 | col1 | col2 | col3 |
1040 |------+---------+------|
1048 #+begin_src R :var tabel=egtable :colnames t
1053 | "col1" | "col2" | "col3" |
1054 |--------+-----------+--------|
1056 | 4 | "schulte" | 6 |
1059 Another example is in the [[*operations%20in%20on%20tables][grades example]].
1060 ** DEFERRED use textConnection to pass tsv to R?
1061 When passing args from the org buffer to R, the following route is
1062 used: arg in buffer -> elisp -> tsv on file -> data frame in R. I
1063 think it would be possible to avoid having to write to file by
1064 constructing an R expression in org-babel-R-assign-elisp, something
1067 #+begin_src emacs-lisp
1068 (org-babel-R-input-command
1069 (format "%s <- read.table(textConnection(\"%s\"), sep=\"\\t\", as.is=TRUE)"
1070 name (orgtbl-to-tsv value '(:sep "\t" :fmt org-babel-R-quote-tsv-field))))
1073 I haven't tried to implement this yet as it's basically just
1074 fiddling with something that works. The only reason for it I can
1075 think of would be efficiency and I haven't tested that.
1077 This Didn't work after an initial test. I still think this is a
1078 good idea (I also think we should try to do something similar when
1079 writing out results frmo R to elisp) however as it wouldn't result
1080 in any functional changes I'm bumping it down to deferred for
1085 #+tblname: quick-test
1088 #+srcname: quick-test-src-blk
1089 #+begin_src R :var vec=quick-test
1099 ** DEFERRED Rework Interaction with Running Processes [2/5]
1100 *** DONE robust to errors interrupting execution
1102 #+srcname: long-runner-ruby
1103 #+begin_src ruby :results silent
1105 :patton_is_an_grumpy
1108 *** DEFERRED use =C-g= keyboard-quit to push processing into the background
1109 This may be possible using the `run-with-timer' command.
1111 I have no idea how this could work...
1113 #+srcname: long-runner-ruby
1114 #+begin_src ruby :results silent
1116 :patton_is_an_grumpy
1119 *** TODO ability to select which of multiple sessions is being used
1120 Increasingly it is looking like we're going to want to run all
1121 source code blocks in comint buffer (sessions). Which will have
1123 1) allowing background execution
1124 2) maintaining state between source-blocks
1125 - allowing inline blocks w/o header arguments
1128 (like ess-switch-process in .R buffers)
1130 Maybe this could be packaged into a header argument, something
1131 like =:R_session= which could accept either the name of the
1132 session to use, or the string =prompt=, in which case we could use
1133 the =ess-switch-process= command to select a new process.
1135 *** TODO evaluation of shell code as background process?
1136 After C-c C-c on an R code block, the process may appear to
1137 block, but C-g can be used to reclaim control of the .org buffer,
1138 without interrupting the R evalution. However I believe this is not
1139 true of bash/sh evaluation. [Haven't tried other languages] Perhaps
1140 a solution is just to background the individual shell commands.
1142 The other languages (aside from emacs lisp) are run through the
1143 shell, so if we find a shell solution it should work for them as
1146 Adding an ampersand seems to be a supported way to run commands in
1147 the background (see [[http://www.emacswiki.org/emacs/ExecuteExternalCommand#toc4][external-commands]]). Although a more extensible
1148 solution may involve the use of the [[elisp:(progn (describe-function 'call-process-region) nil)][call-process-region]] function.
1150 Going to try this out in a new file [[file:lisp/org-babel-proc.el][org-babel-proc.el]]. This should
1151 contain functions for asynchronously running generic shell commands
1152 in the background, and then returning their input.
1154 **** partial update of org-mode buffer
1155 The sleekest solution to this may be using a comint buffer, and
1156 then defining a filter function which would incrementally interpret
1157 the results as they are returned, including insertion into the
1158 org-mode buffer. This may actually cause more problems than it is
1159 worth, what with the complexities of identifying the types of
1160 incrementally returned results, and the need for maintenance of a
1161 process marker in the org buffer.
1163 **** 'working' spinner
1164 It may be nice and not too difficult to place a spinner on/near the
1165 evaluating source code block
1167 *** TODO conversion of output from interactive shell, R (and python) sessions to org-babel buffers
1168 [DED] This would be a nice feature I think. Although an org-babel
1169 purist would say that it's working the wrong way round... After
1170 some interactive work in a *R* buffer, you save the buffer, maybe
1171 edit out some lines, and then convert it to org-babel format for
1172 posterity. Same for a shell session either in a *shell* buffer, or
1173 pasted from another terminal emulator. And python of course.
1174 ** DEFERRED improve the source-block snippet
1175 any real improvement seems somewhat beyond the ability of yasnippet
1178 [[file:~/src/emacs-starter-kit/src/snippets/text-mode/rst-mode/chap::name%20Chapter%20title][file:~/src/emacs-starter-kit/src/snippets/text-mode/rst-mode/chap::name Chapter title]]
1180 ,#name : Chapter title
1183 ${1:$(make-string (string-width text) ?\=)}
1188 [[file:snippets/org-mode/sb][sb -- snippet]]
1190 waiting for guidance from those more familiar with yasnippets
1192 ** REJECTED re-implement R evaluation using ess-command or ess-execute
1193 I don't have any complaints with the current R evaluation code or
1194 behaviour, but I think it would be good to use the ESS functions
1195 from a political point of view. Plus of course it has the normal
1196 benefits of an API (insulates us from any underlying changes etc). [DED]
1198 I'll look into this. I believe that I looked at and rejected these
1199 functions initially but now I can't remember why. I agree with
1200 your overall point about using API's where available. I will take
1201 a look back at these and either switch to using the ess commands,
1202 or at least articulate under this TODO the reasons for using our
1203 custom R-interaction commands. [Eric]
1207 Lets just replace =org-babel-R-input-command= with =ess-execute=.
1209 I tried this, and although it works in some situations, I find that
1210 =ess-command= will often just hang indefinitely without returning
1211 results. Also =ess-execute= will occasionally hang, and pops up
1212 the buffer containing the results of the command's execution, which
1213 is undesirable. For now these functions can not be used. Maybe
1214 someone more familiar with the ESS code can recommend proper usage
1215 of =ess-command= or some other lower-level function which could be
1216 used in place of [[file:lisp/org-babel-R.el::defun%20org-babel%20R%20input%20command%20command][org-babel-R-input-command]].
1220 #+begin_quote ess-command
1221 (ess-command COM &optional BUF SLEEP NO-PROMPT-CHECK)
1223 Send the ESS process command COM and delete the output
1224 from the ESS process buffer. If an optional second argument BUF exists
1225 save the output in that buffer. BUF is erased before use.
1226 COM should have a terminating newline.
1227 Guarantees that the value of .Last.value will be preserved.
1228 When optional third arg SLEEP is non-nil, `(sleep-for (* a SLEEP))'
1229 will be used in a few places where `a' is proportional to `ess-cmd-delay'.
1232 #+begin_quote ess-execute
1233 (ess-execute COMMAND &optional INVERT BUFF MESSAGE)
1235 Send a command to the ESS process.
1236 A newline is automatically added to COMMAND. Prefix arg (or second arg
1237 INVERT) means invert the meaning of
1238 `ess-execute-in-process-buffer'. If INVERT is 'buffer, output is
1239 forced to go to the process buffer. If the output is going to a
1240 buffer, name it *BUFF*. This buffer is erased before use. Optional
1241 fourth arg MESSAGE is text to print at the top of the buffer (defaults
1242 to the command if BUFF is not given.)
1245 *** out current setup
1247 1) The body of the R source code block is wrapped in a function
1248 2) The function is called inside of a =write.table= function call
1249 writing the results to a table
1250 3) The table is read using =org-table-import=
1251 ** DONE org-bable-tangle: no default extension if one already exists
1252 ** DONE take default values for header args from properties
1253 Use file-wide and subtree wide properties to set default values for
1256 [DED] One thing I'm finding when working with R is that an org file
1257 may contain many source blocks, but that I just want to evaluate a
1258 subset of them. Typically this is in order to take up where I left
1259 off: I need to recreate a bunch of variables in the session
1260 environment. I'm thinking maybe we want to use a tag-based
1261 mechanism similar to :export: and :noexport: to control evaluation
1262 on a per-subtree basis.
1264 *** test-header with properties
1270 Ahh... as is so often the case, just had to wrap
1271 `org-babel-params-from-properties' in a `save-match-data' form.
1273 #+tblname: why-def-props-cause-probs
1276 #+srcname: default-props-implementation
1277 #+begin_src emacs-lisp :tangle no :var my-lis=why-def-props-cause-probs :results silent
1278 (+ (length my-lis) def)
1281 ** DONE new reference syntax *inside* source code blocks
1282 This is from an email discussion on the org-mode mailing list with
1283 Sébastien. The goal here is to mimic the source-block reference style
1284 of Noweb. Upon export and/or tangle these references could be
1285 replaced with the actual body of the referenced source-code block.
1287 See the following for an example.
1289 #+srcname: ems-ruby-print-header
1291 puts "---------------------------header---------------------------"
1294 #+srcname: emacs-ruby-print-footer
1296 puts "---------------------------footer---------------------------"
1299 #+srcname: ems-ruby-print-message
1300 #+begin_src ruby :file ruby-noweb.rb
1301 # <<ems-ruby-print-header>>
1303 # <<ems-ruby-print-footer>>
1306 Upon export the previous source-code block would result in a file
1307 being generated at =ruby-noweb.rb= with the following contents
1309 : puts "---------------------------header---------------------------"
1311 : puts "---------------------------footer---------------------------"
1313 the body of a source-code block with all =<<src-name>>= references
1314 expanded can now be returned by `org-babel-expand-noweb-references'.
1315 This function is now called by default on all source-code blocks on
1318 ** DONE re-work tangling system
1319 Sometimes when tangling a file (e.g. when extracting elisp from a
1320 org-mode file) we want to get nearly every source-code block.
1322 Sometimes we want to only extract those source-code blocks which
1323 reference a indicate that they should be extracted (e.g. traditional
1324 literate programming along the Noweb model)
1326 I'm not sure how we can devise a single simple tangling system that
1327 naturally fits both of these use cases.
1330 the =tangle= header argument will default to =no= meaning source-code
1331 blocks will *not* be exported by default. In order for a source-code
1332 block to be tangled it needs to have an output file specified. This
1333 can happen in two ways...
1335 1) a file-wide default output file can be passed to `org-babel-tangle'
1336 which will then be used for all blocks
1337 2) if the value of the =tangle= header argument is anything other than
1338 =no= or =yes= then it is used as the file name
1340 #+srcname: test-new-tangling
1341 #+begin_src emacs-lisp
1342 (org-babel-load-file "test-tangle.org")
1343 (if (string= test-tangle-advert "use org-babel-tangle for all your emacs initialization files!!")
1351 ** DONE =\C-c \C-o= to open results of source block
1352 by adding a =defadvice= to =org-open-at-point= we can use the common
1353 =\C-c \C-o= keybinding to open the results of a source-code block.
1354 This would be especially useful for source-code blocks which generate
1355 graphical results and insert a file link as the results in the
1356 org-mode buffer. (see [[* figure out how to handle graphic output][TODO figure out how to handle graphic output]]).
1357 This could also act reasonably with other results types...
1359 - file :: use org-open-at-point to open the file
1360 - scalar :: open results unquoted in a new buffer
1361 - tabular :: export the table to a new buffer and open that buffer
1363 when called with a prefix argument the block is re-run
1365 #+srcname: task-opening-results-of-blocks
1366 #+begin_src ditaa :results replace :file blue.png :cmdline -r
1377 [[file:blue.png][blue.png]]
1379 #+srcname: task-open-vector
1380 #+begin_src emacs-lisp
1388 #+srcname: task-open-scalar
1389 #+begin_src ruby :results output
1405 ** DONE Stop spaces causing vector output
1406 This simple example of multilingual chaining produces vector output if
1407 there are spaces in the message and scalar otherwise.
1411 #+srcname: msg-from-R(msg=msg-from-python)
1413 paste(msg, "und R", sep=" ")
1417 : org-babel speaks elisp y python und R
1419 #+srcname: msg-from-python(msg=msg-from-elisp)
1424 #+srcname: msg-from-elisp(msg="org-babel speaks")
1425 #+begin_src emacs-lisp
1426 (concat msg " elisp")
1429 ** DONE add =:tangle= family of header arguments
1431 - no :: don't include source-code block when tangling
1432 - yes :: do include source-code block when tangling
1434 this is tested in [[file:test-tangle.org::*Emacs%20Lisp%20initialization%20stuff][test-tangle.org]]
1436 ** DONE extensible library of callable source blocks
1438 This is covered by the [[file:library-of-babel.org][Library of Babel]], which will contain
1439 ready-made source blocks designed to carry out useful common tasks.
1440 *** Initial statement [Eric]
1441 Much of the power of org-R seems to be in it's helper functions for
1442 the quick graphing of tables. Should we try to re-implement these
1443 functions on top of org-babel?
1445 I'm thinking this may be useful both to add features to org-babel-R and
1446 also to potentially suggest extensions of the framework. For example
1447 one that comes to mind is the ability to treat a source-code block
1448 like a function which accepts arguments and returns results. Actually
1449 this can be it's own TODO (see [[* source blocks as functions][source blocks as functions]]).
1450 *** Objectives [Dan]
1451 - We want to provide convenient off-the-shelf actions
1452 (e.g. plotting data) that make use of our new code evaluation
1453 environment but do not require any actual coding.
1454 *** Initial Design proposal [Dan]
1455 - *Input data* will be specified using the same mechanism as :var
1456 references, thus the input data may come from a table, or
1457 another source block, and it is initially available as an elisp
1459 - We introduce a new #+ line, e.g. #+BABELDO. C-c C-c on that
1460 line will apply an *action* to the referenced data.
1461 - *Actions correspond to source blocks*: our library of available
1462 actions will be a library of org-babel source blocks. Thus the
1463 code for executing an action, and the code for dealing with the
1464 output of the action will be the same code as for executing
1465 source blocks in general
1466 - Optionally, the user can have the relevant source block inserted
1467 into the org buffer after the (say) #+BABELDO line. This will
1468 allow the user to fine tune the action by modifying the code
1469 (especially useful for plots).
1470 - So maybe a #+BABELDO line will have header args
1471 - :data (a reference to a table or source code block)
1472 - :action (or should that be :srcname?) which will be something
1473 like :action pie-chart, referring to a source block which will
1474 be executed with the :data referent passed in using a :var arg.
1475 - :showcode or something controlling whether to show the code
1477 *** Modification to design
1478 I'm implementing this, at least initially, as a new interpreter
1479 named 'babel', which has an empty body. 'babel' blocks take
1480 a :srcname header arg, and look for the source-code block with
1481 that name. They then execute the referenced block, after first
1482 appending their own header args on to the target block's header
1485 If the target block is in the library of babel (a.o.t. e.g. the
1486 current buffer), then the code in the block will refer to the
1487 input data with a name dictated by convention (e.g. __data__
1488 (something which is syntactically legal in all languages...). Thus
1489 the babel block will use a :var __data__ = whatever header arg to
1490 reference the data to be plotted.
1492 ** DONE Column names in R input/output
1493 This has been implemented: Automatic on input to R; optional in
1494 output. Note that this equates column names with the header row in
1495 an org table; whereas org actually has a mechanism whereby a row
1496 with a '!' in the first field defines column names. I have not
1497 attempted to support these org table mechanisms yet. See [[*Support%20rownames%20and%20other%20org%20babel%20table%20features][this
1498 DEFERRED todo item]].
1499 ** DONE use example block for large amounts of stdout output?
1500 We're currently `examplizing' with : at the beginning of the line,
1501 but should larger amounts of output be in a
1502 \#+begin_example...\#+end_example block? What's the cutoff? > 1
1503 line? This would be nice as it would allow folding of lengthy
1504 output. Sometimes one will want to see stdout just to check
1505 everything looks OK, and then fold it away.
1507 I'm addressing this in branch 'examplizing-output'.
1508 Yea, that makes sense. (either that or allow folding of large
1509 blocks escaped with =:=).
1511 Proposed cutoff of 10 lines, we can save this value in a user
1512 customizable variable.
1513 *** DONE add ability to remove such results
1514 ** DONE exclusive =exports= params
1516 #+srcname: implement-export-exclusivity
1523 ** DONE LoB: allow output in buffer
1524 ** DONE allow default header arguments by language
1525 org-babel-default-header-args:lang-name
1527 An example of when this is useful is for languages which always return
1528 files as their results (e.g. [[*** ditaa][ditaa]], and [[*** gnuplot][gnuplot]]).
1529 ** DONE singe-function tangling and loading elisp from literate org-mode file [3/3]
1531 This function should tangle the org-mode file for elisp, and then call
1532 `load-file' on the resulting tangled file.
1534 #+srcname: test-loading-embedded-emacs-lisp
1535 #+begin_src emacs-lisp :results replace
1536 (setq test-tangle-advert nil)
1537 (setq test-tangle-loading nil)
1538 (setq results (list :before test-tangle-loading test-tangle-advert))
1539 (org-babel-load-file "test-tangle.org")
1540 (setq results (list (list :after test-tangle-loading test-tangle-advert) results))
1541 (delete-file "test-tangle.el")
1545 #+resname: test-loading-embedded-emacs-lisp
1546 | :before | nil | nil |
1547 | :after | "org-babel tangles" | "use org-babel-tangle for all your emacs initialization files!!" |
1549 *** DONE add optional language limiter to org-babel-tangle
1550 This should check to see if there is any need to re-export
1552 *** DONE ensure that org-babel-tangle returns the path to the tangled file(s)
1554 #+srcname: test-return-value-of-org-babel-tangle
1555 #+begin_src emacs-lisp :results replace
1556 (mapcar #'file-name-nondirectory (org-babel-tangle-file "test-tangle.org" "emacs-lisp"))
1560 | "test-tangle.el" |
1562 *** DONE only tangle the file if it's actually necessary
1563 ** DONE add a function to jump to a source-block by name
1564 I've had an initial stab at that in org-babel-find-named-block
1565 (library-of-babel branch).
1567 At the same time I introduced org-babel-named-src-block-regexp, to
1568 match src-blocks with srcname.
1570 This is now working with the command
1571 `org-babel-goto-named-source-block', all we need is a good key
1574 ** DONE add =:none= session argument (for purely functional execution) [4/4]
1575 This would allow source blocks to be run in their own new process
1577 - These blocks could then also be run in the background (since we can
1578 detach and just wait for the process to signal that it has terminated)
1579 - We wouldn't be drowning in session buffers after running the tests
1580 - we can re-use much of the session code to run in a more /functional/
1583 While session provide a lot of cool features, like persistent
1584 environments, [[* DONE function to bring up inferior-process buffer][pop-to-session]], and hints at exportation for
1585 org-babel-tangle, they also have some down sides and I'm thinking that
1586 session-based execution maybe shouldn't be the default behavior.
1588 Down-sides to sessions
1589 - *much* more complicated than functional evaluation
1590 - maintaining the state of the session has weird issues
1591 - waiting for evaluation to finish
1592 - prompt issues like [[* TODO weird escaped characters in shell prompt break shell evaluation][shell-prompt-escapes-bug]]
1593 - can't run in background
1594 - litter emacs with session buffers
1598 #+srcname: ruby-task-no-session
1599 #+begin_src ruby :results replace output
1605 #+resname: ruby-task-no-session
1610 #+srcname: task-python-none-session
1611 #+begin_src python :session none :results replace value
1617 #+resname: task-python-none-session
1622 #+srcname: task-session-none-sh
1623 #+begin_src sh :results replace
1628 #+resname: task-session-none-sh
1634 #+srcname: task-no-session-R
1635 #+begin_src R :results replace output
1642 #+resname: task-no-session-R
1646 ** DONE fully purge org-babel-R of direct comint interaction
1647 try to remove all code under the [[file:lisp/org-babel-R.el::functions%20for%20evaluation%20of%20R%20code][;; functions for evaluation of R code]] line
1649 ** DONE Create objects in top level (global) environment [5/5]
1652 *** initial requirement statement [DED]
1653 At the moment, objects created by computations performed in the
1654 code block are evaluated in the scope of the
1655 code-block-function-body and therefore disappear when the code
1656 block is evaluated {unless you employ some extra trickery like
1657 assign('name', object, env=globalenv()) }. I think it will be
1658 desirable to also allow for a style wherein objects that are
1659 created in one code block persist in the R global environment and
1660 can be re-used in a separate block.
1662 This is what Sweave does, and while I'm not saying we have to be
1663 the same as Sweave, it wouldn't be hard for us to provide the same
1664 behaviour in this case; if we don't, we risk undeservedly being
1665 written off as an oddity by some.
1667 IOW one aspect of org-babel is that of a sort of functional
1668 meta-programming language. This is crazy, in a very good
1669 way. Nevertheless, wrt R I think there's going to be a lot of value
1670 in providing for a working style in which the objects are stored in
1671 the R session, rather than elisp/org buffer. This will be a very
1672 familiar working style to lots of people.
1674 There are no doubt a number of different ways of accomplishing
1675 this, the simplest being a hack like adding
1678 for(objname in ls())
1679 assign(objname, get(objname), envir=globalenv())
1682 to the source code block function body. (Maybe wrap it in an on.exit() call).
1684 However this may deserve to be thought about more carefully, perhaps
1685 with a view to having a uniform approach across languages. E.g. shell
1686 code blocks have the same semantics at the moment (no persistence of
1687 variables across code blocks), because the body is evaluated in a new
1688 bash shell process rather than a running shell. And I guess the same
1689 is true for python. However, in both these cases, you could imagine
1690 implementing the alternative in which the body is evaluated in a
1691 persistent interactive session. It's just that it's particularly
1692 natural for R, seeing as both ESS and org-babel evaluate commands in a
1693 single persistent R session.
1697 Thanks for bringing this up. I think you are absolutely correct that we
1698 should provide support for a persistent environment (maybe called a
1699 *session*) in which to evaluate code blocks. I think the current setup
1700 demonstrates my personal bias for a functional style of programming
1701 which is certainly not ideal in all contexts.
1703 While the R function you mention does look like an elegant solution, I
1704 think we should choose an implementation that would be the same across
1705 all source code types. Specifically I think we should allow the user to
1706 specify an optional *session* as a header variable (when not present we
1707 assume a default session for each language). The session name could be
1708 used to name a comint buffer (like the *R* buffer) in which all
1709 evaluation would take place (within which variables would retain their
1710 values --at least once I remove some of the functional method wrappings
1711 currently in place-- ).
1713 This would allow multiple environments to be used in the same buffer,
1714 and once this setup was implemented we should be able to fairly easily
1715 implement commands for jumping between source code blocks and the
1716 related session buffers, as well as for dumping the last N commands from
1717 a session into a new or existing source code block.
1719 Please let me know if you foresee any problems with this proposed setup,
1720 or if you think any parts might be confusing for people coming from
1721 Sweave. I'll hopefully find some time to work on this later in the
1724 *** can functional and interpreted/interactive models coexist?
1726 Even though both of these use the same =*R*= buffer the value of =a=
1727 is not preserved because it is assigned inside of a functional
1730 #+srcname: task-R-sessions
1737 #+srcname: task-R-same-session
1742 This functional wrapper was implemented in order to efficiently return
1743 the results of the execution of the entire source code block. However
1744 it inhibits the evaluation of source code blocks in the top level,
1745 which would allow for persistence of variable assignment across
1746 evaluations. How can we allow *both* evaluation in the top level, and
1747 efficient capture of the return value of an entire source code block
1748 in a language independent manner?
1750 Possible solutions...
1751 1) we can't so we will have to implement two types of evaluation
1752 depending on which is appropriate (functional or imperative)
1753 2) we remove the functional wrapper and parse the source code block
1754 into it's top level statements (most often but not always on line
1755 breaks) so that we can isolate the final segment which is our
1757 3) we add some sort of "#+return" line to the code block
1758 4) we take advantage of each languages support for meta-programming
1759 through =eval= type functions, and use said to evaluate the entire
1760 blocks in such a way that their environment can be combined with the
1761 global environment, and their results are still captured.
1762 5) I believe that most modern languages which support interactive
1763 sessions have support for a =last_result= type function, which
1764 returns the result of the last input without re-calculation. If
1765 widely enough present this would be the ideal solution to a
1766 combination of functional and imperative styles.
1768 None of these solutions seem very desirable, but for now I don't see
1769 what else would be possible.
1771 Of these options I was leaning towards (1) and (4) but now believe
1772 that if it is possible option (5) will be ideal.
1774 **** (1) both functional and imperative evaluation
1776 - can take advantage of built in functions for sending regions to the
1778 - retains the proven tested and working functional wrappers
1781 - introduces the complication of keeping track of which type of
1782 evaluation is best suited to a particular context
1783 - the current functional wrappers may require some changes in order to
1784 include the existing global context
1786 **** (4) exploit language meta-programming constructs to explicitly evaluate code
1788 - only one type of evaluation
1791 - some languages may not have sufficient meta-programming constructs
1793 **** (5) exploit some =last_value= functionality if present
1795 Need to ensure that most languages have such a function, those without
1796 will simply have to implement their own similar solution...
1798 | language | =last_value= function |
1799 |------------+-----------------------------|
1803 | shell | see [[* last command for shells][last command for shells]] |
1804 | emacs-lisp | see [[* emacs-lisp will be a special case][special-case]] |
1806 #+srcname: task-last-value
1811 ***** last command for shells
1812 Do this using the =tee= shell command, and continually pipe the output
1815 Got this idea from the following [[http://linux.derkeiler.com/Mailing-Lists/Fedora/2004-01/0898.html][email-thread]].
1817 suggested from mailing list
1819 #+srcname: bash-save-last-output-to-file
1823 bash -c "$line" | tee /tmp/last.out1
1824 mv /tmp/last.out1 /tmp/last.out
1828 another proposed solution from the above thread
1830 #+srcname: bash-save-in-variable
1833 # so - Save Output. Saves output of command in OUT shell variable.
1843 "^M": " | tee /tmp/h_lastcmd.out ^[k"
1846 export __=/tmp/h_lastcmd.out
1848 If you try it, Alt-k will stand for the old Enter; use "command $__" to
1849 access the last output.
1855 Herculano de Lima Einloft Neto
1858 ***** emacs-lisp will be a special case
1859 While it is possible for emacs-lisp to be run in a console type
1860 environment (see the =elim= function) it is *not* possible to run
1861 emacs-lisp in a different *session*. Meaning any variable set top
1862 level of the console environment will be set *everywhere* inside
1863 emacs. For this reason I think that it doesn't make any sense to
1864 worry about session support for emacs-lisp.
1866 *** Further thoughts on 'scripting' vs. functional approaches
1868 These are just thoughts, I don't know how sure I am about this.
1869 And again, perhaps I'm not saying anything very radical, just that
1870 it would be nice to have some options supporting things like
1871 receiving text output in the org buffer.
1873 I can see that you've already gone some way down the road towards
1874 the 'last value' approach, so sorry if my comments come rather
1875 late. I am concerned that we are not giving sufficient attention
1876 to stdout / the text that is returned by the interpreters. In
1877 contrast, many of our potential users will be accustomed to a
1878 'scripting' approach, where they are outputting text at various
1879 points in the code block, not just at the end. I am leaning
1880 towards thinking that we should have 2 modes of evaluation:
1881 'script' mode, and 'functional' mode.
1883 In script mode, evaluation of a code block would result in *all*
1884 text output from that code block appearing as output in the org
1885 buffer, presumably as an #+begin_example...#+end_example. There
1886 could be an :echo option controlling whether the input commands
1887 also appear in the output. [This is like Sweave].
1889 In functional mode, the *result* of the code block is available as
1890 an elisp object, and may appear in the org buffer as an org
1891 table/string, via the mechanisms you have developed already.
1893 One thing I'm wondering about is whether, in script mode, there
1894 simply should not be a return value. Perhaps this is not so
1895 different from what exists: script mode would be new, and what
1896 exists currently would be functional mode.
1898 I think it's likely that, while code evaluation will be exciting
1899 to people, a large majority of our users in a large majority of
1900 their usage will not attempt to actually use the return value from
1901 a source code block in any meaningful way. In that case, it seems
1902 rather restrictive to only allow them to see output from the end
1905 Instead I think the most accessible way to introduce org-babel to
1906 people, at least while they are learning it, is as an immensely
1907 powerful environment in which to embed their 'scripts', which now
1908 also allows them to 'run' their 'scripts'. Especially as such
1909 people are likely to be the least capable of the user-base, a
1910 possible design-rule would be to make the scripting style of usage
1911 easy (default?), perhaps requiring a special option to enable a
1912 functional style. Those who will use the functional style won't
1913 have a problem understanding what's going on, whereas the 'skript
1914 kiddies' might not even know the syntax for defining a function in
1915 their language of choice. And of course we can allow the user to
1916 set a variable in their .emacs controlling the preference, so that
1917 functional users are not inconveniennced by having to provide
1918 header args the whole time.
1920 Please don't get the impression that I am down-valuing the
1921 functional style of org-babel. I am constantly horrified at the
1922 messy 'scripts' that my colleagues produce in perl or R or
1923 whatever! Nevertheless that seems to be how a lot of people work.
1925 I think you were leaning towards the last-value approach because
1926 it offered the possibility of unified code supporting both the
1927 single evaluation environment and the functional style. If you
1928 agree with any of the above then perhaps it will impact upon this
1929 and mean that the code in the two branches has to differ a bit. In
1930 that case, functional mode could perhaps after all evaluate each
1931 code block in its own environment, thus (re)approaching 'true'
1932 functional programming (side-effects are hard to achieve).
1936 echo "There are `wc -l files` files in this directory"
1940 *** even more thoughts on evaluation, results, models and options
1942 Thanks Dan, These comments are invaluable.
1944 What do you think about this as a new list of priorities/requirements
1945 for the execution of source-code blocks.
1948 1) we want the evaluation of the source code block to take place in a
1949 session which can persist state (variables, current directory,
1951 2) source code blocks can specify their session with a header argument
1952 3) each session should correspond to an Emacs comint buffer so that the
1953 user can drop into the session and experiment with live code
1956 1) each source-code block generates some form of results which (as
1957 we have already implemented) is transfered into emacs-lisp
1958 after which it can be inserted into the org-mode buffer, or
1959 used by other source-code blocks
1960 2) when the results are translated into emacs-lisp, forced to be
1961 interpreted as a scalar (dumping their raw values into the
1962 org-mode buffer), as a vector (which is often desirable with R
1963 code blocks), or interpreted on the fly (the default option).
1964 Note that this is very nearly currently implemented through the
1965 [[* DONE results-type header (vector/file)][results-type-header]].
1966 3) there should be *two* means of collecting results from the
1967 execution of a source code block. *Either* the value of the
1968 last statement of the source code block, or the collection of
1969 all that has been passed to STDOUT during the evaluation.
1971 **** header argument or return line (*header argument*)
1973 Rather than using a header argument to specify how the return value
1974 should be passed back, I'm leaning towards the use of a =#+RETURN=
1975 line inside the block. If such a line *is not present* then we
1976 default to using STDOUT to collect results, but if such a line *is
1977 present* then we use it's value as the results of the block. I
1978 think this will allow for the most elegant specification between
1979 functional and script execution. This also cleans up some issues
1980 of implementation and finding which statement is the last
1983 Having given this more thought, I think a header argument is
1984 preferable. The =#+return:= line adds new complicating syntax for
1985 something that does little more than we would accomplish through
1986 the addition of a header argument. The only benefit being that we
1987 know where the final statement starts, which is not an issue in
1988 those languages which contain 'last value' operators.
1990 new header =:results= arguments
1991 - script :: explicitly states that we want to use STDOUT to
1992 initialize our results
1993 - return_last :: stdout is ignored instead the *value* of the final
1994 statement in the block is returned
1995 - echo :: means echo the contents of the source-code block along
1996 with the results (this implies the *script* =:results=
1999 *** DONE rework evaluation lang-by-lang [4/4]
2001 This should include...
2002 - functional results working with the comint buffer
2004 - script :: return the output of STDOUT
2005 - write a macro which runs the first redirection, executes the
2006 body, then runs the second redirection
2007 - last :: return the value of the last statement
2010 - sessions in comint buffers
2012 **** DONE Ruby [4/4]
2013 - [X] functional results working with comint
2014 - [X] script results
2015 - [X] ensure scalar/vector results args are taken into consideration
2016 - [X] ensure callable by other source block
2018 #+srcname: ruby-use-last-output
2019 #+begin_src ruby :results replace
2026 #+resname: ruby-use-last-output
2029 #+srcname: task-call-use-last-output
2030 #+begin_src ruby :var last=ruby-use-last-output :results replace
2031 last.flatten.size + 1
2034 #+resname: task-call-use-last-output
2039 #+srcname: first-ruby-session-task
2040 #+begin_src ruby :session schulte :results silent
2044 #+srcname: second-ruby-session-task
2045 #+begin_src ruby :session schulte :results silent
2049 #+srcname: without-the-right-session
2050 #+begin_src ruby :results silent
2056 - [X] functional results working with comint
2057 - [X] script results
2058 - [X] ensure scalar/vector results args are taken into consideration
2059 - [X] ensure callable by other source block
2061 To redirect output to a file, you can use the =sink()= command.
2064 #+begin_src R :results value vector silent
2071 #+srcname: task-R-use-other-output
2072 #+begin_src R :var twoentyseven=task_R_B() :results replace value
2077 #+resname: task-R-use-other-output
2080 **** DONE Python [4/4]
2081 - [X] functional results working with comint
2082 - [X] script results
2083 - [X] ensure scalar/vector results args are taken into consideration
2084 - [X] ensure callable by other source block
2086 #+srcname: task-new-eval-for-python
2087 #+begin_src python :results silent output scalar
2093 #+srcname: task-use-new-eval
2094 #+begin_src python :var tasking=task-new-eval-for-python() :results replace
2098 #+resname: task-use-new-eval
2101 **** DONE Shells [4/4]
2102 - [X] functional results working with comint
2103 - [X] script results
2104 - [X] ensure scalar/vector results args are taken into consideration
2105 - [X] ensure callable by other source block
2107 #+srcname: task-shell-new-evaluation
2108 #+begin_src sh :results silent value scalar
2113 #+srcname: task-call-other-shell
2114 #+begin_src sh :var other=task-shell-new-evaluation() :results replace scalar
2115 echo $other ' is the old date'
2118 #+resname: task-call-other-shell
2119 : $ Fri Jun 12 13:08:37 PDT 2009 is the old date
2121 *** DONE implement a *session* header argument [4/4]
2122 =:session= header argument to override the default *session* buffer
2126 #+srcname: task-ruby-named-session
2127 #+begin_src ruby :session schulte :results replace
2128 schulte = :in_schulte
2131 #+resname: task-ruby-named-session
2134 #+srcname: another-in-schulte
2135 #+begin_src ruby :session schulte
2139 #+resname: another-in-schulte
2146 #+srcname: python-session-task
2147 #+begin_src python :session what :results silent
2151 #+srcname: python-get-from-session
2152 #+begin_src python :session what :results replace
2156 #+resname: python-get-from-session
2161 #+srcname: task-shell-sessions
2162 #+begin_src sh :session what
2166 #+srcname: task-shell-sessions-what
2167 #+begin_src sh :session what :results replace
2171 #+resname: task-shell-sessions-what
2176 #+srcname: task-R-session
2177 #+begin_src R :session what :results replace
2183 #+resname: task-R-session
2186 #+srcname: another-task-R-session
2187 #+begin_src R :session what :results replace
2191 *** DONE function to bring up inferior-process buffer [4/4]
2193 This should be callable from inside of a source-code block in an
2194 org-mode buffer. It should evaluate the header arguments, then bring
2195 up the inf-proc buffer using =pop-to-buffer=.
2197 For lack of a better place, lets add this to the `org-metadown-hook'
2200 To give this a try, place the cursor on a source block with variables,
2201 (optionally git a prefix argument) then hold meta and press down.
2205 #+srcname: task-ruby-pop-to-session
2206 #+begin_src ruby :var num=9 :var another="something else"
2207 num.times{|n| puts another}
2212 #+srcname: task-python-pop-to-session
2213 #+begin_src python :var num=9 :var another="something else"
2218 #+srcname: task-R-pop-to-session
2219 #+begin_src R :var a=9 :var b=8
2225 #+srcname: task-shell-pop-sessions
2226 #+begin_src sh :var NAME="eric"
2230 *** DEFERRED function to dump last N lines from inf-proc buffer into the current source block
2232 Callable with a prefix argument to specify how many lines should be
2233 dumped into the source-code buffer.
2235 *** REJECTED comint notes
2237 Implementing comint integration in [[file:lisp/org-babel-comint.el][org-babel-comint.el]].
2240 - handling of outputs
2241 - split raw output from process by prompts
2242 - a ring of the outputs, buffer-local, `org-babel-comint-output-ring'
2243 - a switch for dumping all outputs to a buffer
2244 - inputting commands
2246 Lets drop all this language specific stuff, and just use
2247 org-babel-comint to split up our outputs, and return either the last
2248 value of an execution or the combination of values from the
2251 **** comint filter functions
2252 : ;; comint-input-filter-functions hook process-in-a-buffer
2253 : ;; comint-output-filter-functions hook function modes.
2254 : ;; comint-preoutput-filter-functions hook
2255 : ;; comint-input-filter function ...
2257 #+srcname: obc-filter-ruby
2258 #+begin_src ruby :results last
2266 ** DONE Remove protective commas from # comments before evaluating
2267 org inserts protective commas in front of ## comments in language
2268 modes that use them. We need to remove them prior to sending code
2271 #+srcname: testing-removal-of-protective-comas
2273 ,# this one might break it??
2277 ** DONE pass multiple reference arguments into R
2278 Can we do this? I wasn't sure how to supply multiple 'var' header
2279 args. Just delete this if I'm being dense.
2281 This should be working, see the following example...
2283 #+srcname: two-arg-example
2284 #+begin_src R :var n=2 :var m=8
2288 #+resname: two-arg-example
2291 ** DONE ensure that table ranges work
2292 when a table range is passed to org-babel as an argument, it should be
2293 interpreted as a vector.
2296 | 2 | 3 | Fixnum:1 |
2297 | 3 | 4 | Array:123456 |
2301 #+TBLFM: @1$3='(sbe simple-sbe-example (n 4))::@2$3='(sbe task-table-range (n @1$1..@6$1))::@3$3='(sbe task-table-range (n (@1$1..@6$1)))
2303 #+srcname: simple-sbe-example
2304 #+begin_src emacs-lisp
2308 #+srcname: task-table-range
2309 #+begin_src ruby :var n=simple-sbe-example
2313 #+srcname: simple-results
2314 #+begin_src emacs-lisp :var n=task-table-range(n=(1 2 3))
2318 #+resname: simple-results
2321 #+srcname: task-arr-referent
2322 #+begin_src ruby :var ar=(1 2 3)
2326 #+resname: task-arr-referent
2329 ** DONE global variable indicating default to vector output
2330 how about an alist... =org-babel-default-header-args= this may already
2331 exist... just execute the following and all source blocks will default
2334 #+begin_src emacs-lisp
2335 (setq org-babel-default-header-args '((:results . "vector")))
2338 ** DONE name named results if source block is named
2339 currently this isn't happening although it should be
2341 #+srcname: test-naming-named-source-blocks
2342 #+begin_src emacs-lisp
2346 #+resname: test-naming-named-source-blocks
2348 ** DONE (simple caching) check for named results before source blocks
2349 see the TODO comment in [[file:lisp/org-babel-ref.el::TODO%20This%20should%20explicitly%20look%20for%20resname%20lines%20before][org-babel-ref.el#org-babel-ref-resolve-reference]]
2350 ** DONE set =:results silent= when eval with prefix argument
2352 #+begin_src emacs-lisp
2355 ** DONE results-type header (vector/file) [3/3]
2356 In response to a point in Dan's email. We should allow the user to
2357 force scalar or vector results. This could be done with a header
2358 argument, and the default behavior could be controlled through a
2359 configuration variable.
2361 #+srcname: task-trivial-vector
2362 #+begin_src ruby :results replace vector
2369 since it doesn't make sense to turn a vector into a scalar, lets
2370 just add a two values...
2372 - vector :: forces the results to be a vector (potentially 1 dimensional)
2373 - file :: this throws an error if the result isn't a string, and
2374 tries to treat it as a path to a file.
2376 I'm just going to cram all of these into the =:results= header
2377 argument. Then if we allow multiple header arguments it should
2378 work out, for example one possible header argument string could be
2379 =:results replace vector file=, which would *replace* any existing
2380 results forcing the results into an org-mode table, and
2381 interpreting any strings as file paths.
2383 *** DONE multiple =:results= headers
2385 #+srcname: multiple-result-headers
2386 #+begin_src ruby :results replace silent
2392 *** DONE file result types
2393 When inserting into an org-mode buffer create a link with the path
2394 being the value, and optionally the display being the
2395 =file-name-nondirectory= if it exists.
2397 #+srcname: task-file-result
2398 #+begin_src python :results replace file
2403 [[something][something]]
2406 This will be useful because blocks like =ditaa= and =dot= can return
2407 the string path of their files, and can add =file= to their results
2410 *** DONE vector result types
2412 #+srcname: task-force-results
2413 #+begin_src emacs-lisp :results vector
2420 ** DONE results name
2421 In order to do this we will need to start naming our results.
2422 Since the source blocks are named with =#+srcname:= lines we can
2423 name results with =#+resname:= lines (if the source block has no
2424 name then no name is given to the =#+resname:= line on creation,
2425 otherwise the name of the source block is used).
2427 This will have the additional benefit of allowing results and
2428 source blocks to be located in different places in a buffer (and
2429 eventually in different buffers entirely).
2431 #+srcname: developing-resnames
2432 #+begin_src emacs-lisp :results silent
2436 Once source blocks are able to find their own =#+resname:= lines
2439 #+srcname: sbe-w-new-results
2440 #+begin_src emacs-lisp :results replace
2441 (sbe "developing-resnames")
2447 *** TODO change the results insertion functions to use these lines
2449 *** TODO teach references to resolve =#+resname= lines.
2451 ** DONE org-babel tests org-babel [1/1]
2452 since we are accumulating this nice collection of source-code blocks
2453 in the sandbox section we should make use of them as unit tests.
2454 What's more, we should be able to actually use org-babel to run these
2457 We would just need to cycle over every source code block under the
2458 sandbox, run it, and assert that the return value is equal to what we
2461 I have the feeling that this should be possible using only org-babel
2462 functions with minimal or no additional elisp. It would be very cool
2463 for org-babel to be able to test itself.
2465 This is now done, see [[* Tests]].
2467 *** DEFERRED org-babel assertions (may not be necessary)
2468 These could be used to make assertions about the results of a
2469 source-code block. If the assertion fails then the point could be
2470 moved to the block, and error messages and highlighting etc... could
2473 ** DONE make C-c C-c work anywhere within source code block?
2474 This seems like it would be nice to me, but perhaps it would be
2475 inefficient or ugly in implementation? I suppose you could search
2476 forward, and if you find #+end_src before you find #+begin_src,
2477 then you're inside one. [DED]
2479 Agreed, I think inside of the =#+srcname: line= would be useful as
2482 #+srcname: testing-out-cc
2483 #+begin_src emacs-lisp
2487 ** DONE integration with org tables
2488 We should make it easy to call org-babel source blocks from org-mode
2489 table formulas. This is practical now that it is possible to pass
2490 arguments to org-babel source blocks.
2492 See the related [[* (sandbox) integration w/org tables][sandbox]] header for tests/examples.
2494 *** digging in org-table.el
2495 In the past [[file:~/src/org/lisp/org-table.el::org%20table%20el%20The%20table%20editor%20for%20Org%20mode][org-table.el]] has proven difficult to work with.
2497 Should be a hook in [[file:~/src/org/lisp/org-table.el::defun%20org%20table%20eval%20formula%20optional%20arg%20equation][org-table-eval-formula]].
2499 Looks like I need to change this [[file:~/src/org/lisp/org-table.el::if%20lispp][if statement]] (line 2239) into a cond
2502 ** DONE source blocks as functions
2504 Allow source code blocks to be called like functions, with arguments
2505 specified. We are already able to call a source-code block and assign
2506 it's return result to a variable. This would just add the ability to
2507 specify the values of the arguments to the source code block assuming
2508 any exist. For an example see
2510 When a variable appears in a header argument, how do we differentiate
2511 between it's value being a reference or a literal value? I guess this
2512 could work just like a programming language. If it's escaped or in
2513 quotes, then we count it as a literal, otherwise we try to look it up
2516 ** DONE folding of code blocks? [2/2]
2517 [DED] In similar way to using outline-minor-mode for folding function
2518 bodies, can we fold code blocks? #+begin whatever statements are
2519 pretty ugly, and in any case when you're thinking about the overall
2520 game plan you don't necessarily want to see the code for each Step.
2522 *** DONE folding of source code block
2523 Sounds good, and wasn't too hard to implement. Code blocks should
2524 now be fold-able in the same manner as headlines (by pressing TAB
2527 *** REJECTED folding of results
2528 So, lets do a three-stage tab cycle... First fold the src block,
2529 then fold the results, then unfold.
2531 There's no way to tell if the results are a table or not w/o
2532 actually executing the block which would be too expensive of an
2535 ** DONE selective export of text, code, figures
2536 [DED] The org-babel buffer contains everything (code, headings and
2537 notes/prose describing what you're up to, textual/numeric/graphical
2538 code output, etc). However on export to html / LaTeX one might want
2539 to include only a subset of that content. For example you might
2540 want to create a presentation of what you've done which omits the
2543 [EMS] So I think this should be implemented as a property which can
2544 be set globally or on the outline header level (I need to review
2545 the mechanics of org-mode properties). And then as a source block
2546 header argument which will apply only to a specific source code
2547 block. A header argument of =:export= with values of
2549 - =code= :: just show the code in the source code block
2550 - =none= :: don't show the code or the results of the evaluation
2551 - =results= :: just show the results of the code evaluation (don't
2552 show the actual code)
2553 - =both= :: show both the source code, and the results
2555 this will be done in [[* (sandbox) selective export][(sandbox) selective export]].
2557 ** DONE a header argument specifying silent evaluation (no output)
2558 This would be useful across all types of source block. Currently
2559 there is a =:replace t= option to control output, this could be
2560 generalized to an =:output= option which could take the following
2561 options (maybe more)
2563 - =t= :: this would be the default, and would simply insert the
2564 results after the source block
2565 - =replace= :: to replace any results which may already be there
2566 - =silent= :: this would inhibit any insertion of the results
2568 This is now implemented see the example in the [[* silent evaluation][sandbox]]
2570 ** DONE assign variables from tables in R
2571 This is now working (see [[* (sandbox table) R][(sandbox-table)-R]]). Although it's not that
2572 impressive until we are able to print table results from R.
2574 ** DONE insert 2-D R results as tables
2575 everything is working but R and shell
2581 This has already been tackled by Dan in [[file:existing_tools/org-R.el::defconst%20org%20R%20write%20org%20table%20def][org-R:check-dimensions]]. The
2582 functions there should be useful in combination with [[http://cran.r-project.org/doc/manuals/R-data.html#Export-to-text-files][R-export-to-csv]]
2583 as a means of converting multidimensional R objects to emacs lisp.
2585 It may be as simple as first checking if the data is multidimensional,
2586 and then, if so using =write= to write the data out to a temporary
2587 file from which emacs can read the data in using =org-table-import=.
2589 Looking into this further, is seems that there is no such thing as a
2590 scalar in R [[http://tolstoy.newcastle.edu.au/R/help/03a/3733.html][R-scalar-vs-vector]] In that light I am not sure how to
2591 deal with trivial vectors (scalars) in R. I'm tempted to just treat
2592 them as vectors, but then that would lead to a proliferation of
2593 trivial 1-cell tables...
2595 ** DONE allow variable initialization from source blocks
2596 Currently it is possible to initialize a variable from an org-mode
2597 table with a block argument like =table=sandbox= (note that the
2598 variable doesn't have to named =table=) as in the following example
2604 #+begin_src emacs-lisp :var table=sandbox :results replace
2605 (message (format "table = %S" table))
2608 : "table = ((1 2 3) (4 \"schulte\" 6))"
2610 It would be good to allow initialization of variables from the results
2611 of other source blocks in the same manner. This would probably
2612 require the addition of =#+SRCNAME: example= lines for the naming of
2613 source blocks, also the =table=sandbox= syntax may have to be expanded
2614 to specify whether the target is a source code block or a table
2615 (alternately we could just match the first one with the given name
2616 whether it's a table or a source code block).
2618 At least initially I'll try to implement this so that there is no need
2619 to specify whether the reference is to a table or a source-code block.
2620 That seems to be simpler both in terms of use and implementation.
2622 This is now working for emacs-lisp, ruby and python (and mixtures of
2623 the three) source blocks. See the examples in the [[* (sandbox) referencing other source blocks][sandbox]].
2625 This is currently working only with emacs lisp as in the following
2626 example in the [[* emacs lisp source reference][emacs lisp source reference]].
2629 ** TODO Add languages [11/15]
2630 I'm sure there are many more that aren't listed here. Please add
2631 them, and bubble any that you particularly care about up to the top.
2633 Any new language should be implemented in a org-babel-lang.el file.
2634 Follow the pattern set by [[file:lisp/org-babel-script.el][org-babel-script.el]], [[file:lisp/org-babel-shell.el][org-babel-shell.el]] and
2635 [[file:lisp/org-babel-R.el][org-babel-R.el]].
2637 *** STARTED ocaml [2/3]
2639 - [X] Working for the simple case (no arguments, simple output)
2640 - [X] correct handling of vector/list output
2641 - [ ] ability to import arguments
2648 | n -> fib(n - 1) + fib(n - 2) in
2670 [|"ocaml"; "array"|]
2674 | "ocaml" | "array" |
2677 This could probably be added to [[file:lisp/org-babel-script.el][org-babel-script.el]]
2681 - support for sessions
2682 - add more useful header arguments (user, passwd, database, etc...)
2683 - support for more engines (currently only supports mysql)
2684 - what's a reasonable way to drop table data into SQL?
2686 #+srcname: sql-example
2687 #+begin_src sql :engine mysql
2693 | "information_schema" |
2697 Sass is a very nice extension of CSS, which is much nicer to read and
2698 write (see [[http://sass-lang.com/][sass-lang]]).
2700 #+srcname: sass-example
2701 #+begin_src sass :file stylesheet.css :results file
2711 [[file:stylesheet.css][stylesheet.css]]
2714 trivial [[file:lisp/langs/org-babel-css.el][org-babel-css.el]]
2717 (see [[* file result types][file result types]])
2719 #+srcname: implementing-ditaa
2720 #+begin_src ditaa :results replace :file blue.png :cmdline -r
2730 #+resname: implementing-ditaa
2731 [[file:blue.png][blue.png]]
2733 *** DONE gnuplot [7/7]
2734 (see [[* file result types][file result types]])
2736 #+PLOT: title:"Citas" ind:1 deps:(3) type:2d with:histograms set:"yrange [0:]"
2737 #+TBLNAME: gnuplot-data
2738 | independent var | first dependent var | second dependent var |
2739 |-----------------+---------------------+----------------------|
2740 | 0.1 | 0.425 | 0.375 |
2741 | 0.2 | 0.3125 | 0.3375 |
2742 | 0.3 | 0.24999993 | 0.28333338 |
2743 | 0.4 | 0.275 | 0.28125 |
2744 | 0.5 | 0.26 | 0.27 |
2745 | 0.6 | 0.25833338 | 0.24999993 |
2746 | 0.7 | 0.24642845 | 0.23928553 |
2747 | 0.8 | 0.23125 | 0.2375 |
2748 | 0.9 | 0.23333323 | 0.2333332 |
2749 | 1 | 0.2225 | 0.22 |
2750 | 1.1 | 0.20909075 | 0.22272708 |
2751 | 1.2 | 0.19999998 | 0.21458333 |
2752 | 1.3 | 0.19615368 | 0.21730748 |
2754 #+srcname: implementing-gnuplot
2755 #+begin_src gnuplot :var data=gnuplot-data :results silent
2756 set title "Implementing Gnuplot"
2757 plot data using 1:2 with lines
2760 **** DONE add variables
2761 gnuplot 4.2 and up support user defined variables. This is how
2762 we will handle variables with org-babel (meaning we will need to
2763 require gnuplot 4.2 and up for variable support, which can be
2764 install using [[http://www.macports.org/install.php][macports]] on Mac OSX).
2766 - scalar variables should be replaced in the body of the gnuplot code
2767 - vector variables should be exported to tab-separated files, and
2768 the variable names should be replaced with the path to the files
2770 **** DONE direct plotting w/o session
2771 **** DEFERRED gnuplot support for column/row names
2772 This should be implemented along the lines of the [[* STARTED Column (and row) names of tables in R input/output][R-colname-support]].
2774 We can do something similar to the :labels param in org-plot, we just
2775 have to be careful to ensure that each label is aligned with the
2778 This may be walking too close to an entirely prebuilt plotting tool
2779 rather than straight gnuplot code evaluation. For now I think this
2782 **** DONE a =file= header argument
2783 to specify a file holding the results
2785 #+srcname: gnuplot-to-file-implementation
2786 #+begin_src gnuplot :file plot.png :var data=gnuplot-data
2787 plot data using 1:2, data using 1:3 with lines
2791 [[file:plot.png][plot.png]]
2793 **** DONE helpers from org-plot.el
2794 There are a variety of helpers in org-plot which can be fit nicely
2795 into custom gnuplot header arguments.
2797 These should all be in place by now.
2799 **** DEFERRED header argument specifying 3D data
2802 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
2803 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
2804 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
2805 | 0 | 0 | 1 | 1 | 0 | 0 | 1 | 0 | 0 | 0 | 0 | 0 | 1 | 1 | 1 | 0 | 0 |
2806 | 0 | 1 | 0 | 0 | 1 | 0 | 1 | 0 | 0 | 0 | 0 | 1 | 0 | 0 | 0 | 1 | 0 |
2807 | 0 | 1 | 0 | 0 | 1 | 0 | 1 | 0 | 0 | 0 | 0 | 1 | 0 | 0 | 0 | 1 | 0 |
2808 | 0 | 1 | 0 | 0 | 1 | 0 | 1 | 0 | 0 | 0 | 0 | 1 | 0 | 1 | 1 | 1 | 0 |
2809 | 0 | 1 | 0 | 0 | 1 | 0 | 1 | 0 | 0 | 0 | 0 | 1 | 0 | 0 | 0 | 0 | 0 |
2810 | 0 | 1 | 0 | 0 | 1 | 0 | 1 | 0 | 0 | 0 | 0 | 1 | 0 | 0 | 0 | 0 | 0 |
2811 | 0 | 1 | 0 | 0 | 1 | 0 | 1 | 0 | 0 | 0 | 0 | 1 | 0 | 0 | 0 | 1 | 0 |
2812 | 0 | 1 | 0 | 0 | 1 | 0 | 1 | 1 | 0 | 1 | 0 | 1 | 0 | 0 | 0 | 1 | 0 |
2813 | 0 | 0 | 1 | 1 | 0 | 0 | 1 | 0 | 1 | 1 | 0 | 0 | 1 | 1 | 1 | 0 | 0 |
2814 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
2815 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
2817 #+srcname: implementing-gnuplot-grid-plots
2818 #+begin_src gnuplot :vars data=org-grid
2822 **** DONE gnuplot sessions
2823 Working on this, we won't support multiple sessions as `gnuplot-mode'
2824 isn't setup for such things.
2826 Also we can't display results with the default :none session, so for
2827 gnuplot we really want the default behavior to be :default, and to
2828 only run a :none session when explicitly specified.
2830 #+srcname: implementing-gnuplot-sessions
2831 #+begin_src gnuplot :var data=gnuplot-data :session none :file session.png
2832 set title "Implementing Gnuplot Sessions"
2833 plot data using 1:2 with lines
2837 [[file:session.png][session.png]]
2840 (see [[* file result types][file result types]])
2842 #+srcname: implementing-dot-support
2843 #+begin_src dot :file test-dot.png :cmdline -Tpng
2844 digraph data_relationships {
2845 "data_requirement" [shape=Mrecord, label="{DataRequirement|description\lformat\l}"]
2846 "data_product" [shape=Mrecord, label="{DataProduct|name\lversion\lpoc\lformat\l}"]
2847 "data_requirement" -> "data_product"
2852 [[file:test-dot.png][test-dot.png]]
2855 (see [[* file result types][file result types]])
2857 for information on asymptote see http://asymptote.sourceforge.net
2859 #+begin_src asymptote :file asymptote-test.png
2864 real f(real t) {return 1+cos(t);}
2866 path g=polargraph(f,0,2pi,operator ..)--cycle;
2869 xaxis("$x$",above=true);
2870 yaxis("$y$",above=true);
2872 dot("$(a,0)$",(1,0),N);
2873 dot("$(2a,0)$",(2,0),N+E);
2877 [[file:asymptote-test.png][asymptote-test.png]]
2887 ** TODO stripping indentation from source-code blocks
2888 This is a problem in [[file:lisp/org-babel-exp.el][org-babel-exp.el]].
2890 ** TODO failing to write srcname to resname when evaluating a named block
2891 ** DONE Python session evaluation bug
2892 The following block evaluates correctly with :session none
2893 (set :results to output), but fails with session-based evaluation
2894 (with :results value, as below, you see the error message)
2896 I edebug'ed it and it seems fine until [[file:lisp/langs/org-babel-python.el::comint%20session%20evaluation%20org%20babel%20comint%20in%20buffer%20buffer%20let%20full%20body][we go to comint]].
2898 #+begin_src python :session pysession :results value
2900 from subprocess import *
2905 cc = [('58C','NBS'),
2909 for outdir in ['none', 'noscots', 'lax', 'strict']:
2910 outdir = os.path.join('exclusion-study', outdir)
2911 for case, control in cc:
2912 outfile = os.path.join(outdir, '%s-vs-%s-direct' % (case, control))
2913 cmd = 'snptest %s -frequentist 1 -hwe ' % ('-gen_gz' if format == '.gen.gz' else '')
2914 cmd += '-cases %s %s ' % (case + format, case + '.sample')
2916 cmd += '-controls %s %s ' % (control + format, control + '.sample')
2917 cmd += '-exclude_samples %s ' % os.path.join(outdir, 'exclusions')
2918 cmd += '-o %s ' % outfile
2919 cmd += '-chunk %d ' % chunk
2920 cmd += '> %s' % outfile + '.log'
2927 snptest -gen_gz -frequentist 1 -hwe -cases 58C.gen.gz 58C.sample -controls NBS.gen.gz NBS.sample -exclude_samples exclusion-study/none/exclusions -o exclusion-study/none/58C-vs-NBS-direct -chunk 10000 > exclusion-study/none/58C-vs-NBS-direct.log
2928 snptest -gen_gz -frequentist 1 -hwe -cases 58C.gen.gz 58C.sample -controls POBI.gen.gz POBI.sample -exclude_samples exclusion-study/none/exclusions -o exclusion-study/none/58C-vs-POBI-direct -chunk 10000 > exclusion-study/none/58C-vs-POBI-direct.log
2929 snptest -gen_gz -frequentist 1 -hwe -cases NBS.gen.gz NBS.sample -controls POBI.gen.gz POBI.sample -exclude_samples exclusion-study/none/exclusions -o exclusion-study/none/NBS-vs-POBI-direct -chunk 10000 > exclusion-study/none/NBS-vs-POBI-direct.log
2930 snptest -gen_gz -frequentist 1 -hwe -cases 58C.gen.gz 58C.sample -controls NBS.gen.gz NBS.sample -exclude_samples exclusion-study/noscots/exclusions -o exclusion-study/noscots/58C-vs-NBS-direct -chunk 10000 > exclusion-study/noscots/58C-vs-NBS-direct.log
2931 snptest -gen_gz -frequentist 1 -hwe -cases 58C.gen.gz 58C.sample -controls POBI.gen.gz POBI.sample -exclude_samples exclusion-study/noscots/exclusions -o exclusion-study/noscots/58C-vs-POBI-direct -chunk 10000 > exclusion-study/noscots/58C-vs-POBI-direct.log
2932 snptest -gen_gz -frequentist 1 -hwe -cases NBS.gen.gz NBS.sample -controls POBI.gen.gz POBI.sample -exclude_samples exclusion-study/noscots/exclusions -o exclusion-study/noscots/NBS-vs-POBI-direct -chunk 10000 > exclusion-study/noscots/NBS-vs-POBI-direct.log
2933 snptest -gen_gz -frequentist 1 -hwe -cases 58C.gen.gz 58C.sample -controls NBS.gen.gz NBS.sample -exclude_samples exclusion-study/lax/exclusions -o exclusion-study/lax/58C-vs-NBS-direct -chunk 10000 > exclusion-study/lax/58C-vs-NBS-direct.log
2934 snptest -gen_gz -frequentist 1 -hwe -cases 58C.gen.gz 58C.sample -controls POBI.gen.gz POBI.sample -exclude_samples exclusion-study/lax/exclusions -o exclusion-study/lax/58C-vs-POBI-direct -chunk 10000 > exclusion-study/lax/58C-vs-POBI-direct.log
2935 snptest -gen_gz -frequentist 1 -hwe -cases NBS.gen.gz NBS.sample -controls POBI.gen.gz POBI.sample -exclude_samples exclusion-study/lax/exclusions -o exclusion-study/lax/NBS-vs-POBI-direct -chunk 10000 > exclusion-study/lax/NBS-vs-POBI-direct.log
2936 snptest -gen_gz -frequentist 1 -hwe -cases 58C.gen.gz 58C.sample -controls NBS.gen.gz NBS.sample -exclude_samples exclusion-study/strict/exclusions -o exclusion-study/strict/58C-vs-NBS-direct -chunk 10000 > exclusion-study/strict/58C-vs-NBS-direct.log
2937 snptest -gen_gz -frequentist 1 -hwe -cases 58C.gen.gz 58C.sample -controls POBI.gen.gz POBI.sample -exclude_samples exclusion-study/strict/exclusions -o exclusion-study/strict/58C-vs-POBI-direct -chunk 10000 > exclusion-study/strict/58C-vs-POBI-direct.log
2938 snptest -gen_gz -frequentist 1 -hwe -cases NBS.gen.gz NBS.sample -controls POBI.gen.gz POBI.sample -exclude_samples exclusion-study/strict/exclusions -o exclusion-study/strict/NBS-vs-POBI-direct -chunk 10000 > exclusion-study/strict/NBS-vs-POBI-direct.log
2941 ** DONE require users to explicitly turn on each language
2942 As we continue to add more languages to org-babel, many of which will
2943 require new major-modes we need to re-think how languages are added to
2946 Currently we are requiring all available languages in the
2947 [[file:lisp/org-babel-init.el][org-babel-init.el]] file. I think we need to change this to a user
2948 setting so that only the language which have their requirements met
2949 (in terms of system executables and emacs major modes) are loaded. It
2950 is one more step for install, but it seems to me to be the only
2955 we add something like the following to the instillation instructions
2957 #+begin_src emacs-lisp
2958 ;; Uncomment each of the following require lines if you want org-babel
2959 ;; to support that language. Each language has a comment explaining
2960 ;; it's dependencies. See the related files in lisp/langs for more
2961 ;; detailed explanations of requirements.
2963 ;; (require 'org-babel-R) ;; ess-mode
2964 ;; (require 'org-babel-asymptote) ;; asymptote be installed on your system
2965 ;; (require 'org-babel-css) ;; none
2966 ;; (require 'org-babel-ditaa) ;; ditaa be installed on your system
2967 ;; (require 'org-babel-dot) ;; dot be installed on your system
2968 ;; (require 'org-babel-gnuplot) ;; gnuplot-mode
2969 ;; (require 'org-babel-python) ;; python-mode
2970 ;; (require 'org-babel-ruby) ;; inf-ruby mode, ruby and irb must be installed on your system
2971 ;; (require 'org-babel-sql) ;; none
2974 note that =org-babel-sh=, =org-babel-emacs-lisp= are not included in
2975 the list as they can safely be assumed to work on any system.
2978 we should come up with a way to gracefully degrade when support for a
2979 specific language is missing
2981 > To demonstrate creation of documents, open the "test-export.org" file in
2982 > the base of the org-babel directory, and export it as you would any
2983 > other org-mode file. The "exports" header argument controls how
2984 > source-code blocks are exported, with the following options
2986 > - none :: no part of the source-code block is exported in the document
2987 > - results :: only the output of the evaluated block is exported
2988 > - code :: the code itself is exported
2989 > - both :: both the code and results are exported
2991 I have this error showing up:
2993 executing Ruby source code block
2994 apply: Searching for program: no such file or directory, irb
2996 ** DONE problem with newlines in output when :results value
2998 #+begin_src python :results value
2999 '\n'.join(map(str, range(4)))
3008 Whereas I was hoping for
3015 *Note*: to generate the above you can try using the new =raw= results
3018 #+begin_src python :results value raw
3019 '|'+'|\n|'.join(map(str, range(4)))+'|'
3028 This is now working, it doesn't return as a table because the value
3029 returned is technically a string. To return the table mentioned above
3030 try something like the following.
3033 [[0], [1], [2], [3]]
3042 This is some sort of non-printing char / quoting issue I think. Note
3045 #+begin_src python :results value
3046 '\\n'.join(map(str, range(4)))
3053 #+begin_src python :results output
3054 print('\n'.join(map(str, range(4))))
3063 *** collapsing consecutive newlines in string output
3065 This is an example of the same bug
3067 #+srcname: multi-line-string-output
3068 #+begin_src ruby :results output
3069 "the first line ends here
3072 and this is the second one
3077 This doesn't produce anything at all now. I believe that's because
3078 I've changed things so that :results output really does *not* get the
3079 value of the block, only the STDOUT. So if we add a print statement
3082 #+srcname: multi-line-string-output
3083 #+begin_src ruby :results output
3084 print "the first line ends here
3087 and this is the second one
3093 : the first line ends here
3096 : and this is the second one
3100 However, the behaviour with :results value is wrong
3102 #+srcname: multi-line-string-value
3104 "the first line ends here
3107 and this is the second one
3115 ** TODO prompt characters appearing in output with R
3116 #+begin_src R :session *R* :results output
3125 ** TODO o-b-execute-subtree overwrites heading when subtree is folded
3127 Try M-x org-babel-execute-subtree with the subtree folded and
3128 point at the beginning of the heading line.
3132 ** DONE non-orgtbl formatted lists
3135 #+srcname: this-doesn't-match-orgtbl
3136 #+begin_src emacs-lisp :results replace
3137 '((:results . "replace"))
3141 | (:results . "replace") |
3143 #+srcname: this-probably-also-wont-work
3144 #+begin_src emacs-lisp :results replace
3151 ** PROPOSED allow un-named arguments
3160 ## produces no output
3162 It's not essential but would be nice for this to work. To do it
3163 properly, would mean that we'd have to specify rules for how a string
3164 of supplied arguments (some possibly named) interact with the
3165 arguments in the definition (some possibly with defaults) to give
3166 values to the variables in the funbction body.
3167 ** PROPOSED external shell execution can't isolate return values
3168 I have no idea how to do this as of yet. The result is that when
3169 shell functions are run w/o a session there is no difference between
3170 the =output= and =value= result arguments.
3172 Yea, I don't know how to do this either. I searched extensively on
3173 how to isolate the *last* output of a series of shell commands (see
3174 [[* last command for
3175 shells][last command for shells]]). The results of the search were basically
3176 that it was not possible (or at least not accomplish-able with a
3177 reasonable amount of effort).
3179 That fact combined with the tenancy to all ways use standard out in
3180 shell scripts led me to treat these two options (=output= and =value=)
3181 as identical in shell evaluation. Not ideal but maybe good enough for
3184 In the `results' branch I've changed this so that they're not quite
3185 identical: output results in raw stdout contents, whereas value
3186 converts it to elisp, perhaps to a table if it looks tabular. This is
3187 the same for the other languages. [Dan]
3189 ** DONE adding blank line when source-block produces no output
3191 #+srcname: show-org-babel-trim
3193 find . \( -path \*/SCCS -o -path \*/RCS -o -path \*/CVS -o -path \*/MCVS -o -path \*/.svn -o -path \*/.git -o -path \*/.hg -o -path \*/.bzr -o -path \*/_MTN -o -path \*/_darcs -o -path \*/\{arch\} \) -prune -o -type f \( -iname \*.el \) -exec grep -i -nH -e org-babel-trim {} \;
3196 ** DONE Allow source blocks to be recognised when #+ are not first characters on the line
3197 I think Carsten has recently altered the core so that #+ can have
3198 preceding whitespace, at least for literal/code examples. org-babel
3199 should support this.
3201 #+srcname: testing-indentation
3202 #+begin_src emacs-lisp :results silent
3203 (message "i'm indented")
3206 #+srcname: testing-non-indentation
3207 #+begin_src emacs-lisp :results silent
3208 (message "I'm not indented")
3211 #+srcname: i-resolve-references-to-the-indented
3212 #+begin_src emacs-lisp :var speech=testing-indentation :results silent
3213 (message "I said %s" speech)
3216 ** DONE are the org-babel-trim s necessary?
3217 at the end of e.g. org-babel-R-evaluate, org-babel-python-evaluate, but
3218 not org-babel-ruby-evaluate
3220 I think it depends on the language, if we find that extra blank
3221 lines are being inserted in a particular language that is a good
3222 indication that the trim or chomp functions may be appropriate.
3224 org-babel-trim and the related org-babel-chomp are use throughout
3227 #+srcname: show-org-babel-trim-usage
3228 #+begin_src sh :results output
3229 find lisp/ \( -path \*/SCCS -o -path \*/RCS -o -path \*/CVS -o -path \*/MCVS -o -path \*/.svn -o -path \*/.git -o -path \*/.hg -o -path \*/.bzr -o -path \*/_MTN -o -path \*/_darcs -o -path \*/\{arch\} \) -prune -o -type f \( -iname \*.el \) -exec grep -i -nH org-babel-trim {} \;
3234 lisp//langs/org-babel-python.el:49: vars "\n") "\n" (org-babel-trim body) "\n")) ;; then the source block body
3235 lisp//langs/org-babel-python.el:143: (org-remove-indentation (org-babel-trim body)) "[\r\n]")))
3236 lisp//langs/org-babel-python.el:158: #'org-babel-trim
3237 lisp//langs/org-babel-python.el:166: (reverse (mapcar #'org-babel-trim raw)))))))
3238 lisp//langs/org-babel-python.el:169: (output (org-babel-trim (mapconcat #'identity (reverse (cdr results)) "\n")))
3239 lisp//langs/org-babel-python.el:170: (value (org-babel-python-table-or-string (org-babel-trim (car results)))))))))
3240 lisp//langs/org-babel-ruby.el:149: (mapcar #'org-babel-trim raw)))))))
3241 lisp//langs/org-babel-sh.el:148: (mapcar #'org-babel-trim raw)))))))
3242 lisp//langs/org-babel-sh.el:151: (output (org-babel-trim (mapconcat #'org-babel-trim (reverse results) "\n")))
3243 lisp//org-babel-ref.el:161: (mapcar #'org-babel-trim (reverse (cons buffer return)))))
3244 lisp//org-babel.el:198: (with-temp-buffer (insert (org-babel-trim body)) (copy-region-as-kill (point-min) (point-max)))
3245 lisp//org-babel.el:465: (org-babel-trim
3246 lisp//org-babel.el:706:(defun org-babel-trim (string &optional regexp)
3249 #+srcname: show-org-babel-chomp-usage
3250 #+begin_src sh :results output
3251 find lisp/ \( -path \*/SCCS -o -path \*/RCS -o -path \*/CVS -o -path \*/MCVS -o -path \*/.svn -o -path \*/.git -o -path \*/.hg -o -path \*/.bzr -o -path \*/_MTN -o -path \*/_darcs -o -path \*/\{arch\} \) -prune -o -type f \( -iname \*.el \) -exec grep -i -nH org-babel-chomp {} \;
3256 lisp//langs/org-babel-R.el:122: (full-body (mapconcat #'org-babel-chomp
3257 lisp//langs/org-babel-R.el:143: (delete nil (mapcar #'extractor (mapcar #'org-babel-chomp raw))) "\n"))))))))
3258 lisp//langs/org-babel-ruby.el:143: #'org-babel-chomp
3259 lisp//langs/org-babel-sh.el:142: (full-body (mapconcat #'org-babel-chomp
3260 lisp//org-babel-tangle.el:163: (insert (format "\n%s\n" (org-babel-chomp body)))
3261 lisp//org-babel.el:362: (org-babel-chomp (match-string 2 arg)))
3262 lisp//org-babel.el:698:(defun org-babel-chomp (string &optional regexp)
3263 lisp//org-babel.el:707: "Like `org-babel-chomp' only it runs on both the front and back of the string"
3264 lisp//org-babel.el:708: (org-babel-chomp (org-babel-reverse-string
3265 lisp//org-babel.el:709: (org-babel-chomp (org-babel-reverse-string string) regexp)) regexp))
3268 ** DONE LoB is not populated on startup
3269 org-babel-library-of-babel is nil for me on startup. I have to
3270 evaluate the [[file:lisp/org-babel-lob.el::][org-babel-lob-ingest]] line manually.
3272 #+tblname: R-plot-example-data
3279 #+lob: R-plot(data=R-plot-example-data)
3283 I've added a section to [[file:lisp/org-babel-init.el][org-babel-init.el]] which will load the
3284 library of babel on startup.
3286 Note that this needs to be done in [[file:lisp/org-babel-init.el][org-babel-init.el]] rather than in
3287 [[file:lisp/org-babel-lob.el][org-babel-lob.el]], not entirely sure why, something about it being
3290 Also, I'm now having the file closed if it wasn't being visited by
3291 a buffer before being loaded.
3293 ** DONE use new merge function [[file:lisp/org-babel-ref.el::t%20nil%20org%20combine%20plists%20args%20nil][here]]?
3294 And at other occurrences of org-combine-plists?
3295 ** DONE creeping blank lines
3296 There's still inappropriate addition of blank lines in some circumstances.
3298 Hmm, it's a bit confusing. It's to do with o-b-remove-result. LoB
3299 removes the entire (#+resname and result) and starts from scratch,
3300 whereas #+begin_src only removes the result. I haven't worked out
3301 what the correct fix is yet. Maybe the right thing to do is to make
3302 sure that those functions (o-b-remove-result et al.) are neutral
3303 with respect to newlines. Sounds easy, but...
3313 Compare the results of
3314 #+lob: adder(a=5, b=17)
3316 #+resname: python-add(a=5, b=17)
3318 --------------------------------
3326 ---------------------
3327 ** DONE #+srcname arg parsing bug
3328 #+srcname: test-zz(arg=adder(a=1, b=1))
3337 #+srcname: test-zz-nasty(arg=adder(a=adder(a=19,b=adder(a=5,b=2)),b=adder(a=adder(a=1,b=9),b=adder(a=1,b=3))))
3342 #+resname: test-zz-nasty
3347 #+srcname: test-zz-hdr-arg
3348 #+begin_src python :var arg=adder(a=adder(a=19,b=adder(a=5,b=2)),b=adder(a=adder(a=1,b=9),b=adder(a=1,b=3)))
3355 ** DONE Fix nested evaluation and default args
3356 The current parser / evaluator fails with greater levels of nested
3357 function block calls (example below).
3359 *** Initial statement [ded]
3360 If we want to overcome this I think we'd have to redesign some of
3361 the evaluation mechanism. Seeing as we are also facing issues like
3362 dealing with default argument values, and seeing as we now know
3363 how we want the library of babel to behave in addition to the
3364 source blocks, now might be a good time to think about this. It
3365 would be nice to do the full thing at some point, but otoh we may
3366 not consider it a massive priority.
3368 AIui, there are two stages: (i) construct a parse tree, and (ii)
3369 evaluate it and return the value at the root. In the parse tree
3370 each node represents an unevaluated value (either a literal value
3371 or a reference). Node v may have descendent nodes, which represent
3372 values upon which node v's evaluation depends. Once that tree is
3373 constructed, then we evaluate the nodes from the tips towards the
3374 root (a post-order traversal).
3376 [This would also provide a solution for concatenating the STDOUTs
3377 of called blocks, which is a [[*allow%20output%20mode%20to%20return%20stdout%20as%20value][task below]]; we concatenate them in
3378 whatever order the traversal is done in.]
3380 In addition to the variable references (i.e. daughter nodes), each
3381 node would contain the information needed to evaluate that node
3382 (e.g. lang body). Then we would pass a function postorder over the
3383 tree which would call o-b-execute-src-block at each node, finally
3384 returning the value at the root.
3386 Fwiw I made a very tentative small start at stubbing this out in
3387 org-babel-call.el in the 'evaluation' branch. And I've made a start
3388 at sketching a parsing algorithm below.
3389 **** Parse tree algorithm
3390 Seeing as we're just trying to parse a string like
3391 f(a=1,b=g(c=2,d=3)) it shouldn't be too hard. But of course there
3392 are 'proper' parsers written in elisp out there,
3393 e.g. [[http://cedet.sourceforge.net/semantic.shtml][Semantic]]. Perhaps we can find what we need -- our syntax is
3394 pretty much the same as python and R isn't it?
3396 Or, a complete hack, but maybe it would be we easy to transform it
3397 to XML and then parse that with some existing tool?
3399 But if we're doing it ourselves, something very vaguely like this?
3400 (I'm sure there're lots of problems with this)
3402 #+srcname: org-babel-call-parse(call)
3404 ## we are currently reading a reference name: the name of the root function
3405 whereami = "refname"
3406 node = root = Node()
3407 for c in call_string:
3410 whereami = "varname" # now we're reading a variable name
3413 node.daughters = [node.daughters, new]
3416 whereami = "refname"
3418 whereami = "varname"
3425 if whereami = "varname":
3426 node.varnames[varnum] += c
3427 elif whereami = "refname":
3431 *** discussion / investigation
3432 I believe that this issue should be addressed as a bug rather than as
3433 a point for new development. The code in [[file:lisp/org-babel-ref.el][org-babel-ref.el]] already
3434 resolves variable references in a recursive manner which *should* work
3435 in the same manner regardless of the depth of the number of nested
3436 function calls. This recursive evaluation has the effect of
3437 implicitly constructing the parse tree that your are thinking of
3438 constructing explicitly.
3440 Through using some of the commented out debugging statements in
3441 [[file:lisp/org-babel-ref.el][org-babel-ref.el]] I have looked at what may be going wrong in the
3442 current evaluation setup, and it seems that nested variables are being
3443 set using the =:var= header argument, and these variables are being
3444 overridden by the *default* variables which are being entered through
3445 the new functional syntax (see the demonstration header below).
3447 I believe that once this bug is fixed we should be back to fully
3448 resolution of nested arguments. We should capture this functionality
3449 in a test to ensure that we continue to test it as we move forward. I
3450 can take a look at implementing this once I get a chance.
3452 Looks like the problem may be in [[file:lisp/org-babel.el::defun%20org%20babel%20merge%20params%20rest%20plists][org-babel-merge-params]], which seems
3453 to be trampling the provided :vars values.
3455 Nope, now it seems that we are actually looking up the results line,
3456 rather than the actual source-code block, which would make sense given
3457 that the results-line will return the same value regardless of the
3458 arguments supplied. See the output of this [[file:lisp/org-babel-ref.el::message%20type%20S%20type%20debugging][debug-statement]].
3460 We need to be sure that we don't read from a =#+resname:= line when we
3461 have a non-nil set of arguments.
3464 After uncommenting the debugging statements located [[file:lisp/org-babel-ref.el::message%20format%20first%20second%20S%20S%20new%20refere%20new%20referent%20debugging][here]] and more
3465 importantly [[file:lisp/org-babel-ref.el::message%20nested%20args%20S%20args%20debugging][here]], we can see that the current reference code does
3466 evaluate the references correctly, and it uses the =:var= header
3467 argument to set =a=8=, however the default variables specified using
3468 the functional syntax in =adder(a=3, b=2)= is overriding this
3471 ***** doesn't work with functional syntax
3473 #+srcname: adder-func(a=3, b=2)
3478 #+resname: adder-func
3481 #+srcname: after-adder-func(arg=adder-func(a=8))
3486 #+resname: after-adder-func
3489 ***** still does work with =:var= syntax
3491 so it looks like regardless of the syntax used we're not overriding
3492 the default argument values.
3494 #+srcname: adder-header
3495 #+begin_src python :var a=3 :var b=2
3499 #+resname: adder-header
3502 #+srcname: after-adder-header
3503 #+begin_src python :var arg=adder-header(a=8, b=0)
3507 #+resname: after-adder-header
3510 *** Set of test cases
3511 **** Both defaults provided in definition
3512 #+srcname: adder1(a=10,b=20)
3519 ****** DONE Rely on defaults
3527 ******* DONE empty parens () not recognised as lob call
3528 E.g. remove spaces between parens above
3530 updated [[file:lisp/org-babel-lob.el::defvar%20org%20babel%20lob%20one%20liner%20regexp%20lob%20t%20n%20n%20t%20n][org-babel-lob-one-liner-regexp]]
3532 ****** DONE One supplied, one default
3535 #+resname: adder1(a=0)
3542 #+resname: adder1(b=0)
3546 ****** DONE Both supplied
3547 #+lob: adder1(a=1,b=2)
3549 #+resname: adder1(a=1,b=2)
3552 **** One arg lacks default in definition
3553 #+srcname: adder2(a=10,b)
3557 ****** DEFERRED Rely on defaults (one of which is missing)
3562 ## should be error: b has no default
3564 Maybe we should let the programming language handle this case. For
3565 example python spits out an error in the =#+lob= line above. Maybe
3566 rather than catching these errors our-selves we should raise an error
3567 when the source-block returns an error. I'll propose a [[* PROPOSED raise elisp error when source-blocks return errors][task]] for this
3568 idea, I'm not sure how/if it would work...
3570 ****** DEFERRED Default over-ridden
3573 See the above [[* DEFERRED Rely on defaults (one of which is missing)][deferred]] and the new proposed [[* PROPOSED raise elisp error when source-blocks return errors][task]], I think it may be
3574 more flexible to allow the source block language to handle the error.
3577 ## should be error: b has no default
3579 ****** DONE Missing default supplied
3582 #+resname: adder2(b=1)
3589 ****** DONE One over-ridden, one supplied
3590 #+lob: adder2(a=1,b=2)
3592 #+resname: adder2(a=1,b=2)
3597 *** Example that fails
3599 #+srcname: adder(a=0, b=99)
3612 #+srcname: level-one-nesting()
3613 #+begin_src python :var arg=adder(a=one(),b=one())
3617 #+resname: level-one-nesting
3621 #+srcname: level-one-nesting()
3622 #+begin_src python :var arg=adder(a=adder(a=one(),b=one()),b=adder(a=one(),b=one()))
3629 *** DONE deeply nested arguments still fails
3631 #+srcname: deeply-nested-args-bug
3632 #+begin_src python :var arg=adder(a=adder(a=one(),b=one()),b=adder(a=one(),b=one()))
3639 **** Used to result in this error
3640 : supplied params=nil
3641 : new-refere="adder", new-referent="a=adder(a=one(),b=one()),b=adder(a=one(),b=one())"
3642 : args=((:var . "a=adder(a=one()") (:var . "b=one())") (:var . "b=adder(a=one()") (:var . "b=one())"))
3644 : supplied params=((:var . "a=adder(a=one()") (:var . "b=one())") (:var . "b=adder(a=one()") (:var . "b=one())"))
3645 : new-refere="adder", new-referent="a=one("
3646 : args=((:var . "a=one("))
3648 : supplied params=((:var . "a=one("))
3649 : reference 'one(' not found in this buffer
3651 Need to change the regexp in [[file:lisp/org-babel-ref.el::assign%20any%20arguments%20to%20pass%20to%20source%20block][org-babel-ref-resolve-reference]] so that
3652 it only matches when the parenthesis are balanced. Maybe look at
3653 [[http://www.gnu.org/software/emacs/elisp/html_node/List-Motion.html][this]].
3655 *** DONE Still some problems with deeply nested arguments and defaults
3657 **** DONE Parsing / defaults bug
3658 Try inserting a space between 'a=0,' and 'b=0' and comparing results
3659 #+srcname: parsing-defaults-bug()
3660 #+begin_src python :var arg=adder(a=adder(a=0,b=0))
3664 #+resname: parsing-defaults-bug
3668 #+srcname: deeply-nested-args-bug-orig()
3669 #+begin_src python :var arg=adder(a=adder(a=one(),b=one()),b=adder(a=adder(a=3, b=4),b=one()))
3673 #+resname: deeply-nested-args-bug-orig
3678 **** DONE Nesting problem II
3679 This generates parsing errors
3681 Fixed: c2bef96b7f644c05be5a38cad6ad1d28723533aa
3683 #+srcname: deeply-nested-args-bug-II-1()
3684 #+begin_src python :var arg=adder(a=adder(a=one(),b=adder(a=2,b=4)))
3688 #+resname: deeply-nested-args-bug-II-1
3692 #+srcname: deeply-nested-args-bug-II-original()
3693 #+begin_src python :var arg=adder(a=adder(a=one(),b=one()),b=adder(a=one(),b=adder(a=1,b=4)))
3697 #+resname: deeply-nested-args-bug-II-original
3702 **** DONE Why does this give 8?
3703 It was picking up the wrong definition of adder
3704 #+srcname: deeply-nested-args-bug-2()
3705 #+begin_src python :var arg=adder(a=adder(a=one(),b=one()))
3709 #+resname: deeply-nested-args-bug-2
3713 **** DONE Problem with empty argument list
3714 This gives empty list with () and 'no output' with ( )
3716 I think this is OK now.
3719 #+begin_src python :var arg=adder( )
3726 #+srcname: deeply-nested-args-bug-orig()
3727 #+begin_src python :var arg=adder(a=adder(a=one(),b=one()),b=adder(a=adder(a=3, b=4),b=one()))
3731 #+resname: deeply-nested-args-bug-orig
3736 **** DONE Nesting problem II
3737 This generates parsing errors
3739 Fixed: c2bef96b7f644c05be5a38cad6ad1d28723533aa
3741 #+srcname: deeply-nested-args-bug-II-1()
3742 #+begin_src python :var arg=adder(a=adder(a=one(),b=adder(a=2,b=4)))
3746 #+resname: deeply-nested-args-bug-II-1
3750 #+srcname: deeply-nested-args-bug-II-original()
3751 #+begin_src python :var arg=adder(a=adder(a=one(),b=one()),b=adder(a=one(),b=adder(a=1,b=4)))
3755 #+resname: deeply-nested-args-bug-II-original
3760 **** DONE Why does this give 8?
3761 It was picking up the wrong definition of adder
3762 #+srcname: deeply-nested-args-bug-2()
3763 #+begin_src python :var arg=adder(a=adder(a=one(),b=one()))
3767 #+resname: deeply-nested-args-bug-2
3771 **** DONE Problem with empty argument list
3772 This gives empty list with () and 'no output' with ( )
3774 I think this is OK now.
3777 #+begin_src python :var arg=adder( )
3810 #+srcname: test-zz(arg=adder(a=adder(a=19,b=adder(a=5,b=2)),b=adder(a=adder(a=1,b=9),b=adder(a=1,b=3))))
3816 *** DONE Arg lacking default
3817 This would be good thing to address soon. I'm imagining that
3818 e.g. here, the 'caller' block would return the answer 30. I believe
3819 there's a few issues here: i.e. the naked 'a' without a reference
3820 is not understood; the default arg b=6 is not understood.
3822 #+srcname: adder-with-arg-lacking-default(a, b=6)
3829 #+srcname: caller(var=adder-with-arg-lacking-default(a=24))
3830 #+begin_src python :results silent
3834 ** DONE allow srcname to omit function call parentheses
3835 Someone needs to revisit those regexps. Is there an argument for
3836 moving some of the regexps used to match function calls into
3837 defvars? (i.e. in o-b.el and o-b-ref.el)
3839 This seems to work now. It still might be a good idea to separate
3840 out some of the key regexps into defvars at some point.
3842 #+srcname: omit-parens-test
3843 #+begin_src ruby :results output
3852 ** DONE avoid stripping whitespace from output when :results output
3853 This may be partly solved by using o-b-chomp rather than o-b-trim
3854 in the o-b-LANG-evaluate functions.
3855 ** DEFERRED weird escaped characters in shell prompt break shell evaluation
3856 E.g. this doesn't work. Should the shell sessions set a sane prompt
3857 when they start up? Or is it a question of altering
3858 comint-prompt-regexp? Or altering org-babel regexps?
3861 black=30 ; red=31 ; green=32 ; yellow=33 ; blue=34 ; magenta=35 ; cyan=36 ; white=37
3864 export PS1="\[\033[${prompt_col}m\]\w${prompt_char} \[\033[0m\]"
3867 I just pushed a good amount of changes, could you see if your shell
3868 problems still exist?
3870 The problem's still there. Specifically, aIui, at [[file:lisp/langs/org-babel-sh.el::raw%20org%20babel%20comint%20with%20output%20buffer%20org%20babel%20sh%20eoe%20output%20nil%20insert%20full%20body%20comint%20send%20input%20nil%20t][this line]] of
3871 org-babel-sh.el, raw gets the value
3873 ("" "
\e[0m Sun Jun 14 19:26:24 EDT 2009\n" "
\e[0m org_babel_sh_eoe\n" "
\e[0m ")
3875 and therefore (member org-babel-sh-eoe-output ...) fails
3877 I think that `comint-prompt-regexp' needs to be altered to match
3878 the shell prompt. This shouldn't be too difficult to do by hand,
3879 using the `regexp-builder' command and should probably be part of
3880 the user's regular emacs init. I can't think of a way for us to
3881 set this automatically, and we are SOL without a regexp to match
3883 ** DONE function calls in #+srcname: refs
3885 My srcname references don't seem to be working for function
3886 calls. This needs fixing.
3893 srcname function call doesn't work for calling a source block
3894 #+srcname: caller(var1=called())
3907 They do work for a simple reference
3908 #+srcname: caller2(var1=56)
3917 and they do work for :var header arg
3919 #+begin_src python :var var1=called()
3925 ** DONE LoB: with output to buffer, not working in buffers other than library-of-babel.org
3927 I haven't fixed this yet. org-babel-ref-resolve-reference moves
3928 point around, inside a save-excursion. Somehow when it comes to
3929 inserting the results (after possible further recursive calls to
3930 org-babel-ref-resolve-reference), point hasn't gone back to the
3933 #+tblname: test-data
3938 #+lob: R-plot(data=test-data)
3940 #+lob: python-add(a=2, b=9)
3942 #+resname: python-add(a=2, b=9)
3946 I think this got fixed in the bugfixes before merging results into master.
3948 ** DONE cursor movement when evaluating source blocks
3949 E.g. the pie chart example. Despite the save-window-excursion in
3950 org-babel-execute:R. (I never learned how to do this properly: org-R
3951 jumps all over the place...)
3953 I don't see this now [ded]
3955 ** DONE LoB: calls fail if reference has single character name
3956 commit 21d058869df1ff23f4f8cc26f63045ac9c0190e2
3957 **** This doesn't work
3958 #+lob: R-plot(data=X)
3977 #+lob: R-plot(data=XX)
3979 ** DONE make :results replace the default?
3980 I'm tending to think that appending results to pre-existing results
3981 creates mess, and that the cleaner `replace' option should be the
3982 default. E.g. when a source block creates an image, we would want
3983 that to be updated, rather than have a new one be added.
3987 ** DONE ruby evaluation not working under ubuntu emacs 23
3988 With emacs 23.0.91.1 on ubuntu, for C-h f run-ruby I have the
3989 following, which seems to conflict with [[file:lisp/langs/org-babel-ruby.el::let%20session%20buffer%20save%20window%20excursion%20run%20ruby%20nil%20session%20current%20buffer][this line]] in org-babel-ruby.el.
3992 run-ruby is an interactive compiled Lisp function.
3996 Run an inferior Ruby process, input and output via buffer *ruby*.
3997 If there is a process already running in `*ruby*', switch to that buffer.
3998 With argument, allows you to edit the command line (default is value
3999 of `ruby-program-name'). Runs the hooks `inferior-ruby-mode-hook'
4000 (after the `comint-mode-hook' is run).
4001 (Type C-h m in the process buffer for a list of commands.)
4004 So, I may have a non-standard inf-ruby.el. Here's my version of
4008 run-ruby is an interactive Lisp function in `inf-ruby.el'.
4010 (run-ruby &optional COMMAND NAME)
4012 Run an inferior Ruby process, input and output via buffer *ruby*.
4013 If there is a process already running in `*ruby*', switch to that buffer.
4014 With argument, allows you to edit the command line (default is value
4015 of `ruby-program-name'). Runs the hooks `inferior-ruby-mode-hook'
4016 (after the `comint-mode-hook' is run).
4017 (Type C-h m in the process buffer for a list of commands.)
4020 It seems we could either bundle my version of inf-ruby.el (as it's
4021 the newest). Or we could change the use of `run-ruby' so that it
4022 is robust across multiple distributions. I think I'd prefer the
4023 former, unless the older version of inf-ruby is actually bundled
4024 with emacs, in which case maybe we should go out of our way to
4025 support it. Thoughts?
4027 I think for now I'll just include the latest [[file:util/inf-ruby.el][inf-ruby.el]] in the
4028 newly created utility directory. I doubt anyone would have a
4029 problem using the latest version of this file.
4030 ** DONE test failing forcing vector results with =test-forced-vector-results= ruby code block
4031 Note that this only seems to happen the *second* time the test table
4034 #+srcname: bug-trivial-vector
4035 #+begin_src emacs-lisp :results vector silent
4039 #+srcname: bug-forced-vector-results
4040 #+begin_src ruby :var triv=test-trivial-vector :results silent
4044 mysteriously this seems to be fixed...
4045 ** DONE defunct R sessions
4046 Sometimes an old R session will turn defunct, and newly inserted code
4047 will not be evaluated (leading to a hang).
4049 This seems to be fixed by using `inferior-ess-send-input' rather than `comint-send-input'.
4050 ** DONE ruby fails on first call to non-default session
4052 #+srcname: bug-new-session
4053 #+begin_src ruby :session is-new
4057 ** DONE when reading results from =#+resname= line
4059 Errors when trying to read from resname lines.
4061 #+resname: bug-in-resname
4064 #+srcname: bug-in-resname-reader
4065 #+begin_src emacs-lisp :var buggy=bug-in-resname() :results silent
4069 ** DONE R-code broke on "org-babel" rename
4071 #+srcname: bug-R-babels
4076 ** DONE error on trivial R results
4078 So I know it's generally not a good idea to squash error without
4079 handling them, but in this case the error almost always means that
4080 there was no file contents to be read by =org-table-import=, so I
4083 #+srcname: bug-trivial-r1
4084 #+begin_src R :results replace
4085 pie(c(1, 2, 3), labels = c(1, 2, 3))
4088 #+srcname: bug-trivial-r2
4089 #+begin_src R :results replace
4093 #+resname: bug-trivial-r2
4096 #+srcname: bug-trivial-r3
4097 #+begin_src R :results replace
4101 #+resname: bug-trivial-r3
4106 ** DONE ruby new variable creation (multi-line ruby blocks)
4107 Actually it looks like we were dropping all but the last line.
4109 #+srcname: multi-line-ruby-test
4110 #+begin_src ruby :var table=bug-numerical-table :results replace
4112 table.each{|n| total += n}
4119 ** DONE R code execution seems to choke on certain inputs
4120 Currently the R code seems to work on vertical (but not landscape)
4123 #+srcname: little-fake
4124 #+begin_src emacs-lisp
4129 #+begin_src R :var num=little-fake
4136 #+srcname: set-debug-on-error
4137 #+begin_src emacs-lisp :results silent
4138 (setq debug-on-error t)
4141 #+srcname: bug-numerical-table
4142 #+begin_src emacs-lisp :results silent
4149 #+srcname: bug-R-number-evaluation
4150 #+begin_src R :var table=bug-numerical-table
4159 #+tblname: bug-vert-table
4164 #+srcname: bug-R-vertical-table
4165 #+begin_src R :var table=bug-vert-table :results silent
4169 ** DONE org bug/request: prevent certain org behaviour within code blocks
4170 E.g. [[]] gets recognised as a link (when there's text inside the
4171 brackets). This is bad for R code at least, and more generally
4172 could be argued to be inappropriate. Is it difficult to get org to
4173 ignore text in code blocks? [DED]
4175 I believe Carsten addressed this recently on the mailing list with
4176 the comment that it was indeed a difficult issue. I believe this
4177 may be one area where we could wait for an upstream (org-mode) fix.
4179 [Dan] Carsten has fixed this now in the core.
4181 ** DONE with :results replace, non-table output doesn't replace table output
4182 And vice versa. E.g. Try this first with table and then with len(table) [DED]
4183 #+begin_src python :var table=sandbox :results replace
4188 | 4 | "schulte" | 6 |
4191 Yes, this is certainly a problem. I fear that if we begin replacing
4192 anything immediately following a source block (regardless of whether
4193 it matches the type of our current results) we may accidentally delete
4194 hand written portions of the user's org-mode buffer.
4196 I think that the best solution here would be to actually start
4197 labeling results with a line that looks something like...
4201 This would have a couple of benefits...
4202 1) we wouldn't have to worry about possibly deleting non-results
4203 (which is currently an issue)
4204 2) we could reliably replace results even if there are different types
4205 3) we could reference the results of a source-code block in variable
4206 definitions, which would be useful if for example we don't wish to
4207 re-run a source-block every time because it is long-running.
4209 Thoughts? If no-one objects, I believe I will implement the labeling
4212 ** DONE extra quotes for nested string
4213 Well R appears to be reading the tables without issue...
4215 these *should* be quoted
4217 #+begin_src sh :results replace
4222 | "README.markdown" |
4225 | "existing_tools" |
4229 | "test-export.html" |
4230 | "test-export.org" |
4232 #+srcname: test-quotes
4233 #+begin_src ruby :var tab=ls
4239 #+srcname: test-quotes
4240 #+begin_src R :var tab=ls
4246 ** DONE simple ruby arrays not working
4248 As an example eval the following. Adding a line to test
4250 #+tblname: simple-ruby-array
4253 #+srcname: ruby-array-test
4254 #+begin_src ruby :var ar = simple-ruby-array :results silent
4258 ** DONE space trailing language name
4259 fix regexp so it works when there's a space trailing the language name
4261 #+srcname: test-trailing-space
4266 ** DONE Args out of range error
4268 The following block resulted in the error below [DED]. It ran without
4269 error directly in the shell.
4272 for platf in ill aff ; do
4273 for pop in CEU YRI ASI ; do
4274 rm -f $platf/hapmap-genos-$pop-all $platf/hapmap-rs-all
4275 cat $platf/hapmap-genos-$pop-* > $platf/hapmap-genos-$pop-all
4276 cat $platf/hapmap-rs-* > $platf/hapmap-rs-all
4281 executing source block with sh...
4282 finished executing source block
4283 string-equal: Args out of range: "", -1, 0
4285 the error =string-equal: Args out of range: "", -1, 0= looks like what
4286 used to be output when the block returned an empty results string.
4287 This should be fixed in the current version, you should now see the
4288 following message =no result returned by source block=.
4290 ** DONE ruby arrays not recognized as such
4292 Something is wrong in [[file:lisp/org-babel-script.el]] related to the
4293 recognition of ruby arrays as such.
4295 #+begin_src ruby :results replace
4301 #+begin_src python :results replace
4306 ** REJECTED elisp reference fails for literal number
4307 That's a bug in Dan's elisp, not in org-babel.
4308 #+srcname: elisp-test(a=4)
4309 #+begin_src emacs-lisp
4321 Evaluate all the cells in this table for a comprehensive test of the
4322 org-babel functionality.
4324 *Note*: if you have customized =org-babel-default-header-args= then some
4325 of these tests may fail.
4327 #+TBLNAME: org-babel-tests
4328 | functionality | block | arg | expected | results | pass |
4329 |-------------------------+----------------------------+-----+-------------+-------------+------|
4330 | basic evaluation | | | | | pass |
4331 |-------------------------+----------------------------+-----+-------------+-------------+------|
4332 | emacs lisp | basic-elisp | | 5 | 5 | pass |
4333 | shell | basic-shell | | 6 | 6 | pass |
4334 | ruby | basic-ruby | | org-babel | org-babel | pass |
4335 | python | basic-python | | hello world | hello world | pass |
4336 | R | basic-R | | 13 | 13 | pass |
4337 |-------------------------+----------------------------+-----+-------------+-------------+------|
4338 | tables | | | | | pass |
4339 |-------------------------+----------------------------+-----+-------------+-------------+------|
4340 | emacs lisp | table-elisp | | 3 | 3 | pass |
4341 | ruby | table-ruby | | 1-2-3 | 1-2-3 | pass |
4342 | python | table-python | | 5 | 5 | pass |
4343 | R | table-R | | 3.5 | 3.5 | pass |
4344 | R: col names in R | table-R-colnames | | -3 | -3 | pass |
4345 | R: col names in org | table-R-colnames-org | | 169 | 169 | pass |
4346 |-------------------------+----------------------------+-----+-------------+-------------+------|
4347 | source block references | | | | | pass |
4348 |-------------------------+----------------------------+-----+-------------+-------------+------|
4349 | all languages | chained-ref-last | | Array | Array | pass |
4350 |-------------------------+----------------------------+-----+-------------+-------------+------|
4351 | source block functions | | | | | pass |
4352 |-------------------------+----------------------------+-----+-------------+-------------+------|
4353 | emacs lisp | defun-fibb | | fibbd | fibbd | pass |
4354 | run over | Fibonacci | 0 | 1 | 1 | pass |
4355 | a | Fibonacci | 1 | 1 | 1 | pass |
4356 | variety | Fibonacci | 2 | 2 | 2 | pass |
4357 | of | Fibonacci | 3 | 3 | 3 | pass |
4358 | different | Fibonacci | 4 | 5 | 5 | pass |
4359 | arguments | Fibonacci | 5 | 8 | 8 | pass |
4360 |-------------------------+----------------------------+-----+-------------+-------------+------|
4361 | bugs and tasks | | | | | pass |
4362 |-------------------------+----------------------------+-----+-------------+-------------+------|
4363 | simple ruby arrays | ruby-array-test | | 3 | 3 | pass |
4364 | R number evaluation | bug-R-number-evaluation | | 2 | 2 | pass |
4365 | multi-line ruby blocks | multi-line-ruby-test | | 2 | 2 | pass |
4366 | forcing vector results | test-forced-vector-results | | Array | Array | pass |
4367 | deeply nested arguments | deeply-nested-args-bug | | 4 | 4 | pass |
4368 |-------------------------+----------------------------+-----+-------------+-------------+------|
4369 | sessions | | | | | pass |
4370 |-------------------------+----------------------------+-----+-------------+-------------+------|
4371 | set ruby session | set-ruby-session-var | | :set | :set | pass |
4372 | get from ruby session | get-ruby-session-var | | 3 | 3 | pass |
4373 | set python session | set-python-session-var | | set | set | pass |
4374 | get from python session | get-python-session-var | | 4 | 4 | pass |
4375 | set R session | set-R-session-var | | set | set | pass |
4376 | get from R session | get-R-session-var | | 5 | 5 | pass |
4377 #+TBLFM: $5='(if (= (length $3) 1) (progn (message (format "running %S" '(sbe $2 (n $3)))) (sbe $2 (n $3))) (sbe $2))::$6='(if (string= $4 $5) "pass" (format "expected %S but was %S" $4 $5))
4378 #+TBLFM: $5=""::$6=""
4381 The second TBLFM line (followed by replacing '[]' with '') can be used
4382 to blank out the table results, in the absence of a better method.
4386 #+srcname: basic-elisp
4387 #+begin_src emacs-lisp :results silent
4392 #+srcname: basic-shell
4393 #+begin_src sh :results silent
4398 #+srcname: date-simple
4399 #+begin_src sh :results silent
4403 #+srcname: basic-ruby
4404 #+begin_src ruby :results silent
4409 #+srcname: basic-python
4410 #+begin_src python :results silent
4416 #+begin_src R :results silent
4424 #+tblname: test-table
4428 #+tblname: test-table-colnames
4429 | var1 | var2 | var3 |
4430 |------+------+------|
4434 #+srcname: table-elisp
4435 #+begin_src emacs-lisp :results silent :var table=test-table
4436 (length (car table))
4440 #+srcname: table-ruby
4441 #+begin_src ruby :results silent :var table=test-table
4442 table.first.join("-")
4446 #+srcname: table-python
4447 #+begin_src python :var table=test-table
4451 #+srcname: table-R(table=test-table)
4456 #+srcname: table-R-colnames(table=test-table-colnames)
4457 #+begin_src R :results silent
4458 sum(table$var2 - table$var3)
4461 #+srcname: R-square(x=default-name-doesnt-exist)
4462 #+begin_src R :colnames t
4466 This should return 169. The fact that R is able to use the column name
4467 to index the data frame (x$var3) proves that a table with column names
4468 (a header row) has been recognised as input for the R-square function
4469 block, and that the R-square block has output an elisp table with
4470 column names, and that the colnames have again been recognised when
4471 creating the R variables in this block.
4472 #+srcname: table-R-colnames-org(x = R-square(x=test-table-colnames))
4482 Lets pass a references through all of our languages...
4484 Lets start by reversing the table from the previous examples
4486 #+srcname: chained-ref-first
4487 #+begin_src python :var table = test-table
4492 #+resname: chained-ref-first
4496 Take the first part of the list
4498 #+srcname: chained-ref-second
4499 #+begin_src R :var table = chained-ref-first
4503 #+resname: chained-ref-second
4507 Turn the numbers into string
4509 #+srcname: chained-ref-third
4510 #+begin_src emacs-lisp :var table = chained-ref-second
4511 (mapcar (lambda (el) (format "%S" el)) table)
4514 #+resname: chained-ref-third
4517 and Check that it is still a list
4519 #+srcname: chained-ref-last
4520 #+begin_src ruby :var table=chained-ref-third
4525 ** source blocks as functions
4527 #+srcname: defun-fibb
4528 #+begin_src emacs-lisp :results silent
4529 (defun fibbd (n) (if (< n 2) 1 (+ (fibbd (- n 1)) (fibbd (- n 2)))))
4533 #+srcname: fibonacci
4534 #+begin_src emacs-lisp :results silent :var n=7
4544 ** sbe tests (these don't seem to be working...)
4545 Testing the insertion of results into org-mode tables.
4547 #+srcname: multi-line-output
4548 #+begin_src ruby :results replace
4549 "the first line ends here
4552 and this is the second one
4558 : the first line ends here\n\n\n and this is the second one\n\neven a third
4560 #+srcname: multi-line-error
4561 #+begin_src ruby :results replace
4562 raise "oh nooooooooooo"
4568 | the first line ends here... | -:5: warning: parenthesize argument(s) for future version... |
4569 #+TBLFM: $1='(sbe "multi-line-output")::$2='(sbe "multi-line-error")
4571 ** forcing results types tests
4573 #+srcname: test-trivial-vector
4574 #+begin_src emacs-lisp :results vector silent
4578 #+srcname: test-forced-vector-results
4579 #+begin_src ruby :var triv=test-trivial-vector :results silent
4585 #+srcname: set-ruby-session-var
4586 #+begin_src ruby :session :results silent
4591 #+srcname: get-ruby-session-var
4592 #+begin_src ruby :session :results silent
4596 #+srcname: set-python-session-var
4597 #+begin_src python :session
4602 #+srcname: get-python-session-var
4603 #+begin_src python :session
4607 #+srcname: set-R-session-var
4608 #+begin_src R :session
4613 #+srcname: get-R-session-var
4614 #+begin_src R :session
4623 To run these examples evaluate [[file:lisp/org-babel-init.el][org-babel-init.el]]
4625 ** org-babel.el beginning functionality
4627 #+begin_src sh :results replace
4632 : Sun Jul 5 18:54:39 EDT 2009
4639 : Sun Jul 05 18:54:35 -0400 2009
4651 #+begin_src R :results replace
4661 hist(rgamma(20,3,3))
4666 ** org-babel plays with tables
4667 Alright, this should demonstrate both the ability of org-babel to read
4668 tables into a lisp source code block, and to then convert the results
4669 of the source code block into an org table. It's using the classic
4670 "lisp is elegant" demonstration transpose function. To try this
4673 1. evaluate [[file:lisp/org-babel-init.el]] to load org-babel and friends
4674 2. evaluate the transpose definition =\C-c\\C-c= on the beginning of
4676 3. evaluate the next source code block, this should read in the table
4677 because of the =:var table=previous=, then transpose the table, and
4678 finally it should insert the transposed table into the buffer
4679 immediately following the block
4683 #+begin_src emacs-lisp :results silent
4684 (defun transpose (table)
4685 (apply #'mapcar* #'list table))
4693 #+begin_src emacs-lisp :var table=sandbox :results replace
4698 #+begin_src emacs-lisp
4703 | 1 | 2 | 3 | 4 | 5 |
4707 #+begin_src ruby :var table=sandbox :results replace
4708 table.first.join(" - ")
4714 #+begin_src python :var table=sandbox
4719 #+begin_src ruby :var table=sandbox :results replace
4724 : [[1, 2, 3], [4, "schulte", 6]]
4728 | 4 | "schulte" | 6 |
4730 #+begin_src python :var table=sandbox :results replace
4736 | "__add__" | "__class__" | "__contains__" | "__delattr__" | "__delitem__" | "__delslice__" | "__doc__" | "__eq__" | "__format__" | "__ge__" | "__getattribute__" | "__getitem__" | "__getslice__" | "__gt__" | "__hash__" | "__iadd__" | "__imul__" | "__init__" | "__iter__" | "__le__" | "__len__" | "__lt__" | "__mul__" | "__ne__" | "__new__" | "__reduce__" | "__reduce_ex__" | "__repr__" | "__reversed__" | "__rmul__" | "__setattr__" | "__setitem__" | "__setslice__" | "__sizeof__" | "__str__" | "__subclasshook__" | "append" | "count" | "extend" | "index" | "insert" | "pop" | "remove" | "reverse" | "sort" |
4738 *** (sandbox table) R
4740 #+TBLNAME: sandbox_r
4744 #+begin_src R :results replace
4745 x <- c(rnorm(10, mean=-3, sd=1), rnorm(10, mean=3, sd=1))
4749 | -3.35473133869346 |
4751 | -3.32819924928633 |
4752 | -2.97310212756194 |
4753 | -2.09640758369576 |
4754 | -5.06054014378736 |
4755 | -2.20713700711221 |
4756 | -1.37618039712037 |
4757 | -1.95839385821742 |
4758 | -3.90407396475502 |
4759 | 2.51168071590226 |
4760 | 3.96753011570494 |
4761 | 3.31793212627865 |
4762 | 1.99829753972341 |
4763 | 4.00403686419829 |
4764 | 4.63723764452927 |
4765 | 3.94636744261313 |
4766 | 3.58355906547775 |
4767 | 3.01563442274226 |
4770 #+begin_src R var tabel=sandbox_r :results replace
4775 | 4 | "schulte" | 6 |
4778 Now shell commands are converted to tables using =org-table-import=
4779 and if these tables are non-trivial (i.e. have multiple elements) then
4780 they are imported as org-mode tables...
4782 #+begin_src sh :results replace
4786 | "total" | 208 | "" | "" | "" | "" | "" | "" |
4787 | "-rw-r--r--" | 1 | "dan" | "dan" | 57 | 2009 | 15 | "block" |
4788 | "-rw-r--r--" | 1 | "dan" | "dan" | 35147 | 2009 | 15 | "COPYING" |
4789 | "-rw-r--r--" | 1 | "dan" | "dan" | 722 | 2009 | 18 | "examples.org" |
4790 | "drwxr-xr-x" | 4 | "dan" | "dan" | 4096 | 2009 | 19 | "existing_tools" |
4791 | "-rw-r--r--" | 1 | "dan" | "dan" | 2207 | 2009 | 14 | "intro.org" |
4792 | "drwxr-xr-x" | 2 | "dan" | "dan" | 4096 | 2009 | 18 | "org-babel" |
4793 | "-rw-r--r--" | 1 | "dan" | "dan" | 277 | 2009 | 20 | "README.markdown" |
4794 | "-rw-r--r--" | 1 | "dan" | "dan" | 11837 | 2009 | 18 | "rorg.html" |
4795 | "-rw-r--r--" | 1 | "dan" | "dan" | 61829 | 2009 | 19 | "#rorg.org#" |
4796 | "-rw-r--r--" | 1 | "dan" | "dan" | 60190 | 2009 | 19 | "rorg.org" |
4797 | "-rw-r--r--" | 1 | "dan" | "dan" | 972 | 2009 | 11 | "test-export.org" |
4800 ** silent evaluation
4808 #+begin_src ruby :results silent
4812 #+begin_src ruby :results replace
4819 ** (sandbox) referencing other source blocks
4820 Doing this in emacs-lisp first because it's trivial to convert
4821 emacs-lisp results to and from emacs-lisp.
4823 *** emacs lisp source reference
4824 This first example performs a calculation in the first source block
4825 named =top=, the results of this calculation are then saved into the
4826 variable =first= by the header argument =:var first=top=, and it is
4827 used in the calculations of the second source block.
4830 #+begin_src emacs-lisp
4834 #+begin_src emacs-lisp :var first=top :results replace
4840 This example is the same as the previous only the variable being
4841 passed through is a table rather than a number.
4843 #+begin_src emacs-lisp :results silent
4844 (defun transpose (table)
4845 (apply #'mapcar* #'list table))
4848 #+TBLNAME: top_table
4852 #+SRCNAME: second_src_example
4853 #+begin_src emacs-lisp :var table=top_table
4857 #+begin_src emacs-lisp :var table=second_src_example :results replace
4862 | 4 | "schulte" | 6 |
4864 Now working for ruby
4871 #+begin_src ruby :var other=start :results replace
4877 #+SRCNAME: start_two
4882 #+begin_src python :var another=start_two :results replace
4887 Since all variables are converted into Emacs Lisp it is no problem to
4888 reference variables specified in another language.
4890 #+SRCNAME: ruby-block
4895 #+SRCNAME: lisp_block
4896 #+begin_src emacs-lisp :var ruby-variable=ruby-block
4900 #+begin_src python :var lisp_var=lisp_block
4909 #+begin_src R :results replace
4916 #+begin_src R :var other=first_r :results replace
4923 ** (sandbox) selective export
4925 For exportation tests and examples see (including exportation of
4926 inline source code blocks) [[file:test-export.org]]
4929 ** (sandbox) source blocks as functions
4932 #+begin_src emacs-lisp :results silent
4937 #+begin_src emacs-lisp :var n=default :results replace
4943 #+begin_src emacs-lisp :var result=triple(n=3, m=98) :results replace
4949 The following just demonstrates the ability to assign variables to
4950 literal values, which was not implemented until recently.
4952 #+begin_src ruby :var num="eric" :results replace
4959 ** (sandbox) inline source blocks
4961 This is an inline source code block src_ruby{1 + 6}. And another
4962 source block with text output src_emacs-lisp{"eric"}.
4964 This is an inline source code block with header
4965 arguments. src_ruby[:var n=fibbd( n = 0 )]{n}
4968 ** (sandbox) integration w/org tables
4970 #+begin_src emacs-lisp :results silent
4971 (defun fibbd (n) (if (< n 2) 1 (+ (fibbd (- n 1)) (fibbd (- n 2)))))
4975 #+begin_src emacs-lisp :var n=4 :results silent
4979 #+begin_src emacs-lisp :results silent
4980 (mapcar #'fibbd '(0 1 2 3 4 5 6 7 8))
4983 Something is not working here. The function `sbe ' works fine when
4984 called from outside of the table (see the source block below), but
4985 produces an error when called from inside the table. I think there
4986 must be some narrowing going on during intra-table emacs-lisp
4989 | original | fibbd |
4990 |----------+-------|
5001 #+TBLFM: $2='(sbe "fibbd" (n $1))
5005 #+begin_src emacs-lisp :results silent
5006 (sbe 'fibbd (n "8"))
5011 LocalWords: DBlocks dblocks org-babel el eric fontification