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:t
4 #+STARTUP: oddeven hideblocks
6 Through Org-Babel Org-Mode can communicate with programming languages.
7 Code contained in source-code blocks can be evaluated and data can
8 pass seamlessly between different programming languages, Org-Mode
9 constructs (tables, file links, example text) and interactive comint
13 - The [[* Introduction][Introduction]] :: provides a brief overview of the design and use
14 of Org-Babel including tutorials and examples.
15 - In [[* Getting started][Getting Started]] :: find instructions for installing org-babel
16 into your emacs configuration.
17 - The [[* Tasks][Tasks]] :: section contains current and past tasks roughly ordered
18 by TODO state, then importance or date-completed. This would be
19 a good place to suggest ideas for development.
20 - The [[* Bugs][Bugs]] :: section contains bug reports.
21 - The [[* Tests][Tests]] :: section consists of a large table which can be
22 evaluated to run Org-Babel's functional test suite. This
23 provides a good overview of the current functionality with
24 pointers to example source blocks.
25 - The [[* Sandbox][Sandbox]] :: demonstrates much of the early/basic functionality
26 through commented source-code blocks.
28 Also see the [[file:library-of-babel.org][Library of Babel]], an extensible collection of ready-made
29 and easily-shortcut-callable source-code blocks for handling common
34 Org-Babel enables *communication* between programming languages and
38 - communication between programs :: Data passes seamlessly between
39 different programming languages, Org-Mode constructs (tables,
40 file links, example text) and interactive comint buffers.
41 - communication between people :: Data and calculations are embedded
42 in the same document as notes explanations and reports.
44 ** communication between programs
46 Org-Mode supports embedded blocks of source code (in any language)
47 inside of Org documents. Org-Babel allows these blocks of code to be
48 executed from within Org-Mode with natural handling of their inputs
52 with both scalar, file, and table output
54 *** reading information from tables
56 *** reading information from other source blocks (disk usage in your home directory)
58 This will work for Linux and Mac users, not so sure about shell
59 commands for windows users.
61 To run place the cursor on the =#+begin_src= line of the source block
62 labeled directory-pie and press =\C-c\C-c=.
64 #+srcname: directories
65 #+begin_src bash :results replace
66 cd ~ && du -sc * |grep -v total
69 #+resname: directories
71 | 11882808 | "Documents" |
72 | 8210024 | "Downloads" |
73 | 879800 | "Library" |
76 | 5307664 | "Pictures" |
85 #+srcname: directory-pie
86 #+begin_src R :var dirs = directories :session R-pie-example
87 pie(dirs[,1], labels = dirs[,2])
95 *** operations in/on tables
97 #+tblname: grades-table
98 | student | grade | letter |
99 |---------+-------+--------|
106 #+TBLFM: $2='(sbe random-score-generator)::$3='(sbe assign-grade (score $2))
108 #+srcname: assign-grade
109 #+begin_src ruby :var score=99
120 #+srcname: random-score-generator
125 #+srcname: show-distribution
126 #+begin_src R :var grades=grades-table :session *R*
134 ** communication between people
135 Quick overview of Org-Mode's exportation abilities, with links to the
136 online Org-Mode documentation, a focus on source-code blocks, and the
137 exportation options provided by Org-Babel.
139 *** Interactive tutorial
140 This would demonstrate applicability to Reproducible Research, and
141 Literate Programming.
143 *** Tests embedded in documentation
144 org-babels own functional tests are contained in a large org-mode
145 table, allowing the test suite to be run be evaluation of the table
146 and the results to be collected in the same table.
148 *** Emacs initialization files stored in Org-Mode buffers
149 Using `org-babel-tangle' it is possible to embed your Emacs
150 initialization into org-mode files. This allows for folding,
151 note-taking, todo's etc... embedded with the source-code of your Emacs
152 initialization, and through org-mode's publishing features aids in
153 sharing your customizations with others.
155 It may be worthwhile to create a fork of Phil Hagelberg's
156 [[http://github.com/technomancy/emacs-starter-kit/tree/master][emacs-starter-kit]] which uses literate org-mode files for all of the
157 actual elisp customization. These org-mode files could then be
158 exported to html and used to populate the repositories wiki on [[http://github.com/][github]].
163 *** code evaluation (comint buffer sessions and external processes)
164 There are two main ways to evaluate source blocks with org-babel.
166 - external :: By default (if the =:session= header argument is not
167 present) all source code blocks are evaluated in
168 external processes. In these cases an external process
169 is used to evaluate the source-code blocks.
170 - session :: Session based evaluation uses persistent sessions in
171 comint buffers. Sessions can be used across multiple
172 source blocks setting and accessing variables in the
175 Evaluating source blocks in sessions also allows for
176 interaction with the code. To jump to the session of a
177 source block use the `org-babel-pop-to-session' command
178 or press =M-[down]= while inside of a source code block.
179 When called with a prefix argument
180 `org-babel-pop-to-session' will evaluate all header
181 arguments before jumping to the source-code block.
183 *** results (values and outputs)
184 Either the *value* or the *output* of source code blocks can be
185 collected after evaluation.
187 - value :: The default way to collect results from a source-code block
188 is to return the value of the last statement in the block.
189 This can be thought of as the return value of the block.
190 In this case any printed output of the block is ignored.
191 This can be though of a similar to a "functional" value of
193 - output :: Another way of generating results from a source-code block
194 is to collect the output generated by the execution of the
195 block. In this case all printed output is collected
196 throughout the execution of the block. This can be
197 thought of as similar to a "script" style of evaluation.
201 Add the following lines to your .emacs, replacing the path as
202 appropriate. A good place to check that things are up and running
203 would then be [[#sandbox][the sandbox]].
204 #+begin_src emacs-lisp
205 (add-to-list 'load-path "/path/to/org-babel/lisp")
206 (require 'org-babel-init)
211 ** PROPOSED implement fully recursive evaluation machinery
212 The current parser / evaluator fails with greater levels of nested
213 function block calls (example below). If we want to overcome this I
214 think we'd have to redesign some of the evaluation
215 mechanism. Seeing as we are also facing issues like dealing with
216 default argument values, and seeing as we now know how we want the
217 library of babel to behave in addition to the source blocks, now
218 might be a good time to think about this. It would be nice to do
219 the full thing at some point, but otoh we may not consider it a
222 AIui, there are two stages: (i) construct a parse tree, and (ii)
223 evaluate it and return the value at the root. In the parse tree
224 each node represents an unevaluated value (either a literal value
225 or a reference). Node v may have descendent nodes, which represent
226 values upon which node v's evaluation depends. Once that tree is
227 constructed, then we evaluate the nodes from the tips towards the
228 root (a post-order traversal).
230 [This would also provide a solution for concatenating the STDOUTs
231 of called blocks, which is a [[*allow%20output%20mode%20to%20return%20stdout%20as%20value][task below]]; we concatenate them in
232 whatever order the traversal is done in.]
234 In addition to the variable references (i.e. daughter nodes), each
235 node would contain the information needed to evaluate that node
236 (e.g. lang body). Then we would pass a function postorder over the
237 tree which would call o-b-execute-src-block at each node, finally
238 returning the value at the root.
240 Fwiw I made a very tentative small start at stubbing this out in
241 org-babel-call.el in the 'evaluation' branch. And I've made a start
242 at sketching a parsing algorithm below.
245 I believe that this issue should be addressed as a bug rather than as
246 a point for new development. The code in [[file:lisp/org-babel-ref.el][org-babel-ref.el]] already
247 resolves variable references in a recursive manner which *should* work
248 in the same manner regardless of the depth of the number of nested
249 function calls. This recursive evaluation has the effect of
250 implicitly constructing the parse tree that your are thinking of
251 constructing explicitly.
253 Through using some of the commented out debugging statements in
254 [[file:lisp/org-babel-ref.el][org-babel-ref.el]] I have looked at what may be going wrong in the
255 current evaluation setup, and it seems that nested variables are being
256 set using the =:var= header argument, and these variables are being
257 overridden by the *default* variables which are being entered through
258 the new functional syntax (see the demonstration header below).
260 I believe that once this bug is fixed we should be back to fully
261 resolution of nested arguments. We should capture this functionality
262 in a test to ensure that we continue to test it as we move forward. I
263 can take a look at implementing this once I get a chance.
267 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
268 importantly [[file:lisp/org-babel-ref.el::message%20nested%20args%20S%20args%20debugging][here]], we can see that the current reference code does
269 evaluate the references correctly, and it uses the =:var= header
270 argument to set =a=8=, however the default variables specified using
271 the functional syntax in =adder(a=3, b=2)= is overriding this
274 #+srcname: adder(a=3, b=2)
283 #+srcname: after-adder(arg=adder(a=8))
288 #+resname: after-adder
291 *** Parse tree algorithm
292 Seeing as we're just trying to parse a string like
293 f(a=1,b=g(c=2,d=3)) it shouldn't be too hard. But of course there
294 are 'proper' parsers written in elisp out there,
295 e.g. [[http://cedet.sourceforge.net/semantic.shtml][Semantic]]. Perhaps we can find what we need -- our syntax is
296 pretty much the same as python and R isn't it?
298 Or, a complete hack, but maybe it would be we easy to transform it
299 to XML and then parse that with some existing tool?
301 But if we're doing it ourselves, something very vaguely like this?
302 (I'm sure there're lots of problems with this)
304 #+srcname: org-babel-call-parse(call)
306 ## we are currently reading a reference name: the name of the root function
309 for c in call_string:
312 whereami = "varname" # now we're reading a variable name
315 node.daughters = [node.daughters, new]
327 if whereami = "varname":
328 node.varnames[varnum] += c
329 elif whereami = "refname":
333 *** Example that fails
335 #+srcname: adder(a=0, b=0)
341 #+begin_src python :results silent
346 #+srcname: level-one-nesting
347 #+begin_src python :var arg=adder(a=one(),b=one())
352 #+srcname: level-one-nesting
353 #+begin_src python :var arg=adder(a=adder(a=one(),b=one()),b=adder(a=one(),b=one()))
357 ** PROPOSED Default args
358 This would be good thing to address soon. I'm imagining that
359 e.g. here, the 'caller' block would return the answer 30. I believe
360 there's a few issues here: i.e. the naked 'a' without a reference
361 is not understood; the default arg b=6 is not understood.
363 #+srcname: adder(a, b=6)
368 #+srcname: caller(var=adder(a=24))
373 ** PROPOSED allow `anonymous' function block with function call args?
374 My question here is simply whether we're going to allow
375 #+begin_src python(arg=ref)
379 but with preference given to
380 #+srcname blockname(arg=ref)
381 ** PROPOSED allow :result as synonym for :results?
382 ** PROPOSED allow 'output mode to return stdout as value?
383 Maybe we should allow this. In fact, if block x is called
384 with :results output, and it references blocks y and z, then
385 shouldn't the output of x contain a concatenation of the outputs of
386 y and z, together with x's own output? That would raise the
387 question of what happens if y is defined with :results output and z
388 with :results value. I guess z's (possibly vector/tabular) output
389 would be inside a literal example block containing the whole lot.
390 ** PROPOSED optional timestamp for output
391 Add option to place an (inactive) timestamp at the #+resname, to
392 record when that output was generated.
394 *** source code block timestamps (optional addition)
395 [Eric] If we did this would we then want to place a timestamp on the
396 source-code block, so that we would know if the results are
397 current or out of date? This would have the effect of caching the
398 results of calculations and then only re-running if the
399 source-code has changed. For the caching to work we would need to
400 check not only the timestamp on a source-code block, but also the
401 timestamps of any tables or source-code blocks referenced by the
402 original source-code block.
404 [Dan] I do remember getting frustrated by Sweave always having to
405 re-do everything, so this could be desirable, as long as it's easy
406 to over-ride of course. I'm not sure it should be the default
407 behaviour unless we are very confident that it works well.
409 **** maintaining source-code block timestamps
410 It may make sense to add a hook to `org-edit-special' which could
411 update the source-code blocks timestamp. If the user edits the
412 contents of a source-code block directly I can think of no
413 efficient way of maintaining the timestamp.
415 ** TODO make tangle files read-only?
416 With a file-local variable setting, yea that makes sense. Maybe
417 the header should reference the related org-mode file.
418 ** TODO take default values for header args from properties
419 Use file-wide and subtree wide properties to set default values for
421 ** TODO support for working with =*Org Edit Src Example*= buffers [2/4]
422 *** TODO optionally evaluate header references when we switch to =*Org Edit Src*= buffer
423 That seems to imply that the header references need to be evaluated
424 and transformed into the target language object when we hit C-c ' to
425 enter the *Org Edit Src* buffer [DED]
427 Good point, I heartily agree that this should be supported [Eric]
429 (or at least before the first time we attempt to evaluate code in that
430 buffer -- I suppose there might be an argument for lazy evaluation, in
431 case someone hits C-c ' but is "just looking" and not actually
432 evaluating anything.) Of course if evaluating the reference is
433 computationally intensive then the user might have to wait before they
434 get the *Org Edit Src* buffer. [DED]
436 I fear that it may be hard to anticipate when the references will be
437 needed, some major-modes do on-the-fly evaluation while the buffer is
438 being edited. I think that we should either do this before the buffer
439 is opened or not at all, specifically I think we should resolve
440 references if the user calls C-c ' with a prefix argument. Does that
441 sound reasonable? [Eric]
445 [Dan] So now that we have org-src-mode and org-src-mode-hook, I guess
446 org-babel should do this by using the hook to make sure that, when C-c
447 C-' is issued on a source block, any references are resolved and
448 assignments are made in the appropriate session.
449 *** TODO set buffer-local-process variables appropriately [DED]
450 I think something like this would be great. You've probably
451 already thought of this, but just to note it down: it would be really
452 nice if org-babel's notion of a buffer's 'session/process' played
453 nicely with ESS's notion of the buffer's session/process. ESS keeps
454 the current process name for a buffer in a buffer-local variable
455 ess-local-process-name. So one thing we will probably want to do is
456 make sure that the *Org Edit Src Example* buffer sets that variable
459 I had not thought of that, but I agree whole heartedly. [Eric]
461 Once this is done every variable should be able to dump regions into
462 their inferior-process buffer using major-mode functions.
463 *** DEFERRED send code to inferior process
464 Another thought on this topic: I think we will want users to send
465 chunks of code to the interpreter from within the *Org Edit Src*
466 buffer, and I think that's what you have in mind already. In ESS that
467 is done using the ess-eval-* functions. [DED]
469 I think we can leave this up to the major-mode in the source code
470 buffer, as almost every source-code major mode will have functions for
471 doing things like sending regions to the inferior process. If
472 anything we might need to set the value of the buffer local inferior
473 process variable. [Eric]
475 *** DONE some possible requests/proposed changes for Carsten [4/4]
476 While I remember, some possible requests/proposed changes for Carsten
477 come to mind in that regard:
479 **** DONE Remap C-x C-s to save the source to the org buffer?
480 I've done this personally and I find it essential. I'm using
481 #+begin_src emacs-lisp
482 (defun org-edit-src-save ()
483 "Update the parent org buffer with the edited source code, save
484 the parent org-buffer, and return to the source code edit
493 (define-key org-exit-edit-mode-map "\C-x\C-s" 'org-edit-src-save)
497 I think this is great, but I think it should be implemented in the
500 **** DEFERRED Rename buffer and minor mode?
501 Something shorter than *Org Edit Src Example* for the buffer
502 name. org-babel is bringing org's source code interaction to a
503 level of maturity where the 'example' is no longer
504 appropriate. And if further keybindings are going to be added to
505 the minor mode then maybe org-edit-src-mode is a better name than
508 Maybe we should name the buffer with a combination of the source
509 code and the session. I think that makes sense.
511 [ES] Are you also suggesting a new org-edit-src minor mode?
512 [DED] org-exit-edit-mode is a minor mode that already exists:
514 Minor mode installing a single key binding, "C-c '" to exit special edit.
516 org-edit-src-save now has a binding in that mode, so I guess all
517 I'm saying at this stage is that it's a bit of a misnomer. But
518 perhaps we will also have more functionality to add to that minor
519 mode, making it even more of a misnomer. Perhaps something like
520 org-src-mode would be better.
521 **** DONE Changed minor mode name and added hooks
523 **** DONE a hook called when the src edit buffer is created
524 This should be implemented in the org-mode core
526 ** TODO resolve references to other org buffers/files
527 This would allow source blocks to call upon tables, source-blocks,
528 and results in other org buffers/files.
531 - [[file:lisp/org-babel-ref.el::TODO%20allow%20searching%20for%20names%20in%20other%20buffers][org-babel-ref.el:searching-in-other-buffers]]
532 - [[file:lisp/org-babel.el::defun%20org-babel%20find%20named%20result%20name][org-babel.el#org-babel-find-named-result]]
533 ** TODO resolve references to other non-org files
534 - tabular data in .csv, .tsv etc format
535 - files of interpreted code: anything stopping us giving such files
536 similar status to a source code block?
537 - Would be nice to allow org and non-org files to be remote
538 ** TODO figure out how to handle errors during evaluation
539 I expect it will be hard to do this properly, but ultimately it
540 would be nice to be able to specify somewhere to receive STDERR,
541 and to be warned if it is non-empty.
543 Probably simpler in non-session evaluation than session? At least
544 the mechanism will be different I guess.
546 R has a try function, with error handling, along the lines of
547 python. I bet ruby does too. Maybe more of an issue for functional
548 style; in my proposed scripting style the error just gets dumped to
549 the org buffer and the user is thus alerted.
550 ** TODO figure out how to handle graphic output
551 This is listed under [[* graphical output][graphical output]] in out objectives.
553 This should take advantage of the =:results file= option, and
554 languages which almost always produce graphical output should set
555 =:results file= to true by default (this is currently done for the
556 gnuplot and ditaa languages). That would handle placing these results
557 in the buffer. Then if there is a combination of =silent= and =file=
558 =:results= headers we could drop the results to a temp buffer and pop
561 Display of file results is addressed in the [[* =\C-c \C-o= to open results of source block][open-results-task]].
563 *** TODO R graphics to screen means session evaluation
564 If R graphical output is going to screen then evaluation must be
565 in a session, otherwise the graphics will disappear as soon as the
568 *** Adding to a discussion started in email
569 I'm not deeply wedded to these ideas, just noting them down. I'm
570 probably just thinking of R and haven't really thought about how
571 this fits with the other graphics-generating languages.
573 > I used the approach below to get graphical file output
574 > today, which is one idea at least. Maybe it could be linked up with
575 > your :results file variable. (Or do we need a :results image for R?)
578 I don't think we need a special image results variable, but I may be
579 missing what the code below accomplishes. Would the task I added about
580 adding org-open-at-point functionality to source code blocks take care
583 Dan: I'm not sure. I think the ability for a script to generate both
584 text and graphical output might be a natural expectation, at least for
590 > #+srcname: cohort-scatter-plots-2d(org_babel_graphical_output_file="cohort-scatter-plots-2d.png")
592 > if(exists("org_babel_output_file"))
593 > png(filename=org_babel_graphical_output_file, width=1000, height=1000)
594 > ## plotting code in here
595 > if(exists("org_babel_graphical_output_file")) dev.off()
598 Dan: Yes, the results :file option is nice for dealing with graphical
599 output, and that could well be enough. Something based on the scheme
600 above would have a couple of points in its favour:
601 1. It's easy to switch between output going to on-screen graphics and
602 output going to file: Output will go to screen unless a string variable
603 with a standard name (e.g. ""org_babel_graphical_output_file"")
604 exists in which case it will go to the file indicated by the value
606 2. The block can return a result / script output, as well as produce
609 In interactive use we might want to allow the user to choose between
610 screen and file output. In non-interactive use such as export, it
611 would be file output (subject to the :exports directives).
613 ** TODO =\C-c \C-o= to open results of source block
614 by adding a =defadvice= to =org-open-at-point= we can use the common
615 =\C-c \C-o= keybinding to open the results of a source-code block.
616 This would be especially useful for source-code blocks which generate
617 graphical results and insert a file link as the results in the
618 org-mode buffer. (see [[* figure out how to handle graphic output][TODO figure out how to handle graphic output]]).
619 This could also act reasonably with other results types...
621 - file :: use org-open-at-point to open the file
622 - scalar :: open results unquoted in a new buffer
623 - tabular :: export the table to a new buffer and open that buffer
624 ** TODO Finalise behaviour regarding vector/scalar output
625 *** DONE Stop spaces causing vector output
626 This simple example of multilingual chaining produces vector output if
627 there are spaces in the message and scalar otherwise.
631 #+srcname: msg-from-R(msg=msg-from-python)
633 paste(msg, "und R", sep=" ")
637 : org-babel speaks elisp y python und R
639 #+srcname: msg-from-python(msg=msg-from-elisp)
644 #+srcname: msg-from-elisp(msg="org-babel speaks")
645 #+begin_src emacs-lisp
646 (concat msg " elisp")
648 ** STARTED share org-babel [1/4]
649 how should we share org-babel?
651 *** DONE post to org-mode
652 *** TODO post to ess mailing list
653 *** TODO create a org-babel page on worg
654 *** TODO create a short screencast demonstrating org-babel in action
657 we need to think up some good examples
659 **** interactive tutorials
660 This could be a place to use [[* org-babel assertions][org-babel assertions]].
662 for example the first step of a tutorial could assert that the version
663 of the software-package (or whatever) is equal to some value, then
664 source-code blocks could be used with confidence (and executed
665 directly from) the rest of the tutorial.
667 **** answering a text-book question w/code example
668 org-babel is an ideal environment enabling both the development and
669 demonstrationg of the code snippets required as answers to many
672 **** something using tables
673 maybe something along the lines of calculations from collected grades
676 Maybe something like the following which outputs sizes of directories
677 under the home directory, and then instead of the trivial =emacs-lisp=
678 block we could use an R block to create a nice pie chart of the
682 #+begin_src bash :results replace
686 #+begin_src emacs-lisp :var sizes=sizes :results replace
690 ** TODO command line execution
691 Allow source code blocks to be called form the command line. This
692 will be easy using the =sbe= function in [[file:lisp/org-babel-table.el][org-babel-table.el]].
694 This will rely upon [[* resolve references to other buffers][resolve references to other buffers]].
696 ** TODO inline source code blocks [3/5]
697 Like the =\R{ code }= blocks
699 not sure what the format should be, maybe just something simple
700 like =src_lang[]{}= where lang is the name of the source code
701 language to be evaluated, =[]= is optional and contains any header
702 arguments and ={}= contains the code.
704 (see [[* (sandbox) inline source blocks][the-sandbox]])
706 *** DONE evaluation with \C-c\C-c
707 Putting aside the header argument issue for now we can just run these
708 with the following default header arguments
709 - =:results= :: silent
710 - =:exports= :: results
712 *** DONE inline exportation
713 Need to add an interblock hook (or some such) through org-exp-blocks
714 *** DONE header arguments
715 We should make it possible to use header arguments.
717 *** TODO fontification
718 we should color these blocks differently
720 *** TODO refine html exportation
721 should use a span class, and should show original source in tool-tip
723 ** TODO LoB: re-implement plotting and analysis functions from org-R
724 I'll do this soon, now that we things are a bit more settled and we
725 have column names in R.
726 ** PROPOSED conversion between org-babel and noweb (e.g. .Rnw) format
727 I haven't thought about this properly. Just noting it down. What
728 Sweave uses is called "R noweb" (.Rnw).
730 I found a good description of noweb in the following article (see
731 the [[http://www.cs.tufts.edu/~nr/pubs/lpsimp.pdf][pdf]]).
733 I think there are two parts to noweb, the construction of
734 documentation and the extraction of source-code (with notangle).
736 *documentation*: org-mode handles all of our documentation needs in
737 a manner that I believe is superior to noweb.
739 *source extraction* At this point I don't see anyone writing large
740 applications with 100% of the source code contained in org-babel
741 files, rather I see org-babel files containing things like
742 - notes with active code chunks
743 - interactive tutorials
744 - requirements documents with code running test suites
745 - and of course experimental reports with the code to run the
746 experiment, and perform analysis
748 Basically I think the scope of the programs written in org-babel
749 (at least initially) will be small enough that it wont require the
750 addition of a tangle type program to extract all of the source code
751 into a running application.
753 On the other hand, since we already have named blocks of source
754 code which reference other blocks on which they rely, this
755 shouldn't be too hard to implement either on our own, or possibly
756 relying on something like noweb/notangle.
758 ** PROPOSED support for passing paths to files between source blocks
759 Maybe this should be it's own result type (in addition to scalars and
760 vectors). The reason being that some source-code blocks (for example
761 ditaa or anything that results in the creation of a file) may want to
762 pass a file path back to org-mode which could then be inserted into
763 the org-mode buffer as a link to the file...
765 This would allow for display of images upon export providing
766 functionality similar to =org-exp-blocks= only in a more general
768 ** DEFERRED Support rownames and other org babel table features?
770 The full org table features are detailed in the manual [[http://orgmode.org/manual/Advanced-features.html#Advanced-features][here]].
773 Perhaps add a :rownames header arg. This would be an integer
774 (usually 1) which would have the effect of post-processing all the
775 variables created in the R session in the following way: if the
776 integer is j, set the row names to the contents of column j and
777 delete column j. Perhaps it is artificial to allow this integer to
778 take any value other than 1. The default would be nil which would
779 mean no such behaviour.
781 Actually I don't know about that. If multiple variables are passed
782 in, it's not appropriate to alter them all in the same way. The
783 rownames specification would normally refer to just one of the
784 variables. For now maybe just say this has to be done in R. E.g.
786 #+TBLNAME: sample-sizes
787 | collection | size | exclude | include | exclude2 | include2 |
788 |-----------------+------+---------+---------+----------+----------|
789 | 58C | 2936 | 8 | 2928 | 256 | 2680 |
790 | MS | 5852 | 771 | 5081 | 771 | 5081 |
791 | NBS | 2929 | 64 | 2865 | 402 | 2527 |
792 | POBI | 2717 | 1 | 2716 | 1 | 2716 |
793 | 58C+MS+NBS+POBI | | | 13590 | | 13004 |
794 #+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)
796 #+srcname: make-size-table(size=sample-sizes)
798 rownames(size) <- size[,1]
804 [I don't think it's as problematic as this makes out]
805 This is non-trivial, but may be worth doing, in particular to
806 develop a nice framework for sending data to/from R.
808 In R, indexing vector elements, and rows and columns, using
809 strings rather than integers is an important part of the
811 - elements of a vector may have names
812 - matrices and data.frames may have "column names" and "row names"
813 which can be used for indexing
814 - In a data frame, row names *must* be unique
822 > mat <- matrix(1:4, nrow=2, ncol=2, dimnames=list(c("r1","r2"), c("c1","c2")))
827 > # The names are separate from the data: they do not interfere with operations on the data
834 > df <- data.frame(var1=1:26, var2=26:1, row.names=letters)
836 [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
842 So it's tempting to try to provide support for this in org-babel. For example
843 - allow R to refer to columns of a :var reference by their names
844 - When appropriate, results from R appear in the org buffer with "named
847 However none (?) of the other languages we are currently supporting
848 really have a native matrix type, let alone "column names" or "row
849 names". Names are used in e.g. python and perl to refer to entries
852 It currently seems to me that support for this in org-babel would
853 require setting rules about when org tables are considered to have
854 named columns/fields, and ensuring that (a) languages with a notion
855 of named columns/fields use them appropriately and (b) languages
856 with no such notion do not treat then as data.
858 - Org allows something that *looks* like column names to be separated
860 - Org also allows a row to *function* as column names when special
861 markers are placed in the first column. An hline is unnecessary
862 (indeed hlines are purely cosmetic in org [correct?]
863 - Org does not have a notion of "row names" [correct?]
865 The full org table functionality exeplified [[http://orgmode.org/manual/Advanced-features.html#Advanced-features][here]] has features that
866 we would not support in e.g. R (like names for the row below).
868 **** Initial statement: allow tables with hline to be passed as args into R
869 This doesn't seem to work at the moment (example below). It would
870 also be nice to have a natural way for the column names of the org
871 table to become the column names of the R data frame, and to have
872 the option to specify that the first column is to be used as row
873 names in R (these must be unique). But this might require a bit of
878 | col1 | col2 | col3 |
879 |------+---------+------|
887 #+begin_src R :var tabel=egtable :colnames t
892 | "col1" | "col2" | "col3" |
893 |--------+-----------+--------|
895 | 4 | "schulte" | 6 |
898 Another example is in the [[*operations%20in%20on%20tables][grades example]].
900 ** DEFERRED use textConnection to pass tsv to R?
901 When passing args from the org buffer to R, the following route is
902 used: arg in buffer -> elisp -> tsv on file -> data frame in R. I
903 think it would be possible to avoid having to write to file by
904 constructing an R expression in org-babel-R-assign-elisp, something
907 #+begin_src emacs-lisp
908 (org-babel-R-input-command
909 (format "%s <- read.table(textConnection(\"%s\"), sep=\"\\t\", as.is=TRUE)"
910 name (orgtbl-to-tsv value '(:sep "\t" :fmt org-babel-R-quote-tsv-field))))
913 I haven't tried to implement this yet as it's basically just
914 fiddling with something that works. The only reason for it I can
915 think of would be efficiency and I haven't tested that.
917 This Didn't work after an initial test. I still think this is a
918 good idea (I also think we should try to do something similar when
919 writing out results frmo R to elisp) however as it wouldn't result
920 in any functional changes I'm bumping it down to deferred for
925 #+tblname: quick-test
928 #+srcname: quick-test-src-blk
929 #+begin_src R :var vec=quick-test
939 ** DEFERRED Rework Interaction with Running Processes [2/5]
940 *** DONE robust to errors interrupting execution
942 #+srcname: long-runner-ruby
943 #+begin_src ruby :results silent
948 *** DEFERRED use =C-g= keyboard-quit to push processing into the background
949 This may be possible using the `run-with-timer' command.
951 I have no idea how this could work...
953 #+srcname: long-runner-ruby
954 #+begin_src ruby :results silent
959 *** TODO ability to select which of multiple sessions is being used
960 Increasingly it is looking like we're going to want to run all
961 source code blocks in comint buffer (sessions). Which will have
963 1) allowing background execution
964 2) maintaining state between source-blocks
965 - allowing inline blocks w/o header arguments
968 (like ess-switch-process in .R buffers)
970 Maybe this could be packaged into a header argument, something
971 like =:R_session= which could accept either the name of the
972 session to use, or the string =prompt=, in which case we could use
973 the =ess-switch-process= command to select a new process.
975 *** TODO evaluation of shell code as background process?
976 After C-c C-c on an R code block, the process may appear to
977 block, but C-g can be used to reclaim control of the .org buffer,
978 without interrupting the R evalution. However I believe this is not
979 true of bash/sh evaluation. [Haven't tried other languages] Perhaps
980 a solution is just to background the individual shell commands.
982 The other languages (aside from emacs lisp) are run through the
983 shell, so if we find a shell solution it should work for them as
986 Adding an ampersand seems to be a supported way to run commands in
987 the background (see [[http://www.emacswiki.org/emacs/ExecuteExternalCommand#toc4][external-commands]]). Although a more extensible
988 solution may involve the use of the [[elisp:(progn (describe-function 'call-process-region) nil)][call-process-region]] function.
990 Going to try this out in a new file [[file:lisp/org-babel-proc.el][org-babel-proc.el]]. This should
991 contain functions for asynchronously running generic shell commands
992 in the background, and then returning their input.
994 **** partial update of org-mode buffer
995 The sleekest solution to this may be using a comint buffer, and
996 then defining a filter function which would incrementally interpret
997 the results as they are returned, including insertion into the
998 org-mode buffer. This may actually cause more problems than it is
999 worth, what with the complexities of identifying the types of
1000 incrementally returned results, and the need for maintenance of a
1001 process marker in the org buffer.
1003 **** 'working' spinner
1004 It may be nice and not too difficult to place a spinner on/near the
1005 evaluating source code block
1007 *** TODO conversion of output from interactive shell, R (and python) sessions to org-babel buffers
1008 [DED] This would be a nice feature I think. Although an org-babel
1009 purist would say that it's working the wrong way round... After
1010 some interactive work in a *R* buffer, you save the buffer, maybe
1011 edit out some lines, and then convert it to org-babel format for
1012 posterity. Same for a shell session either in a *shell* buffer, or
1013 pasted from another terminal emulator. And python of course.
1015 ** DEFERRED improve the source-block snippet
1016 any real improvement seems somewhat beyond the ability of yasnippet
1019 [[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]]
1021 ,#name : Chapter title
1024 ${1:$(make-string (string-width text) ?\=)}
1029 [[file:snippets/org-mode/sb][sb -- snippet]]
1031 waiting for guidance from those more familiar with yasnippets
1033 ** REJECTED re-implement R evaluation using ess-command or ess-execute
1034 I don't have any complaints with the current R evaluation code or
1035 behaviour, but I think it would be good to use the ESS functions
1036 from a political point of view. Plus of course it has the normal
1037 benefits of an API (insulates us from any underlying changes etc). [DED]
1039 I'll look into this. I believe that I looked at and rejected these
1040 functions initially but now I can't remember why. I agree with
1041 your overall point about using API's where available. I will take
1042 a look back at these and either switch to using the ess commands,
1043 or at least articulate under this TODO the reasons for using our
1044 custom R-interaction commands. [Eric]
1048 Lets just replace =org-babel-R-input-command= with =ess-execute=.
1050 I tried this, and although it works in some situations, I find that
1051 =ess-command= will often just hang indefinitely without returning
1052 results. Also =ess-execute= will occasionally hang, and pops up
1053 the buffer containing the results of the command's execution, which
1054 is undesirable. For now these functions can not be used. Maybe
1055 someone more familiar with the ESS code can recommend proper usage
1056 of =ess-command= or some other lower-level function which could be
1057 used in place of [[file:lisp/org-babel-R.el::defun%20org-babel%20R%20input%20command%20command][org-babel-R-input-command]].
1061 #+begin_quote ess-command
1062 (ess-command COM &optional BUF SLEEP NO-PROMPT-CHECK)
1064 Send the ESS process command COM and delete the output
1065 from the ESS process buffer. If an optional second argument BUF exists
1066 save the output in that buffer. BUF is erased before use.
1067 COM should have a terminating newline.
1068 Guarantees that the value of .Last.value will be preserved.
1069 When optional third arg SLEEP is non-nil, `(sleep-for (* a SLEEP))'
1070 will be used in a few places where `a' is proportional to `ess-cmd-delay'.
1073 #+begin_quote ess-execute
1074 (ess-execute COMMAND &optional INVERT BUFF MESSAGE)
1076 Send a command to the ESS process.
1077 A newline is automatically added to COMMAND. Prefix arg (or second arg
1078 INVERT) means invert the meaning of
1079 `ess-execute-in-process-buffer'. If INVERT is 'buffer, output is
1080 forced to go to the process buffer. If the output is going to a
1081 buffer, name it *BUFF*. This buffer is erased before use. Optional
1082 fourth arg MESSAGE is text to print at the top of the buffer (defaults
1083 to the command if BUFF is not given.)
1086 *** out current setup
1088 1) The body of the R source code block is wrapped in a function
1089 2) The function is called inside of a =write.table= function call
1090 writing the results to a table
1091 3) The table is read using =org-table-import=
1093 ** DONE extensible library of callable source blocks
1095 This is covered by the [[file:library-of-babel.org][Library of Babel]], which will contain
1096 ready-made source blocks designed to carry out useful common tasks.
1097 *** Initial statement [Eric]
1098 Much of the power of org-R seems to be in it's helper functions for
1099 the quick graphing of tables. Should we try to re-implement these
1100 functions on top of org-babel?
1102 I'm thinking this may be useful both to add features to org-babel-R and
1103 also to potentially suggest extensions of the framework. For example
1104 one that comes to mind is the ability to treat a source-code block
1105 like a function which accepts arguments and returns results. Actually
1106 this can be it's own TODO (see [[* source blocks as functions][source blocks as functions]]).
1107 *** Objectives [Dan]
1108 - We want to provide convenient off-the-shelf actions
1109 (e.g. plotting data) that make use of our new code evaluation
1110 environment but do not require any actual coding.
1111 *** Initial Design proposal [Dan]
1112 - *Input data* will be specified using the same mechanism as :var
1113 references, thus the input data may come from a table, or
1114 another source block, and it is initially available as an elisp
1116 - We introduce a new #+ line, e.g. #+BABELDO. C-c C-c on that
1117 line will apply an *action* to the referenced data.
1118 - *Actions correspond to source blocks*: our library of available
1119 actions will be a library of org-babel source blocks. Thus the
1120 code for executing an action, and the code for dealing with the
1121 output of the action will be the same code as for executing
1122 source blocks in general
1123 - Optionally, the user can have the relevant source block inserted
1124 into the org buffer after the (say) #+BABELDO line. This will
1125 allow the user to fine tune the action by modifying the code
1126 (especially useful for plots).
1127 - So maybe a #+BABELDO line will have header args
1128 - :data (a reference to a table or source code block)
1129 - :action (or should that be :srcname?) which will be something
1130 like :action pie-chart, referring to a source block which will
1131 be executed with the :data referent passed in using a :var arg.
1132 - :showcode or something controlling whether to show the code
1134 *** Modification to design
1135 I'm implementing this, at least initially, as a new interpreter
1136 named 'babel', which has an empty body. 'babel' blocks take
1137 a :srcname header arg, and look for the source-code block with
1138 that name. They then execute the referenced block, after first
1139 appending their own header args on to the target block's header
1142 If the target block is in the library of babel (a.o.t. e.g. the
1143 current buffer), then the code in the block will refer to the
1144 input data with a name dictated by convention (e.g. __data__
1145 (something which is syntactically legal in all languages...). Thus
1146 the babel block will use a :var __data__ = whatever header arg to
1147 reference the data to be plotted.
1149 ** DONE Column names in R input/output
1150 This has been implemented: Automatic on input to R; optional in
1151 output. Note that this equates column names with the header row in
1152 an org table; whereas org actually has a mechanism whereby a row
1153 with a '!' in the first field defines column names. I have not
1154 attempted to support these org table mechanisms yet. See [[*Support%20rownames%20and%20other%20org%20babel%20table%20features][this
1155 DEFERRED todo item]].
1156 ** DONE use example block for large amounts of stdout output?
1157 We're currently `examplizing' with : at the beginning of the line,
1158 but should larger amounts of output be in a
1159 \#+begin_example...\#+end_example block? What's the cutoff? > 1
1160 line? This would be nice as it would allow folding of lengthy
1161 output. Sometimes one will want to see stdout just to check
1162 everything looks OK, and then fold it away.
1164 I'm addressing this in branch 'examplizing-output'.
1165 Yea, that makes sense. (either that or allow folding of large
1166 blocks escaped with =:=).
1168 Proposed cutoff of 10 lines, we can save this value in a user
1169 customizable variable.
1170 *** DONE add ability to remove such results
1171 ** DONE exclusive =exports= params
1173 #+srcname: implement-export-exclusivity
1180 ** DONE LoB: allow output in buffer
1181 ** DONE allow default header arguments by language
1182 org-babel-default-header-args:lang-name
1184 An example of when this is useful is for languages which always return
1185 files as their results (e.g. [[*** ditaa][ditaa]], and [[*** gnuplot][gnuplot]]).
1186 ** DONE singe-function tangling and loading elisp from literate org-mode file [3/3]
1188 This function should tangle the org-mode file for elisp, and then call
1189 `load-file' on the resulting tangled file.
1191 #+srcname: test-loading-embedded-emacs-lisp
1192 #+begin_src emacs-lisp :results replace
1193 (setq test-tangle-advert nil)
1194 (setq test-tangle-loading nil)
1195 (setq results (list :before test-tangle-loading test-tangle-advert))
1196 (org-babel-load-file "test-tangle.org")
1197 (setq results (list (list :after test-tangle-loading test-tangle-advert) results))
1198 (delete-file "test-tangle.el")
1202 #+resname: test-loading-embedded-emacs-lisp
1203 | :before | nil | nil |
1204 | :after | "org-babel tangles" | "use org-babel-tangle for all your emacs initialization files!!" |
1206 *** DONE add optional language limiter to org-babel-tangle
1207 This should check to see if there is any need to re-export
1209 *** DONE ensure that org-babel-tangle returns the path to the tangled file(s)
1211 #+srcname: test-return-value-of-org-babel-tangle
1212 #+begin_src emacs-lisp :results replace
1213 (mapcar #'file-name-nondirectory (org-babel-tangle-file "test-tangle.org" "emacs-lisp"))
1217 | "test-tangle.el" |
1219 *** DONE only tangle the file if it's actually necessary
1220 ** DONE add a function to jump to a source-block by name
1221 I've had an initial stab at that in org-babel-find-named-block
1222 (library-of-babel branch).
1224 At the same time I introduced org-babel-named-src-block-regexp, to
1225 match src-blocks with srcname.
1227 This is now working with the command
1228 `org-babel-goto-named-source-block', all we need is a good key
1231 ** DONE add =:none= session argument (for purely functional execution) [4/4]
1232 This would allow source blocks to be run in their own new process
1234 - These blocks could then also be run in the background (since we can
1235 detach and just wait for the process to signal that it has terminated)
1236 - We wouldn't be drowning in session buffers after running the tests
1237 - we can re-use much of the session code to run in a more /functional/
1240 While session provide a lot of cool features, like persistent
1241 environments, [[* DONE function to bring up inferior-process buffer][pop-to-session]], and hints at exportation for
1242 org-babel-tangle, they also have some down sides and I'm thinking that
1243 session-based execution maybe shouldn't be the default behavior.
1245 Down-sides to sessions
1246 - *much* more complicated than functional evaluation
1247 - maintaining the state of the session has weird issues
1248 - waiting for evaluation to finish
1249 - prompt issues like [[* TODO weird escaped characters in shell prompt break shell evaluation][shell-prompt-escapes-bug]]
1250 - can't run in background
1251 - litter emacs with session buffers
1255 #+srcname: ruby-task-no-session
1256 #+begin_src ruby :results replace output
1262 #+resname: ruby-task-no-session
1267 #+srcname: task-python-none-session
1268 #+begin_src python :session none :results replace value
1274 #+resname: task-python-none-session
1279 #+srcname: task-session-none-sh
1280 #+begin_src sh :results replace
1285 #+resname: task-session-none-sh
1291 #+srcname: task-no-session-R
1292 #+begin_src R :results replace output
1299 #+resname: task-no-session-R
1303 ** DONE fully purge org-babel-R of direct comint interaction
1304 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
1306 ** DONE Create objects in top level (global) environment [5/5]
1309 *** initial requirement statement [DED]
1310 At the moment, objects created by computations performed in the
1311 code block are evaluated in the scope of the
1312 code-block-function-body and therefore disappear when the code
1313 block is evaluated {unless you employ some extra trickery like
1314 assign('name', object, env=globalenv()) }. I think it will be
1315 desirable to also allow for a style wherein objects that are
1316 created in one code block persist in the R global environment and
1317 can be re-used in a separate block.
1319 This is what Sweave does, and while I'm not saying we have to be
1320 the same as Sweave, it wouldn't be hard for us to provide the same
1321 behaviour in this case; if we don't, we risk undeservedly being
1322 written off as an oddity by some.
1324 IOW one aspect of org-babel is that of a sort of functional
1325 meta-programming language. This is crazy, in a very good
1326 way. Nevertheless, wrt R I think there's going to be a lot of value
1327 in providing for a working style in which the objects are stored in
1328 the R session, rather than elisp/org buffer. This will be a very
1329 familiar working style to lots of people.
1331 There are no doubt a number of different ways of accomplishing
1332 this, the simplest being a hack like adding
1335 for(objname in ls())
1336 assign(objname, get(objname), envir=globalenv())
1339 to the source code block function body. (Maybe wrap it in an on.exit() call).
1341 However this may deserve to be thought about more carefully, perhaps
1342 with a view to having a uniform approach across languages. E.g. shell
1343 code blocks have the same semantics at the moment (no persistence of
1344 variables across code blocks), because the body is evaluated in a new
1345 bash shell process rather than a running shell. And I guess the same
1346 is true for python. However, in both these cases, you could imagine
1347 implementing the alternative in which the body is evaluated in a
1348 persistent interactive session. It's just that it's particularly
1349 natural for R, seeing as both ESS and org-babel evaluate commands in a
1350 single persistent R session.
1354 Thanks for bringing this up. I think you are absolutely correct that we
1355 should provide support for a persistent environment (maybe called a
1356 *session*) in which to evaluate code blocks. I think the current setup
1357 demonstrates my personal bias for a functional style of programming
1358 which is certainly not ideal in all contexts.
1360 While the R function you mention does look like an elegant solution, I
1361 think we should choose an implementation that would be the same across
1362 all source code types. Specifically I think we should allow the user to
1363 specify an optional *session* as a header variable (when not present we
1364 assume a default session for each language). The session name could be
1365 used to name a comint buffer (like the *R* buffer) in which all
1366 evaluation would take place (within which variables would retain their
1367 values --at least once I remove some of the functional method wrappings
1368 currently in place-- ).
1370 This would allow multiple environments to be used in the same buffer,
1371 and once this setup was implemented we should be able to fairly easily
1372 implement commands for jumping between source code blocks and the
1373 related session buffers, as well as for dumping the last N commands from
1374 a session into a new or existing source code block.
1376 Please let me know if you foresee any problems with this proposed setup,
1377 or if you think any parts might be confusing for people coming from
1378 Sweave. I'll hopefully find some time to work on this later in the
1381 *** can functional and interpreted/interactive models coexist?
1383 Even though both of these use the same =*R*= buffer the value of =a=
1384 is not preserved because it is assigned inside of a functional
1387 #+srcname: task-R-sessions
1394 #+srcname: task-R-same-session
1399 This functional wrapper was implemented in order to efficiently return
1400 the results of the execution of the entire source code block. However
1401 it inhibits the evaluation of source code blocks in the top level,
1402 which would allow for persistence of variable assignment across
1403 evaluations. How can we allow *both* evaluation in the top level, and
1404 efficient capture of the return value of an entire source code block
1405 in a language independent manner?
1407 Possible solutions...
1408 1) we can't so we will have to implement two types of evaluation
1409 depending on which is appropriate (functional or imperative)
1410 2) we remove the functional wrapper and parse the source code block
1411 into it's top level statements (most often but not always on line
1412 breaks) so that we can isolate the final segment which is our
1414 3) we add some sort of "#+return" line to the code block
1415 4) we take advantage of each languages support for meta-programming
1416 through =eval= type functions, and use said to evaluate the entire
1417 blocks in such a way that their environment can be combined with the
1418 global environment, and their results are still captured.
1419 5) I believe that most modern languages which support interactive
1420 sessions have support for a =last_result= type function, which
1421 returns the result of the last input without re-calculation. If
1422 widely enough present this would be the ideal solution to a
1423 combination of functional and imperative styles.
1425 None of these solutions seem very desirable, but for now I don't see
1426 what else would be possible.
1428 Of these options I was leaning towards (1) and (4) but now believe
1429 that if it is possible option (5) will be ideal.
1431 **** (1) both functional and imperative evaluation
1433 - can take advantage of built in functions for sending regions to the
1435 - retains the proven tested and working functional wrappers
1438 - introduces the complication of keeping track of which type of
1439 evaluation is best suited to a particular context
1440 - the current functional wrappers may require some changes in order to
1441 include the existing global context
1443 **** (4) exploit language meta-programming constructs to explicitly evaluate code
1445 - only one type of evaluation
1448 - some languages may not have sufficient meta-programming constructs
1450 **** (5) exploit some =last_value= functionality if present
1452 Need to ensure that most languages have such a function, those without
1453 will simply have to implement their own similar solution...
1455 | language | =last_value= function |
1456 |------------+-----------------------------|
1460 | shell | see [[* last command for shells][last command for shells]] |
1461 | emacs-lisp | see [[* emacs-lisp will be a special case][special-case]] |
1463 #+srcname: task-last-value
1468 ***** last command for shells
1469 Do this using the =tee= shell command, and continually pipe the output
1472 Got this idea from the following [[http://linux.derkeiler.com/Mailing-Lists/Fedora/2004-01/0898.html][email-thread]].
1474 suggested from mailing list
1476 #+srcname: bash-save-last-output-to-file
1480 bash -c "$line" | tee /tmp/last.out1
1481 mv /tmp/last.out1 /tmp/last.out
1485 another proposed solution from the above thread
1487 #+srcname: bash-save-in-variable
1490 # so - Save Output. Saves output of command in OUT shell variable.
1500 "^M": " | tee /tmp/h_lastcmd.out ^[k"
1503 export __=/tmp/h_lastcmd.out
1505 If you try it, Alt-k will stand for the old Enter; use "command $__" to
1506 access the last output.
1512 Herculano de Lima Einloft Neto
1515 ***** emacs-lisp will be a special case
1516 While it is possible for emacs-lisp to be run in a console type
1517 environment (see the =elim= function) it is *not* possible to run
1518 emacs-lisp in a different *session*. Meaning any variable set top
1519 level of the console environment will be set *everywhere* inside
1520 emacs. For this reason I think that it doesn't make any sense to
1521 worry about session support for emacs-lisp.
1523 *** Further thoughts on 'scripting' vs. functional approaches
1525 These are just thoughts, I don't know how sure I am about this.
1526 And again, perhaps I'm not saying anything very radical, just that
1527 it would be nice to have some options supporting things like
1528 receiving text output in the org buffer.
1530 I can see that you've already gone some way down the road towards
1531 the 'last value' approach, so sorry if my comments come rather
1532 late. I am concerned that we are not giving sufficient attention
1533 to stdout / the text that is returned by the interpreters. In
1534 contrast, many of our potential users will be accustomed to a
1535 'scripting' approach, where they are outputting text at various
1536 points in the code block, not just at the end. I am leaning
1537 towards thinking that we should have 2 modes of evaluation:
1538 'script' mode, and 'functional' mode.
1540 In script mode, evaluation of a code block would result in *all*
1541 text output from that code block appearing as output in the org
1542 buffer, presumably as an #+begin_example...#+end_example. There
1543 could be an :echo option controlling whether the input commands
1544 also appear in the output. [This is like Sweave].
1546 In functional mode, the *result* of the code block is available as
1547 an elisp object, and may appear in the org buffer as an org
1548 table/string, via the mechanisms you have developed already.
1550 One thing I'm wondering about is whether, in script mode, there
1551 simply should not be a return value. Perhaps this is not so
1552 different from what exists: script mode would be new, and what
1553 exists currently would be functional mode.
1555 I think it's likely that, while code evaluation will be exciting
1556 to people, a large majority of our users in a large majority of
1557 their usage will not attempt to actually use the return value from
1558 a source code block in any meaningful way. In that case, it seems
1559 rather restrictive to only allow them to see output from the end
1562 Instead I think the most accessible way to introduce org-babel to
1563 people, at least while they are learning it, is as an immensely
1564 powerful environment in which to embed their 'scripts', which now
1565 also allows them to 'run' their 'scripts'. Especially as such
1566 people are likely to be the least capable of the user-base, a
1567 possible design-rule would be to make the scripting style of usage
1568 easy (default?), perhaps requiring a special option to enable a
1569 functional style. Those who will use the functional style won't
1570 have a problem understanding what's going on, whereas the 'skript
1571 kiddies' might not even know the syntax for defining a function in
1572 their language of choice. And of course we can allow the user to
1573 set a variable in their .emacs controlling the preference, so that
1574 functional users are not inconveniennced by having to provide
1575 header args the whole time.
1577 Please don't get the impression that I am down-valuing the
1578 functional style of org-babel. I am constantly horrified at the
1579 messy 'scripts' that my colleagues produce in perl or R or
1580 whatever! Nevertheless that seems to be how a lot of people work.
1582 I think you were leaning towards the last-value approach because
1583 it offered the possibility of unified code supporting both the
1584 single evaluation environment and the functional style. If you
1585 agree with any of the above then perhaps it will impact upon this
1586 and mean that the code in the two branches has to differ a bit. In
1587 that case, functional mode could perhaps after all evaluate each
1588 code block in its own environment, thus (re)approaching 'true'
1589 functional programming (side-effects are hard to achieve).
1593 echo "There are `wc -l files` files in this directory"
1597 *** even more thoughts on evaluation, results, models and options
1599 Thanks Dan, These comments are invaluable.
1601 What do you think about this as a new list of priorities/requirements
1602 for the execution of source-code blocks.
1605 1) we want the evaluation of the source code block to take place in a
1606 session which can persist state (variables, current directory,
1608 2) source code blocks can specify their session with a header argument
1609 3) each session should correspond to an Emacs comint buffer so that the
1610 user can drop into the session and experiment with live code
1613 1) each source-code block generates some form of results which (as
1614 we have already implemented) is transfered into emacs-lisp
1615 after which it can be inserted into the org-mode buffer, or
1616 used by other source-code blocks
1617 2) when the results are translated into emacs-lisp, forced to be
1618 interpreted as a scalar (dumping their raw values into the
1619 org-mode buffer), as a vector (which is often desirable with R
1620 code blocks), or interpreted on the fly (the default option).
1621 Note that this is very nearly currently implemented through the
1622 [[* DONE results-type header (vector/file)][results-type-header]].
1623 3) there should be *two* means of collecting results from the
1624 execution of a source code block. *Either* the value of the
1625 last statement of the source code block, or the collection of
1626 all that has been passed to STDOUT during the evaluation.
1628 **** header argument or return line (*header argument*)
1630 Rather than using a header argument to specify how the return value
1631 should be passed back, I'm leaning towards the use of a =#+RETURN=
1632 line inside the block. If such a line *is not present* then we
1633 default to using STDOUT to collect results, but if such a line *is
1634 present* then we use it's value as the results of the block. I
1635 think this will allow for the most elegant specification between
1636 functional and script execution. This also cleans up some issues
1637 of implementation and finding which statement is the last
1640 Having given this more thought, I think a header argument is
1641 preferable. The =#+return:= line adds new complicating syntax for
1642 something that does little more than we would accomplish through
1643 the addition of a header argument. The only benefit being that we
1644 know where the final statement starts, which is not an issue in
1645 those languages which contain 'last value' operators.
1647 new header =:results= arguments
1648 - script :: explicitly states that we want to use STDOUT to
1649 initialize our results
1650 - return_last :: stdout is ignored instead the *value* of the final
1651 statement in the block is returned
1652 - echo :: means echo the contents of the source-code block along
1653 with the results (this implies the *script* =:results=
1656 *** DONE rework evaluation lang-by-lang [4/4]
1658 This should include...
1659 - functional results working with the comint buffer
1661 - script :: return the output of STDOUT
1662 - write a macro which runs the first redirection, executes the
1663 body, then runs the second redirection
1664 - last :: return the value of the last statement
1667 - sessions in comint buffers
1669 **** DONE Ruby [4/4]
1670 - [X] functional results working with comint
1671 - [X] script results
1672 - [X] ensure scalar/vector results args are taken into consideration
1673 - [X] ensure callable by other source block
1675 #+srcname: ruby-use-last-output
1676 #+begin_src ruby :results replace
1683 #+resname: ruby-use-last-output
1686 #+srcname: task-call-use-last-output
1687 #+begin_src ruby :var last=ruby-use-last-output :results replace
1688 last.flatten.size + 1
1691 #+resname: task-call-use-last-output
1696 #+srcname: first-ruby-session-task
1697 #+begin_src ruby :session schulte :results silent
1701 #+srcname: second-ruby-session-task
1702 #+begin_src ruby :session schulte :results silent
1706 #+srcname: without-the-right-session
1707 #+begin_src ruby :results silent
1713 - [X] functional results working with comint
1714 - [X] script results
1715 - [X] ensure scalar/vector results args are taken into consideration
1716 - [X] ensure callable by other source block
1718 To redirect output to a file, you can use the =sink()= command.
1721 #+begin_src R :results value vector silent
1728 #+srcname: task-R-use-other-output
1729 #+begin_src R :var twoentyseven=task_R_B() :results replace value
1734 #+resname: task-R-use-other-output
1737 **** DONE Python [4/4]
1738 - [X] functional results working with comint
1739 - [X] script results
1740 - [X] ensure scalar/vector results args are taken into consideration
1741 - [X] ensure callable by other source block
1743 #+srcname: task-new-eval-for-python
1744 #+begin_src python :results silent output scalar
1750 #+srcname: task-use-new-eval
1751 #+begin_src python :var tasking=task-new-eval-for-python() :results replace
1755 #+resname: task-use-new-eval
1758 **** DONE Shells [4/4]
1759 - [X] functional results working with comint
1760 - [X] script results
1761 - [X] ensure scalar/vector results args are taken into consideration
1762 - [X] ensure callable by other source block
1764 #+srcname: task-shell-new-evaluation
1765 #+begin_src sh :results silent value scalar
1770 #+srcname: task-call-other-shell
1771 #+begin_src sh :var other=task-shell-new-evaluation() :results replace scalar
1772 echo $other ' is the old date'
1775 #+resname: task-call-other-shell
1776 : $ Fri Jun 12 13:08:37 PDT 2009 is the old date
1778 *** DONE implement a *session* header argument [4/4]
1779 =:session= header argument to override the default *session* buffer
1783 #+srcname: task-ruby-named-session
1784 #+begin_src ruby :session schulte :results replace
1785 schulte = :in_schulte
1788 #+resname: task-ruby-named-session
1791 #+srcname: another-in-schulte
1792 #+begin_src ruby :session schulte
1796 #+resname: another-in-schulte
1803 #+srcname: python-session-task
1804 #+begin_src python :session what :results silent
1808 #+srcname: python-get-from-session
1809 #+begin_src python :session what :results replace
1813 #+resname: python-get-from-session
1818 #+srcname: task-shell-sessions
1819 #+begin_src sh :session what
1823 #+srcname: task-shell-sessions-what
1824 #+begin_src sh :session what :results replace
1828 #+resname: task-shell-sessions-what
1833 #+srcname: task-R-session
1834 #+begin_src R :session what :results replace
1840 #+resname: task-R-session
1843 #+srcname: another-task-R-session
1844 #+begin_src R :session what :results replace
1848 *** DONE function to bring up inferior-process buffer [4/4]
1850 This should be callable from inside of a source-code block in an
1851 org-mode buffer. It should evaluate the header arguments, then bring
1852 up the inf-proc buffer using =pop-to-buffer=.
1854 For lack of a better place, lets add this to the `org-metadown-hook'
1857 To give this a try, place the cursor on a source block with variables,
1858 (optionally git a prefix argument) then hold meta and press down.
1862 #+srcname: task-ruby-pop-to-session
1863 #+begin_src ruby :var num=9 :var another="something else"
1864 num.times{|n| puts another}
1869 #+srcname: task-python-pop-to-session
1870 #+begin_src python :var num=9 :var another="something else"
1875 #+srcname: task-R-pop-to-session
1876 #+begin_src R :var a=9 :var b=8
1882 #+srcname: task-shell-pop-sessions
1883 #+begin_src sh :var NAME="eric"
1887 *** DEFERRED function to dump last N lines from inf-proc buffer into the current source block
1889 Callable with a prefix argument to specify how many lines should be
1890 dumped into the source-code buffer.
1892 *** REJECTED comint notes
1894 Implementing comint integration in [[file:lisp/org-babel-comint.el][org-babel-comint.el]].
1897 - handling of outputs
1898 - split raw output from process by prompts
1899 - a ring of the outputs, buffer-local, `org-babel-comint-output-ring'
1900 - a switch for dumping all outputs to a buffer
1901 - inputting commands
1903 Lets drop all this language specific stuff, and just use
1904 org-babel-comint to split up our outputs, and return either the last
1905 value of an execution or the combination of values from the
1908 **** comint filter functions
1909 : ;; comint-input-filter-functions hook process-in-a-buffer
1910 : ;; comint-output-filter-functions hook function modes.
1911 : ;; comint-preoutput-filter-functions hook
1912 : ;; comint-input-filter function ...
1914 #+srcname: obc-filter-ruby
1915 #+begin_src ruby :results last
1923 ** DONE Remove protective commas from # comments before evaluating
1924 org inserts protective commas in front of ## comments in language
1925 modes that use them. We need to remove them prior to sending code
1928 #+srcname: testing-removal-of-protective-comas
1930 ,# this one might break it??
1934 ** DONE pass multiple reference arguments into R
1935 Can we do this? I wasn't sure how to supply multiple 'var' header
1936 args. Just delete this if I'm being dense.
1938 This should be working, see the following example...
1940 #+srcname: two-arg-example
1941 #+begin_src R :var n=2 :var m=8
1945 #+resname: two-arg-example
1948 ** DONE ensure that table ranges work
1949 when a table range is passed to org-babel as an argument, it should be
1950 interpreted as a vector.
1953 | 2 | 3 | Fixnum:1 |
1954 | 3 | 4 | Array:123456 |
1958 #+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)))
1960 #+srcname: simple-sbe-example
1961 #+begin_src emacs-lisp
1965 #+srcname: task-table-range
1966 #+begin_src ruby :var n=simple-sbe-example
1970 #+srcname: simple-results
1971 #+begin_src emacs-lisp :var n=task-table-range(n=(1 2 3))
1975 #+resname: simple-results
1978 #+srcname: task-arr-referent
1979 #+begin_src ruby :var ar=(1 2 3)
1983 #+resname: task-arr-referent
1986 ** DONE global variable indicating default to vector output
1987 how about an alist... =org-babel-default-header-args= this may already
1988 exist... just execute the following and all source blocks will default
1991 #+begin_src emacs-lisp
1992 (setq org-babel-default-header-args '((:results . "vector")))
1995 ** DONE name named results if source block is named
1996 currently this isn't happening although it should be
1998 #+srcname: test-naming-named-source-blocks
1999 #+begin_src emacs-lisp
2003 #+resname: test-naming-named-source-blocks
2005 ** DONE (simple caching) check for named results before source blocks
2006 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]]
2007 ** DONE set =:results silent= when eval with prefix argument
2009 #+begin_src emacs-lisp
2012 ** DONE results-type header (vector/file) [3/3]
2013 In response to a point in Dan's email. We should allow the user to
2014 force scalar or vector results. This could be done with a header
2015 argument, and the default behavior could be controlled through a
2016 configuration variable.
2018 #+srcname: task-trivial-vector
2019 #+begin_src ruby :results replace vector
2026 since it doesn't make sense to turn a vector into a scalar, lets
2027 just add a two values...
2029 - vector :: forces the results to be a vector (potentially 1 dimensional)
2030 - file :: this throws an error if the result isn't a string, and
2031 tries to treat it as a path to a file.
2033 I'm just going to cram all of these into the =:results= header
2034 argument. Then if we allow multiple header arguments it should
2035 work out, for example one possible header argument string could be
2036 =:results replace vector file=, which would *replace* any existing
2037 results forcing the results into an org-mode table, and
2038 interpreting any strings as file paths.
2040 *** DONE multiple =:results= headers
2042 #+srcname: multiple-result-headers
2043 #+begin_src ruby :results replace silent
2049 *** DONE file result types
2050 When inserting into an org-mode buffer create a link with the path
2051 being the value, and optionally the display being the
2052 =file-name-nondirectory= if it exists.
2054 #+srcname: task-file-result
2055 #+begin_src python :results replace file
2060 [[something][something]]
2063 This will be useful because blocks like =ditaa= and =dot= can return
2064 the string path of their files, and can add =file= to their results
2067 *** DONE vector result types
2069 #+srcname: task-force-results
2070 #+begin_src emacs-lisp :results vector
2077 ** DONE results name
2078 In order to do this we will need to start naming our results.
2079 Since the source blocks are named with =#+srcname:= lines we can
2080 name results with =#+resname:= lines (if the source block has no
2081 name then no name is given to the =#+resname:= line on creation,
2082 otherwise the name of the source block is used).
2084 This will have the additional benefit of allowing results and
2085 source blocks to be located in different places in a buffer (and
2086 eventually in different buffers entirely).
2088 #+srcname: developing-resnames
2089 #+begin_src emacs-lisp :results silent
2093 Once source blocks are able to find their own =#+resname:= lines
2096 #+srcname: sbe-w-new-results
2097 #+begin_src emacs-lisp :results replace
2098 (sbe "developing-resnames")
2104 *** TODO change the results insertion functions to use these lines
2106 *** TODO teach references to resolve =#+resname= lines.
2108 ** DONE org-babel tests org-babel [1/1]
2109 since we are accumulating this nice collection of source-code blocks
2110 in the sandbox section we should make use of them as unit tests.
2111 What's more, we should be able to actually use org-babel to run these
2114 We would just need to cycle over every source code block under the
2115 sandbox, run it, and assert that the return value is equal to what we
2118 I have the feeling that this should be possible using only org-babel
2119 functions with minimal or no additional elisp. It would be very cool
2120 for org-babel to be able to test itself.
2122 This is now done, see [[* Tests]].
2124 *** DEFERRED org-babel assertions (may not be necessary)
2125 These could be used to make assertions about the results of a
2126 source-code block. If the assertion fails then the point could be
2127 moved to the block, and error messages and highlighting etc... could
2130 ** DONE make C-c C-c work anywhere within source code block?
2131 This seems like it would be nice to me, but perhaps it would be
2132 inefficient or ugly in implementation? I suppose you could search
2133 forward, and if you find #+end_src before you find #+begin_src,
2134 then you're inside one. [DED]
2136 Agreed, I think inside of the =#+srcname: line= would be useful as
2139 #+srcname: testing-out-cc
2140 #+begin_src emacs-lisp
2144 ** DONE integration with org tables
2145 We should make it easy to call org-babel source blocks from org-mode
2146 table formulas. This is practical now that it is possible to pass
2147 arguments to org-babel source blocks.
2149 See the related [[* (sandbox) integration w/org tables][sandbox]] header for tests/examples.
2151 *** digging in org-table.el
2152 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.
2154 Should be a hook in [[file:~/src/org/lisp/org-table.el::defun%20org%20table%20eval%20formula%20optional%20arg%20equation][org-table-eval-formula]].
2156 Looks like I need to change this [[file:~/src/org/lisp/org-table.el::if%20lispp][if statement]] (line 2239) into a cond
2159 ** DONE source blocks as functions
2161 Allow source code blocks to be called like functions, with arguments
2162 specified. We are already able to call a source-code block and assign
2163 it's return result to a variable. This would just add the ability to
2164 specify the values of the arguments to the source code block assuming
2165 any exist. For an example see
2167 When a variable appears in a header argument, how do we differentiate
2168 between it's value being a reference or a literal value? I guess this
2169 could work just like a programming language. If it's escaped or in
2170 quotes, then we count it as a literal, otherwise we try to look it up
2173 ** DONE folding of code blocks? [2/2]
2174 [DED] In similar way to using outline-minor-mode for folding function
2175 bodies, can we fold code blocks? #+begin whatever statements are
2176 pretty ugly, and in any case when you're thinking about the overall
2177 game plan you don't necessarily want to see the code for each Step.
2179 *** DONE folding of source code block
2180 Sounds good, and wasn't too hard to implement. Code blocks should
2181 now be fold-able in the same manner as headlines (by pressing TAB
2184 *** REJECTED folding of results
2185 So, lets do a three-stage tab cycle... First fold the src block,
2186 then fold the results, then unfold.
2188 There's no way to tell if the results are a table or not w/o
2189 actually executing the block which would be too expensive of an
2192 ** DONE selective export of text, code, figures
2193 [DED] The org-babel buffer contains everything (code, headings and
2194 notes/prose describing what you're up to, textual/numeric/graphical
2195 code output, etc). However on export to html / LaTeX one might want
2196 to include only a subset of that content. For example you might
2197 want to create a presentation of what you've done which omits the
2200 [EMS] So I think this should be implemented as a property which can
2201 be set globally or on the outline header level (I need to review
2202 the mechanics of org-mode properties). And then as a source block
2203 header argument which will apply only to a specific source code
2204 block. A header argument of =:export= with values of
2206 - =code= :: just show the code in the source code block
2207 - =none= :: don't show the code or the results of the evaluation
2208 - =results= :: just show the results of the code evaluation (don't
2209 show the actual code)
2210 - =both= :: show both the source code, and the results
2212 this will be done in [[* (sandbox) selective export][(sandbox) selective export]].
2214 ** DONE a header argument specifying silent evaluation (no output)
2215 This would be useful across all types of source block. Currently
2216 there is a =:replace t= option to control output, this could be
2217 generalized to an =:output= option which could take the following
2218 options (maybe more)
2220 - =t= :: this would be the default, and would simply insert the
2221 results after the source block
2222 - =replace= :: to replace any results which may already be there
2223 - =silent= :: this would inhibit any insertion of the results
2225 This is now implemented see the example in the [[* silent evaluation][sandbox]]
2227 ** DONE assign variables from tables in R
2228 This is now working (see [[* (sandbox table) R][(sandbox-table)-R]]). Although it's not that
2229 impressive until we are able to print table results from R.
2231 ** DONE insert 2-D R results as tables
2232 everything is working but R and shell
2238 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
2239 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]]
2240 as a means of converting multidimensional R objects to emacs lisp.
2242 It may be as simple as first checking if the data is multidimensional,
2243 and then, if so using =write= to write the data out to a temporary
2244 file from which emacs can read the data in using =org-table-import=.
2246 Looking into this further, is seems that there is no such thing as a
2247 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
2248 deal with trivial vectors (scalars) in R. I'm tempted to just treat
2249 them as vectors, but then that would lead to a proliferation of
2250 trivial 1-cell tables...
2252 ** DONE allow variable initialization from source blocks
2253 Currently it is possible to initialize a variable from an org-mode
2254 table with a block argument like =table=sandbox= (note that the
2255 variable doesn't have to named =table=) as in the following example
2261 #+begin_src emacs-lisp :var table=sandbox :results replace
2262 (message (format "table = %S" table))
2265 : "table = ((1 2 3) (4 \"schulte\" 6))"
2267 It would be good to allow initialization of variables from the results
2268 of other source blocks in the same manner. This would probably
2269 require the addition of =#+SRCNAME: example= lines for the naming of
2270 source blocks, also the =table=sandbox= syntax may have to be expanded
2271 to specify whether the target is a source code block or a table
2272 (alternately we could just match the first one with the given name
2273 whether it's a table or a source code block).
2275 At least initially I'll try to implement this so that there is no need
2276 to specify whether the reference is to a table or a source-code block.
2277 That seems to be simpler both in terms of use and implementation.
2279 This is now working for emacs-lisp, ruby and python (and mixtures of
2280 the three) source blocks. See the examples in the [[* (sandbox) referencing other source blocks][sandbox]].
2282 This is currently working only with emacs lisp as in the following
2283 example in the [[* emacs lisp source reference][emacs lisp source reference]].
2286 ** TODO Add languages [2/6]
2287 I'm sure there are many more that aren't listed here. Please add
2288 them, and bubble any that you particularly care about up to the top.
2290 Any new language should be implemented in a org-babel-lang.el file.
2291 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
2292 [[file:lisp/org-babel-R.el][org-babel-R.el]].
2295 This could probably be added to [[file:lisp/org-babel-script.el][org-babel-script.el]]
2300 (see [[* file result types][file result types]])
2302 #+srcname: implementing-ditaa
2303 #+begin_src ditaa :results replace :file blue.png :cmdline -r
2313 #+resname: implementing-ditaa
2314 [[file:blue.png][blue.png]]
2316 *** DONE gnuplot [7/7]
2317 (see [[* file result types][file result types]])
2319 #+PLOT: title:"Citas" ind:1 deps:(3) type:2d with:histograms set:"yrange [0:]"
2320 #+TBLNAME: gnuplot-data
2321 | independent var | first dependent var | second dependent var |
2322 |-----------------+---------------------+----------------------|
2323 | 0.1 | 0.425 | 0.375 |
2324 | 0.2 | 0.3125 | 0.3375 |
2325 | 0.3 | 0.24999993 | 0.28333338 |
2326 | 0.4 | 0.275 | 0.28125 |
2327 | 0.5 | 0.26 | 0.27 |
2328 | 0.6 | 0.25833338 | 0.24999993 |
2329 | 0.7 | 0.24642845 | 0.23928553 |
2330 | 0.8 | 0.23125 | 0.2375 |
2331 | 0.9 | 0.23333323 | 0.2333332 |
2332 | 1 | 0.2225 | 0.22 |
2333 | 1.1 | 0.20909075 | 0.22272708 |
2334 | 1.2 | 0.19999998 | 0.21458333 |
2335 | 1.3 | 0.19615368 | 0.21730748 |
2337 #+srcname: implementing-gnuplot
2338 #+begin_src gnuplot :var data=gnuplot-data :results silent
2339 set title "Implementing Gnuplot"
2340 plot data using 1:2 with lines
2343 **** DONE add variables
2344 gnuplot 4.2 and up support user defined variables. This is how
2345 we will handle variables with org-babel (meaning we will need to
2346 require gnuplot 4.2 and up for variable support, which can be
2347 install using [[http://www.macports.org/install.php][macports]] on Mac OSX).
2349 - scalar variables should be replaced in the body of the gnuplot code
2350 - vector variables should be exported to tab-separated files, and
2351 the variable names should be replaced with the path to the files
2353 **** DONE direct plotting w/o session
2354 **** DEFERRED gnuplot support for column/row names
2355 This should be implemented along the lines of the [[* STARTED Column (and row) names of tables in R input/output][R-colname-support]].
2357 We can do something similar to the :labels param in org-plot, we just
2358 have to be careful to ensure that each label is aligned with the
2361 This may be walking too close to an entirely prebuilt plotting tool
2362 rather than straight gnuplot code evaluation. For now I think this
2365 **** DONE a =file= header argument
2366 to specify a file holding the results
2368 #+srcname: gnuplot-to-file-implementation
2369 #+begin_src gnuplot :file plot.png :var data=gnuplot-data
2370 plot data using 1:2, data using 1:3 with lines
2374 [[file:plot.png][plot.png]]
2376 **** DONE helpers from org-plot.el
2377 There are a variety of helpers in org-plot which can be fit nicely
2378 into custom gnuplot header arguments.
2380 These should all be in place by now.
2382 **** DEFERRED header argument specifying 3D data
2385 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
2386 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
2387 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
2388 | 0 | 0 | 1 | 1 | 0 | 0 | 1 | 0 | 0 | 0 | 0 | 0 | 1 | 1 | 1 | 0 | 0 |
2389 | 0 | 1 | 0 | 0 | 1 | 0 | 1 | 0 | 0 | 0 | 0 | 1 | 0 | 0 | 0 | 1 | 0 |
2390 | 0 | 1 | 0 | 0 | 1 | 0 | 1 | 0 | 0 | 0 | 0 | 1 | 0 | 0 | 0 | 1 | 0 |
2391 | 0 | 1 | 0 | 0 | 1 | 0 | 1 | 0 | 0 | 0 | 0 | 1 | 0 | 1 | 1 | 1 | 0 |
2392 | 0 | 1 | 0 | 0 | 1 | 0 | 1 | 0 | 0 | 0 | 0 | 1 | 0 | 0 | 0 | 0 | 0 |
2393 | 0 | 1 | 0 | 0 | 1 | 0 | 1 | 0 | 0 | 0 | 0 | 1 | 0 | 0 | 0 | 0 | 0 |
2394 | 0 | 1 | 0 | 0 | 1 | 0 | 1 | 0 | 0 | 0 | 0 | 1 | 0 | 0 | 0 | 1 | 0 |
2395 | 0 | 1 | 0 | 0 | 1 | 0 | 1 | 1 | 0 | 1 | 0 | 1 | 0 | 0 | 0 | 1 | 0 |
2396 | 0 | 0 | 1 | 1 | 0 | 0 | 1 | 0 | 1 | 1 | 0 | 0 | 1 | 1 | 1 | 0 | 0 |
2397 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
2398 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
2400 #+srcname: implementing-gnuplot-grid-plots
2401 #+begin_src gnuplot :vars data=org-grid
2405 **** DONE gnuplot sessions
2406 Working on this, we won't support multiple sessions as `gnuplot-mode'
2407 isn't setup for such things.
2409 Also we can't display results with the default :none session, so for
2410 gnuplot we really want the default behavior to be :default, and to
2411 only run a :none session when explicitly specified.
2413 #+srcname: implementing-gnuplot-sessions
2414 #+begin_src gnuplot :var data=gnuplot-data :session none :file session.png
2415 set title "Implementing Gnuplot Sessions"
2416 plot data using 1:2 with lines
2420 [[file:session.png][session.png]]
2423 (see [[* file result types][file result types]])
2426 (see [[* file result types][file result types]])
2430 ** TODO allow srcname to omit function call parentheses
2431 Someone needs to revisit those regexps. Is there an argument for
2432 moving some of the regexps used to match function calls into
2433 defvars? (i.e. in o-b.el and o-b-ref.el)
2435 ** TODO creeping blank lines
2436 There's still inappropriate addition of blank lines in some circumstances.
2438 Hmm, it's a bit confusing. It's to do with o-b-remove-result. LoB
2439 removes the entire (#+resname and result) and starts from scratch,
2440 whereas #+begin_src only removes the result. I haven't worked out
2441 what the correct fix is yet. Maybe the right thing to do is to make
2442 sure that those functions (o-b-remove-result et al.) are neutral
2443 with respect to newlines. Sounds easy, but...
2453 Compare the results of
2454 #+lob: python-add(a=5, b=17)
2456 #+resname: python-add(a=5, b=17)
2458 --------------------------------
2466 ---------------------
2467 ** TODO problem with newlines in output when :results value
2468 #+begin_src python :results value
2469 '\n'.join(map(str, range(4)))
2475 Whereas I was hoping for
2482 This is some sort of non-printing char / quoting issue I think. Note
2485 #+begin_src python :results value
2486 '\\n'.join(map(str, range(4)))
2493 #+begin_src python :results output
2494 print('\n'.join(map(str, range(4))))
2503 *** collapsing consecutive newlines in string output
2505 This is an example of the same bug
2507 #+srcname: multi-line-string-output
2508 #+begin_src ruby :results output
2509 "the first line ends here
2512 and this is the second one
2517 This doesn't produce anything at all now. I believe that's because
2518 I've changed things so that :results output really does *not* get the
2519 value of the block, only the STDOUT. So if we add a print statement
2522 #+srcname: multi-line-string-output
2523 #+begin_src ruby :results output
2524 print "the first line ends here
2527 and this is the second one
2533 : the first line ends here
2536 : and this is the second one
2540 However, the behaviour with :results value is wrong
2542 #+srcname: multi-line-string-value
2544 "the first line ends here
2547 and this is the second one
2555 ** TODO prompt characters appearing in output with R
2556 #+begin_src R :session *R* :results output
2565 ** TODO o-b-execute-subtree overwrites heading when subtree is folded
2567 Try M-x org-babel-execute-subtree with the subtree folded and
2568 point at the beginning of the heading line.
2573 ** TODO Allow source blocks to be recognised when #+ are not first characters on the line
2574 I think Carsten has recently altered the core so that #+ can have
2575 preceding whitespace, at least for literal/code examples. org-babel
2576 should support this.
2578 ** TODO non-orgtbl formatted lists
2581 #+srcname: this-doesn't-match-orgtbl
2582 #+begin_src emacs-lisp :results replace
2583 '((:results . "replace"))
2586 #+resname: this-doesn't-match-orgtbl
2588 ** PROPOSED external shell execution can't isolate return values
2589 I have no idea how to do this as of yet. The result is that when
2590 shell functions are run w/o a session there is no difference between
2591 the =output= and =value= result arguments.
2593 Yea, I don't know how to do this either. I searched extensively on
2594 how to isolate the *last* output of a series of shell commands (see
2595 [[* last command for
2596 shells][last command for shells]]). The results of the search were basically
2597 that it was not possible (or at least not accomplish-able with a
2598 reasonable amount of effort).
2600 That fact combined with the tenancy to all ways use standard out in
2601 shell scripts led me to treat these two options (=output= and =value=)
2602 as identical in shell evaluation. Not ideal but maybe good enough for
2605 In the `results' branch I've changed this so that they're not quite
2606 identical: output results in raw stdout contents, whereas value
2607 converts it to elisp, perhaps to a table if it looks tabular. This is
2608 the same for the other languages. [Dan]
2610 ** TODO are the org-babel-trim s necessary?
2611 at the end of e.g. org-babel-R-evaluate, org-babel-python-evaluate, but
2612 not org-babel-ruby-evaluate
2613 ** TODO use new merge function [[file:lisp/org-babel-ref.el::t%20nil%20org%20combine%20plists%20args%20nil][here]]?
2614 And at other occurrences of org-combine-plists?
2615 ** TODO LoB is not populated on startup
2616 org-babel-library-of-babel is nil for me on startup. I have to
2617 evaluate the [[file:lisp/org-babel-lob.el::][org-babel-lob-ingest]] line manually.
2619 ** DONE avoid stripping whitespace from output when :results output
2620 This may be partly solved by using o-b-chomp rather than o-b-trim
2621 in the o-b-LANG-evaluate functions.
2622 ** DEFERRED weird escaped characters in shell prompt break shell evaluation
2623 E.g. this doesn't work. Should the shell sessions set a sane prompt
2624 when they start up? Or is it a question of altering
2625 comint-prompt-regexp? Or altering org-babel regexps?
2628 black=30 ; red=31 ; green=32 ; yellow=33 ; blue=34 ; magenta=35 ; cyan=36 ; white=37
2631 export PS1="\[\033[${prompt_col}m\]\w${prompt_char} \[\033[0m\]"
2634 I just pushed a good amount of changes, could you see if your shell
2635 problems still exist?
2637 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
2638 org-babel-sh.el, raw gets the value
2640 ("" "
\e[0m Sun Jun 14 19:26:24 EDT 2009\n" "
\e[0m org_babel_sh_eoe\n" "
\e[0m ")
2642 and therefore (member org-babel-sh-eoe-output ...) fails
2644 I think that `comint-prompt-regexp' needs to be altered to match
2645 the shell prompt. This shouldn't be too difficult to do by hand,
2646 using the `regexp-builder' command and should probably be part of
2647 the user's regular emacs init. I can't think of a way for us to
2648 set this automatically, and we are SOL without a regexp to match
2650 ** DONE function calls in #+srcname: refs
2652 My srcname references don't seem to be working for function
2653 calls. This needs fixing.
2660 srcname function call doesn't work for calling a source block
2661 #+srcname: caller(var1=called())
2674 They do work for a simple reference
2675 #+srcname: caller2(var1=56)
2684 and they do work for :var header arg
2686 #+begin_src python :var var1=called()
2692 ** DONE LoB: with output to buffer, not working in buffers other than library-of-babel.org
2694 I haven't fixed this yet. org-babel-ref-resolve-reference moves
2695 point around, inside a save-excursion. Somehow when it comes to
2696 inserting the results (after possible further recursive calls to
2697 org-babel-ref-resolve-reference), point hasn't gone back to the
2700 #+tblname: test-data
2705 #+lob: R-plot(data=test-data)
2707 #+lob: python-add(a=2, b=9)
2709 #+resname: python-add(a=2, b=9)
2713 I think this got fixed in the bugfixes before merging results into master.
2715 ** DONE cursor movement when evaluating source blocks
2716 E.g. the pie chart example. Despite the save-window-excursion in
2717 org-babel-execute:R. (I never learned how to do this properly: org-R
2718 jumps all over the place...)
2720 I don't see this now [ded]
2722 ** DONE LoB: calls fail if reference has single character name
2723 commit 21d058869df1ff23f4f8cc26f63045ac9c0190e2
2724 **** This doesn't work
2725 #+lob: R-plot(data=X)
2744 #+lob: R-plot(data=XX)
2746 ** DONE make :results replace the default?
2747 I'm tending to think that appending results to pre-existing results
2748 creates mess, and that the cleaner `replace' option should be the
2749 default. E.g. when a source block creates an image, we would want
2750 that to be updated, rather than have a new one be added.
2754 ** DONE ruby evaluation not working under ubuntu emacs 23
2755 With emacs 23.0.91.1 on ubuntu, for C-h f run-ruby I have the
2756 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.
2759 run-ruby is an interactive compiled Lisp function.
2763 Run an inferior Ruby process, input and output via buffer *ruby*.
2764 If there is a process already running in `*ruby*', switch to that buffer.
2765 With argument, allows you to edit the command line (default is value
2766 of `ruby-program-name'). Runs the hooks `inferior-ruby-mode-hook'
2767 (after the `comint-mode-hook' is run).
2768 (Type C-h m in the process buffer for a list of commands.)
2771 So, I may have a non-standard inf-ruby.el. Here's my version of
2775 run-ruby is an interactive Lisp function in `inf-ruby.el'.
2777 (run-ruby &optional COMMAND NAME)
2779 Run an inferior Ruby process, input and output via buffer *ruby*.
2780 If there is a process already running in `*ruby*', switch to that buffer.
2781 With argument, allows you to edit the command line (default is value
2782 of `ruby-program-name'). Runs the hooks `inferior-ruby-mode-hook'
2783 (after the `comint-mode-hook' is run).
2784 (Type C-h m in the process buffer for a list of commands.)
2787 It seems we could either bundle my version of inf-ruby.el (as it's
2788 the newest). Or we could change the use of `run-ruby' so that it
2789 is robust across multiple distributions. I think I'd prefer the
2790 former, unless the older version of inf-ruby is actually bundled
2791 with emacs, in which case maybe we should go out of our way to
2792 support it. Thoughts?
2794 I think for now I'll just include the latest [[file:util/inf-ruby.el][inf-ruby.el]] in the
2795 newly created utility directory. I doubt anyone would have a
2796 problem using the latest version of this file.
2797 ** DONE test failing forcing vector results with =test-forced-vector-results= ruby code block
2798 Note that this only seems to happen the *second* time the test table
2801 #+srcname: bug-trivial-vector
2802 #+begin_src emacs-lisp :results vector silent
2806 #+srcname: bug-forced-vector-results
2807 #+begin_src ruby :var triv=test-trivial-vector :results silent
2811 mysteriously this seems to be fixed...
2812 ** DONE defunct R sessions
2813 Sometimes an old R session will turn defunct, and newly inserted code
2814 will not be evaluated (leading to a hang).
2816 This seems to be fixed by using `inferior-ess-send-input' rather than `comint-send-input'.
2817 ** DONE ruby fails on first call to non-default session
2819 #+srcname: bug-new-session
2820 #+begin_src ruby :session is-new
2824 ** DONE when reading results from =#+resname= line
2826 Errors when trying to read from resname lines.
2828 #+resname: bug-in-resname
2831 #+srcname: bug-in-resname-reader
2832 #+begin_src emacs-lisp :var buggy=bug-in-resname() :results silent
2836 ** DONE R-code broke on "org-babel" rename
2838 #+srcname: bug-R-babels
2843 ** DONE error on trivial R results
2845 So I know it's generally not a good idea to squash error without
2846 handling them, but in this case the error almost always means that
2847 there was no file contents to be read by =org-table-import=, so I
2850 #+srcname: bug-trivial-r1
2851 #+begin_src R :results replace
2852 pie(c(1, 2, 3), labels = c(1, 2, 3))
2855 #+srcname: bug-trivial-r2
2856 #+begin_src R :results replace
2860 #+resname: bug-trivial-r2
2863 #+srcname: bug-trivial-r3
2864 #+begin_src R :results replace
2868 #+resname: bug-trivial-r3
2873 ** DONE ruby new variable creation (multi-line ruby blocks)
2874 Actually it looks like we were dropping all but the last line.
2876 #+srcname: multi-line-ruby-test
2877 #+begin_src ruby :var table=bug-numerical-table :results replace
2879 table.each{|n| total += n}
2886 ** DONE R code execution seems to choke on certain inputs
2887 Currently the R code seems to work on vertical (but not landscape)
2890 #+srcname: little-fake
2891 #+begin_src emacs-lisp
2896 #+begin_src R :var num=little-fake
2903 #+srcname: set-debug-on-error
2904 #+begin_src emacs-lisp :results silent
2905 (setq debug-on-error t)
2908 #+srcname: bug-numerical-table
2909 #+begin_src emacs-lisp :results silent
2916 #+srcname: bug-R-number-evaluation
2917 #+begin_src R :var table=bug-numerical-table
2926 #+tblname: bug-vert-table
2931 #+srcname: bug-R-vertical-table
2932 #+begin_src R :var table=bug-vert-table :results silent
2936 ** DONE org bug/request: prevent certain org behaviour within code blocks
2937 E.g. [[]] gets recognised as a link (when there's text inside the
2938 brackets). This is bad for R code at least, and more generally
2939 could be argued to be inappropriate. Is it difficult to get org to
2940 ignore text in code blocks? [DED]
2942 I believe Carsten addressed this recently on the mailing list with
2943 the comment that it was indeed a difficult issue. I believe this
2944 may be one area where we could wait for an upstream (org-mode) fix.
2946 [Dan] Carsten has fixed this now in the core.
2948 ** DONE with :results replace, non-table output doesn't replace table output
2949 And vice versa. E.g. Try this first with table and then with len(table) [DED]
2950 #+begin_src python :var table=sandbox :results replace
2955 | 4 | "schulte" | 6 |
2958 Yes, this is certainly a problem. I fear that if we begin replacing
2959 anything immediately following a source block (regardless of whether
2960 it matches the type of our current results) we may accidentally delete
2961 hand written portions of the user's org-mode buffer.
2963 I think that the best solution here would be to actually start
2964 labeling results with a line that looks something like...
2968 This would have a couple of benefits...
2969 1) we wouldn't have to worry about possibly deleting non-results
2970 (which is currently an issue)
2971 2) we could reliably replace results even if there are different types
2972 3) we could reference the results of a source-code block in variable
2973 definitions, which would be useful if for example we don't wish to
2974 re-run a source-block every time because it is long-running.
2976 Thoughts? If no-one objects, I believe I will implement the labeling
2979 ** DONE extra quotes for nested string
2980 Well R appears to be reading the tables without issue...
2982 these *should* be quoted
2984 #+begin_src sh :results replace
2989 | "README.markdown" |
2992 | "existing_tools" |
2996 | "test-export.html" |
2997 | "test-export.org" |
2999 #+srcname: test-quotes
3000 #+begin_src ruby :var tab=ls
3006 #+srcname: test-quotes
3007 #+begin_src R :var tab=ls
3013 ** DONE simple ruby arrays not working
3015 As an example eval the following. Adding a line to test
3017 #+tblname: simple-ruby-array
3020 #+srcname: ruby-array-test
3021 #+begin_src ruby :var ar = simple-ruby-array :results silent
3025 ** DONE space trailing language name
3026 fix regexp so it works when there's a space trailing the language name
3028 #+srcname: test-trailing-space
3033 ** DONE Args out of range error
3035 The following block resulted in the error below [DED]. It ran without
3036 error directly in the shell.
3039 for platf in ill aff ; do
3040 for pop in CEU YRI ASI ; do
3041 rm -f $platf/hapmap-genos-$pop-all $platf/hapmap-rs-all
3042 cat $platf/hapmap-genos-$pop-* > $platf/hapmap-genos-$pop-all
3043 cat $platf/hapmap-rs-* > $platf/hapmap-rs-all
3048 executing source block with sh...
3049 finished executing source block
3050 string-equal: Args out of range: "", -1, 0
3052 the error =string-equal: Args out of range: "", -1, 0= looks like what
3053 used to be output when the block returned an empty results string.
3054 This should be fixed in the current version, you should now see the
3055 following message =no result returned by source block=.
3057 ** DONE ruby arrays not recognized as such
3059 Something is wrong in [[file:lisp/org-babel-script.el]] related to the
3060 recognition of ruby arrays as such.
3062 #+begin_src ruby :results replace
3068 #+begin_src python :results replace
3073 ** REJECTED elisp reference fails for literal number
3074 That's a bug in Dan's elisp, not in org-babel.
3075 #+srcname: elisp-test(a=4)
3076 #+begin_src emacs-lisp
3086 Evaluate all the cells in this table for a comprehensive test of the
3087 org-babel functionality.
3089 *Note*: if you have customized =org-babel-default-header-args= then some
3090 of these tests may fail.
3092 #+TBLNAME: org-babel-tests
3093 | functionality | block | arg | expected | results | pass |
3094 |-------------------------+----------------------------+-----+-------------+-------------+------|
3095 | basic evaluation | | | | | pass |
3096 |-------------------------+----------------------------+-----+-------------+-------------+------|
3097 | emacs lisp | basic-elisp | | 5 | 5 | pass |
3098 | shell | basic-shell | | 6 | 6 | pass |
3099 | ruby | basic-ruby | | org-babel | org-babel | pass |
3100 | python | basic-python | | hello world | hello world | pass |
3101 | R | basic-R | | 13 | 13 | pass |
3102 |-------------------------+----------------------------+-----+-------------+-------------+------|
3103 | tables | | | | | pass |
3104 |-------------------------+----------------------------+-----+-------------+-------------+------|
3105 | emacs lisp | table-elisp | | 3 | 3 | pass |
3106 | ruby | table-ruby | | 1-2-3 | 1-2-3 | pass |
3107 | python | table-python | | 5 | 5 | pass |
3108 | R | table-R | | 3.5 | 3.5 | pass |
3109 | R: col names in R | table-R-colnames | | -3 | -3 | pass |
3110 | R: col names in org | table-R-colnames-org | | 169 | 169 | pass |
3111 |-------------------------+----------------------------+-----+-------------+-------------+------|
3112 | source block references | | | | | pass |
3113 |-------------------------+----------------------------+-----+-------------+-------------+------|
3114 | all languages | chained-ref-last | | Array | Array | pass |
3115 |-------------------------+----------------------------+-----+-------------+-------------+------|
3116 | source block functions | | | | | pass |
3117 |-------------------------+----------------------------+-----+-------------+-------------+------|
3118 | emacs lisp | defun-fibb | | fibbd | fibbd | pass |
3119 | run over | Fibonacci | 0 | 1 | 1 | pass |
3120 | a | Fibonacci | 1 | 1 | 1 | pass |
3121 | variety | Fibonacci | 2 | 2 | 2 | pass |
3122 | of | Fibonacci | 3 | 3 | 3 | pass |
3123 | different | Fibonacci | 4 | 5 | 5 | pass |
3124 | arguments | Fibonacci | 5 | 8 | 8 | pass |
3125 |-------------------------+----------------------------+-----+-------------+-------------+------|
3126 | bugs and tasks | | | | | pass |
3127 |-------------------------+----------------------------+-----+-------------+-------------+------|
3128 | simple ruby arrays | ruby-array-test | | 3 | 3 | pass |
3129 | R number evaluation | bug-R-number-evaluation | | 2 | 2 | pass |
3130 | multi-line ruby blocks | multi-line-ruby-test | | 2 | 2 | pass |
3131 | forcing vector results | test-forced-vector-results | | Array | Array | pass |
3132 |-------------------------+----------------------------+-----+-------------+-------------+------|
3133 | sessions | | | | | pass |
3134 |-------------------------+----------------------------+-----+-------------+-------------+------|
3135 | set ruby session | set-ruby-session-var | | :set | :set | pass |
3136 | get from ruby session | get-ruby-session-var | | 3 | 3 | pass |
3137 | set python session | set-python-session-var | | set | set | pass |
3138 | get from python session | get-python-session-var | | 4 | 4 | pass |
3139 | set R session | set-R-session-var | | set | set | pass |
3140 | get from R session | get-R-session-var | | 5 | 5 | pass |
3141 #+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))
3142 #+TBLFM: $5=""::$6=""
3145 The second TBLFM line (followed by replacing '[]' with '') can be used
3146 to blank out the table results, in the absence of a better method.
3150 #+srcname: basic-elisp
3151 #+begin_src emacs-lisp :results silent
3156 #+srcname: basic-shell
3157 #+begin_src sh :results silent
3162 #+srcname: date-simple
3163 #+begin_src sh :results silent
3167 #+srcname: basic-ruby
3168 #+begin_src ruby :results silent
3173 #+srcname: basic-python
3174 #+begin_src python :results silent
3180 #+begin_src R :results silent
3188 #+tblname: test-table
3192 #+tblname: test-table-colnames
3193 | var1 | var2 | var3 |
3194 |------+------+------|
3198 #+srcname: table-elisp
3199 #+begin_src emacs-lisp :results silent :var table=test-table
3200 (length (car table))
3204 #+srcname: table-ruby
3205 #+begin_src ruby :results silent :var table=test-table
3206 table.first.join("-")
3210 #+srcname: table-python
3211 #+begin_src python :var table=test-table
3215 #+srcname: table-R(table=test-table)
3220 #+srcname: table-R-colnames(table=test-table-colnames)
3221 #+begin_src R :results silent
3222 sum(table$var2 - table$var3)
3225 #+srcname: R-square(x=default-name-doesnt-exist)
3226 #+begin_src R :colnames t
3230 This should return 169. The fact that R is able to use the column name
3231 to index the data frame (x$var3) proves that a table with column names
3232 (a header row) has been recognised as input for the R-square function
3233 block, and that the R-square block has output an elisp table with
3234 column names, and that the colnames have again been recognised when
3235 creating the R variables in this block.
3236 #+srcname: table-R-colnames-org(x = R-square(x=test-table-colnames))
3246 Lets pass a references through all of our languages...
3248 Lets start by reversing the table from the previous examples
3250 #+srcname: chained-ref-first
3251 #+begin_src python :var table = test-table
3256 #+resname: chained-ref-first
3260 Take the first part of the list
3262 #+srcname: chained-ref-second
3263 #+begin_src R :var table = chained-ref-first
3267 #+resname: chained-ref-second
3271 Turn the numbers into string
3273 #+srcname: chained-ref-third
3274 #+begin_src emacs-lisp :var table = chained-ref-second
3275 (mapcar (lambda (el) (format "%S" el)) table)
3278 #+resname: chained-ref-third
3281 and Check that it is still a list
3283 #+srcname: chained-ref-last
3284 #+begin_src ruby :var table=chained-ref-third
3289 ** source blocks as functions
3291 #+srcname: defun-fibb
3292 #+begin_src emacs-lisp :results silent
3293 (defun fibbd (n) (if (< n 2) 1 (+ (fibbd (- n 1)) (fibbd (- n 2)))))
3297 #+srcname: fibonacci
3298 #+begin_src emacs-lisp :results silent :var n=7
3308 ** sbe tests (these don't seem to be working...)
3309 Testing the insertion of results into org-mode tables.
3311 #+srcname: multi-line-output
3312 #+begin_src ruby :results replace
3313 "the first line ends here
3316 and this is the second one
3322 : the first line ends here\n\n\n and this is the second one\n\neven a third
3324 #+srcname: multi-line-error
3325 #+begin_src ruby :results replace
3326 raise "oh nooooooooooo"
3332 | the first line ends here... | -:5: warning: parenthesize argument(s) for future version... |
3333 #+TBLFM: $1='(sbe "multi-line-output")::$2='(sbe "multi-line-error")
3335 ** forcing results types tests
3337 #+srcname: test-trivial-vector
3338 #+begin_src emacs-lisp :results vector silent
3342 #+srcname: test-forced-vector-results
3343 #+begin_src ruby :var triv=test-trivial-vector :results silent
3349 #+srcname: set-ruby-session-var
3350 #+begin_src ruby :session :results silent
3355 #+srcname: get-ruby-session-var
3356 #+begin_src ruby :session :results silent
3360 #+srcname: set-python-session-var
3361 #+begin_src python :session
3366 #+srcname: get-python-session-var
3367 #+begin_src python :session
3371 #+srcname: set-R-session-var
3372 #+begin_src R :session
3377 #+srcname: get-R-session-var
3378 #+begin_src R :session
3387 To run these examples evaluate [[file:lisp/org-babel-init.el][org-babel-init.el]]
3389 ** org-babel.el beginning functionality
3391 #+begin_src sh :results replace
3396 : Sun Jul 5 18:54:39 EDT 2009
3403 : Sun Jul 05 18:54:35 -0400 2009
3415 #+begin_src R :results replace
3425 hist(rgamma(20,3,3))
3430 ** org-babel plays with tables
3431 Alright, this should demonstrate both the ability of org-babel to read
3432 tables into a lisp source code block, and to then convert the results
3433 of the source code block into an org table. It's using the classic
3434 "lisp is elegant" demonstration transpose function. To try this
3437 1. evaluate [[file:lisp/org-babel-init.el]] to load org-babel and friends
3438 2. evaluate the transpose definition =\C-c\\C-c= on the beginning of
3440 3. evaluate the next source code block, this should read in the table
3441 because of the =:var table=previous=, then transpose the table, and
3442 finally it should insert the transposed table into the buffer
3443 immediately following the block
3447 #+begin_src emacs-lisp :results silent
3448 (defun transpose (table)
3449 (apply #'mapcar* #'list table))
3457 #+begin_src emacs-lisp :var table=sandbox :results replace
3462 #+begin_src emacs-lisp
3467 | 1 | 2 | 3 | 4 | 5 |
3471 #+begin_src ruby :var table=sandbox :results replace
3472 table.first.join(" - ")
3478 #+begin_src python :var table=sandbox
3483 #+begin_src ruby :var table=sandbox :results replace
3488 : [[1, 2, 3], [4, "schulte", 6]]
3492 | 4 | "schulte" | 6 |
3494 #+begin_src python :var table=sandbox :results replace
3500 | "__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" |
3502 *** (sandbox table) R
3504 #+TBLNAME: sandbox_r
3508 #+begin_src R :results replace
3509 x <- c(rnorm(10, mean=-3, sd=1), rnorm(10, mean=3, sd=1))
3513 | -3.35473133869346 |
3515 | -3.32819924928633 |
3516 | -2.97310212756194 |
3517 | -2.09640758369576 |
3518 | -5.06054014378736 |
3519 | -2.20713700711221 |
3520 | -1.37618039712037 |
3521 | -1.95839385821742 |
3522 | -3.90407396475502 |
3523 | 2.51168071590226 |
3524 | 3.96753011570494 |
3525 | 3.31793212627865 |
3526 | 1.99829753972341 |
3527 | 4.00403686419829 |
3528 | 4.63723764452927 |
3529 | 3.94636744261313 |
3530 | 3.58355906547775 |
3531 | 3.01563442274226 |
3534 #+begin_src R var tabel=sandbox_r :results replace
3539 | 4 | "schulte" | 6 |
3542 Now shell commands are converted to tables using =org-table-import=
3543 and if these tables are non-trivial (i.e. have multiple elements) then
3544 they are imported as org-mode tables...
3546 #+begin_src sh :results replace
3550 | "total" | 208 | "" | "" | "" | "" | "" | "" |
3551 | "-rw-r--r--" | 1 | "dan" | "dan" | 57 | 2009 | 15 | "block" |
3552 | "-rw-r--r--" | 1 | "dan" | "dan" | 35147 | 2009 | 15 | "COPYING" |
3553 | "-rw-r--r--" | 1 | "dan" | "dan" | 722 | 2009 | 18 | "examples.org" |
3554 | "drwxr-xr-x" | 4 | "dan" | "dan" | 4096 | 2009 | 19 | "existing_tools" |
3555 | "-rw-r--r--" | 1 | "dan" | "dan" | 2207 | 2009 | 14 | "intro.org" |
3556 | "drwxr-xr-x" | 2 | "dan" | "dan" | 4096 | 2009 | 18 | "org-babel" |
3557 | "-rw-r--r--" | 1 | "dan" | "dan" | 277 | 2009 | 20 | "README.markdown" |
3558 | "-rw-r--r--" | 1 | "dan" | "dan" | 11837 | 2009 | 18 | "rorg.html" |
3559 | "-rw-r--r--" | 1 | "dan" | "dan" | 61829 | 2009 | 19 | "#rorg.org#" |
3560 | "-rw-r--r--" | 1 | "dan" | "dan" | 60190 | 2009 | 19 | "rorg.org" |
3561 | "-rw-r--r--" | 1 | "dan" | "dan" | 972 | 2009 | 11 | "test-export.org" |
3564 ** silent evaluation
3572 #+begin_src ruby :results silent
3576 #+begin_src ruby :results replace
3583 ** (sandbox) referencing other source blocks
3584 Doing this in emacs-lisp first because it's trivial to convert
3585 emacs-lisp results to and from emacs-lisp.
3587 *** emacs lisp source reference
3588 This first example performs a calculation in the first source block
3589 named =top=, the results of this calculation are then saved into the
3590 variable =first= by the header argument =:var first=top=, and it is
3591 used in the calculations of the second source block.
3594 #+begin_src emacs-lisp
3598 #+begin_src emacs-lisp :var first=top :results replace
3604 This example is the same as the previous only the variable being
3605 passed through is a table rather than a number.
3607 #+begin_src emacs-lisp :results silent
3608 (defun transpose (table)
3609 (apply #'mapcar* #'list table))
3612 #+TBLNAME: top_table
3616 #+SRCNAME: second_src_example
3617 #+begin_src emacs-lisp :var table=top_table
3621 #+begin_src emacs-lisp :var table=second_src_example :results replace
3626 | 4 | "schulte" | 6 |
3628 Now working for ruby
3635 #+begin_src ruby :var other=start :results replace
3641 #+SRCNAME: start_two
3646 #+begin_src python :var another=start_two :results replace
3651 Since all variables are converted into Emacs Lisp it is no problem to
3652 reference variables specified in another language.
3654 #+SRCNAME: ruby-block
3659 #+SRCNAME: lisp_block
3660 #+begin_src emacs-lisp :var ruby-variable=ruby-block
3664 #+begin_src python :var lisp_var=lisp_block
3673 #+begin_src R :results replace
3680 #+begin_src R :var other=first_r :results replace
3687 ** (sandbox) selective export
3689 For exportation tests and examples see (including exportation of
3690 inline source code blocks) [[file:test-export.org]]
3693 ** (sandbox) source blocks as functions
3696 #+begin_src emacs-lisp :results silent
3701 #+begin_src emacs-lisp :var n=default :results replace
3707 #+begin_src emacs-lisp :var result=triple(n=3, m=98) :results replace
3713 The following just demonstrates the ability to assign variables to
3714 literal values, which was not implemented until recently.
3716 #+begin_src ruby :var num="eric" :results replace
3723 ** (sandbox) inline source blocks
3725 This is an inline source code block src_ruby{1 + 6}. And another
3726 source block with text output src_emacs-lisp{"eric"}.
3728 This is an inline source code block with header
3729 arguments. src_ruby[:var n=fibbd( n = 0 )]{n}
3732 ** (sandbox) integration w/org tables
3734 #+begin_src emacs-lisp :results silent
3735 (defun fibbd (n) (if (< n 2) 1 (+ (fibbd (- n 1)) (fibbd (- n 2)))))
3739 #+begin_src emacs-lisp :var n=4 :results silent
3743 #+begin_src emacs-lisp :results silent
3744 (mapcar #'fibbd '(0 1 2 3 4 5 6 7 8))
3747 Something is not working here. The function `sbe ' works fine when
3748 called from outside of the table (see the source block below), but
3749 produces an error when called from inside the table. I think there
3750 must be some narrowing going on during intra-table emacs-lisp
3753 | original | fibbd |
3754 |----------+-------|
3765 #+TBLFM: $2='(sbe "fibbd" (n $1))
3769 #+begin_src emacs-lisp :results silent
3770 (sbe 'fibbd (n "8"))
3775 LocalWords: DBlocks dblocks org-babel el eric fontification