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
87 pie(dirs[,1], labels = dirs[,2])
90 *** operations in/on tables
92 #+tblname: grades-table
93 | student | grade | letter |
94 |---------+-------+--------|
101 #+TBLFM: $2='(sbe random-score-generator)::$3='(sbe assign-grade (score $2))
103 #+srcname: assign-grade
104 #+begin_src ruby :var score=99
115 #+srcname: random-score-generator
120 #+srcname: show-distribution
121 #+begin_src R :var grades=grades-table
126 ** communication between people
127 Quick overview of Org-Mode's exportation abilities, with links to the
128 online Org-Mode documentation, a focus on source-code blocks, and the
129 exportation options provided by Org-Babel.
131 *** Interactive tutorial
132 This would demonstrate applicability to Reproducible Research, and
133 Literate Programming.
135 *** Tests embedded in documentation
136 org-babels own functional tests are contained in a large org-mode
137 table, allowing the test suite to be run be evaluation of the table
138 and the results to be collected in the same table.
140 *** Emacs initialization files stored in Org-Mode buffers
141 Using `org-babel-tangle' it is possible to embed your Emacs
142 initialization into org-mode files. This allows for folding,
143 note-taking, todo's etc... embedded with the source-code of your Emacs
144 initialization, and through org-mode's publishing features aids in
145 sharing your customizations with others.
147 It may be worthwhile to create a fork of Phil Hagelberg's
148 [[http://github.com/technomancy/emacs-starter-kit/tree/master][emacs-starter-kit]] which uses literate org-mode files for all of the
149 actual elisp customization. These org-mode files could then be
150 exported to html and used to populate the repositories wiki on [[http://github.com/][github]].
155 *** code evaluation (comint buffer sessions and external processes)
156 There are two main ways to evaluate source blocks with org-babel.
158 - external :: By default (if the =:session= header argument is not
159 present) all source code blocks are evaluated in
160 external processes. In these cases an external process
161 is used to evaluate the source-code blocks.
162 - session :: Session based evaluation uses persistent sessions in
163 comint buffers. Sessions can be used across multiple
164 source blocks setting and accessing variables in the
167 Evaluating source blocks in sessions also allows for
168 interaction with the code. To jump to the session of a
169 source block use the `org-babel-pop-to-session' command
170 or press =M-[down]= while inside of a source code block.
171 When called with a prefix argument
172 `org-babel-pop-to-session' will evaluate all header
173 arguments before jumping to the source-code block.
175 *** results (values and outputs)
176 Either the *value* or the *output* of source code blocks can be
177 collected after evaluation.
179 - value :: The default way to collect results from a source-code block
180 is to return the value of the last statement in the block.
181 This can be thought of as the return value of the block.
182 In this case any printed output of the block is ignored.
183 This can be though of a similar to a "functional" value of
185 - output :: Another way of generating results from a source-code block
186 is to collect the output generated by the execution of the
187 block. In this case all printed output is collected
188 throughout the execution of the block. This can be
189 thought of as similar to a "script" style of evaluation.
193 Add the following lines to your .emacs, replacing the path as
194 appropriate. A good place to check that things are up and running
195 would then be [[#sandbox][the sandbox]].
196 #+begin_src emacs-lisp
197 (add-to-list 'load-path "/path/to/org-babel/lisp")
198 (require 'org-babel-init)
203 ** TODO make tangle files read-only?
204 With a file-local variable setting?
205 ** TODO fold source blocks on #+srcname line?
206 I.e., as well as on #+begin_src line.
207 ** TODO support for working with =*Org Edit Src Example*= buffers [2/4]
208 *** TODO optionally evaluate header references when we switch to =*Org Edit Src*= buffer
209 That seems to imply that the header references need to be evaluated
210 and transformed into the target language object when we hit C-c ' to
211 enter the *Org Edit Src* buffer [DED]
213 Good point, I heartily agree that this should be supported [Eric]
215 (or at least before the first time we attempt to evaluate code in that
216 buffer -- I suppose there might be an argument for lazy evaluation, in
217 case someone hits C-c ' but is "just looking" and not actually
218 evaluating anything.) Of course if evaluating the reference is
219 computationally intensive then the user might have to wait before they
220 get the *Org Edit Src* buffer. [DED]
222 I fear that it may be hard to anticipate when the references will be
223 needed, some major-modes do on-the-fly evaluation while the buffer is
224 being edited. I think that we should either do this before the buffer
225 is opened or not at all, specifically I think we should resolve
226 references if the user calls C-c ' with a prefix argument. Does that
227 sound reasonable? [Eric]
231 [Dan] So now that we have org-src-mode and org-src-mode-hook, I guess
232 org-babel should do this by using the hook to make sure that, when C-c
233 C-' is issued on a source block, any references are resolved and
234 assignments are made in the appropriate session.
235 *** TODO set buffer-local-process variables appropriately [DED]
236 I think something like this would be great. You've probably
237 already thought of this, but just to note it down: it would be really
238 nice if org-babel's notion of a buffer's 'session/process' played
239 nicely with ESS's notion of the buffer's session/process. ESS keeps
240 the current process name for a buffer in a buffer-local variable
241 ess-local-process-name. So one thing we will probably want to do is
242 make sure that the *Org Edit Src Example* buffer sets that variable
245 I had not thought of that, but I agree whole heartedly. [Eric]
247 Once this is done every variable should be able to dump regions into
248 their inferior-process buffer using major-mode functions.
249 *** DEFERRED send code to inferior process
250 Another thought on this topic: I think we will want users to send
251 chunks of code to the interpreter from within the *Org Edit Src*
252 buffer, and I think that's what you have in mind already. In ESS that
253 is done using the ess-eval-* functions. [DED]
255 I think we can leave this up to the major-mode in the source code
256 buffer, as almost every source-code major mode will have functions for
257 doing things like sending regions to the inferior process. If
258 anything we might need to set the value of the buffer local inferior
259 process variable. [Eric]
261 *** DONE some possible requests/proposed changes for Carsten [4/4]
262 While I remember, some possible requests/proposed changes for Carsten
263 come to mind in that regard:
265 **** DONE Remap C-x C-s to save the source to the org buffer?
266 I've done this personally and I find it essential. I'm using
267 #+begin_src emacs-lisp
268 (defun org-edit-src-save ()
269 "Update the parent org buffer with the edited source code, save
270 the parent org-buffer, and return to the source code edit
279 (define-key org-exit-edit-mode-map "\C-x\C-s" 'org-edit-src-save)
283 I think this is great, but I think it should be implemented in the
286 **** DEFERRED Rename buffer and minor mode?
287 Something shorter than *Org Edit Src Example* for the buffer
288 name. org-babel is bringing org's source code interaction to a
289 level of maturity where the 'example' is no longer
290 appropriate. And if further keybindings are going to be added to
291 the minor mode then maybe org-edit-src-mode is a better name than
294 Maybe we should name the buffer with a combination of the source
295 code and the session. I think that makes sense.
297 [ES] Are you also suggesting a new org-edit-src minor mode?
298 [DED] org-exit-edit-mode is a minor mode that already exists:
300 Minor mode installing a single key binding, "C-c '" to exit special edit.
302 org-edit-src-save now has a binding in that mode, so I guess all
303 I'm saying at this stage is that it's a bit of a misnomer. But
304 perhaps we will also have more functionality to add to that minor
305 mode, making it even more of a misnomer. Perhaps something like
306 org-src-mode would be better.
307 **** DONE Changed minor mode name and added hooks
309 **** DEFERRED a hook called when the src edit buffer is created
310 This should be implemented in the org-mode core
312 ** TODO resolve references to other org buffers/files
313 This would allow source blocks to call upon tables, source-blocks,
314 and results in other org buffers/files.
317 - [[file:lisp/org-babel-ref.el::TODO%20allow%20searching%20for%20names%20in%20other%20buffers][org-babel-ref.el:searching-in-other-buffers]]
318 - [[file:lisp/org-babel.el::defun%20org-babel%20find%20named%20result%20name][org-babel.el#org-babel-find-named-result]]
319 ** TODO resolve references to other non-org files
320 - tabular data in .csv, .tsv etc format
321 - files of interpreted code: anything stopping us giving such files
322 similar status to a source code block?
323 - Would be nice to allow org and non-org files to be remote
324 ** TODO figure out how to handle errors during evaluation
325 R has a try function, with error handling, along the lines of
326 python. I bet ruby does too. Maybe more of an issue for functional
327 style; in my proposed scripting style the error just gets dumped to
328 the org buffer and the user is thus alerted.
329 ** TODO figure out how to handle graphic output
330 This is listed under [[* graphical output][graphical output]] in out objectives.
332 This should take advantage of the =:results file= option, and
333 languages which almost always produce graphical output should set
334 =:results file= to true by default. That would handle placing these
335 results in the buffer. Then if there is a combination of =silent= and
336 =file= =:results= headers we could drop the results to a temp buffer
337 and pop open that buffer...
338 ** TODO =\C-c \C-o= to open results of source block
339 by adding a =defadvice= to =org-open-at-point= we can use the common
340 =\C-c \C-o= keybinding to open the results of a source-code block.
341 This would be especially useful for source-code blocks which generate
342 graphical results and insert a file link as the results in the
343 org-mode buffer. (see [[* figure out how to handle graphic output][TODO figure out how to handle graphic output]]).
344 This could also act reasonably with other results types...
346 - file :: use org-open-at-point to open the file
347 - scalar :: open results unquoted in a new buffer
348 - tabular :: export the table to a new buffer and open that buffer
350 ** TODO Finalise behaviour regarding vector/scalar output
351 *** DONE Stop spaces causing vector output
352 This simple example of multilingual chaining produces vector output if
353 there are spaces in the message and scalar otherwise.
357 #+begin_src R :var msg=msg-from-python
358 paste(msg, "und R", sep=" ")
362 : org-babel speaks elisp y python und R
364 #+srcname: msg-from-python
365 #+begin_src python :var msg=msg-from-elisp
369 #+srcname: msg-from-elisp
370 #+begin_src emacs-lisp :var msg="org-babel speaks"
371 (concat msg " elisp")
373 ** STARTED share org-babel [1/4]
374 how should we share org-babel?
376 *** DONE post to org-mode
377 *** TODO post to ess mailing list
378 *** TODO create a org-babel page on worg
379 *** TODO create a short screencast demonstrating org-babel in action
382 we need to think up some good examples
384 **** interactive tutorials
385 This could be a place to use [[* org-babel assertions][org-babel assertions]].
387 for example the first step of a tutorial could assert that the version
388 of the software-package (or whatever) is equal to some value, then
389 source-code blocks could be used with confidence (and executed
390 directly from) the rest of the tutorial.
392 **** answering a text-book question w/code example
393 org-babel is an ideal environment enabling both the development and
394 demonstrationg of the code snippets required as answers to many
397 **** something using tables
398 maybe something along the lines of calculations from collected grades
401 Maybe something like the following which outputs sizes of directories
402 under the home directory, and then instead of the trivial =emacs-lisp=
403 block we could use an R block to create a nice pie chart of the
407 #+begin_src bash :results replace
411 #+begin_src emacs-lisp :var sizes=sizes :results replace
415 ** TODO command line execution
416 Allow source code blocks to be called form the command line. This
417 will be easy using the =sbe= function in [[file:lisp/org-babel-table.el][org-babel-table.el]].
419 This will rely upon [[* resolve references to other buffers][resolve references to other buffers]].
421 ** TODO inline source code blocks [3/5]
422 Like the =\R{ code }= blocks
424 not sure what the format should be, maybe just something simple
425 like =src_lang[]{}= where lang is the name of the source code
426 language to be evaluated, =[]= is optional and contains any header
427 arguments and ={}= contains the code.
429 (see [[* (sandbox) inline source blocks][the-sandbox]])
431 *** DONE evaluation with \C-c\C-c
432 Putting aside the header argument issue for now we can just run these
433 with the following default header arguments
434 - =:results= :: silent
435 - =:exports= :: results
437 *** DONE inline exportation
438 Need to add an interblock hook (or some such) through org-exp-blocks
439 *** DONE header arguments
440 We should make it possible to use header arguments.
442 *** TODO fontification
443 we should color these blocks differently
445 *** TODO refine html exportation
446 should use a span class, and should show original source in tool-tip
448 ** TODO formulate general rules for handling vectors and tables / matrices with names
449 This is non-trivial, but may be worth doing, in particular to
450 develop a nice framework for sending data to/from R.
452 In R, indexing vector elements, and rows and columns, using
453 strings rather than integers is an important part of the
455 - elements of a vector may have names
456 - matrices and data.frames may have "column names" and "row names"
457 which can be used for indexing
458 - In a data frame, row names *must* be unique
466 > mat <- matrix(1:4, nrow=2, ncol=2, dimnames=list(c("r1","r2"), c("c1","c2")))
471 > # The names are separate from the data: they do not interfere with operations on the data
478 > df <- data.frame(var1=1:26, var2=26:1, row.names=letters)
480 [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
486 So it's tempting to try to provide support for this in org-babel. For example
487 - allow R to refer to columns of a :var reference by their names
488 - When appropriate, results from R appear in the org buffer with "named
491 However none (?) of the other languages we are currently supporting
492 really have a native matrix type, let alone "column names" or "row
493 names". Names are used in e.g. python and perl to refer to entries
496 It currently seems to me that support for this in org-babel would
497 require setting rules about when org tables are considered to have
498 named columns/fields, and ensuring that (a) languages with a notion
499 of named columns/fields use them appropriately and (b) languages
500 with no such notion do not treat then as data.
502 - Org allows something that *looks* like column names to be separated
504 - Org also allows a row to *function* as column names when special
505 markers are placed in the first column. An hline is unnecessary
506 (indeed hlines are purely cosmetic in org [correct?]
507 - Org does not have a notion of "row names" [correct?]
509 The full org table functionality exeplified [[http://orgmode.org/manual/Advanced-features.html#Advanced-features][here]] has features that
510 we would not support in e.g. R (like names for the row below).
512 *** Initial statement: allow tables with hline to be passed as args into R
513 This doesn't seem to work at the moment (example below). It would
514 also be nice to have a natural way for the column names of the org
515 table to become the column names of the R data frame, and to have
516 the option to specify that the first column is to be used as row
517 names in R (these must be unique). But this might require a bit of
522 | col1 | col2 | col3 |
523 |------+---------+------|
531 #+begin_src R var tabel=egtable
536 | "col1" | "col2" | "col3" |
537 |--------+-----------+--------|
539 | 4 | "schulte" | 6 |
542 Another example is in the [[*operations%20in%20on%20tables][grades example]].
544 ** TODO re-implement helper functions from org-R
545 *** Initial statement [Eric]
546 Much of the power of org-R seems to be in it's helper functions for
547 the quick graphing of tables. Should we try to re-implement these
548 functions on top of org-babel?
550 I'm thinking this may be useful both to add features to org-babel-R and
551 also to potentially suggest extensions of the framework. For example
552 one that comes to mind is the ability to treat a source-code block
553 like a function which accepts arguments and returns results. Actually
554 this can be it's own TODO (see [[* source blocks as functions][source blocks as functions]]).
556 - We want to provide convenient off-the-shelf actions
557 (e.g. plotting data) that make use of our new code evaluation
558 environment but do not require any actual coding.
559 *** Initial Design proposal [Dan]
560 - *Input data* will be specified using the same mechanism as :var
561 references, thus the input data may come from a table, or
562 another source block, and it is initially available as an elisp
564 - We introduce a new #+ line, e.g. #+BABELDO. C-c C-c on that
565 line will apply an *action* to the referenced data.
566 - *Actions correspond to source blocks*: our library of available
567 actions will be a library of org-babel source blocks. Thus the
568 code for executing an action, and the code for dealing with the
569 output of the action will be the same code as for executing
570 source blocks in general
571 - Optionally, the user can have the relevant source block inserted
572 into the org buffer after the (say) #+BABELDO line. This will
573 allow the user to fine tune the action by modifying the code
574 (especially useful for plots).
575 - So maybe a #+BABELDO line will have header args
576 - :data (a reference to a table or source code block)
577 - :action (or should that be :srcname?) which will be something
578 like :action pie-chart, referring to a source block which will
579 be executed with the :data referent passed in using a :var arg.
580 - :showcode or something controlling whether to show the code
582 *** Modification to design
583 I'm implementing this, at least initially, as a new interpreter
584 named 'babel', which has an empty body. 'babel' blocks take
585 a :srcname header arg, and look for the source-code block with
586 that name. They then execute the referenced block, after first
587 appending their own header args on to the target block's header
590 If the target block is in the library of babel (a.o.t. e.g. the
591 current buffer), then the code in the block will refer to the
592 input data with a name dictated by convention (e.g. __data__
593 (something which is syntactically legal in all languages...). Thus
594 the babel block will use a :var __data__ = whatever header arg to
595 reference the data to be plotted.
598 This is covered by the [[file:library-of-babel.org][Library of Babel]], which will contain
599 ready-made source blocks designed to carry out useful common tasks.
600 ** PROPOSED Are we happy with current behaviour regarding vector/scalar output?
601 This simple example of multilingual chaining produces vector output if
602 there are spaces in the message and scalar otherwise.
604 #+begin_src R :var msg=msg-from-python
605 paste(msg, "und_R", sep="_")
608 #+srcname: msg-from-python
609 #+begin_src python :var msg=msg-from-elisp
613 #+srcname: msg-from-elisp
614 #+begin_src emacs-lisp :var msg="org-babel_speaks"
615 (concat msg "_elisp")
618 ** PROPOSED conversion between org-babel and noweb (e.g. .Rnw) format
619 I haven't thought about this properly. Just noting it down. What
620 Sweave uses is called "R noweb" (.Rnw).
622 I found a good description of noweb in the following article (see
623 the [[http://www.cs.tufts.edu/~nr/pubs/lpsimp.pdf][pdf]]).
625 I think there are two parts to noweb, the construction of
626 documentation and the extraction of source-code (with notangle).
628 *documentation*: org-mode handles all of our documentation needs in
629 a manner that I believe is superior to noweb.
631 *source extraction* At this point I don't see anyone writing large
632 applications with 100% of the source code contained in org-babel
633 files, rather I see org-babel files containing things like
634 - notes with active code chunks
635 - interactive tutorials
636 - requirements documents with code running test suites
637 - and of course experimental reports with the code to run the
638 experiment, and perform analysis
640 Basically I think the scope of the programs written in org-babel
641 (at least initially) will be small enough that it wont require the
642 addition of a tangle type program to extract all of the source code
643 into a running application.
645 On the other hand, since we already have named blocks of source
646 code which reference other blocks on which they rely, this
647 shouldn't be too hard to implement either on our own, or possibly
648 relying on something like noweb/notangle.
650 ** PROPOSED support for passing paths to files between source blocks
651 Maybe this should be it's own result type (in addition to scalars and
652 vectors). The reason being that some source-code blocks (for example
653 ditaa or anything that results in the creation of a file) may want to
654 pass a file path back to org-mode which could then be inserted into
655 the org-mode buffer as a link to the file...
657 This would allow for display of images upon export providing
658 functionality similar to =org-exp-blocks= only in a more general
661 ** DEFERRED use textConnection to pass tsv to R?
662 When passing args from the org buffer to R, the following route is
663 used: arg in buffer -> elisp -> tsv on file -> data frame in R. I
664 think it would be possible to avoid having to write to file by
665 constructing an R expression in org-babel-R-assign-elisp, something
668 #+begin_src emacs-lisp
669 (org-babel-R-input-command
670 (format "%s <- read.table(textConnection(\"%s\"), sep=\"\\t\", as.is=TRUE)"
671 name (orgtbl-to-tsv value '(:sep "\t" :fmt org-babel-R-quote-tsv-field))))
674 I haven't tried to implement this yet as it's basically just
675 fiddling with something that works. The only reason for it I can
676 think of would be efficiency and I haven't tested that.
678 This Didn't work after an initial test. I still think this is a
679 good idea (I also think we should try to do something similar when
680 writing out results frmo R to elisp) however as it wouldn't result
681 in any functional changes I'm bumping it down to deferred for
686 #+tblname: quick-test
689 #+srcname: quick-test-src-blk
690 #+begin_src R :var vec=quick-test
696 ** DEFERRED re-implement R evaluation using ess-command or ess-execute
697 I don't have any complaints with the current R evaluation code or
698 behaviour, but I think it would be good to use the ESS functions
699 from a political point of view. Plus of course it has the normal
700 benefits of an API (insulates us from any underlying changes etc). [DED]
702 I'll look into this. I believe that I looked at and rejected these
703 functions initially but now I can't remember why. I agree with
704 your overall point about using API's where available. I will take
705 a look back at these and either switch to using the ess commands,
706 or at least articulate under this TODO the reasons for using our
707 custom R-interaction commands. [Eric]
711 Lets just replace =org-babel-R-input-command= with =ess-execute=.
713 I tried this, and although it works in some situations, I find that
714 =ess-command= will often just hang indefinitely without returning
715 results. Also =ess-execute= will occasionally hang, and pops up
716 the buffer containing the results of the command's execution, which
717 is undesirable. For now these functions can not be used. Maybe
718 someone more familiar with the ESS code can recommend proper usage
719 of =ess-command= or some other lower-level function which could be
720 used in place of [[file:lisp/org-babel-R.el::defun%20org-babel%20R%20input%20command%20command][org-babel-R-input-command]].
724 #+begin_quote ess-command
725 (ess-command COM &optional BUF SLEEP NO-PROMPT-CHECK)
727 Send the ESS process command COM and delete the output
728 from the ESS process buffer. If an optional second argument BUF exists
729 save the output in that buffer. BUF is erased before use.
730 COM should have a terminating newline.
731 Guarantees that the value of .Last.value will be preserved.
732 When optional third arg SLEEP is non-nil, `(sleep-for (* a SLEEP))'
733 will be used in a few places where `a' is proportional to `ess-cmd-delay'.
736 #+begin_quote ess-execute
737 (ess-execute COMMAND &optional INVERT BUFF MESSAGE)
739 Send a command to the ESS process.
740 A newline is automatically added to COMMAND. Prefix arg (or second arg
741 INVERT) means invert the meaning of
742 `ess-execute-in-process-buffer'. If INVERT is 'buffer, output is
743 forced to go to the process buffer. If the output is going to a
744 buffer, name it *BUFF*. This buffer is erased before use. Optional
745 fourth arg MESSAGE is text to print at the top of the buffer (defaults
746 to the command if BUFF is not given.)
749 *** out current setup
751 1) The body of the R source code block is wrapped in a function
752 2) The function is called inside of a =write.table= function call
753 writing the results to a table
754 3) The table is read using =org-table-import=
756 ** DEFERRED Rework Interaction with Running Processes [2/5]
757 *** DONE robust to errors interrupting execution
759 #+srcname: long-runner-ruby
760 #+begin_src ruby :results silent
765 *** DEFERRED use =C-g= keyboard-quit to push processing into the background
766 This may be possible using the `run-with-timer' command.
768 I have no idea how this could work...
770 #+srcname: long-runner-ruby
771 #+begin_src ruby :results silent
776 *** TODO ability to select which of multiple sessions is being used
777 Increasingly it is looking like we're going to want to run all
778 source code blocks in comint buffer (sessions). Which will have
780 1) allowing background execution
781 2) maintaining state between source-blocks
782 - allowing inline blocks w/o header arguments
785 (like ess-switch-process in .R buffers)
787 Maybe this could be packaged into a header argument, something
788 like =:R_session= which could accept either the name of the
789 session to use, or the string =prompt=, in which case we could use
790 the =ess-switch-process= command to select a new process.
792 *** TODO evaluation of shell code as background process?
793 After C-c C-c on an R code block, the process may appear to
794 block, but C-g can be used to reclaim control of the .org buffer,
795 without interrupting the R evalution. However I believe this is not
796 true of bash/sh evaluation. [Haven't tried other languages] Perhaps
797 a solution is just to background the individual shell commands.
799 The other languages (aside from emacs lisp) are run through the
800 shell, so if we find a shell solution it should work for them as
803 Adding an ampersand seems to be a supported way to run commands in
804 the background (see [[http://www.emacswiki.org/emacs/ExecuteExternalCommand#toc4][external-commands]]). Although a more extensible
805 solution may involve the use of the [[elisp:(progn (describe-function 'call-process-region) nil)][call-process-region]] function.
807 Going to try this out in a new file [[file:lisp/org-babel-proc.el][org-babel-proc.el]]. This should
808 contain functions for asynchronously running generic shell commands
809 in the background, and then returning their input.
811 **** partial update of org-mode buffer
812 The sleekest solution to this may be using a comint buffer, and
813 then defining a filter function which would incrementally interpret
814 the results as they are returned, including insertion into the
815 org-mode buffer. This may actually cause more problems than it is
816 worth, what with the complexities of identifying the types of
817 incrementally returned results, and the need for maintenance of a
818 process marker in the org buffer.
820 **** 'working' spinner
821 It may be nice and not too difficult to place a spinner on/near the
822 evaluating source code block
824 *** TODO conversion of output from interactive shell, R (and python) sessions to org-babel buffers
825 [DED] This would be a nice feature I think. Although an org-babel
826 purist would say that it's working the wrong way round... After
827 some interactive work in a *R* buffer, you save the buffer, maybe
828 edit out some lines, and then convert it to org-babel format for
829 posterity. Same for a shell session either in a *shell* buffer, or
830 pasted from another terminal emulator. And python of course.
832 ** DEFERRED improve the source-block snippet
833 any real improvement seems somewhat beyond the ability of yasnippet
836 [[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]]
838 ,#name : Chapter title
841 ${1:$(make-string (string-width text) ?\=)}
846 [[file:snippets/org-mode/sb][sb -- snippet]]
848 waiting for guidance from those more familiar with yasnippets
849 ** DONE LoB: allow output in buffer
850 ** DONE allow default header arguments by language
851 org-babel-default-header-args:lang-name
853 An example of when this is useful is for languages which always return
854 files as their results (e.g. [[*** ditaa][ditaa]], and [[*** gnuplot][gnuplot]]).
855 ** DONE singe-function tangling and loading elisp from literate org-mode file [3/3]
857 This function should tangle the org-mode file for elisp, and then call
858 `load-file' on the resulting tangled file.
860 #+srcname: test-loading-embedded-emacs-lisp
861 #+begin_src emacs-lisp :results replace
862 (setq test-tangle-advert nil)
863 (setq test-tangle-loading nil)
864 (setq results (list :before test-tangle-loading test-tangle-advert))
865 (org-babel-load-file "test-tangle.org")
866 (setq results (list (list :after test-tangle-loading test-tangle-advert) results))
867 (delete-file "test-tangle.el")
871 #+resname: test-loading-embedded-emacs-lisp
872 | :before | nil | nil |
873 | :after | "org-babel tangles" | "use org-babel-tangle for all your emacs initialization files!!" |
875 *** DONE add optional language limiter to org-babel-tangle
876 This should check to see if there is any need to re-export
878 *** DONE ensure that org-babel-tangle returns the path to the tangled file(s)
880 #+srcname: test-return-value-of-org-babel-tangle
881 #+begin_src emacs-lisp :results replace
882 (mapcar #'file-name-nondirectory (org-babel-tangle-file "test-tangle.org" "emacs-lisp"))
888 *** DONE only tangle the file if it's actually necessary
889 ** DONE add a function to jump to a source-block by name
890 I've had an initial stab at that in org-babel-find-named-block
891 (library-of-babel branch).
893 At the same time I introduced org-babel-named-src-block-regexp, to
894 match src-blocks with srcname.
896 This is now working with the command
897 `org-babel-goto-named-source-block', all we need is a good key
900 ** DONE add =:none= session argument (for purely functional execution) [4/4]
901 This would allow source blocks to be run in their own new process
903 - These blocks could then also be run in the background (since we can
904 detach and just wait for the process to signal that it has terminated)
905 - We wouldn't be drowning in session buffers after running the tests
906 - we can re-use much of the session code to run in a more /functional/
909 While session provide a lot of cool features, like persistent
910 environments, [[* DONE function to bring up inferior-process buffer][pop-to-session]], and hints at exportation for
911 org-babel-tangle, they also have some down sides and I'm thinking that
912 session-based execution maybe shouldn't be the default behavior.
914 Down-sides to sessions
915 - *much* more complicated than functional evaluation
916 - maintaining the state of the session has weird issues
917 - waiting for evaluation to finish
918 - prompt issues like [[* TODO weird escaped characters in shell prompt break shell evaluation][shell-prompt-escapes-bug]]
919 - can't run in background
920 - litter emacs with session buffers
924 #+srcname: ruby-task-no-session
925 #+begin_src ruby :results replace output
931 #+resname: ruby-task-no-session
936 #+srcname: task-python-none-session
937 #+begin_src python :session none :results replace value
943 #+resname: task-python-none-session
948 #+srcname: task-session-none-sh
949 #+begin_src sh :results replace
954 #+resname: task-session-none-sh
960 #+srcname: task-no-session-R
961 #+begin_src R :results replace output
968 #+resname: task-no-session-R
972 ** DONE fully purge org-babel-R of direct comint interaction
973 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
975 ** DONE Create objects in top level (global) environment [5/5]
978 *** initial requirement statement [DED]
979 At the moment, objects created by computations performed in the
980 code block are evaluated in the scope of the
981 code-block-function-body and therefore disappear when the code
982 block is evaluated {unless you employ some extra trickery like
983 assign('name', object, env=globalenv()) }. I think it will be
984 desirable to also allow for a style wherein objects that are
985 created in one code block persist in the R global environment and
986 can be re-used in a separate block.
988 This is what Sweave does, and while I'm not saying we have to be
989 the same as Sweave, it wouldn't be hard for us to provide the same
990 behaviour in this case; if we don't, we risk undeservedly being
991 written off as an oddity by some.
993 IOW one aspect of org-babel is that of a sort of functional
994 meta-programming language. This is crazy, in a very good
995 way. Nevertheless, wrt R I think there's going to be a lot of value
996 in providing for a working style in which the objects are stored in
997 the R session, rather than elisp/org buffer. This will be a very
998 familiar working style to lots of people.
1000 There are no doubt a number of different ways of accomplishing
1001 this, the simplest being a hack like adding
1004 for(objname in ls())
1005 assign(objname, get(objname), envir=globalenv())
1008 to the source code block function body. (Maybe wrap it in an on.exit() call).
1010 However this may deserve to be thought about more carefully, perhaps
1011 with a view to having a uniform approach across languages. E.g. shell
1012 code blocks have the same semantics at the moment (no persistence of
1013 variables across code blocks), because the body is evaluated in a new
1014 bash shell process rather than a running shell. And I guess the same
1015 is true for python. However, in both these cases, you could imagine
1016 implementing the alternative in which the body is evaluated in a
1017 persistent interactive session. It's just that it's particularly
1018 natural for R, seeing as both ESS and org-babel evaluate commands in a
1019 single persistent R session.
1023 Thanks for bringing this up. I think you are absolutely correct that we
1024 should provide support for a persistent environment (maybe called a
1025 *session*) in which to evaluate code blocks. I think the current setup
1026 demonstrates my personal bias for a functional style of programming
1027 which is certainly not ideal in all contexts.
1029 While the R function you mention does look like an elegant solution, I
1030 think we should choose an implementation that would be the same across
1031 all source code types. Specifically I think we should allow the user to
1032 specify an optional *session* as a header variable (when not present we
1033 assume a default session for each language). The session name could be
1034 used to name a comint buffer (like the *R* buffer) in which all
1035 evaluation would take place (within which variables would retain their
1036 values --at least once I remove some of the functional method wrappings
1037 currently in place-- ).
1039 This would allow multiple environments to be used in the same buffer,
1040 and once this setup was implemented we should be able to fairly easily
1041 implement commands for jumping between source code blocks and the
1042 related session buffers, as well as for dumping the last N commands from
1043 a session into a new or existing source code block.
1045 Please let me know if you foresee any problems with this proposed setup,
1046 or if you think any parts might be confusing for people coming from
1047 Sweave. I'll hopefully find some time to work on this later in the
1050 *** can functional and interpreted/interactive models coexist?
1052 Even though both of these use the same =*R*= buffer the value of =a=
1053 is not preserved because it is assigned inside of a functional
1056 #+srcname: task-R-sessions
1063 #+srcname: task-R-same-session
1068 This functional wrapper was implemented in order to efficiently return
1069 the results of the execution of the entire source code block. However
1070 it inhibits the evaluation of source code blocks in the top level,
1071 which would allow for persistence of variable assignment across
1072 evaluations. How can we allow *both* evaluation in the top level, and
1073 efficient capture of the return value of an entire source code block
1074 in a language independent manner?
1076 Possible solutions...
1077 1) we can't so we will have to implement two types of evaluation
1078 depending on which is appropriate (functional or imperative)
1079 2) we remove the functional wrapper and parse the source code block
1080 into it's top level statements (most often but not always on line
1081 breaks) so that we can isolate the final segment which is our
1083 3) we add some sort of "#+return" line to the code block
1084 4) we take advantage of each languages support for meta-programming
1085 through =eval= type functions, and use said to evaluate the entire
1086 blocks in such a way that their environment can be combined with the
1087 global environment, and their results are still captured.
1088 5) I believe that most modern languages which support interactive
1089 sessions have support for a =last_result= type function, which
1090 returns the result of the last input without re-calculation. If
1091 widely enough present this would be the ideal solution to a
1092 combination of functional and imperative styles.
1094 None of these solutions seem very desirable, but for now I don't see
1095 what else would be possible.
1097 Of these options I was leaning towards (1) and (4) but now believe
1098 that if it is possible option (5) will be ideal.
1100 **** (1) both functional and imperative evaluation
1102 - can take advantage of built in functions for sending regions to the
1104 - retains the proven tested and working functional wrappers
1107 - introduces the complication of keeping track of which type of
1108 evaluation is best suited to a particular context
1109 - the current functional wrappers may require some changes in order to
1110 include the existing global context
1112 **** (4) exploit language meta-programming constructs to explicitly evaluate code
1114 - only one type of evaluation
1117 - some languages may not have sufficient meta-programming constructs
1119 **** (5) exploit some =last_value= functionality if present
1121 Need to ensure that most languages have such a function, those without
1122 will simply have to implement their own similar solution...
1124 | language | =last_value= function |
1125 |------------+-----------------------------|
1129 | shell | see [[* last command for shells][last command for shells]] |
1130 | emacs-lisp | see [[* emacs-lisp will be a special case][special-case]] |
1132 #+srcname: task-last-value
1137 ***** last command for shells
1138 Do this using the =tee= shell command, and continually pipe the output
1141 Got this idea from the following [[http://linux.derkeiler.com/Mailing-Lists/Fedora/2004-01/0898.html][email-thread]].
1143 suggested from mailing list
1145 #+srcname: bash-save-last-output-to-file
1149 bash -c "$line" | tee /tmp/last.out1
1150 mv /tmp/last.out1 /tmp/last.out
1154 another proposed solution from the above thread
1156 #+srcname: bash-save-in-variable
1159 # so - Save Output. Saves output of command in OUT shell variable.
1169 "^M": " | tee /tmp/h_lastcmd.out ^[k"
1172 export __=/tmp/h_lastcmd.out
1174 If you try it, Alt-k will stand for the old Enter; use "command $__" to
1175 access the last output.
1181 Herculano de Lima Einloft Neto
1184 ***** emacs-lisp will be a special case
1185 While it is possible for emacs-lisp to be run in a console type
1186 environment (see the =elim= function) it is *not* possible to run
1187 emacs-lisp in a different *session*. Meaning any variable set top
1188 level of the console environment will be set *everywhere* inside
1189 emacs. For this reason I think that it doesn't make any sense to
1190 worry about session support for emacs-lisp.
1192 *** Further thoughts on 'scripting' vs. functional approaches
1194 These are just thoughts, I don't know how sure I am about this.
1195 And again, perhaps I'm not saying anything very radical, just that
1196 it would be nice to have some options supporting things like
1197 receiving text output in the org buffer.
1199 I can see that you've already gone some way down the road towards
1200 the 'last value' approach, so sorry if my comments come rather
1201 late. I am concerned that we are not giving sufficient attention
1202 to stdout / the text that is returned by the interpreters. In
1203 contrast, many of our potential users will be accustomed to a
1204 'scripting' approach, where they are outputting text at various
1205 points in the code block, not just at the end. I am leaning
1206 towards thinking that we should have 2 modes of evaluation:
1207 'script' mode, and 'functional' mode.
1209 In script mode, evaluation of a code block would result in *all*
1210 text output from that code block appearing as output in the org
1211 buffer, presumably as an #+begin_example...#+end_example. There
1212 could be an :echo option controlling whether the input commands
1213 also appear in the output. [This is like Sweave].
1215 In functional mode, the *result* of the code block is available as
1216 an elisp object, and may appear in the org buffer as an org
1217 table/string, via the mechanisms you have developed already.
1219 One thing I'm wondering about is whether, in script mode, there
1220 simply should not be a return value. Perhaps this is not so
1221 different from what exists: script mode would be new, and what
1222 exists currently would be functional mode.
1224 I think it's likely that, while code evaluation will be exciting
1225 to people, a large majority of our users in a large majority of
1226 their usage will not attempt to actually use the return value from
1227 a source code block in any meaningful way. In that case, it seems
1228 rather restrictive to only allow them to see output from the end
1231 Instead I think the most accessible way to introduce org-babel to
1232 people, at least while they are learning it, is as an immensely
1233 powerful environment in which to embed their 'scripts', which now
1234 also allows them to 'run' their 'scripts'. Especially as such
1235 people are likely to be the least capable of the user-base, a
1236 possible design-rule would be to make the scripting style of usage
1237 easy (default?), perhaps requiring a special option to enable a
1238 functional style. Those who will use the functional style won't
1239 have a problem understanding what's going on, whereas the 'skript
1240 kiddies' might not even know the syntax for defining a function in
1241 their language of choice. And of course we can allow the user to
1242 set a variable in their .emacs controlling the preference, so that
1243 functional users are not inconveniennced by having to provide
1244 header args the whole time.
1246 Please don't get the impression that I am down-valuing the
1247 functional style of org-babel. I am constantly horrified at the
1248 messy 'scripts' that my colleagues produce in perl or R or
1249 whatever! Nevertheless that seems to be how a lot of people work.
1251 I think you were leaning towards the last-value approach because
1252 it offered the possibility of unified code supporting both the
1253 single evaluation environment and the functional style. If you
1254 agree with any of the above then perhaps it will impact upon this
1255 and mean that the code in the two branches has to differ a bit. In
1256 that case, functional mode could perhaps after all evaluate each
1257 code block in its own environment, thus (re)approaching 'true'
1258 functional programming (side-effects are hard to achieve).
1262 echo "There are `wc -l files` files in this directory"
1266 *** even more thoughts on evaluation, results, models and options
1268 Thanks Dan, These comments are invaluable.
1270 What do you think about this as a new list of priorities/requirements
1271 for the execution of source-code blocks.
1274 1) we want the evaluation of the source code block to take place in a
1275 session which can persist state (variables, current directory,
1277 2) source code blocks can specify their session with a header argument
1278 3) each session should correspond to an Emacs comint buffer so that the
1279 user can drop into the session and experiment with live code
1282 1) each source-code block generates some form of results which (as
1283 we have already implemented) is transfered into emacs-lisp
1284 after which it can be inserted into the org-mode buffer, or
1285 used by other source-code blocks
1286 2) when the results are translated into emacs-lisp, forced to be
1287 interpreted as a scalar (dumping their raw values into the
1288 org-mode buffer), as a vector (which is often desirable with R
1289 code blocks), or interpreted on the fly (the default option).
1290 Note that this is very nearly currently implemented through the
1291 [[* DONE results-type header (vector/file)][results-type-header]].
1292 3) there should be *two* means of collecting results from the
1293 execution of a source code block. *Either* the value of the
1294 last statement of the source code block, or the collection of
1295 all that has been passed to STDOUT during the evaluation.
1297 **** header argument or return line (*header argument*)
1299 Rather than using a header argument to specify how the return value
1300 should be passed back, I'm leaning towards the use of a =#+RETURN=
1301 line inside the block. If such a line *is not present* then we
1302 default to using STDOUT to collect results, but if such a line *is
1303 present* then we use it's value as the results of the block. I
1304 think this will allow for the most elegant specification between
1305 functional and script execution. This also cleans up some issues
1306 of implementation and finding which statement is the last
1309 Having given this more thought, I think a header argument is
1310 preferable. The =#+return:= line adds new complicating syntax for
1311 something that does little more than we would accomplish through
1312 the addition of a header argument. The only benefit being that we
1313 know where the final statement starts, which is not an issue in
1314 those languages which contain 'last value' operators.
1316 new header =:results= arguments
1317 - script :: explicitly states that we want to use STDOUT to
1318 initialize our results
1319 - return_last :: stdout is ignored instead the *value* of the final
1320 statement in the block is returned
1321 - echo :: means echo the contents of the source-code block along
1322 with the results (this implies the *script* =:results=
1325 *** DONE rework evaluation lang-by-lang [4/4]
1327 This should include...
1328 - functional results working with the comint buffer
1330 - script :: return the output of STDOUT
1331 - write a macro which runs the first redirection, executes the
1332 body, then runs the second redirection
1333 - last :: return the value of the last statement
1336 - sessions in comint buffers
1338 **** DONE Ruby [4/4]
1339 - [X] functional results working with comint
1340 - [X] script results
1341 - [X] ensure scalar/vector results args are taken into consideration
1342 - [X] ensure callable by other source block
1344 #+srcname: ruby-use-last-output
1345 #+begin_src ruby :results replace
1352 #+resname: ruby-use-last-output
1355 #+srcname: task-call-use-last-output
1356 #+begin_src ruby :var last=ruby-use-last-output :results replace
1357 last.flatten.size + 1
1360 #+resname: task-call-use-last-output
1365 #+srcname: first-ruby-session-task
1366 #+begin_src ruby :session schulte :results silent
1370 #+srcname: second-ruby-session-task
1371 #+begin_src ruby :session schulte :results silent
1375 #+srcname: without-the-right-session
1376 #+begin_src ruby :results silent
1382 - [X] functional results working with comint
1383 - [X] script results
1384 - [X] ensure scalar/vector results args are taken into consideration
1385 - [X] ensure callable by other source block
1387 To redirect output to a file, you can use the =sink()= command.
1390 #+begin_src R :results value vector silent
1397 #+srcname: task-R-use-other-output
1398 #+begin_src R :var twoentyseven=task_R_B() :results replace value
1403 #+resname: task-R-use-other-output
1406 **** DONE Python [4/4]
1407 - [X] functional results working with comint
1408 - [X] script results
1409 - [X] ensure scalar/vector results args are taken into consideration
1410 - [X] ensure callable by other source block
1412 #+srcname: task-new-eval-for-python
1413 #+begin_src python :results silent output scalar
1419 #+srcname: task-use-new-eval
1420 #+begin_src python :var tasking=task-new-eval-for-python() :results replace
1424 #+resname: task-use-new-eval
1427 **** DONE Shells [4/4]
1428 - [X] functional results working with comint
1429 - [X] script results
1430 - [X] ensure scalar/vector results args are taken into consideration
1431 - [X] ensure callable by other source block
1433 #+srcname: task-shell-new-evaluation
1434 #+begin_src sh :results silent value scalar
1439 #+srcname: task-call-other-shell
1440 #+begin_src sh :var other=task-shell-new-evaluation() :results replace scalar
1441 echo $other ' is the old date'
1444 #+resname: task-call-other-shell
1445 : $ Fri Jun 12 13:08:37 PDT 2009 is the old date
1447 *** DONE implement a *session* header argument [4/4]
1448 =:session= header argument to override the default *session* buffer
1452 #+srcname: task-ruby-named-session
1453 #+begin_src ruby :session schulte :results replace
1454 schulte = :in_schulte
1457 #+resname: task-ruby-named-session
1460 #+srcname: another-in-schulte
1461 #+begin_src ruby :session schulte
1465 #+resname: another-in-schulte
1472 #+srcname: python-session-task
1473 #+begin_src python :session what :results silent
1477 #+srcname: python-get-from-session
1478 #+begin_src python :session what :results replace
1482 #+resname: python-get-from-session
1487 #+srcname: task-shell-sessions
1488 #+begin_src sh :session what
1492 #+srcname: task-shell-sessions-what
1493 #+begin_src sh :session what :results replace
1497 #+resname: task-shell-sessions-what
1502 #+srcname: task-R-session
1503 #+begin_src R :session what :results replace
1509 #+resname: task-R-session
1512 #+srcname: another-task-R-session
1513 #+begin_src R :session what :results replace
1517 *** DONE function to bring up inferior-process buffer [4/4]
1519 This should be callable from inside of a source-code block in an
1520 org-mode buffer. It should evaluate the header arguments, then bring
1521 up the inf-proc buffer using =pop-to-buffer=.
1523 For lack of a better place, lets add this to the `org-metadown-hook'
1526 To give this a try, place the cursor on a source block with variables,
1527 (optionally git a prefix argument) then hold meta and press down.
1531 #+srcname: task-ruby-pop-to-session
1532 #+begin_src ruby :var num=9 :var another="something else"
1533 num.times{|n| puts another}
1538 #+srcname: task-python-pop-to-session
1539 #+begin_src python :var num=9 :var another="something else"
1544 #+srcname: task-R-pop-to-session
1545 #+begin_src R :var a=9 :var b=8
1551 #+srcname: task-shell-pop-sessions
1552 #+begin_src sh :var NAME="eric"
1556 *** DEFERRED function to dump last N lines from inf-proc buffer into the current source block
1558 Callable with a prefix argument to specify how many lines should be
1559 dumped into the source-code buffer.
1561 *** REJECTED comint notes
1563 Implementing comint integration in [[file:lisp/org-babel-comint.el][org-babel-comint.el]].
1566 - handling of outputs
1567 - split raw output from process by prompts
1568 - a ring of the outputs, buffer-local, `org-babel-comint-output-ring'
1569 - a switch for dumping all outputs to a buffer
1570 - inputting commands
1572 Lets drop all this language specific stuff, and just use
1573 org-babel-comint to split up our outputs, and return either the last
1574 value of an execution or the combination of values from the
1577 **** comint filter functions
1578 : ;; comint-input-filter-functions hook process-in-a-buffer
1579 : ;; comint-output-filter-functions hook function modes.
1580 : ;; comint-preoutput-filter-functions hook
1581 : ;; comint-input-filter function ...
1583 #+srcname: obc-filter-ruby
1584 #+begin_src ruby :results last
1592 ** DONE Remove protective commas from # comments before evaluating
1593 org inserts protective commas in front of ## comments in language
1594 modes that use them. We need to remove them prior to sending code
1597 #+srcname: testing-removal-of-protective-comas
1599 ,# this one might break it??
1603 ** DONE pass multiple reference arguments into R
1604 Can we do this? I wasn't sure how to supply multiple 'var' header
1605 args. Just delete this if I'm being dense.
1607 This should be working, see the following example...
1609 #+srcname: two-arg-example
1610 #+begin_src R :var n=2 :var m=8
1614 #+resname: two-arg-example
1617 ** DONE ensure that table ranges work
1618 when a table range is passed to org-babel as an argument, it should be
1619 interpreted as a vector.
1622 | 2 | 3 | Fixnum:1 |
1623 | 3 | 4 | Array:123456 |
1627 #+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)))
1629 #+srcname: simple-sbe-example
1630 #+begin_src emacs-lisp
1634 #+srcname: task-table-range
1635 #+begin_src ruby :var n=simple-sbe-example
1639 #+srcname: simple-results
1640 #+begin_src emacs-lisp :var n=task-table-range(n=(1 2 3))
1644 #+resname: simple-results
1647 #+srcname: task-arr-referent
1648 #+begin_src ruby :var ar=(1 2 3)
1652 #+resname: task-arr-referent
1655 ** DONE global variable indicating default to vector output
1656 how about an alist... =org-babel-default-header-args= this may already
1657 exist... just execute the following and all source blocks will default
1660 #+begin_src emacs-lisp
1661 (setq org-babel-default-header-args '((:results . "vector")))
1664 ** DONE name named results if source block is named
1665 currently this isn't happening although it should be
1667 #+srcname: test-naming-named-source-blocks
1668 #+begin_src emacs-lisp
1672 #+resname: test-naming-named-source-blocks
1674 ** DONE (simple caching) check for named results before source blocks
1675 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]]
1676 ** DONE set =:results silent= when eval with prefix argument
1678 #+begin_src emacs-lisp
1681 ** DONE results-type header (vector/file) [3/3]
1682 In response to a point in Dan's email. We should allow the user to
1683 force scalar or vector results. This could be done with a header
1684 argument, and the default behavior could be controlled through a
1685 configuration variable.
1687 #+srcname: task-trivial-vector
1688 #+begin_src ruby :results replace vector
1695 since it doesn't make sense to turn a vector into a scalar, lets
1696 just add a two values...
1698 - vector :: forces the results to be a vector (potentially 1 dimensional)
1699 - file :: this throws an error if the result isn't a string, and
1700 tries to treat it as a path to a file.
1702 I'm just going to cram all of these into the =:results= header
1703 argument. Then if we allow multiple header arguments it should
1704 work out, for example one possible header argument string could be
1705 =:results replace vector file=, which would *replace* any existing
1706 results forcing the results into an org-mode table, and
1707 interpreting any strings as file paths.
1709 *** DONE multiple =:results= headers
1711 #+srcname: multiple-result-headers
1712 #+begin_src ruby :results replace silent
1718 *** DONE file result types
1719 When inserting into an org-mode buffer create a link with the path
1720 being the value, and optionally the display being the
1721 =file-name-nondirectory= if it exists.
1723 #+srcname: task-file-result
1724 #+begin_src python :results replace file
1729 [[something][something]]
1732 This will be useful because blocks like =ditaa= and =dot= can return
1733 the string path of their files, and can add =file= to their results
1736 *** DONE vector result types
1738 #+srcname: task-force-results
1739 #+begin_src emacs-lisp :results vector
1746 ** DONE results name
1747 In order to do this we will need to start naming our results.
1748 Since the source blocks are named with =#+srcname:= lines we can
1749 name results with =#+resname:= lines (if the source block has no
1750 name then no name is given to the =#+resname:= line on creation,
1751 otherwise the name of the source block is used).
1753 This will have the additional benefit of allowing results and
1754 source blocks to be located in different places in a buffer (and
1755 eventually in different buffers entirely).
1757 #+srcname: developing-resnames
1758 #+begin_src emacs-lisp :results silent
1762 Once source blocks are able to find their own =#+resname:= lines
1765 #+srcname: sbe-w-new-results
1766 #+begin_src emacs-lisp :results replace
1767 (sbe "developing-resnames")
1773 *** TODO change the results insertion functions to use these lines
1775 *** TODO teach references to resolve =#+resname= lines.
1777 ** DONE org-babel tests org-babel [1/1]
1778 since we are accumulating this nice collection of source-code blocks
1779 in the sandbox section we should make use of them as unit tests.
1780 What's more, we should be able to actually use org-babel to run these
1783 We would just need to cycle over every source code block under the
1784 sandbox, run it, and assert that the return value is equal to what we
1787 I have the feeling that this should be possible using only org-babel
1788 functions with minimal or no additional elisp. It would be very cool
1789 for org-babel to be able to test itself.
1791 This is now done, see [[* Tests]].
1793 *** DEFERRED org-babel assertions (may not be necessary)
1794 These could be used to make assertions about the results of a
1795 source-code block. If the assertion fails then the point could be
1796 moved to the block, and error messages and highlighting etc... could
1799 ** DONE make C-c C-c work anywhere within source code block?
1800 This seems like it would be nice to me, but perhaps it would be
1801 inefficient or ugly in implementation? I suppose you could search
1802 forward, and if you find #+end_src before you find #+begin_src,
1803 then you're inside one. [DED]
1805 Agreed, I think inside of the =#+srcname: line= would be useful as
1808 #+srcname: testing-out-cc
1809 #+begin_src emacs-lisp
1813 ** DONE integration with org tables
1814 We should make it easy to call org-babel source blocks from org-mode
1815 table formulas. This is practical now that it is possible to pass
1816 arguments to org-babel source blocks.
1818 See the related [[* (sandbox) integration w/org tables][sandbox]] header for tests/examples.
1820 *** digging in org-table.el
1821 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.
1823 Should be a hook in [[file:~/src/org/lisp/org-table.el::defun%20org%20table%20eval%20formula%20optional%20arg%20equation][org-table-eval-formula]].
1825 Looks like I need to change this [[file:~/src/org/lisp/org-table.el::if%20lispp][if statement]] (line 2239) into a cond
1828 ** DONE source blocks as functions
1830 Allow source code blocks to be called like functions, with arguments
1831 specified. We are already able to call a source-code block and assign
1832 it's return result to a variable. This would just add the ability to
1833 specify the values of the arguments to the source code block assuming
1834 any exist. For an example see
1836 When a variable appears in a header argument, how do we differentiate
1837 between it's value being a reference or a literal value? I guess this
1838 could work just like a programming language. If it's escaped or in
1839 quotes, then we count it as a literal, otherwise we try to look it up
1842 ** DONE folding of code blocks? [2/2]
1843 [DED] In similar way to using outline-minor-mode for folding function
1844 bodies, can we fold code blocks? #+begin whatever statements are
1845 pretty ugly, and in any case when you're thinking about the overall
1846 game plan you don't necessarily want to see the code for each Step.
1848 *** DONE folding of source code block
1849 Sounds good, and wasn't too hard to implement. Code blocks should
1850 now be fold-able in the same manner as headlines (by pressing TAB
1853 *** REJECTED folding of results
1854 So, lets do a three-stage tab cycle... First fold the src block,
1855 then fold the results, then unfold.
1857 There's no way to tell if the results are a table or not w/o
1858 actually executing the block which would be too expensive of an
1861 ** DONE selective export of text, code, figures
1862 [DED] The org-babel buffer contains everything (code, headings and
1863 notes/prose describing what you're up to, textual/numeric/graphical
1864 code output, etc). However on export to html / LaTeX one might want
1865 to include only a subset of that content. For example you might
1866 want to create a presentation of what you've done which omits the
1869 [EMS] So I think this should be implemented as a property which can
1870 be set globally or on the outline header level (I need to review
1871 the mechanics of org-mode properties). And then as a source block
1872 header argument which will apply only to a specific source code
1873 block. A header argument of =:export= with values of
1875 - =code= :: just show the code in the source code block
1876 - =none= :: don't show the code or the results of the evaluation
1877 - =results= :: just show the results of the code evaluation (don't
1878 show the actual code)
1879 - =both= :: show both the source code, and the results
1881 this will be done in [[* (sandbox) selective export][(sandbox) selective export]].
1883 ** DONE a header argument specifying silent evaluation (no output)
1884 This would be useful across all types of source block. Currently
1885 there is a =:replace t= option to control output, this could be
1886 generalized to an =:output= option which could take the following
1887 options (maybe more)
1889 - =t= :: this would be the default, and would simply insert the
1890 results after the source block
1891 - =replace= :: to replace any results which may already be there
1892 - =silent= :: this would inhibit any insertion of the results
1894 This is now implemented see the example in the [[* silent evaluation][sandbox]]
1896 ** DONE assign variables from tables in R
1897 This is now working (see [[* (sandbox table) R][(sandbox-table)-R]]). Although it's not that
1898 impressive until we are able to print table results from R.
1900 ** DONE insert 2-D R results as tables
1901 everything is working but R and shell
1907 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
1908 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]]
1909 as a means of converting multidimensional R objects to emacs lisp.
1911 It may be as simple as first checking if the data is multidimensional,
1912 and then, if so using =write= to write the data out to a temporary
1913 file from which emacs can read the data in using =org-table-import=.
1915 Looking into this further, is seems that there is no such thing as a
1916 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
1917 deal with trivial vectors (scalars) in R. I'm tempted to just treat
1918 them as vectors, but then that would lead to a proliferation of
1919 trivial 1-cell tables...
1921 ** DONE allow variable initialization from source blocks
1922 Currently it is possible to initialize a variable from an org-mode
1923 table with a block argument like =table=sandbox= (note that the
1924 variable doesn't have to named =table=) as in the following example
1930 #+begin_src emacs-lisp :var table=sandbox :results replace
1931 (message (format "table = %S" table))
1934 : "table = ((1 2 3) (4 \"schulte\" 6))"
1936 It would be good to allow initialization of variables from the results
1937 of other source blocks in the same manner. This would probably
1938 require the addition of =#+SRCNAME: example= lines for the naming of
1939 source blocks, also the =table=sandbox= syntax may have to be expanded
1940 to specify whether the target is a source code block or a table
1941 (alternately we could just match the first one with the given name
1942 whether it's a table or a source code block).
1944 At least initially I'll try to implement this so that there is no need
1945 to specify whether the reference is to a table or a source-code block.
1946 That seems to be simpler both in terms of use and implementation.
1948 This is now working for emacs-lisp, ruby and python (and mixtures of
1949 the three) source blocks. See the examples in the [[* (sandbox) referencing other source blocks][sandbox]].
1951 This is currently working only with emacs lisp as in the following
1952 example in the [[* emacs lisp source reference][emacs lisp source reference]].
1955 ** TODO Add languages [1/6]
1956 I'm sure there are many more that aren't listed here. Please add
1957 them, and bubble any that you particularly care about up to the top.
1959 Any new language should be implemented in a org-babel-lang.el file.
1960 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
1961 [[file:lisp/org-babel-R.el][org-babel-R.el]].
1964 This could probably be added to [[file:lisp/org-babel-script.el][org-babel-script.el]]
1969 (see [[* file result types][file result types]])
1971 #+srcname: implementing-ditaa
1972 #+begin_src ditaa :results replace :file blue.png :cmdline -r
1982 #+resname: implementing-ditaa
1983 [[file:blue.png][blue.png]]
1986 (see [[* file result types][file result types]])
1988 - a required =file= header argument
1989 - a =cmdline= header argument
1991 - scalar variables should be replaced in the body of the gnuplot code
1992 - vector variables should be exported to tab-separated files, and
1993 the variable names should be replaced with the path to the files
1996 (see [[* file result types][file result types]])
1999 (see [[* file result types][file result types]])
2003 ** TODO Allow source blocks to be recognised when #+ are not first characters on the line
2004 I think Carsten has recently altered the core so that #+ can have
2005 preceding whitespace, at least for literal/code examples. org-babel
2006 should support this.
2008 ** TODO non-orgtbl formatted lists
2011 #+srcname: this-doesn't-match-orgtbl
2012 #+begin_src emacs-lisp :results replace
2013 '((:results . "replace"))
2016 #+resname: this-doesn't-match-orgtbl
2018 ** TODO collapsing consecutive newlines in string output
2020 #+srcname: multi-line-string-output
2021 #+begin_src ruby :results output
2022 "the first line ends here
2025 and this is the second one
2030 #+resname: multi-line-string-output
2032 ** TODO cursor movement when evaluating source blocks
2033 E.g. the pie chart example. Despite the save-window-excursion in
2034 org-babel-execute:R. (I never learned how to do this properly: org-R
2035 jumps all over the place...)
2037 ** PROPOSED external shell execution can't isolate return values
2038 I have no idea how to do this as of yet. The result is that when
2039 shell functions are run w/o a session there is no difference between
2040 the =output= and =value= result arguments.
2042 Yea, I don't know how to do this either. I searched extensively on
2043 how to isolate the *last* output of a series of shell commands (see
2044 [[* last command for
2045 shells][last command for shells]]). The results of the search were basically
2046 that it was not possible (or at least not accomplish-able with a
2047 reasonable amount of effort).
2049 That fact combined with the tenancy to all ways use standard out in
2050 shell scripts led me to treat these two options (=output= and =value=)
2051 as identical in shell evaluation. Not ideal but maybe good enough for
2054 In the `results' branch I've changed this so that they're not quite
2055 identical: output results in raw stdout contents, whereas value
2056 converts it to elisp, perhaps to a table if it looks tabular. This is
2057 the same for the other languages. [Dan]
2059 ** TODO are the org-babel-trim s necessary?
2060 at the end of e.g. org-babel-R-evaluate, org-babel-python-evaluate, but
2061 not org-babel-ruby-evaluate
2062 ** results branch bugs
2063 *** TODO elisp reference fails for literal number
2064 #+srcname: elisp-test(a=4)
2065 #+begin_src emacs-lisp
2068 *** TODO use new merge function [[file:lisp/org-babel-ref.el::t%20nil%20org%20combine%20plists%20args%20nil][here]] and [[file:lisp/org-babel.el::params%20org%20combine%20plists%20params%20third%20info][here]]?
2069 And at other occurrences of org-combine-plists?
2070 *** TODO LoB: with output to buffer, not working in buffers other than library-of-babel.org
2071 I haven't fixed this yet. org-babel-ref-resolve-reference moves
2072 point around, inside a save-excursion. Somehow when it comes to
2073 inserting the results (after possible further recursive calls to
2074 org-babel-ref-resolve-reference), point hasn't gone back to the
2076 *** TODO LoB: output to buffer adds creeping blank lines
2077 Compare the results of
2078 #+lob: python-add(a=5, b=17)
2080 #+resname: python-add(a=5, b=17)
2082 --------------------------------
2090 ---------------------
2092 Hmm, it's a bit confusing. I think it's to do with the fact that
2093 LoB removes the entire (#+resname and result) and starts from
2094 scratch, whereas #+begin_src only removes the result. I haven't
2095 worked out what the correct fix is yet.
2096 ** DEFERRED weird escaped characters in shell prompt break shell evaluation
2097 E.g. this doesn't work. Should the shell sessions set a sane prompt
2098 when they start up? Or is it a question of altering
2099 comint-prompt-regexp? Or altering org-babel regexps?
2102 black=30 ; red=31 ; green=32 ; yellow=33 ; blue=34 ; magenta=35 ; cyan=36 ; white=37
2105 export PS1="\[\033[${prompt_col}m\]\w${prompt_char} \[\033[0m\]"
2108 I just pushed a good amount of changes, could you see if your shell
2109 problems still exist?
2111 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
2112 org-babel-sh.el, raw gets the value
2114 ("" "
\e[0m Sun Jun 14 19:26:24 EDT 2009\n" "
\e[0m org_babel_sh_eoe\n" "
\e[0m ")
2116 and therefore (member org-babel-sh-eoe-output ...) fails
2118 I think that `comint-prompt-regexp' needs to be altered to match
2119 the shell prompt. This shouldn't be too difficult to do by hand,
2120 using the `regexp-builder' command and should probably be part of
2121 the user's regular emacs init. I can't think of a way for us to
2122 set this automatically, and we are SOL without a regexp to match
2125 ** DONE LoB: calls fail if reference has single character name
2126 commit 21d058869df1ff23f4f8cc26f63045ac9c0190e2
2127 **** This doesn't work
2128 #+lob: R-plot(data=X)
2147 #+lob: R-plot(data=XX)
2149 ** DONE make :results replace the default?
2150 I'm tending to think that appending results to pre-existing results
2151 creates mess, and that the cleaner `replace' option should be the
2152 default. E.g. when a source block creates an image, we would want
2153 that to be updated, rather than have a new one be added.
2157 ** DONE ruby evaluation not working under ubuntu emacs 23
2158 With emacs 23.0.91.1 on ubuntu, for C-h f run-ruby I have the
2159 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.
2162 run-ruby is an interactive compiled Lisp function.
2166 Run an inferior Ruby process, input and output via buffer *ruby*.
2167 If there is a process already running in `*ruby*', switch to that buffer.
2168 With argument, allows you to edit the command line (default is value
2169 of `ruby-program-name'). Runs the hooks `inferior-ruby-mode-hook'
2170 (after the `comint-mode-hook' is run).
2171 (Type C-h m in the process buffer for a list of commands.)
2174 So, I may have a non-standard inf-ruby.el. Here's my version of
2178 run-ruby is an interactive Lisp function in `inf-ruby.el'.
2180 (run-ruby &optional COMMAND NAME)
2182 Run an inferior Ruby process, input and output via buffer *ruby*.
2183 If there is a process already running in `*ruby*', switch to that buffer.
2184 With argument, allows you to edit the command line (default is value
2185 of `ruby-program-name'). Runs the hooks `inferior-ruby-mode-hook'
2186 (after the `comint-mode-hook' is run).
2187 (Type C-h m in the process buffer for a list of commands.)
2190 It seems we could either bundle my version of inf-ruby.el (as it's
2191 the newest). Or we could change the use of `run-ruby' so that it
2192 is robust across multiple distributions. I think I'd prefer the
2193 former, unless the older version of inf-ruby is actually bundled
2194 with emacs, in which case maybe we should go out of our way to
2195 support it. Thoughts?
2197 I think for now I'll just include the latest [[file:util/inf-ruby.el][inf-ruby.el]] in the
2198 newly created utility directory. I doubt anyone would have a
2199 problem using the latest version of this file.
2200 ** DONE test failing forcing vector results with =test-forced-vector-results= ruby code block
2201 Note that this only seems to happen the *second* time the test table
2204 #+srcname: bug-trivial-vector
2205 #+begin_src emacs-lisp :results vector silent
2209 #+srcname: bug-forced-vector-results
2210 #+begin_src ruby :var triv=test-trivial-vector :results silent
2214 mysteriously this seems to be fixed...
2215 ** DONE defunct R sessions
2216 Sometimes an old R session will turn defunct, and newly inserted code
2217 will not be evaluated (leading to a hang).
2219 This seems to be fixed by using `inferior-ess-send-input' rather than `comint-send-input'.
2220 ** DONE ruby fails on first call to non-default session
2222 #+srcname: bug-new-session
2223 #+begin_src ruby :session is-new
2227 ** DONE when reading results from =#+resname= line
2229 Errors when trying to read from resname lines.
2231 #+resname: bug-in-resname
2234 #+srcname: bug-in-resname-reader
2235 #+begin_src emacs-lisp :var buggy=bug-in-resname() :results silent
2239 ** DONE R-code broke on "org-babel" rename
2241 #+srcname: bug-R-babels
2246 ** DONE error on trivial R results
2248 So I know it's generally not a good idea to squash error without
2249 handling them, but in this case the error almost always means that
2250 there was no file contents to be read by =org-table-import=, so I
2253 #+srcname: bug-trivial-r1
2254 #+begin_src R :results replace
2255 pie(c(1, 2, 3), labels = c(1, 2, 3))
2258 #+srcname: bug-trivial-r2
2259 #+begin_src R :results replace
2263 #+resname: bug-trivial-r2
2266 #+srcname: bug-trivial-r3
2267 #+begin_src R :results replace
2271 #+resname: bug-trivial-r3
2276 ** DONE ruby new variable creation (multi-line ruby blocks)
2277 Actually it looks like we were dropping all but the last line.
2279 #+srcname: multi-line-ruby-test
2280 #+begin_src ruby :var table=bug-numerical-table :results replace
2282 table.each{|n| total += n}
2289 ** DONE R code execution seems to choke on certain inputs
2290 Currently the R code seems to work on vertical (but not landscape)
2293 #+srcname: little-fake
2294 #+begin_src emacs-lisp
2299 #+begin_src R :var num=little-fake
2306 #+srcname: set-debug-on-error
2307 #+begin_src emacs-lisp :results silent
2308 (setq debug-on-error t)
2311 #+srcname: bug-numerical-table
2312 #+begin_src emacs-lisp :results silent
2319 #+srcname: bug-R-number-evaluation
2320 #+begin_src R :var table=bug-numerical-table
2329 #+tblname: bug-vert-table
2334 #+srcname: bug-R-vertical-table
2335 #+begin_src R :var table=bug-vert-table :results silent
2339 ** DONE org bug/request: prevent certain org behaviour within code blocks
2340 E.g. [[]] gets recognised as a link (when there's text inside the
2341 brackets). This is bad for R code at least, and more generally
2342 could be argued to be inappropriate. Is it difficult to get org to
2343 ignore text in code blocks? [DED]
2345 I believe Carsten addressed this recently on the mailing list with
2346 the comment that it was indeed a difficult issue. I believe this
2347 may be one area where we could wait for an upstream (org-mode) fix.
2349 [Dan] Carsten has fixed this now in the core.
2351 ** DONE with :results replace, non-table output doesn't replace table output
2352 And vice versa. E.g. Try this first with table and then with len(table) [DED]
2353 #+begin_src python :var table=sandbox :results replace
2358 | 4 | "schulte" | 6 |
2361 Yes, this is certainly a problem. I fear that if we begin replacing
2362 anything immediately following a source block (regardless of whether
2363 it matches the type of our current results) we may accidentally delete
2364 hand written portions of the user's org-mode buffer.
2366 I think that the best solution here would be to actually start
2367 labeling results with a line that looks something like...
2371 This would have a couple of benefits...
2372 1) we wouldn't have to worry about possibly deleting non-results
2373 (which is currently an issue)
2374 2) we could reliably replace results even if there are different types
2375 3) we could reference the results of a source-code block in variable
2376 definitions, which would be useful if for example we don't wish to
2377 re-run a source-block every time because it is long-running.
2379 Thoughts? If no-one objects, I believe I will implement the labeling
2382 ** DONE extra quotes for nested string
2383 Well R appears to be reading the tables without issue...
2385 these *should* be quoted
2387 #+begin_src sh :results replace
2392 | "README.markdown" |
2395 | "existing_tools" |
2399 | "test-export.html" |
2400 | "test-export.org" |
2402 #+srcname: test-quotes
2403 #+begin_src ruby :var tab=ls
2409 #+srcname: test-quotes
2410 #+begin_src R :var tab=ls
2416 ** DONE simple ruby arrays not working
2418 As an example eval the following. Adding a line to test
2420 #+tblname: simple-ruby-array
2423 #+srcname: ruby-array-test
2424 #+begin_src ruby :var ar = simple-ruby-array :results silent
2431 ** DONE space trailing language name
2432 fix regexp so it works when there's a space trailing the language name
2434 #+srcname: test-trailing-space
2439 ** DONE Args out of range error
2441 The following block resulted in the error below [DED]. It ran without
2442 error directly in the shell.
2445 for platf in ill aff ; do
2446 for pop in CEU YRI ASI ; do
2447 rm -f $platf/hapmap-genos-$pop-all $platf/hapmap-rs-all
2448 cat $platf/hapmap-genos-$pop-* > $platf/hapmap-genos-$pop-all
2449 cat $platf/hapmap-rs-* > $platf/hapmap-rs-all
2454 executing source block with sh...
2455 finished executing source block
2456 string-equal: Args out of range: "", -1, 0
2458 the error =string-equal: Args out of range: "", -1, 0= looks like what
2459 used to be output when the block returned an empty results string.
2460 This should be fixed in the current version, you should now see the
2461 following message =no result returned by source block=.
2463 ** DONE ruby arrays not recognized as such
2465 Something is wrong in [[file:lisp/org-babel-script.el]] related to the
2466 recognition of ruby arrays as such.
2468 #+begin_src ruby :results replace
2474 #+begin_src python :results replace
2485 Evaluate all the cells in this table for a comprehensive test of the
2486 org-babel functionality.
2488 *Note*: if you have customized =org-babel-default-header-args= then some
2489 of these tests may fail.
2491 #+TBLNAME: org-babel-tests
2492 | functionality | block | arg | expected | results | pass |
2493 |-------------------------+----------------------------+-----+-------------+-------------+------|
2494 | basic evaluation | | | | | pass |
2495 |-------------------------+----------------------------+-----+-------------+-------------+------|
2496 | emacs lisp | basic-elisp | | 5 | 5 | pass |
2497 | shell | basic-shell | | 6 | 6 | pass |
2498 | ruby | basic-ruby | | org-babel | org-babel | pass |
2499 | python | basic-python | | hello world | hello world | pass |
2500 | R | basic-R | | 13 | 13 | pass |
2501 |-------------------------+----------------------------+-----+-------------+-------------+------|
2502 | tables | | | | | pass |
2503 |-------------------------+----------------------------+-----+-------------+-------------+------|
2504 | emacs lisp | table-elisp | | 3 | 3 | pass |
2505 | ruby | table-ruby | | 1-2-3 | 1-2-3 | pass |
2506 | python | table-python | | 5 | 5 | pass |
2507 | R | table-R | | 3.5 | 3.5 | pass |
2508 |-------------------------+----------------------------+-----+-------------+-------------+------|
2509 | source block references | | | | | pass |
2510 |-------------------------+----------------------------+-----+-------------+-------------+------|
2511 | all languages | chained-ref-last | | Array | Array | pass |
2512 |-------------------------+----------------------------+-----+-------------+-------------+------|
2513 | source block functions | | | | | pass |
2514 |-------------------------+----------------------------+-----+-------------+-------------+------|
2515 | emacs lisp | defun-fibb | | fibbd | fibbd | pass |
2516 | run over | Fibonacci | 0 | 1 | 1 | pass |
2517 | a | Fibonacci | 1 | 1 | 1 | pass |
2518 | variety | Fibonacci | 2 | 2 | 2 | pass |
2519 | of | Fibonacci | 3 | 3 | 3 | pass |
2520 | different | Fibonacci | 4 | 5 | 5 | pass |
2521 | arguments | Fibonacci | 5 | 8 | 8 | pass |
2522 |-------------------------+----------------------------+-----+-------------+-------------+------|
2523 | bugs and tasks | | | | | pass |
2524 |-------------------------+----------------------------+-----+-------------+-------------+------|
2525 | simple ruby arrays | ruby-array-test | | 3 | 3 | pass |
2526 | R number evaluation | bug-R-number-evaluation | | 2 | 2 | pass |
2527 | multi-line ruby blocks | multi-line-ruby-test | | 2 | 2 | pass |
2528 | forcing vector results | test-forced-vector-results | | Array | Array | pass |
2529 |-------------------------+----------------------------+-----+-------------+-------------+------|
2530 | sessions | | | | | pass |
2531 |-------------------------+----------------------------+-----+-------------+-------------+------|
2532 | set ruby session | set-ruby-session-var | | :set | :set | pass |
2533 | get from ruby session | get-ruby-session-var | | 3 | 3 | pass |
2534 | set python session | set-python-session-var | | set | set | pass |
2535 | get from python session | get-python-session-var | | 4 | 4 | pass |
2536 | set R session | set-R-session-var | | set | set | pass |
2537 | get from R session | get-R-session-var | | 5 | 5 | pass |
2538 #+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))
2542 #+srcname: basic-elisp
2543 #+begin_src emacs-lisp :results silent
2552 #+srcname: basic-shell
2553 #+begin_src sh :results silent
2560 #+srcname: date-simple
2561 #+begin_src sh :results silent
2565 #+srcname: basic-ruby
2566 #+begin_src ruby :results silent
2573 #+srcname: basic-python
2574 #+begin_src python :results silent
2582 #+begin_src R :results silent
2593 #+tblname: test-table
2597 #+srcname: table-elisp
2598 #+begin_src emacs-lisp :results silent :var table=test-table
2599 (length (car table))
2605 #+srcname: table-ruby
2606 #+begin_src ruby :results silent :var table=test-table
2607 table.first.join("-")
2613 #+srcname: table-python
2614 #+begin_src python :var table=test-table
2622 #+begin_src R :var table=test-table
2632 Lets pass a references through all of our languages...
2634 Lets start by reversing the table from the previous examples
2636 #+srcname: chained-ref-first
2637 #+begin_src python :var table = test-table
2642 #+resname: chained-ref-first
2646 Take the first part of the list
2648 #+srcname: chained-ref-second
2649 #+begin_src R :var table = chained-ref-first
2653 #+resname: chained-ref-second
2657 Turn the numbers into string
2659 #+srcname: chained-ref-third
2660 #+begin_src emacs-lisp :var table = chained-ref-second
2661 (mapcar (lambda (el) (format "%S" el)) table)
2664 #+resname: chained-ref-third
2667 and Check that it is still a list
2669 #+srcname: chained-ref-last
2670 #+begin_src ruby :var table=chained-ref-third
2678 ** source blocks as functions
2680 #+srcname: defun-fibb
2681 #+begin_src emacs-lisp :results silent
2682 (defun fibbd (n) (if (< n 2) 1 (+ (fibbd (- n 1)) (fibbd (- n 2)))))
2688 #+srcname: fibonacci
2689 #+begin_src emacs-lisp :results silent :var n=7
2712 ** sbe tests (these don't seem to be working...)
2713 Testing the insertion of results into org-mode tables.
2715 #+srcname: multi-line-output
2716 #+begin_src ruby :results replace
2717 "the first line ends here
2720 and this is the second one
2726 : the first line ends here\n\n\n and this is the second one\n\neven a third
2728 #+srcname: multi-line-error
2729 #+begin_src ruby :results replace
2730 raise "oh nooooooooooo"
2736 | the first line ends here... | -:5: warning: parenthesize argument(s) for future version... |
2737 #+TBLFM: $1='(sbe "multi-line-output")::$2='(sbe "multi-line-error")
2739 ** forcing results types tests
2741 #+srcname: test-trivial-vector
2742 #+begin_src emacs-lisp :results vector silent
2746 #+srcname: test-forced-vector-results
2747 #+begin_src ruby :var triv=test-trivial-vector :results silent
2753 #+srcname: set-ruby-session-var
2754 #+begin_src ruby :session :results silent
2759 #+srcname: get-ruby-session-var
2760 #+begin_src ruby :session :results silent
2764 #+srcname: set-python-session-var
2765 #+begin_src python :session
2770 #+srcname: get-python-session-var
2771 #+begin_src python :session
2775 #+srcname: set-R-session-var
2776 #+begin_src R :session
2781 #+srcname: get-R-session-var
2782 #+begin_src R :session
2791 To run these examples evaluate [[file:lisp/org-babel-init.el][org-babel-init.el]]
2793 ** org-babel.el beginning functionality
2795 #+begin_src sh :results replace
2800 : Sun Jul 5 18:54:39 EDT 2009
2807 : Sun Jul 05 18:54:35 -0400 2009
2819 #+begin_src R :results replace
2829 hist(rgamma(20,3,3))
2834 ** org-babel plays with tables
2835 Alright, this should demonstrate both the ability of org-babel to read
2836 tables into a lisp source code block, and to then convert the results
2837 of the source code block into an org table. It's using the classic
2838 "lisp is elegant" demonstration transpose function. To try this
2841 1. evaluate [[file:lisp/org-babel-init.el]] to load org-babel and friends
2842 2. evaluate the transpose definition =\C-c\\C-c= on the beginning of
2844 3. evaluate the next source code block, this should read in the table
2845 because of the =:var table=previous=, then transpose the table, and
2846 finally it should insert the transposed table into the buffer
2847 immediately following the block
2851 #+begin_src emacs-lisp :results silent
2852 (defun transpose (table)
2853 (apply #'mapcar* #'list table))
2861 #+begin_src emacs-lisp :var table=sandbox :results replace
2866 #+begin_src emacs-lisp
2871 | 1 | 2 | 3 | 4 | 5 |
2875 #+begin_src ruby :var table=sandbox :results replace
2876 table.first.join(" - ")
2882 #+begin_src python :var table=sandbox
2887 #+begin_src ruby :var table=sandbox :results replace
2892 : [[1, 2, 3], [4, "schulte", 6]]
2896 | 4 | "schulte" | 6 |
2898 #+begin_src python :var table=sandbox :results replace
2904 | "__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" |
2906 *** (sandbox table) R
2908 #+TBLNAME: sandbox_r
2912 #+begin_src R :results replace
2913 x <- c(rnorm(10, mean=-3, sd=1), rnorm(10, mean=3, sd=1))
2917 | -3.35473133869346 |
2919 | -3.32819924928633 |
2920 | -2.97310212756194 |
2921 | -2.09640758369576 |
2922 | -5.06054014378736 |
2923 | -2.20713700711221 |
2924 | -1.37618039712037 |
2925 | -1.95839385821742 |
2926 | -3.90407396475502 |
2927 | 2.51168071590226 |
2928 | 3.96753011570494 |
2929 | 3.31793212627865 |
2930 | 1.99829753972341 |
2931 | 4.00403686419829 |
2932 | 4.63723764452927 |
2933 | 3.94636744261313 |
2934 | 3.58355906547775 |
2935 | 3.01563442274226 |
2938 #+begin_src R var tabel=sandbox_r :results replace
2943 | 4 | "schulte" | 6 |
2946 Now shell commands are converted to tables using =org-table-import=
2947 and if these tables are non-trivial (i.e. have multiple elements) then
2948 they are imported as org-mode tables...
2950 #+begin_src sh :results replace
2954 | "total" | 208 | "" | "" | "" | "" | "" | "" |
2955 | "-rw-r--r--" | 1 | "dan" | "dan" | 57 | 2009 | 15 | "block" |
2956 | "-rw-r--r--" | 1 | "dan" | "dan" | 35147 | 2009 | 15 | "COPYING" |
2957 | "-rw-r--r--" | 1 | "dan" | "dan" | 722 | 2009 | 18 | "examples.org" |
2958 | "drwxr-xr-x" | 4 | "dan" | "dan" | 4096 | 2009 | 19 | "existing_tools" |
2959 | "-rw-r--r--" | 1 | "dan" | "dan" | 2207 | 2009 | 14 | "intro.org" |
2960 | "drwxr-xr-x" | 2 | "dan" | "dan" | 4096 | 2009 | 18 | "org-babel" |
2961 | "-rw-r--r--" | 1 | "dan" | "dan" | 277 | 2009 | 20 | "README.markdown" |
2962 | "-rw-r--r--" | 1 | "dan" | "dan" | 11837 | 2009 | 18 | "rorg.html" |
2963 | "-rw-r--r--" | 1 | "dan" | "dan" | 61829 | 2009 | 19 | "#rorg.org#" |
2964 | "-rw-r--r--" | 1 | "dan" | "dan" | 60190 | 2009 | 19 | "rorg.org" |
2965 | "-rw-r--r--" | 1 | "dan" | "dan" | 972 | 2009 | 11 | "test-export.org" |
2968 ** silent evaluation
2976 #+begin_src ruby :results silent
2980 #+begin_src ruby :results replace
2987 ** (sandbox) referencing other source blocks
2988 Doing this in emacs-lisp first because it's trivial to convert
2989 emacs-lisp results to and from emacs-lisp.
2991 *** emacs lisp source reference
2992 This first example performs a calculation in the first source block
2993 named =top=, the results of this calculation are then saved into the
2994 variable =first= by the header argument =:var first=top=, and it is
2995 used in the calculations of the second source block.
2998 #+begin_src emacs-lisp
3002 #+begin_src emacs-lisp :var first=top :results replace
3008 This example is the same as the previous only the variable being
3009 passed through is a table rather than a number.
3011 #+begin_src emacs-lisp :results silent
3012 (defun transpose (table)
3013 (apply #'mapcar* #'list table))
3016 #+TBLNAME: top_table
3020 #+SRCNAME: second_src_example
3021 #+begin_src emacs-lisp :var table=top_table
3025 #+begin_src emacs-lisp :var table=second_src_example :results replace
3030 | 4 | "schulte" | 6 |
3032 Now working for ruby
3039 #+begin_src ruby :var other=start :results replace
3045 #+SRCNAME: start_two
3050 #+begin_src python :var another=start_two :results replace
3055 Since all variables are converted into Emacs Lisp it is no problem to
3056 reference variables specified in another language.
3058 #+SRCNAME: ruby-block
3063 #+SRCNAME: lisp_block
3064 #+begin_src emacs-lisp :var ruby-variable=ruby-block
3068 #+begin_src python :var lisp_var=lisp_block
3077 #+begin_src R :results replace
3084 #+begin_src R :var other=first_r :results replace
3091 ** (sandbox) selective export
3093 For exportation tests and examples see (including exportation of
3094 inline source code blocks) [[file:test-export.org]]
3097 ** (sandbox) source blocks as functions
3100 #+begin_src emacs-lisp :results silent
3105 #+begin_src emacs-lisp :var n=default :results replace
3111 #+begin_src emacs-lisp :var result=triple(n=3, m=98) :results replace
3117 The following just demonstrates the ability to assign variables to
3118 literal values, which was not implemented until recently.
3120 #+begin_src ruby :var num="eric" :results replace
3127 ** (sandbox) inline source blocks
3129 This is an inline source code block src_ruby{1 + 6}. And another
3130 source block with text output src_emacs-lisp{"eric"}.
3132 This is an inline source code block with header
3133 arguments. src_ruby[:var n=fibbd( n = 0 )]{n}
3136 ** (sandbox) integration w/org tables
3138 #+begin_src emacs-lisp :results silent
3139 (defun fibbd (n) (if (< n 2) 1 (+ (fibbd (- n 1)) (fibbd (- n 2)))))
3143 #+begin_src emacs-lisp :var n=4 :results silent
3147 #+begin_src emacs-lisp :results silent
3148 (mapcar #'fibbd '(0 1 2 3 4 5 6 7 8))
3151 Something is not working here. The function `sbe ' works fine when
3152 called from outside of the table (see the source block below), but
3153 produces an error when called from inside the table. I think there
3154 must be some narrowing going on during intra-table emacs-lisp
3157 | original | fibbd |
3158 |----------+-------|
3169 #+TBLFM: $2='(sbe "fibbd" (n $1))
3173 #+begin_src emacs-lisp :results silent
3174 (sbe 'fibbd (n "8"))
3179 LocalWords: DBlocks dblocks org-babel el eric fontification