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 ** PROPOSED optional timestamp for output
204 Add option to place an (inactive) timestamp at the #+resname, to
205 record when that output was generated.
206 ** PROPOSED use example block for large amounts of stdout output?
207 We're currently `examplizing' with : at the beginning of the line,
208 but should larger amounts of output be in a
209 \#+begin_example...\#+end_example block? What's the cutoff? > 1
210 line? This would be nice as it would allow folding of lengthy
211 output. Sometimes one will want to see stdout just to check
212 everything looks OK, and then fold it away.
214 ** TODO make tangle files read-only?
215 With a file-local variable setting, yea that makes sense. Maybe
216 the header should reference the related org-mode file.
217 ** TODO take default values for header args from properties
218 Use file-wide and subtree wide properties to set default values for
220 ** TODO support for working with =*Org Edit Src Example*= buffers [2/4]
221 *** TODO optionally evaluate header references when we switch to =*Org Edit Src*= buffer
222 That seems to imply that the header references need to be evaluated
223 and transformed into the target language object when we hit C-c ' to
224 enter the *Org Edit Src* buffer [DED]
226 Good point, I heartily agree that this should be supported [Eric]
228 (or at least before the first time we attempt to evaluate code in that
229 buffer -- I suppose there might be an argument for lazy evaluation, in
230 case someone hits C-c ' but is "just looking" and not actually
231 evaluating anything.) Of course if evaluating the reference is
232 computationally intensive then the user might have to wait before they
233 get the *Org Edit Src* buffer. [DED]
235 I fear that it may be hard to anticipate when the references will be
236 needed, some major-modes do on-the-fly evaluation while the buffer is
237 being edited. I think that we should either do this before the buffer
238 is opened or not at all, specifically I think we should resolve
239 references if the user calls C-c ' with a prefix argument. Does that
240 sound reasonable? [Eric]
244 [Dan] So now that we have org-src-mode and org-src-mode-hook, I guess
245 org-babel should do this by using the hook to make sure that, when C-c
246 C-' is issued on a source block, any references are resolved and
247 assignments are made in the appropriate session.
248 *** TODO set buffer-local-process variables appropriately [DED]
249 I think something like this would be great. You've probably
250 already thought of this, but just to note it down: it would be really
251 nice if org-babel's notion of a buffer's 'session/process' played
252 nicely with ESS's notion of the buffer's session/process. ESS keeps
253 the current process name for a buffer in a buffer-local variable
254 ess-local-process-name. So one thing we will probably want to do is
255 make sure that the *Org Edit Src Example* buffer sets that variable
258 I had not thought of that, but I agree whole heartedly. [Eric]
260 Once this is done every variable should be able to dump regions into
261 their inferior-process buffer using major-mode functions.
262 *** DEFERRED send code to inferior process
263 Another thought on this topic: I think we will want users to send
264 chunks of code to the interpreter from within the *Org Edit Src*
265 buffer, and I think that's what you have in mind already. In ESS that
266 is done using the ess-eval-* functions. [DED]
268 I think we can leave this up to the major-mode in the source code
269 buffer, as almost every source-code major mode will have functions for
270 doing things like sending regions to the inferior process. If
271 anything we might need to set the value of the buffer local inferior
272 process variable. [Eric]
274 *** DONE some possible requests/proposed changes for Carsten [4/4]
275 While I remember, some possible requests/proposed changes for Carsten
276 come to mind in that regard:
278 **** DONE Remap C-x C-s to save the source to the org buffer?
279 I've done this personally and I find it essential. I'm using
280 #+begin_src emacs-lisp
281 (defun org-edit-src-save ()
282 "Update the parent org buffer with the edited source code, save
283 the parent org-buffer, and return to the source code edit
292 (define-key org-exit-edit-mode-map "\C-x\C-s" 'org-edit-src-save)
296 I think this is great, but I think it should be implemented in the
299 **** DEFERRED Rename buffer and minor mode?
300 Something shorter than *Org Edit Src Example* for the buffer
301 name. org-babel is bringing org's source code interaction to a
302 level of maturity where the 'example' is no longer
303 appropriate. And if further keybindings are going to be added to
304 the minor mode then maybe org-edit-src-mode is a better name than
307 Maybe we should name the buffer with a combination of the source
308 code and the session. I think that makes sense.
310 [ES] Are you also suggesting a new org-edit-src minor mode?
311 [DED] org-exit-edit-mode is a minor mode that already exists:
313 Minor mode installing a single key binding, "C-c '" to exit special edit.
315 org-edit-src-save now has a binding in that mode, so I guess all
316 I'm saying at this stage is that it's a bit of a misnomer. But
317 perhaps we will also have more functionality to add to that minor
318 mode, making it even more of a misnomer. Perhaps something like
319 org-src-mode would be better.
320 **** DONE Changed minor mode name and added hooks
322 **** DEFERRED a hook called when the src edit buffer is created
323 This should be implemented in the org-mode core
325 ** TODO resolve references to other org buffers/files
326 This would allow source blocks to call upon tables, source-blocks,
327 and results in other org buffers/files.
330 - [[file:lisp/org-babel-ref.el::TODO%20allow%20searching%20for%20names%20in%20other%20buffers][org-babel-ref.el:searching-in-other-buffers]]
331 - [[file:lisp/org-babel.el::defun%20org-babel%20find%20named%20result%20name][org-babel.el#org-babel-find-named-result]]
332 ** TODO resolve references to other non-org files
333 - tabular data in .csv, .tsv etc format
334 - files of interpreted code: anything stopping us giving such files
335 similar status to a source code block?
336 - Would be nice to allow org and non-org files to be remote
337 ** TODO figure out how to handle errors during evaluation
338 R has a try function, with error handling, along the lines of
339 python. I bet ruby does too. Maybe more of an issue for functional
340 style; in my proposed scripting style the error just gets dumped to
341 the org buffer and the user is thus alerted.
342 ** TODO figure out how to handle graphic output
343 This is listed under [[* graphical output][graphical output]] in out objectives.
345 This should take advantage of the =:results file= option, and
346 languages which almost always produce graphical output should set
347 =:results file= to true by default. That would handle placing these
348 results in the buffer. Then if there is a combination of =silent= and
349 =file= =:results= headers we could drop the results to a temp buffer
350 and pop open that buffer...
351 ** TODO =\C-c \C-o= to open results of source block
352 by adding a =defadvice= to =org-open-at-point= we can use the common
353 =\C-c \C-o= keybinding to open the results of a source-code block.
354 This would be especially useful for source-code blocks which generate
355 graphical results and insert a file link as the results in the
356 org-mode buffer. (see [[* figure out how to handle graphic output][TODO figure out how to handle graphic output]]).
357 This could also act reasonably with other results types...
359 - file :: use org-open-at-point to open the file
360 - scalar :: open results unquoted in a new buffer
361 - tabular :: export the table to a new buffer and open that buffer
363 ** TODO Finalise behaviour regarding vector/scalar output
364 *** DONE Stop spaces causing vector output
365 This simple example of multilingual chaining produces vector output if
366 there are spaces in the message and scalar otherwise.
370 #+begin_src R :var msg=msg-from-python
371 paste(msg, "und R", sep=" ")
375 : org-babel speaks elisp y python und R
377 #+srcname: msg-from-python
378 #+begin_src python :var msg=msg-from-elisp
382 #+srcname: msg-from-elisp
383 #+begin_src emacs-lisp :var msg="org-babel speaks"
384 (concat msg " elisp")
386 ** STARTED share org-babel [1/4]
387 how should we share org-babel?
389 *** DONE post to org-mode
390 *** TODO post to ess mailing list
391 *** TODO create a org-babel page on worg
392 *** TODO create a short screencast demonstrating org-babel in action
395 we need to think up some good examples
397 **** interactive tutorials
398 This could be a place to use [[* org-babel assertions][org-babel assertions]].
400 for example the first step of a tutorial could assert that the version
401 of the software-package (or whatever) is equal to some value, then
402 source-code blocks could be used with confidence (and executed
403 directly from) the rest of the tutorial.
405 **** answering a text-book question w/code example
406 org-babel is an ideal environment enabling both the development and
407 demonstrationg of the code snippets required as answers to many
410 **** something using tables
411 maybe something along the lines of calculations from collected grades
414 Maybe something like the following which outputs sizes of directories
415 under the home directory, and then instead of the trivial =emacs-lisp=
416 block we could use an R block to create a nice pie chart of the
420 #+begin_src bash :results replace
424 #+begin_src emacs-lisp :var sizes=sizes :results replace
428 ** TODO command line execution
429 Allow source code blocks to be called form the command line. This
430 will be easy using the =sbe= function in [[file:lisp/org-babel-table.el][org-babel-table.el]].
432 This will rely upon [[* resolve references to other buffers][resolve references to other buffers]].
434 ** TODO inline source code blocks [3/5]
435 Like the =\R{ code }= blocks
437 not sure what the format should be, maybe just something simple
438 like =src_lang[]{}= where lang is the name of the source code
439 language to be evaluated, =[]= is optional and contains any header
440 arguments and ={}= contains the code.
442 (see [[* (sandbox) inline source blocks][the-sandbox]])
444 *** DONE evaluation with \C-c\C-c
445 Putting aside the header argument issue for now we can just run these
446 with the following default header arguments
447 - =:results= :: silent
448 - =:exports= :: results
450 *** DONE inline exportation
451 Need to add an interblock hook (or some such) through org-exp-blocks
452 *** DONE header arguments
453 We should make it possible to use header arguments.
455 *** TODO fontification
456 we should color these blocks differently
458 *** TODO refine html exportation
459 should use a span class, and should show original source in tool-tip
461 ** TODO formulate general rules for handling vectors and tables / matrices with names
462 This is non-trivial, but may be worth doing, in particular to
463 develop a nice framework for sending data to/from R.
465 In R, indexing vector elements, and rows and columns, using
466 strings rather than integers is an important part of the
468 - elements of a vector may have names
469 - matrices and data.frames may have "column names" and "row names"
470 which can be used for indexing
471 - In a data frame, row names *must* be unique
479 > mat <- matrix(1:4, nrow=2, ncol=2, dimnames=list(c("r1","r2"), c("c1","c2")))
484 > # The names are separate from the data: they do not interfere with operations on the data
491 > df <- data.frame(var1=1:26, var2=26:1, row.names=letters)
493 [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
499 So it's tempting to try to provide support for this in org-babel. For example
500 - allow R to refer to columns of a :var reference by their names
501 - When appropriate, results from R appear in the org buffer with "named
504 However none (?) of the other languages we are currently supporting
505 really have a native matrix type, let alone "column names" or "row
506 names". Names are used in e.g. python and perl to refer to entries
509 It currently seems to me that support for this in org-babel would
510 require setting rules about when org tables are considered to have
511 named columns/fields, and ensuring that (a) languages with a notion
512 of named columns/fields use them appropriately and (b) languages
513 with no such notion do not treat then as data.
515 - Org allows something that *looks* like column names to be separated
517 - Org also allows a row to *function* as column names when special
518 markers are placed in the first column. An hline is unnecessary
519 (indeed hlines are purely cosmetic in org [correct?]
520 - Org does not have a notion of "row names" [correct?]
522 The full org table functionality exeplified [[http://orgmode.org/manual/Advanced-features.html#Advanced-features][here]] has features that
523 we would not support in e.g. R (like names for the row below).
525 *** Initial statement: allow tables with hline to be passed as args into R
526 This doesn't seem to work at the moment (example below). It would
527 also be nice to have a natural way for the column names of the org
528 table to become the column names of the R data frame, and to have
529 the option to specify that the first column is to be used as row
530 names in R (these must be unique). But this might require a bit of
535 | col1 | col2 | col3 |
536 |------+---------+------|
544 #+begin_src R var tabel=egtable
549 | "col1" | "col2" | "col3" |
550 |--------+-----------+--------|
552 | 4 | "schulte" | 6 |
555 Another example is in the [[*operations%20in%20on%20tables][grades example]].
557 ** TODO re-implement helper functions from org-R
558 *** Initial statement [Eric]
559 Much of the power of org-R seems to be in it's helper functions for
560 the quick graphing of tables. Should we try to re-implement these
561 functions on top of org-babel?
563 I'm thinking this may be useful both to add features to org-babel-R and
564 also to potentially suggest extensions of the framework. For example
565 one that comes to mind is the ability to treat a source-code block
566 like a function which accepts arguments and returns results. Actually
567 this can be it's own TODO (see [[* source blocks as functions][source blocks as functions]]).
569 - We want to provide convenient off-the-shelf actions
570 (e.g. plotting data) that make use of our new code evaluation
571 environment but do not require any actual coding.
572 *** Initial Design proposal [Dan]
573 - *Input data* will be specified using the same mechanism as :var
574 references, thus the input data may come from a table, or
575 another source block, and it is initially available as an elisp
577 - We introduce a new #+ line, e.g. #+BABELDO. C-c C-c on that
578 line will apply an *action* to the referenced data.
579 - *Actions correspond to source blocks*: our library of available
580 actions will be a library of org-babel source blocks. Thus the
581 code for executing an action, and the code for dealing with the
582 output of the action will be the same code as for executing
583 source blocks in general
584 - Optionally, the user can have the relevant source block inserted
585 into the org buffer after the (say) #+BABELDO line. This will
586 allow the user to fine tune the action by modifying the code
587 (especially useful for plots).
588 - So maybe a #+BABELDO line will have header args
589 - :data (a reference to a table or source code block)
590 - :action (or should that be :srcname?) which will be something
591 like :action pie-chart, referring to a source block which will
592 be executed with the :data referent passed in using a :var arg.
593 - :showcode or something controlling whether to show the code
595 *** Modification to design
596 I'm implementing this, at least initially, as a new interpreter
597 named 'babel', which has an empty body. 'babel' blocks take
598 a :srcname header arg, and look for the source-code block with
599 that name. They then execute the referenced block, after first
600 appending their own header args on to the target block's header
603 If the target block is in the library of babel (a.o.t. e.g. the
604 current buffer), then the code in the block will refer to the
605 input data with a name dictated by convention (e.g. __data__
606 (something which is syntactically legal in all languages...). Thus
607 the babel block will use a :var __data__ = whatever header arg to
608 reference the data to be plotted.
611 This is covered by the [[file:library-of-babel.org][Library of Babel]], which will contain
612 ready-made source blocks designed to carry out useful common tasks.
613 ** PROPOSED Are we happy with current behaviour regarding vector/scalar output?
614 This simple example of multilingual chaining produces vector output if
615 there are spaces in the message and scalar otherwise.
617 #+begin_src R :var msg=msg-from-python
618 paste(msg, "und_R", sep="_")
621 #+srcname: msg-from-python
622 #+begin_src python :var msg=msg-from-elisp
626 #+srcname: msg-from-elisp
627 #+begin_src emacs-lisp :var msg="org-babel_speaks"
628 (concat msg "_elisp")
631 ** PROPOSED conversion between org-babel and noweb (e.g. .Rnw) format
632 I haven't thought about this properly. Just noting it down. What
633 Sweave uses is called "R noweb" (.Rnw).
635 I found a good description of noweb in the following article (see
636 the [[http://www.cs.tufts.edu/~nr/pubs/lpsimp.pdf][pdf]]).
638 I think there are two parts to noweb, the construction of
639 documentation and the extraction of source-code (with notangle).
641 *documentation*: org-mode handles all of our documentation needs in
642 a manner that I believe is superior to noweb.
644 *source extraction* At this point I don't see anyone writing large
645 applications with 100% of the source code contained in org-babel
646 files, rather I see org-babel files containing things like
647 - notes with active code chunks
648 - interactive tutorials
649 - requirements documents with code running test suites
650 - and of course experimental reports with the code to run the
651 experiment, and perform analysis
653 Basically I think the scope of the programs written in org-babel
654 (at least initially) will be small enough that it wont require the
655 addition of a tangle type program to extract all of the source code
656 into a running application.
658 On the other hand, since we already have named blocks of source
659 code which reference other blocks on which they rely, this
660 shouldn't be too hard to implement either on our own, or possibly
661 relying on something like noweb/notangle.
663 ** PROPOSED support for passing paths to files between source blocks
664 Maybe this should be it's own result type (in addition to scalars and
665 vectors). The reason being that some source-code blocks (for example
666 ditaa or anything that results in the creation of a file) may want to
667 pass a file path back to org-mode which could then be inserted into
668 the org-mode buffer as a link to the file...
670 This would allow for display of images upon export providing
671 functionality similar to =org-exp-blocks= only in a more general
674 ** DEFERRED use textConnection to pass tsv to R?
675 When passing args from the org buffer to R, the following route is
676 used: arg in buffer -> elisp -> tsv on file -> data frame in R. I
677 think it would be possible to avoid having to write to file by
678 constructing an R expression in org-babel-R-assign-elisp, something
681 #+begin_src emacs-lisp
682 (org-babel-R-input-command
683 (format "%s <- read.table(textConnection(\"%s\"), sep=\"\\t\", as.is=TRUE)"
684 name (orgtbl-to-tsv value '(:sep "\t" :fmt org-babel-R-quote-tsv-field))))
687 I haven't tried to implement this yet as it's basically just
688 fiddling with something that works. The only reason for it I can
689 think of would be efficiency and I haven't tested that.
691 This Didn't work after an initial test. I still think this is a
692 good idea (I also think we should try to do something similar when
693 writing out results frmo R to elisp) however as it wouldn't result
694 in any functional changes I'm bumping it down to deferred for
699 #+tblname: quick-test
702 #+srcname: quick-test-src-blk
703 #+begin_src R :var vec=quick-test
709 ** DEFERRED re-implement R evaluation using ess-command or ess-execute
710 I don't have any complaints with the current R evaluation code or
711 behaviour, but I think it would be good to use the ESS functions
712 from a political point of view. Plus of course it has the normal
713 benefits of an API (insulates us from any underlying changes etc). [DED]
715 I'll look into this. I believe that I looked at and rejected these
716 functions initially but now I can't remember why. I agree with
717 your overall point about using API's where available. I will take
718 a look back at these and either switch to using the ess commands,
719 or at least articulate under this TODO the reasons for using our
720 custom R-interaction commands. [Eric]
724 Lets just replace =org-babel-R-input-command= with =ess-execute=.
726 I tried this, and although it works in some situations, I find that
727 =ess-command= will often just hang indefinitely without returning
728 results. Also =ess-execute= will occasionally hang, and pops up
729 the buffer containing the results of the command's execution, which
730 is undesirable. For now these functions can not be used. Maybe
731 someone more familiar with the ESS code can recommend proper usage
732 of =ess-command= or some other lower-level function which could be
733 used in place of [[file:lisp/org-babel-R.el::defun%20org-babel%20R%20input%20command%20command][org-babel-R-input-command]].
737 #+begin_quote ess-command
738 (ess-command COM &optional BUF SLEEP NO-PROMPT-CHECK)
740 Send the ESS process command COM and delete the output
741 from the ESS process buffer. If an optional second argument BUF exists
742 save the output in that buffer. BUF is erased before use.
743 COM should have a terminating newline.
744 Guarantees that the value of .Last.value will be preserved.
745 When optional third arg SLEEP is non-nil, `(sleep-for (* a SLEEP))'
746 will be used in a few places where `a' is proportional to `ess-cmd-delay'.
749 #+begin_quote ess-execute
750 (ess-execute COMMAND &optional INVERT BUFF MESSAGE)
752 Send a command to the ESS process.
753 A newline is automatically added to COMMAND. Prefix arg (or second arg
754 INVERT) means invert the meaning of
755 `ess-execute-in-process-buffer'. If INVERT is 'buffer, output is
756 forced to go to the process buffer. If the output is going to a
757 buffer, name it *BUFF*. This buffer is erased before use. Optional
758 fourth arg MESSAGE is text to print at the top of the buffer (defaults
759 to the command if BUFF is not given.)
762 *** out current setup
764 1) The body of the R source code block is wrapped in a function
765 2) The function is called inside of a =write.table= function call
766 writing the results to a table
767 3) The table is read using =org-table-import=
769 ** DEFERRED Rework Interaction with Running Processes [2/5]
770 *** DONE robust to errors interrupting execution
772 #+srcname: long-runner-ruby
773 #+begin_src ruby :results silent
778 *** DEFERRED use =C-g= keyboard-quit to push processing into the background
779 This may be possible using the `run-with-timer' command.
781 I have no idea how this could work...
783 #+srcname: long-runner-ruby
784 #+begin_src ruby :results silent
789 *** TODO ability to select which of multiple sessions is being used
790 Increasingly it is looking like we're going to want to run all
791 source code blocks in comint buffer (sessions). Which will have
793 1) allowing background execution
794 2) maintaining state between source-blocks
795 - allowing inline blocks w/o header arguments
798 (like ess-switch-process in .R buffers)
800 Maybe this could be packaged into a header argument, something
801 like =:R_session= which could accept either the name of the
802 session to use, or the string =prompt=, in which case we could use
803 the =ess-switch-process= command to select a new process.
805 *** TODO evaluation of shell code as background process?
806 After C-c C-c on an R code block, the process may appear to
807 block, but C-g can be used to reclaim control of the .org buffer,
808 without interrupting the R evalution. However I believe this is not
809 true of bash/sh evaluation. [Haven't tried other languages] Perhaps
810 a solution is just to background the individual shell commands.
812 The other languages (aside from emacs lisp) are run through the
813 shell, so if we find a shell solution it should work for them as
816 Adding an ampersand seems to be a supported way to run commands in
817 the background (see [[http://www.emacswiki.org/emacs/ExecuteExternalCommand#toc4][external-commands]]). Although a more extensible
818 solution may involve the use of the [[elisp:(progn (describe-function 'call-process-region) nil)][call-process-region]] function.
820 Going to try this out in a new file [[file:lisp/org-babel-proc.el][org-babel-proc.el]]. This should
821 contain functions for asynchronously running generic shell commands
822 in the background, and then returning their input.
824 **** partial update of org-mode buffer
825 The sleekest solution to this may be using a comint buffer, and
826 then defining a filter function which would incrementally interpret
827 the results as they are returned, including insertion into the
828 org-mode buffer. This may actually cause more problems than it is
829 worth, what with the complexities of identifying the types of
830 incrementally returned results, and the need for maintenance of a
831 process marker in the org buffer.
833 **** 'working' spinner
834 It may be nice and not too difficult to place a spinner on/near the
835 evaluating source code block
837 *** TODO conversion of output from interactive shell, R (and python) sessions to org-babel buffers
838 [DED] This would be a nice feature I think. Although an org-babel
839 purist would say that it's working the wrong way round... After
840 some interactive work in a *R* buffer, you save the buffer, maybe
841 edit out some lines, and then convert it to org-babel format for
842 posterity. Same for a shell session either in a *shell* buffer, or
843 pasted from another terminal emulator. And python of course.
845 ** DEFERRED improve the source-block snippet
846 any real improvement seems somewhat beyond the ability of yasnippet
849 [[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]]
851 ,#name : Chapter title
854 ${1:$(make-string (string-width text) ?\=)}
859 [[file:snippets/org-mode/sb][sb -- snippet]]
861 waiting for guidance from those more familiar with yasnippets
862 ** DONE LoB: allow output in buffer
863 ** DONE allow default header arguments by language
864 org-babel-default-header-args:lang-name
866 An example of when this is useful is for languages which always return
867 files as their results (e.g. [[*** ditaa][ditaa]], and [[*** gnuplot][gnuplot]]).
868 ** DONE singe-function tangling and loading elisp from literate org-mode file [3/3]
870 This function should tangle the org-mode file for elisp, and then call
871 `load-file' on the resulting tangled file.
873 #+srcname: test-loading-embedded-emacs-lisp
874 #+begin_src emacs-lisp :results replace
875 (setq test-tangle-advert nil)
876 (setq test-tangle-loading nil)
877 (setq results (list :before test-tangle-loading test-tangle-advert))
878 (org-babel-load-file "test-tangle.org")
879 (setq results (list (list :after test-tangle-loading test-tangle-advert) results))
880 (delete-file "test-tangle.el")
884 #+resname: test-loading-embedded-emacs-lisp
885 | :before | nil | nil |
886 | :after | "org-babel tangles" | "use org-babel-tangle for all your emacs initialization files!!" |
888 *** DONE add optional language limiter to org-babel-tangle
889 This should check to see if there is any need to re-export
891 *** DONE ensure that org-babel-tangle returns the path to the tangled file(s)
893 #+srcname: test-return-value-of-org-babel-tangle
894 #+begin_src emacs-lisp :results replace
895 (mapcar #'file-name-nondirectory (org-babel-tangle-file "test-tangle.org" "emacs-lisp"))
901 *** DONE only tangle the file if it's actually necessary
902 ** DONE add a function to jump to a source-block by name
903 I've had an initial stab at that in org-babel-find-named-block
904 (library-of-babel branch).
906 At the same time I introduced org-babel-named-src-block-regexp, to
907 match src-blocks with srcname.
909 This is now working with the command
910 `org-babel-goto-named-source-block', all we need is a good key
913 ** DONE add =:none= session argument (for purely functional execution) [4/4]
914 This would allow source blocks to be run in their own new process
916 - These blocks could then also be run in the background (since we can
917 detach and just wait for the process to signal that it has terminated)
918 - We wouldn't be drowning in session buffers after running the tests
919 - we can re-use much of the session code to run in a more /functional/
922 While session provide a lot of cool features, like persistent
923 environments, [[* DONE function to bring up inferior-process buffer][pop-to-session]], and hints at exportation for
924 org-babel-tangle, they also have some down sides and I'm thinking that
925 session-based execution maybe shouldn't be the default behavior.
927 Down-sides to sessions
928 - *much* more complicated than functional evaluation
929 - maintaining the state of the session has weird issues
930 - waiting for evaluation to finish
931 - prompt issues like [[* TODO weird escaped characters in shell prompt break shell evaluation][shell-prompt-escapes-bug]]
932 - can't run in background
933 - litter emacs with session buffers
937 #+srcname: ruby-task-no-session
938 #+begin_src ruby :results replace output
944 #+resname: ruby-task-no-session
949 #+srcname: task-python-none-session
950 #+begin_src python :session none :results replace value
956 #+resname: task-python-none-session
961 #+srcname: task-session-none-sh
962 #+begin_src sh :results replace
967 #+resname: task-session-none-sh
973 #+srcname: task-no-session-R
974 #+begin_src R :results replace output
981 #+resname: task-no-session-R
985 ** DONE fully purge org-babel-R of direct comint interaction
986 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
988 ** DONE Create objects in top level (global) environment [5/5]
991 *** initial requirement statement [DED]
992 At the moment, objects created by computations performed in the
993 code block are evaluated in the scope of the
994 code-block-function-body and therefore disappear when the code
995 block is evaluated {unless you employ some extra trickery like
996 assign('name', object, env=globalenv()) }. I think it will be
997 desirable to also allow for a style wherein objects that are
998 created in one code block persist in the R global environment and
999 can be re-used in a separate block.
1001 This is what Sweave does, and while I'm not saying we have to be
1002 the same as Sweave, it wouldn't be hard for us to provide the same
1003 behaviour in this case; if we don't, we risk undeservedly being
1004 written off as an oddity by some.
1006 IOW one aspect of org-babel is that of a sort of functional
1007 meta-programming language. This is crazy, in a very good
1008 way. Nevertheless, wrt R I think there's going to be a lot of value
1009 in providing for a working style in which the objects are stored in
1010 the R session, rather than elisp/org buffer. This will be a very
1011 familiar working style to lots of people.
1013 There are no doubt a number of different ways of accomplishing
1014 this, the simplest being a hack like adding
1017 for(objname in ls())
1018 assign(objname, get(objname), envir=globalenv())
1021 to the source code block function body. (Maybe wrap it in an on.exit() call).
1023 However this may deserve to be thought about more carefully, perhaps
1024 with a view to having a uniform approach across languages. E.g. shell
1025 code blocks have the same semantics at the moment (no persistence of
1026 variables across code blocks), because the body is evaluated in a new
1027 bash shell process rather than a running shell. And I guess the same
1028 is true for python. However, in both these cases, you could imagine
1029 implementing the alternative in which the body is evaluated in a
1030 persistent interactive session. It's just that it's particularly
1031 natural for R, seeing as both ESS and org-babel evaluate commands in a
1032 single persistent R session.
1036 Thanks for bringing this up. I think you are absolutely correct that we
1037 should provide support for a persistent environment (maybe called a
1038 *session*) in which to evaluate code blocks. I think the current setup
1039 demonstrates my personal bias for a functional style of programming
1040 which is certainly not ideal in all contexts.
1042 While the R function you mention does look like an elegant solution, I
1043 think we should choose an implementation that would be the same across
1044 all source code types. Specifically I think we should allow the user to
1045 specify an optional *session* as a header variable (when not present we
1046 assume a default session for each language). The session name could be
1047 used to name a comint buffer (like the *R* buffer) in which all
1048 evaluation would take place (within which variables would retain their
1049 values --at least once I remove some of the functional method wrappings
1050 currently in place-- ).
1052 This would allow multiple environments to be used in the same buffer,
1053 and once this setup was implemented we should be able to fairly easily
1054 implement commands for jumping between source code blocks and the
1055 related session buffers, as well as for dumping the last N commands from
1056 a session into a new or existing source code block.
1058 Please let me know if you foresee any problems with this proposed setup,
1059 or if you think any parts might be confusing for people coming from
1060 Sweave. I'll hopefully find some time to work on this later in the
1063 *** can functional and interpreted/interactive models coexist?
1065 Even though both of these use the same =*R*= buffer the value of =a=
1066 is not preserved because it is assigned inside of a functional
1069 #+srcname: task-R-sessions
1076 #+srcname: task-R-same-session
1081 This functional wrapper was implemented in order to efficiently return
1082 the results of the execution of the entire source code block. However
1083 it inhibits the evaluation of source code blocks in the top level,
1084 which would allow for persistence of variable assignment across
1085 evaluations. How can we allow *both* evaluation in the top level, and
1086 efficient capture of the return value of an entire source code block
1087 in a language independent manner?
1089 Possible solutions...
1090 1) we can't so we will have to implement two types of evaluation
1091 depending on which is appropriate (functional or imperative)
1092 2) we remove the functional wrapper and parse the source code block
1093 into it's top level statements (most often but not always on line
1094 breaks) so that we can isolate the final segment which is our
1096 3) we add some sort of "#+return" line to the code block
1097 4) we take advantage of each languages support for meta-programming
1098 through =eval= type functions, and use said to evaluate the entire
1099 blocks in such a way that their environment can be combined with the
1100 global environment, and their results are still captured.
1101 5) I believe that most modern languages which support interactive
1102 sessions have support for a =last_result= type function, which
1103 returns the result of the last input without re-calculation. If
1104 widely enough present this would be the ideal solution to a
1105 combination of functional and imperative styles.
1107 None of these solutions seem very desirable, but for now I don't see
1108 what else would be possible.
1110 Of these options I was leaning towards (1) and (4) but now believe
1111 that if it is possible option (5) will be ideal.
1113 **** (1) both functional and imperative evaluation
1115 - can take advantage of built in functions for sending regions to the
1117 - retains the proven tested and working functional wrappers
1120 - introduces the complication of keeping track of which type of
1121 evaluation is best suited to a particular context
1122 - the current functional wrappers may require some changes in order to
1123 include the existing global context
1125 **** (4) exploit language meta-programming constructs to explicitly evaluate code
1127 - only one type of evaluation
1130 - some languages may not have sufficient meta-programming constructs
1132 **** (5) exploit some =last_value= functionality if present
1134 Need to ensure that most languages have such a function, those without
1135 will simply have to implement their own similar solution...
1137 | language | =last_value= function |
1138 |------------+-----------------------------|
1142 | shell | see [[* last command for shells][last command for shells]] |
1143 | emacs-lisp | see [[* emacs-lisp will be a special case][special-case]] |
1145 #+srcname: task-last-value
1150 ***** last command for shells
1151 Do this using the =tee= shell command, and continually pipe the output
1154 Got this idea from the following [[http://linux.derkeiler.com/Mailing-Lists/Fedora/2004-01/0898.html][email-thread]].
1156 suggested from mailing list
1158 #+srcname: bash-save-last-output-to-file
1162 bash -c "$line" | tee /tmp/last.out1
1163 mv /tmp/last.out1 /tmp/last.out
1167 another proposed solution from the above thread
1169 #+srcname: bash-save-in-variable
1172 # so - Save Output. Saves output of command in OUT shell variable.
1182 "^M": " | tee /tmp/h_lastcmd.out ^[k"
1185 export __=/tmp/h_lastcmd.out
1187 If you try it, Alt-k will stand for the old Enter; use "command $__" to
1188 access the last output.
1194 Herculano de Lima Einloft Neto
1197 ***** emacs-lisp will be a special case
1198 While it is possible for emacs-lisp to be run in a console type
1199 environment (see the =elim= function) it is *not* possible to run
1200 emacs-lisp in a different *session*. Meaning any variable set top
1201 level of the console environment will be set *everywhere* inside
1202 emacs. For this reason I think that it doesn't make any sense to
1203 worry about session support for emacs-lisp.
1205 *** Further thoughts on 'scripting' vs. functional approaches
1207 These are just thoughts, I don't know how sure I am about this.
1208 And again, perhaps I'm not saying anything very radical, just that
1209 it would be nice to have some options supporting things like
1210 receiving text output in the org buffer.
1212 I can see that you've already gone some way down the road towards
1213 the 'last value' approach, so sorry if my comments come rather
1214 late. I am concerned that we are not giving sufficient attention
1215 to stdout / the text that is returned by the interpreters. In
1216 contrast, many of our potential users will be accustomed to a
1217 'scripting' approach, where they are outputting text at various
1218 points in the code block, not just at the end. I am leaning
1219 towards thinking that we should have 2 modes of evaluation:
1220 'script' mode, and 'functional' mode.
1222 In script mode, evaluation of a code block would result in *all*
1223 text output from that code block appearing as output in the org
1224 buffer, presumably as an #+begin_example...#+end_example. There
1225 could be an :echo option controlling whether the input commands
1226 also appear in the output. [This is like Sweave].
1228 In functional mode, the *result* of the code block is available as
1229 an elisp object, and may appear in the org buffer as an org
1230 table/string, via the mechanisms you have developed already.
1232 One thing I'm wondering about is whether, in script mode, there
1233 simply should not be a return value. Perhaps this is not so
1234 different from what exists: script mode would be new, and what
1235 exists currently would be functional mode.
1237 I think it's likely that, while code evaluation will be exciting
1238 to people, a large majority of our users in a large majority of
1239 their usage will not attempt to actually use the return value from
1240 a source code block in any meaningful way. In that case, it seems
1241 rather restrictive to only allow them to see output from the end
1244 Instead I think the most accessible way to introduce org-babel to
1245 people, at least while they are learning it, is as an immensely
1246 powerful environment in which to embed their 'scripts', which now
1247 also allows them to 'run' their 'scripts'. Especially as such
1248 people are likely to be the least capable of the user-base, a
1249 possible design-rule would be to make the scripting style of usage
1250 easy (default?), perhaps requiring a special option to enable a
1251 functional style. Those who will use the functional style won't
1252 have a problem understanding what's going on, whereas the 'skript
1253 kiddies' might not even know the syntax for defining a function in
1254 their language of choice. And of course we can allow the user to
1255 set a variable in their .emacs controlling the preference, so that
1256 functional users are not inconveniennced by having to provide
1257 header args the whole time.
1259 Please don't get the impression that I am down-valuing the
1260 functional style of org-babel. I am constantly horrified at the
1261 messy 'scripts' that my colleagues produce in perl or R or
1262 whatever! Nevertheless that seems to be how a lot of people work.
1264 I think you were leaning towards the last-value approach because
1265 it offered the possibility of unified code supporting both the
1266 single evaluation environment and the functional style. If you
1267 agree with any of the above then perhaps it will impact upon this
1268 and mean that the code in the two branches has to differ a bit. In
1269 that case, functional mode could perhaps after all evaluate each
1270 code block in its own environment, thus (re)approaching 'true'
1271 functional programming (side-effects are hard to achieve).
1275 echo "There are `wc -l files` files in this directory"
1279 *** even more thoughts on evaluation, results, models and options
1281 Thanks Dan, These comments are invaluable.
1283 What do you think about this as a new list of priorities/requirements
1284 for the execution of source-code blocks.
1287 1) we want the evaluation of the source code block to take place in a
1288 session which can persist state (variables, current directory,
1290 2) source code blocks can specify their session with a header argument
1291 3) each session should correspond to an Emacs comint buffer so that the
1292 user can drop into the session and experiment with live code
1295 1) each source-code block generates some form of results which (as
1296 we have already implemented) is transfered into emacs-lisp
1297 after which it can be inserted into the org-mode buffer, or
1298 used by other source-code blocks
1299 2) when the results are translated into emacs-lisp, forced to be
1300 interpreted as a scalar (dumping their raw values into the
1301 org-mode buffer), as a vector (which is often desirable with R
1302 code blocks), or interpreted on the fly (the default option).
1303 Note that this is very nearly currently implemented through the
1304 [[* DONE results-type header (vector/file)][results-type-header]].
1305 3) there should be *two* means of collecting results from the
1306 execution of a source code block. *Either* the value of the
1307 last statement of the source code block, or the collection of
1308 all that has been passed to STDOUT during the evaluation.
1310 **** header argument or return line (*header argument*)
1312 Rather than using a header argument to specify how the return value
1313 should be passed back, I'm leaning towards the use of a =#+RETURN=
1314 line inside the block. If such a line *is not present* then we
1315 default to using STDOUT to collect results, but if such a line *is
1316 present* then we use it's value as the results of the block. I
1317 think this will allow for the most elegant specification between
1318 functional and script execution. This also cleans up some issues
1319 of implementation and finding which statement is the last
1322 Having given this more thought, I think a header argument is
1323 preferable. The =#+return:= line adds new complicating syntax for
1324 something that does little more than we would accomplish through
1325 the addition of a header argument. The only benefit being that we
1326 know where the final statement starts, which is not an issue in
1327 those languages which contain 'last value' operators.
1329 new header =:results= arguments
1330 - script :: explicitly states that we want to use STDOUT to
1331 initialize our results
1332 - return_last :: stdout is ignored instead the *value* of the final
1333 statement in the block is returned
1334 - echo :: means echo the contents of the source-code block along
1335 with the results (this implies the *script* =:results=
1338 *** DONE rework evaluation lang-by-lang [4/4]
1340 This should include...
1341 - functional results working with the comint buffer
1343 - script :: return the output of STDOUT
1344 - write a macro which runs the first redirection, executes the
1345 body, then runs the second redirection
1346 - last :: return the value of the last statement
1349 - sessions in comint buffers
1351 **** DONE Ruby [4/4]
1352 - [X] functional results working with comint
1353 - [X] script results
1354 - [X] ensure scalar/vector results args are taken into consideration
1355 - [X] ensure callable by other source block
1357 #+srcname: ruby-use-last-output
1358 #+begin_src ruby :results replace
1365 #+resname: ruby-use-last-output
1368 #+srcname: task-call-use-last-output
1369 #+begin_src ruby :var last=ruby-use-last-output :results replace
1370 last.flatten.size + 1
1373 #+resname: task-call-use-last-output
1378 #+srcname: first-ruby-session-task
1379 #+begin_src ruby :session schulte :results silent
1383 #+srcname: second-ruby-session-task
1384 #+begin_src ruby :session schulte :results silent
1388 #+srcname: without-the-right-session
1389 #+begin_src ruby :results silent
1395 - [X] functional results working with comint
1396 - [X] script results
1397 - [X] ensure scalar/vector results args are taken into consideration
1398 - [X] ensure callable by other source block
1400 To redirect output to a file, you can use the =sink()= command.
1403 #+begin_src R :results value vector silent
1410 #+srcname: task-R-use-other-output
1411 #+begin_src R :var twoentyseven=task_R_B() :results replace value
1416 #+resname: task-R-use-other-output
1419 **** DONE Python [4/4]
1420 - [X] functional results working with comint
1421 - [X] script results
1422 - [X] ensure scalar/vector results args are taken into consideration
1423 - [X] ensure callable by other source block
1425 #+srcname: task-new-eval-for-python
1426 #+begin_src python :results silent output scalar
1432 #+srcname: task-use-new-eval
1433 #+begin_src python :var tasking=task-new-eval-for-python() :results replace
1437 #+resname: task-use-new-eval
1440 **** DONE Shells [4/4]
1441 - [X] functional results working with comint
1442 - [X] script results
1443 - [X] ensure scalar/vector results args are taken into consideration
1444 - [X] ensure callable by other source block
1446 #+srcname: task-shell-new-evaluation
1447 #+begin_src sh :results silent value scalar
1452 #+srcname: task-call-other-shell
1453 #+begin_src sh :var other=task-shell-new-evaluation() :results replace scalar
1454 echo $other ' is the old date'
1457 #+resname: task-call-other-shell
1458 : $ Fri Jun 12 13:08:37 PDT 2009 is the old date
1460 *** DONE implement a *session* header argument [4/4]
1461 =:session= header argument to override the default *session* buffer
1465 #+srcname: task-ruby-named-session
1466 #+begin_src ruby :session schulte :results replace
1467 schulte = :in_schulte
1470 #+resname: task-ruby-named-session
1473 #+srcname: another-in-schulte
1474 #+begin_src ruby :session schulte
1478 #+resname: another-in-schulte
1485 #+srcname: python-session-task
1486 #+begin_src python :session what :results silent
1490 #+srcname: python-get-from-session
1491 #+begin_src python :session what :results replace
1495 #+resname: python-get-from-session
1500 #+srcname: task-shell-sessions
1501 #+begin_src sh :session what
1505 #+srcname: task-shell-sessions-what
1506 #+begin_src sh :session what :results replace
1510 #+resname: task-shell-sessions-what
1515 #+srcname: task-R-session
1516 #+begin_src R :session what :results replace
1522 #+resname: task-R-session
1525 #+srcname: another-task-R-session
1526 #+begin_src R :session what :results replace
1530 *** DONE function to bring up inferior-process buffer [4/4]
1532 This should be callable from inside of a source-code block in an
1533 org-mode buffer. It should evaluate the header arguments, then bring
1534 up the inf-proc buffer using =pop-to-buffer=.
1536 For lack of a better place, lets add this to the `org-metadown-hook'
1539 To give this a try, place the cursor on a source block with variables,
1540 (optionally git a prefix argument) then hold meta and press down.
1544 #+srcname: task-ruby-pop-to-session
1545 #+begin_src ruby :var num=9 :var another="something else"
1546 num.times{|n| puts another}
1551 #+srcname: task-python-pop-to-session
1552 #+begin_src python :var num=9 :var another="something else"
1557 #+srcname: task-R-pop-to-session
1558 #+begin_src R :var a=9 :var b=8
1564 #+srcname: task-shell-pop-sessions
1565 #+begin_src sh :var NAME="eric"
1569 *** DEFERRED function to dump last N lines from inf-proc buffer into the current source block
1571 Callable with a prefix argument to specify how many lines should be
1572 dumped into the source-code buffer.
1574 *** REJECTED comint notes
1576 Implementing comint integration in [[file:lisp/org-babel-comint.el][org-babel-comint.el]].
1579 - handling of outputs
1580 - split raw output from process by prompts
1581 - a ring of the outputs, buffer-local, `org-babel-comint-output-ring'
1582 - a switch for dumping all outputs to a buffer
1583 - inputting commands
1585 Lets drop all this language specific stuff, and just use
1586 org-babel-comint to split up our outputs, and return either the last
1587 value of an execution or the combination of values from the
1590 **** comint filter functions
1591 : ;; comint-input-filter-functions hook process-in-a-buffer
1592 : ;; comint-output-filter-functions hook function modes.
1593 : ;; comint-preoutput-filter-functions hook
1594 : ;; comint-input-filter function ...
1596 #+srcname: obc-filter-ruby
1597 #+begin_src ruby :results last
1605 ** DONE Remove protective commas from # comments before evaluating
1606 org inserts protective commas in front of ## comments in language
1607 modes that use them. We need to remove them prior to sending code
1610 #+srcname: testing-removal-of-protective-comas
1612 ,# this one might break it??
1616 ** DONE pass multiple reference arguments into R
1617 Can we do this? I wasn't sure how to supply multiple 'var' header
1618 args. Just delete this if I'm being dense.
1620 This should be working, see the following example...
1622 #+srcname: two-arg-example
1623 #+begin_src R :var n=2 :var m=8
1627 #+resname: two-arg-example
1630 ** DONE ensure that table ranges work
1631 when a table range is passed to org-babel as an argument, it should be
1632 interpreted as a vector.
1635 | 2 | 3 | Fixnum:1 |
1636 | 3 | 4 | Array:123456 |
1640 #+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)))
1642 #+srcname: simple-sbe-example
1643 #+begin_src emacs-lisp
1647 #+srcname: task-table-range
1648 #+begin_src ruby :var n=simple-sbe-example
1652 #+srcname: simple-results
1653 #+begin_src emacs-lisp :var n=task-table-range(n=(1 2 3))
1657 #+resname: simple-results
1660 #+srcname: task-arr-referent
1661 #+begin_src ruby :var ar=(1 2 3)
1665 #+resname: task-arr-referent
1668 ** DONE global variable indicating default to vector output
1669 how about an alist... =org-babel-default-header-args= this may already
1670 exist... just execute the following and all source blocks will default
1673 #+begin_src emacs-lisp
1674 (setq org-babel-default-header-args '((:results . "vector")))
1677 ** DONE name named results if source block is named
1678 currently this isn't happening although it should be
1680 #+srcname: test-naming-named-source-blocks
1681 #+begin_src emacs-lisp
1685 #+resname: test-naming-named-source-blocks
1687 ** DONE (simple caching) check for named results before source blocks
1688 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]]
1689 ** DONE set =:results silent= when eval with prefix argument
1691 #+begin_src emacs-lisp
1694 ** DONE results-type header (vector/file) [3/3]
1695 In response to a point in Dan's email. We should allow the user to
1696 force scalar or vector results. This could be done with a header
1697 argument, and the default behavior could be controlled through a
1698 configuration variable.
1700 #+srcname: task-trivial-vector
1701 #+begin_src ruby :results replace vector
1708 since it doesn't make sense to turn a vector into a scalar, lets
1709 just add a two values...
1711 - vector :: forces the results to be a vector (potentially 1 dimensional)
1712 - file :: this throws an error if the result isn't a string, and
1713 tries to treat it as a path to a file.
1715 I'm just going to cram all of these into the =:results= header
1716 argument. Then if we allow multiple header arguments it should
1717 work out, for example one possible header argument string could be
1718 =:results replace vector file=, which would *replace* any existing
1719 results forcing the results into an org-mode table, and
1720 interpreting any strings as file paths.
1722 *** DONE multiple =:results= headers
1724 #+srcname: multiple-result-headers
1725 #+begin_src ruby :results replace silent
1731 *** DONE file result types
1732 When inserting into an org-mode buffer create a link with the path
1733 being the value, and optionally the display being the
1734 =file-name-nondirectory= if it exists.
1736 #+srcname: task-file-result
1737 #+begin_src python :results replace file
1742 [[something][something]]
1745 This will be useful because blocks like =ditaa= and =dot= can return
1746 the string path of their files, and can add =file= to their results
1749 *** DONE vector result types
1751 #+srcname: task-force-results
1752 #+begin_src emacs-lisp :results vector
1759 ** DONE results name
1760 In order to do this we will need to start naming our results.
1761 Since the source blocks are named with =#+srcname:= lines we can
1762 name results with =#+resname:= lines (if the source block has no
1763 name then no name is given to the =#+resname:= line on creation,
1764 otherwise the name of the source block is used).
1766 This will have the additional benefit of allowing results and
1767 source blocks to be located in different places in a buffer (and
1768 eventually in different buffers entirely).
1770 #+srcname: developing-resnames
1771 #+begin_src emacs-lisp :results silent
1775 Once source blocks are able to find their own =#+resname:= lines
1778 #+srcname: sbe-w-new-results
1779 #+begin_src emacs-lisp :results replace
1780 (sbe "developing-resnames")
1786 *** TODO change the results insertion functions to use these lines
1788 *** TODO teach references to resolve =#+resname= lines.
1790 ** DONE org-babel tests org-babel [1/1]
1791 since we are accumulating this nice collection of source-code blocks
1792 in the sandbox section we should make use of them as unit tests.
1793 What's more, we should be able to actually use org-babel to run these
1796 We would just need to cycle over every source code block under the
1797 sandbox, run it, and assert that the return value is equal to what we
1800 I have the feeling that this should be possible using only org-babel
1801 functions with minimal or no additional elisp. It would be very cool
1802 for org-babel to be able to test itself.
1804 This is now done, see [[* Tests]].
1806 *** DEFERRED org-babel assertions (may not be necessary)
1807 These could be used to make assertions about the results of a
1808 source-code block. If the assertion fails then the point could be
1809 moved to the block, and error messages and highlighting etc... could
1812 ** DONE make C-c C-c work anywhere within source code block?
1813 This seems like it would be nice to me, but perhaps it would be
1814 inefficient or ugly in implementation? I suppose you could search
1815 forward, and if you find #+end_src before you find #+begin_src,
1816 then you're inside one. [DED]
1818 Agreed, I think inside of the =#+srcname: line= would be useful as
1821 #+srcname: testing-out-cc
1822 #+begin_src emacs-lisp
1826 ** DONE integration with org tables
1827 We should make it easy to call org-babel source blocks from org-mode
1828 table formulas. This is practical now that it is possible to pass
1829 arguments to org-babel source blocks.
1831 See the related [[* (sandbox) integration w/org tables][sandbox]] header for tests/examples.
1833 *** digging in org-table.el
1834 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.
1836 Should be a hook in [[file:~/src/org/lisp/org-table.el::defun%20org%20table%20eval%20formula%20optional%20arg%20equation][org-table-eval-formula]].
1838 Looks like I need to change this [[file:~/src/org/lisp/org-table.el::if%20lispp][if statement]] (line 2239) into a cond
1841 ** DONE source blocks as functions
1843 Allow source code blocks to be called like functions, with arguments
1844 specified. We are already able to call a source-code block and assign
1845 it's return result to a variable. This would just add the ability to
1846 specify the values of the arguments to the source code block assuming
1847 any exist. For an example see
1849 When a variable appears in a header argument, how do we differentiate
1850 between it's value being a reference or a literal value? I guess this
1851 could work just like a programming language. If it's escaped or in
1852 quotes, then we count it as a literal, otherwise we try to look it up
1855 ** DONE folding of code blocks? [2/2]
1856 [DED] In similar way to using outline-minor-mode for folding function
1857 bodies, can we fold code blocks? #+begin whatever statements are
1858 pretty ugly, and in any case when you're thinking about the overall
1859 game plan you don't necessarily want to see the code for each Step.
1861 *** DONE folding of source code block
1862 Sounds good, and wasn't too hard to implement. Code blocks should
1863 now be fold-able in the same manner as headlines (by pressing TAB
1866 *** REJECTED folding of results
1867 So, lets do a three-stage tab cycle... First fold the src block,
1868 then fold the results, then unfold.
1870 There's no way to tell if the results are a table or not w/o
1871 actually executing the block which would be too expensive of an
1874 ** DONE selective export of text, code, figures
1875 [DED] The org-babel buffer contains everything (code, headings and
1876 notes/prose describing what you're up to, textual/numeric/graphical
1877 code output, etc). However on export to html / LaTeX one might want
1878 to include only a subset of that content. For example you might
1879 want to create a presentation of what you've done which omits the
1882 [EMS] So I think this should be implemented as a property which can
1883 be set globally or on the outline header level (I need to review
1884 the mechanics of org-mode properties). And then as a source block
1885 header argument which will apply only to a specific source code
1886 block. A header argument of =:export= with values of
1888 - =code= :: just show the code in the source code block
1889 - =none= :: don't show the code or the results of the evaluation
1890 - =results= :: just show the results of the code evaluation (don't
1891 show the actual code)
1892 - =both= :: show both the source code, and the results
1894 this will be done in [[* (sandbox) selective export][(sandbox) selective export]].
1896 ** DONE a header argument specifying silent evaluation (no output)
1897 This would be useful across all types of source block. Currently
1898 there is a =:replace t= option to control output, this could be
1899 generalized to an =:output= option which could take the following
1900 options (maybe more)
1902 - =t= :: this would be the default, and would simply insert the
1903 results after the source block
1904 - =replace= :: to replace any results which may already be there
1905 - =silent= :: this would inhibit any insertion of the results
1907 This is now implemented see the example in the [[* silent evaluation][sandbox]]
1909 ** DONE assign variables from tables in R
1910 This is now working (see [[* (sandbox table) R][(sandbox-table)-R]]). Although it's not that
1911 impressive until we are able to print table results from R.
1913 ** DONE insert 2-D R results as tables
1914 everything is working but R and shell
1920 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
1921 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]]
1922 as a means of converting multidimensional R objects to emacs lisp.
1924 It may be as simple as first checking if the data is multidimensional,
1925 and then, if so using =write= to write the data out to a temporary
1926 file from which emacs can read the data in using =org-table-import=.
1928 Looking into this further, is seems that there is no such thing as a
1929 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
1930 deal with trivial vectors (scalars) in R. I'm tempted to just treat
1931 them as vectors, but then that would lead to a proliferation of
1932 trivial 1-cell tables...
1934 ** DONE allow variable initialization from source blocks
1935 Currently it is possible to initialize a variable from an org-mode
1936 table with a block argument like =table=sandbox= (note that the
1937 variable doesn't have to named =table=) as in the following example
1943 #+begin_src emacs-lisp :var table=sandbox :results replace
1944 (message (format "table = %S" table))
1947 : "table = ((1 2 3) (4 \"schulte\" 6))"
1949 It would be good to allow initialization of variables from the results
1950 of other source blocks in the same manner. This would probably
1951 require the addition of =#+SRCNAME: example= lines for the naming of
1952 source blocks, also the =table=sandbox= syntax may have to be expanded
1953 to specify whether the target is a source code block or a table
1954 (alternately we could just match the first one with the given name
1955 whether it's a table or a source code block).
1957 At least initially I'll try to implement this so that there is no need
1958 to specify whether the reference is to a table or a source-code block.
1959 That seems to be simpler both in terms of use and implementation.
1961 This is now working for emacs-lisp, ruby and python (and mixtures of
1962 the three) source blocks. See the examples in the [[* (sandbox) referencing other source blocks][sandbox]].
1964 This is currently working only with emacs lisp as in the following
1965 example in the [[* emacs lisp source reference][emacs lisp source reference]].
1968 ** TODO Add languages [1/6]
1969 I'm sure there are many more that aren't listed here. Please add
1970 them, and bubble any that you particularly care about up to the top.
1972 Any new language should be implemented in a org-babel-lang.el file.
1973 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
1974 [[file:lisp/org-babel-R.el][org-babel-R.el]].
1977 This could probably be added to [[file:lisp/org-babel-script.el][org-babel-script.el]]
1982 (see [[* file result types][file result types]])
1984 #+srcname: implementing-ditaa
1985 #+begin_src ditaa :results replace :file blue.png :cmdline -r
1995 #+resname: implementing-ditaa
1996 [[file:blue.png][blue.png]]
1999 (see [[* file result types][file result types]])
2001 - a =file= header argument
2002 - a =cmdline= header argument
2004 - scalar variables should be replaced in the body of the gnuplot code
2005 - vector variables should be exported to tab-separated files, and
2006 the variable names should be replaced with the path to the files
2008 #+PLOT: title:"Citas" ind:1 deps:(3) type:2d with:histograms set:"yrange [0:]"
2009 #+TBLNAME: gnuplot-data
2010 | independent var | first dependent var | second dependent var |
2011 |-----------------+---------------------+----------------------|
2012 | 0.1 | 0.425 | 0.375 |
2013 | 0.2 | 0.3125 | 0.3375 |
2014 | 0.3 | 0.24999993 | 0.28333338 |
2015 | 0.4 | 0.275 | 0.28125 |
2016 | 0.5 | 0.26 | 0.27 |
2017 | 0.6 | 0.25833338 | 0.24999993 |
2018 | 0.7 | 0.24642845 | 0.23928553 |
2019 | 0.8 | 0.23125 | 0.2375 |
2020 | 0.9 | 0.23333323 | 0.2333332 |
2021 | 1 | 0.2225 | 0.22 |
2022 | 1.1 | 0.20909075 | 0.22272708 |
2023 | 1.2 | 0.19999998 | 0.21458333 |
2024 | 1.3 | 0.19615368 | 0.21730748 |
2026 #+srcname: implementing-gnuplot
2027 #+begin_src gnuplot :var data=gnuplot-data
2028 set title "Implementing Gnuplot"
2029 plot "data" using 1:2 with lines
2033 (see [[* file result types][file result types]])
2036 (see [[* file result types][file result types]])
2040 ** TODO o-b-execute-subtree overwrites heading when subtree is folded
2042 Try M-x org-babel-execute-subtree with the subtree folded and
2043 point at the beginning of the heading line.
2048 ** TODO Allow source blocks to be recognised when #+ are not first characters on the line
2049 I think Carsten has recently altered the core so that #+ can have
2050 preceding whitespace, at least for literal/code examples. org-babel
2051 should support this.
2053 ** TODO non-orgtbl formatted lists
2056 #+srcname: this-doesn't-match-orgtbl
2057 #+begin_src emacs-lisp :results replace
2058 '((:results . "replace"))
2061 #+resname: this-doesn't-match-orgtbl
2063 ** TODO collapsing consecutive newlines in string output
2065 #+srcname: multi-line-string-output
2066 #+begin_src ruby :results output
2067 "the first line ends here
2070 and this is the second one
2075 #+resname: multi-line-string-output
2077 ** TODO cursor movement when evaluating source blocks
2078 E.g. the pie chart example. Despite the save-window-excursion in
2079 org-babel-execute:R. (I never learned how to do this properly: org-R
2080 jumps all over the place...)
2082 ** PROPOSED external shell execution can't isolate return values
2083 I have no idea how to do this as of yet. The result is that when
2084 shell functions are run w/o a session there is no difference between
2085 the =output= and =value= result arguments.
2087 Yea, I don't know how to do this either. I searched extensively on
2088 how to isolate the *last* output of a series of shell commands (see
2089 [[* last command for
2090 shells][last command for shells]]). The results of the search were basically
2091 that it was not possible (or at least not accomplish-able with a
2092 reasonable amount of effort).
2094 That fact combined with the tenancy to all ways use standard out in
2095 shell scripts led me to treat these two options (=output= and =value=)
2096 as identical in shell evaluation. Not ideal but maybe good enough for
2099 In the `results' branch I've changed this so that they're not quite
2100 identical: output results in raw stdout contents, whereas value
2101 converts it to elisp, perhaps to a table if it looks tabular. This is
2102 the same for the other languages. [Dan]
2104 ** TODO are the org-babel-trim s necessary?
2105 at the end of e.g. org-babel-R-evaluate, org-babel-python-evaluate, but
2106 not org-babel-ruby-evaluate
2107 ** TODO elisp reference fails for literal number
2108 #+srcname: elisp-test(a=4)
2109 #+begin_src emacs-lisp
2112 ** TODO use new merge function [[file:lisp/org-babel-ref.el::t%20nil%20org%20combine%20plists%20args%20nil][here]]?
2113 And at other occurrences of org-combine-plists?
2114 ** TODO LoB: with output to buffer, not working in buffers other than library-of-babel.org
2115 I haven't fixed this yet. org-babel-ref-resolve-reference moves
2116 point around, inside a save-excursion. Somehow when it comes to
2117 inserting the results (after possible further recursive calls to
2118 org-babel-ref-resolve-reference), point hasn't gone back to the
2120 ** TODO creeping blank lines
2121 There's still inappropriate addition of blank lines in some circumstances. E.g.
2127 Compare the results of
2128 #+lob: python-add(a=5, b=17)
2130 #+resname: python-add(a=5, b=17)
2132 --------------------------------
2140 ---------------------
2142 Hmm, it's a bit confusing. I think it's to do with the fact that
2143 LoB removes the entire (#+resname and result) and starts from
2144 scratch, whereas #+begin_src only removes the result. I haven't
2145 worked out what the correct fix is yet.
2146 ** DEFERRED weird escaped characters in shell prompt break shell evaluation
2147 E.g. this doesn't work. Should the shell sessions set a sane prompt
2148 when they start up? Or is it a question of altering
2149 comint-prompt-regexp? Or altering org-babel regexps?
2152 black=30 ; red=31 ; green=32 ; yellow=33 ; blue=34 ; magenta=35 ; cyan=36 ; white=37
2155 export PS1="\[\033[${prompt_col}m\]\w${prompt_char} \[\033[0m\]"
2158 I just pushed a good amount of changes, could you see if your shell
2159 problems still exist?
2161 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
2162 org-babel-sh.el, raw gets the value
2164 ("" "
\e[0m Sun Jun 14 19:26:24 EDT 2009\n" "
\e[0m org_babel_sh_eoe\n" "
\e[0m ")
2166 and therefore (member org-babel-sh-eoe-output ...) fails
2168 I think that `comint-prompt-regexp' needs to be altered to match
2169 the shell prompt. This shouldn't be too difficult to do by hand,
2170 using the `regexp-builder' command and should probably be part of
2171 the user's regular emacs init. I can't think of a way for us to
2172 set this automatically, and we are SOL without a regexp to match
2175 ** DONE LoB: calls fail if reference has single character name
2176 commit 21d058869df1ff23f4f8cc26f63045ac9c0190e2
2177 **** This doesn't work
2178 #+lob: R-plot(data=X)
2197 #+lob: R-plot(data=XX)
2199 ** DONE make :results replace the default?
2200 I'm tending to think that appending results to pre-existing results
2201 creates mess, and that the cleaner `replace' option should be the
2202 default. E.g. when a source block creates an image, we would want
2203 that to be updated, rather than have a new one be added.
2207 ** DONE ruby evaluation not working under ubuntu emacs 23
2208 With emacs 23.0.91.1 on ubuntu, for C-h f run-ruby I have the
2209 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.
2212 run-ruby is an interactive compiled Lisp function.
2216 Run an inferior Ruby process, input and output via buffer *ruby*.
2217 If there is a process already running in `*ruby*', switch to that buffer.
2218 With argument, allows you to edit the command line (default is value
2219 of `ruby-program-name'). Runs the hooks `inferior-ruby-mode-hook'
2220 (after the `comint-mode-hook' is run).
2221 (Type C-h m in the process buffer for a list of commands.)
2224 So, I may have a non-standard inf-ruby.el. Here's my version of
2228 run-ruby is an interactive Lisp function in `inf-ruby.el'.
2230 (run-ruby &optional COMMAND NAME)
2232 Run an inferior Ruby process, input and output via buffer *ruby*.
2233 If there is a process already running in `*ruby*', switch to that buffer.
2234 With argument, allows you to edit the command line (default is value
2235 of `ruby-program-name'). Runs the hooks `inferior-ruby-mode-hook'
2236 (after the `comint-mode-hook' is run).
2237 (Type C-h m in the process buffer for a list of commands.)
2240 It seems we could either bundle my version of inf-ruby.el (as it's
2241 the newest). Or we could change the use of `run-ruby' so that it
2242 is robust across multiple distributions. I think I'd prefer the
2243 former, unless the older version of inf-ruby is actually bundled
2244 with emacs, in which case maybe we should go out of our way to
2245 support it. Thoughts?
2247 I think for now I'll just include the latest [[file:util/inf-ruby.el][inf-ruby.el]] in the
2248 newly created utility directory. I doubt anyone would have a
2249 problem using the latest version of this file.
2250 ** DONE test failing forcing vector results with =test-forced-vector-results= ruby code block
2251 Note that this only seems to happen the *second* time the test table
2254 #+srcname: bug-trivial-vector
2255 #+begin_src emacs-lisp :results vector silent
2259 #+srcname: bug-forced-vector-results
2260 #+begin_src ruby :var triv=test-trivial-vector :results silent
2264 mysteriously this seems to be fixed...
2265 ** DONE defunct R sessions
2266 Sometimes an old R session will turn defunct, and newly inserted code
2267 will not be evaluated (leading to a hang).
2269 This seems to be fixed by using `inferior-ess-send-input' rather than `comint-send-input'.
2270 ** DONE ruby fails on first call to non-default session
2272 #+srcname: bug-new-session
2273 #+begin_src ruby :session is-new
2277 ** DONE when reading results from =#+resname= line
2279 Errors when trying to read from resname lines.
2281 #+resname: bug-in-resname
2284 #+srcname: bug-in-resname-reader
2285 #+begin_src emacs-lisp :var buggy=bug-in-resname() :results silent
2289 ** DONE R-code broke on "org-babel" rename
2291 #+srcname: bug-R-babels
2296 ** DONE error on trivial R results
2298 So I know it's generally not a good idea to squash error without
2299 handling them, but in this case the error almost always means that
2300 there was no file contents to be read by =org-table-import=, so I
2303 #+srcname: bug-trivial-r1
2304 #+begin_src R :results replace
2305 pie(c(1, 2, 3), labels = c(1, 2, 3))
2308 #+srcname: bug-trivial-r2
2309 #+begin_src R :results replace
2313 #+resname: bug-trivial-r2
2316 #+srcname: bug-trivial-r3
2317 #+begin_src R :results replace
2321 #+resname: bug-trivial-r3
2326 ** DONE ruby new variable creation (multi-line ruby blocks)
2327 Actually it looks like we were dropping all but the last line.
2329 #+srcname: multi-line-ruby-test
2330 #+begin_src ruby :var table=bug-numerical-table :results replace
2332 table.each{|n| total += n}
2339 ** DONE R code execution seems to choke on certain inputs
2340 Currently the R code seems to work on vertical (but not landscape)
2343 #+srcname: little-fake
2344 #+begin_src emacs-lisp
2349 #+begin_src R :var num=little-fake
2356 #+srcname: set-debug-on-error
2357 #+begin_src emacs-lisp :results silent
2358 (setq debug-on-error t)
2361 #+srcname: bug-numerical-table
2362 #+begin_src emacs-lisp :results silent
2369 #+srcname: bug-R-number-evaluation
2370 #+begin_src R :var table=bug-numerical-table
2379 #+tblname: bug-vert-table
2384 #+srcname: bug-R-vertical-table
2385 #+begin_src R :var table=bug-vert-table :results silent
2389 ** DONE org bug/request: prevent certain org behaviour within code blocks
2390 E.g. [[]] gets recognised as a link (when there's text inside the
2391 brackets). This is bad for R code at least, and more generally
2392 could be argued to be inappropriate. Is it difficult to get org to
2393 ignore text in code blocks? [DED]
2395 I believe Carsten addressed this recently on the mailing list with
2396 the comment that it was indeed a difficult issue. I believe this
2397 may be one area where we could wait for an upstream (org-mode) fix.
2399 [Dan] Carsten has fixed this now in the core.
2401 ** DONE with :results replace, non-table output doesn't replace table output
2402 And vice versa. E.g. Try this first with table and then with len(table) [DED]
2403 #+begin_src python :var table=sandbox :results replace
2408 | 4 | "schulte" | 6 |
2411 Yes, this is certainly a problem. I fear that if we begin replacing
2412 anything immediately following a source block (regardless of whether
2413 it matches the type of our current results) we may accidentally delete
2414 hand written portions of the user's org-mode buffer.
2416 I think that the best solution here would be to actually start
2417 labeling results with a line that looks something like...
2421 This would have a couple of benefits...
2422 1) we wouldn't have to worry about possibly deleting non-results
2423 (which is currently an issue)
2424 2) we could reliably replace results even if there are different types
2425 3) we could reference the results of a source-code block in variable
2426 definitions, which would be useful if for example we don't wish to
2427 re-run a source-block every time because it is long-running.
2429 Thoughts? If no-one objects, I believe I will implement the labeling
2432 ** DONE extra quotes for nested string
2433 Well R appears to be reading the tables without issue...
2435 these *should* be quoted
2437 #+begin_src sh :results replace
2442 | "README.markdown" |
2445 | "existing_tools" |
2449 | "test-export.html" |
2450 | "test-export.org" |
2452 #+srcname: test-quotes
2453 #+begin_src ruby :var tab=ls
2459 #+srcname: test-quotes
2460 #+begin_src R :var tab=ls
2466 ** DONE simple ruby arrays not working
2468 As an example eval the following. Adding a line to test
2470 #+tblname: simple-ruby-array
2473 #+srcname: ruby-array-test
2474 #+begin_src ruby :var ar = simple-ruby-array :results silent
2478 ** DONE space trailing language name
2479 fix regexp so it works when there's a space trailing the language name
2481 #+srcname: test-trailing-space
2486 ** DONE Args out of range error
2488 The following block resulted in the error below [DED]. It ran without
2489 error directly in the shell.
2492 for platf in ill aff ; do
2493 for pop in CEU YRI ASI ; do
2494 rm -f $platf/hapmap-genos-$pop-all $platf/hapmap-rs-all
2495 cat $platf/hapmap-genos-$pop-* > $platf/hapmap-genos-$pop-all
2496 cat $platf/hapmap-rs-* > $platf/hapmap-rs-all
2501 executing source block with sh...
2502 finished executing source block
2503 string-equal: Args out of range: "", -1, 0
2505 the error =string-equal: Args out of range: "", -1, 0= looks like what
2506 used to be output when the block returned an empty results string.
2507 This should be fixed in the current version, you should now see the
2508 following message =no result returned by source block=.
2510 ** DONE ruby arrays not recognized as such
2512 Something is wrong in [[file:lisp/org-babel-script.el]] related to the
2513 recognition of ruby arrays as such.
2515 #+begin_src ruby :results replace
2521 #+begin_src python :results replace
2533 Evaluate all the cells in this table for a comprehensive test of the
2534 org-babel functionality.
2536 *Note*: if you have customized =org-babel-default-header-args= then some
2537 of these tests may fail.
2539 #+TBLNAME: org-babel-tests
2540 | functionality | block | arg | expected | results | pass |
2541 |-------------------------+----------------------------+-----+-------------+-------------+------|
2542 | basic evaluation | | | | | pass |
2543 |-------------------------+----------------------------+-----+-------------+-------------+------|
2544 | emacs lisp | basic-elisp | | 5 | 5 | pass |
2545 | shell | basic-shell | | 6 | 6 | pass |
2546 | ruby | basic-ruby | | org-babel | org-babel | pass |
2547 | python | basic-python | | hello world | hello world | pass |
2548 | R | basic-R | | 13 | 13 | pass |
2549 |-------------------------+----------------------------+-----+-------------+-------------+------|
2550 | tables | | | | | pass |
2551 |-------------------------+----------------------------+-----+-------------+-------------+------|
2552 | emacs lisp | table-elisp | | 3 | 3 | pass |
2553 | ruby | table-ruby | | 1-2-3 | 1-2-3 | pass |
2554 | python | table-python | | 5 | 5 | pass |
2555 | R | table-R | | 3.5 | 3.5 | pass |
2556 |-------------------------+----------------------------+-----+-------------+-------------+------|
2557 | source block references | | | | | pass |
2558 |-------------------------+----------------------------+-----+-------------+-------------+------|
2559 | all languages | chained-ref-last | | Array | Array | pass |
2560 |-------------------------+----------------------------+-----+-------------+-------------+------|
2561 | source block functions | | | | | pass |
2562 |-------------------------+----------------------------+-----+-------------+-------------+------|
2563 | emacs lisp | defun-fibb | | fibbd | fibbd | pass |
2564 | run over | Fibonacci | 0 | 1 | 1 | pass |
2565 | a | Fibonacci | 1 | 1 | 1 | pass |
2566 | variety | Fibonacci | 2 | 2 | 2 | pass |
2567 | of | Fibonacci | 3 | 3 | 3 | pass |
2568 | different | Fibonacci | 4 | 5 | 5 | pass |
2569 | arguments | Fibonacci | 5 | 8 | 8 | pass |
2570 |-------------------------+----------------------------+-----+-------------+-------------+------|
2571 | bugs and tasks | | | | | pass |
2572 |-------------------------+----------------------------+-----+-------------+-------------+------|
2573 | simple ruby arrays | ruby-array-test | | 3 | 3 | pass |
2574 | R number evaluation | bug-R-number-evaluation | | 2 | 2 | pass |
2575 | multi-line ruby blocks | multi-line-ruby-test | | 2 | 2 | pass |
2576 | forcing vector results | test-forced-vector-results | | Array | Array | pass |
2577 |-------------------------+----------------------------+-----+-------------+-------------+------|
2578 | sessions | | | | | pass |
2579 |-------------------------+----------------------------+-----+-------------+-------------+------|
2580 | set ruby session | set-ruby-session-var | | :set | :set | pass |
2581 | get from ruby session | get-ruby-session-var | | 3 | 3 | pass |
2582 | set python session | set-python-session-var | | set | set | pass |
2583 | get from python session | get-python-session-var | | 4 | 4 | pass |
2584 | set R session | set-R-session-var | | set | set | pass |
2585 | get from R session | get-R-session-var | | 5 | 5 | pass |
2586 #+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))
2587 #+TBLFM: $5=""::$6=""
2589 The second line (followed by replacing '[]' with '') can be used to blank out the table results, in the absence of a better method.
2593 #+srcname: basic-elisp
2594 #+begin_src emacs-lisp :results silent
2599 #+srcname: basic-shell
2600 #+begin_src sh :results silent
2605 #+srcname: date-simple
2606 #+begin_src sh :results silent
2610 #+srcname: basic-ruby
2611 #+begin_src ruby :results silent
2616 #+srcname: basic-python
2617 #+begin_src python :results silent
2623 #+begin_src R :results silent
2631 #+tblname: test-table
2635 #+srcname: table-elisp
2636 #+begin_src emacs-lisp :results silent :var table=test-table
2637 (length (car table))
2641 #+srcname: table-ruby
2642 #+begin_src ruby :results silent :var table=test-table
2643 table.first.join("-")
2647 #+srcname: table-python
2648 #+begin_src python :var table=test-table
2654 #+begin_src R :var table=test-table
2661 Lets pass a references through all of our languages...
2663 Lets start by reversing the table from the previous examples
2665 #+srcname: chained-ref-first
2666 #+begin_src python :var table = test-table
2671 #+resname: chained-ref-first
2675 Take the first part of the list
2677 #+srcname: chained-ref-second
2678 #+begin_src R :var table = chained-ref-first
2682 #+resname: chained-ref-second
2686 Turn the numbers into string
2688 #+srcname: chained-ref-third
2689 #+begin_src emacs-lisp :var table = chained-ref-second
2690 (mapcar (lambda (el) (format "%S" el)) table)
2693 #+resname: chained-ref-third
2696 and Check that it is still a list
2698 #+srcname: chained-ref-last
2699 #+begin_src ruby :var table=chained-ref-third
2704 ** source blocks as functions
2706 #+srcname: defun-fibb
2707 #+begin_src emacs-lisp :results silent
2708 (defun fibbd (n) (if (< n 2) 1 (+ (fibbd (- n 1)) (fibbd (- n 2)))))
2712 #+srcname: fibonacci
2713 #+begin_src emacs-lisp :results silent :var n=7
2723 ** sbe tests (these don't seem to be working...)
2724 Testing the insertion of results into org-mode tables.
2726 #+srcname: multi-line-output
2727 #+begin_src ruby :results replace
2728 "the first line ends here
2731 and this is the second one
2737 : the first line ends here\n\n\n and this is the second one\n\neven a third
2739 #+srcname: multi-line-error
2740 #+begin_src ruby :results replace
2741 raise "oh nooooooooooo"
2747 | the first line ends here... | -:5: warning: parenthesize argument(s) for future version... |
2748 #+TBLFM: $1='(sbe "multi-line-output")::$2='(sbe "multi-line-error")
2750 ** forcing results types tests
2752 #+srcname: test-trivial-vector
2753 #+begin_src emacs-lisp :results vector silent
2757 #+srcname: test-forced-vector-results
2758 #+begin_src ruby :var triv=test-trivial-vector :results silent
2764 #+srcname: set-ruby-session-var
2765 #+begin_src ruby :session :results silent
2770 #+srcname: get-ruby-session-var
2771 #+begin_src ruby :session :results silent
2775 #+srcname: set-python-session-var
2776 #+begin_src python :session
2781 #+srcname: get-python-session-var
2782 #+begin_src python :session
2786 #+srcname: set-R-session-var
2787 #+begin_src R :session
2792 #+srcname: get-R-session-var
2793 #+begin_src R :session
2802 To run these examples evaluate [[file:lisp/org-babel-init.el][org-babel-init.el]]
2804 ** org-babel.el beginning functionality
2806 #+begin_src sh :results replace
2811 : Sun Jul 5 18:54:39 EDT 2009
2818 : Sun Jul 05 18:54:35 -0400 2009
2830 #+begin_src R :results replace
2840 hist(rgamma(20,3,3))
2845 ** org-babel plays with tables
2846 Alright, this should demonstrate both the ability of org-babel to read
2847 tables into a lisp source code block, and to then convert the results
2848 of the source code block into an org table. It's using the classic
2849 "lisp is elegant" demonstration transpose function. To try this
2852 1. evaluate [[file:lisp/org-babel-init.el]] to load org-babel and friends
2853 2. evaluate the transpose definition =\C-c\\C-c= on the beginning of
2855 3. evaluate the next source code block, this should read in the table
2856 because of the =:var table=previous=, then transpose the table, and
2857 finally it should insert the transposed table into the buffer
2858 immediately following the block
2862 #+begin_src emacs-lisp :results silent
2863 (defun transpose (table)
2864 (apply #'mapcar* #'list table))
2872 #+begin_src emacs-lisp :var table=sandbox :results replace
2877 #+begin_src emacs-lisp
2882 | 1 | 2 | 3 | 4 | 5 |
2886 #+begin_src ruby :var table=sandbox :results replace
2887 table.first.join(" - ")
2893 #+begin_src python :var table=sandbox
2898 #+begin_src ruby :var table=sandbox :results replace
2903 : [[1, 2, 3], [4, "schulte", 6]]
2907 | 4 | "schulte" | 6 |
2909 #+begin_src python :var table=sandbox :results replace
2915 | "__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" |
2917 *** (sandbox table) R
2919 #+TBLNAME: sandbox_r
2923 #+begin_src R :results replace
2924 x <- c(rnorm(10, mean=-3, sd=1), rnorm(10, mean=3, sd=1))
2928 | -3.35473133869346 |
2930 | -3.32819924928633 |
2931 | -2.97310212756194 |
2932 | -2.09640758369576 |
2933 | -5.06054014378736 |
2934 | -2.20713700711221 |
2935 | -1.37618039712037 |
2936 | -1.95839385821742 |
2937 | -3.90407396475502 |
2938 | 2.51168071590226 |
2939 | 3.96753011570494 |
2940 | 3.31793212627865 |
2941 | 1.99829753972341 |
2942 | 4.00403686419829 |
2943 | 4.63723764452927 |
2944 | 3.94636744261313 |
2945 | 3.58355906547775 |
2946 | 3.01563442274226 |
2949 #+begin_src R var tabel=sandbox_r :results replace
2954 | 4 | "schulte" | 6 |
2957 Now shell commands are converted to tables using =org-table-import=
2958 and if these tables are non-trivial (i.e. have multiple elements) then
2959 they are imported as org-mode tables...
2961 #+begin_src sh :results replace
2965 | "total" | 208 | "" | "" | "" | "" | "" | "" |
2966 | "-rw-r--r--" | 1 | "dan" | "dan" | 57 | 2009 | 15 | "block" |
2967 | "-rw-r--r--" | 1 | "dan" | "dan" | 35147 | 2009 | 15 | "COPYING" |
2968 | "-rw-r--r--" | 1 | "dan" | "dan" | 722 | 2009 | 18 | "examples.org" |
2969 | "drwxr-xr-x" | 4 | "dan" | "dan" | 4096 | 2009 | 19 | "existing_tools" |
2970 | "-rw-r--r--" | 1 | "dan" | "dan" | 2207 | 2009 | 14 | "intro.org" |
2971 | "drwxr-xr-x" | 2 | "dan" | "dan" | 4096 | 2009 | 18 | "org-babel" |
2972 | "-rw-r--r--" | 1 | "dan" | "dan" | 277 | 2009 | 20 | "README.markdown" |
2973 | "-rw-r--r--" | 1 | "dan" | "dan" | 11837 | 2009 | 18 | "rorg.html" |
2974 | "-rw-r--r--" | 1 | "dan" | "dan" | 61829 | 2009 | 19 | "#rorg.org#" |
2975 | "-rw-r--r--" | 1 | "dan" | "dan" | 60190 | 2009 | 19 | "rorg.org" |
2976 | "-rw-r--r--" | 1 | "dan" | "dan" | 972 | 2009 | 11 | "test-export.org" |
2979 ** silent evaluation
2987 #+begin_src ruby :results silent
2991 #+begin_src ruby :results replace
2998 ** (sandbox) referencing other source blocks
2999 Doing this in emacs-lisp first because it's trivial to convert
3000 emacs-lisp results to and from emacs-lisp.
3002 *** emacs lisp source reference
3003 This first example performs a calculation in the first source block
3004 named =top=, the results of this calculation are then saved into the
3005 variable =first= by the header argument =:var first=top=, and it is
3006 used in the calculations of the second source block.
3009 #+begin_src emacs-lisp
3013 #+begin_src emacs-lisp :var first=top :results replace
3019 This example is the same as the previous only the variable being
3020 passed through is a table rather than a number.
3022 #+begin_src emacs-lisp :results silent
3023 (defun transpose (table)
3024 (apply #'mapcar* #'list table))
3027 #+TBLNAME: top_table
3031 #+SRCNAME: second_src_example
3032 #+begin_src emacs-lisp :var table=top_table
3036 #+begin_src emacs-lisp :var table=second_src_example :results replace
3041 | 4 | "schulte" | 6 |
3043 Now working for ruby
3050 #+begin_src ruby :var other=start :results replace
3056 #+SRCNAME: start_two
3061 #+begin_src python :var another=start_two :results replace
3066 Since all variables are converted into Emacs Lisp it is no problem to
3067 reference variables specified in another language.
3069 #+SRCNAME: ruby-block
3074 #+SRCNAME: lisp_block
3075 #+begin_src emacs-lisp :var ruby-variable=ruby-block
3079 #+begin_src python :var lisp_var=lisp_block
3088 #+begin_src R :results replace
3095 #+begin_src R :var other=first_r :results replace
3102 ** (sandbox) selective export
3104 For exportation tests and examples see (including exportation of
3105 inline source code blocks) [[file:test-export.org]]
3108 ** (sandbox) source blocks as functions
3111 #+begin_src emacs-lisp :results silent
3116 #+begin_src emacs-lisp :var n=default :results replace
3122 #+begin_src emacs-lisp :var result=triple(n=3, m=98) :results replace
3128 The following just demonstrates the ability to assign variables to
3129 literal values, which was not implemented until recently.
3131 #+begin_src ruby :var num="eric" :results replace
3138 ** (sandbox) inline source blocks
3140 This is an inline source code block src_ruby{1 + 6}. And another
3141 source block with text output src_emacs-lisp{"eric"}.
3143 This is an inline source code block with header
3144 arguments. src_ruby[:var n=fibbd( n = 0 )]{n}
3147 ** (sandbox) integration w/org tables
3149 #+begin_src emacs-lisp :results silent
3150 (defun fibbd (n) (if (< n 2) 1 (+ (fibbd (- n 1)) (fibbd (- n 2)))))
3154 #+begin_src emacs-lisp :var n=4 :results silent
3158 #+begin_src emacs-lisp :results silent
3159 (mapcar #'fibbd '(0 1 2 3 4 5 6 7 8))
3162 Something is not working here. The function `sbe ' works fine when
3163 called from outside of the table (see the source block below), but
3164 produces an error when called from inside the table. I think there
3165 must be some narrowing going on during intra-table emacs-lisp
3168 | original | fibbd |
3169 |----------+-------|
3180 #+TBLFM: $2='(sbe "fibbd" (n $1))
3184 #+begin_src emacs-lisp :results silent
3185 (sbe 'fibbd (n "8"))
3190 LocalWords: DBlocks dblocks org-babel el eric fontification