1 #+TITLE: org-babel --- facilitating communication between programming languages and people
2 #+SEQ_TODO: PROPOSED TODO STARTED | DONE DEFERRED REJECTED
3 #+OPTIONS: H:3 num:nil toc:t
4 #+STARTUP: oddeven hideblocks
6 Through Org-Babel Org-Mode can communicate with programming languages.
7 Code contained in source-code blocks can be evaluated and data can
8 pass seamlessly between different programming languages, Org-Mode
9 constructs (tables, file links, example text) and interactive comint
13 - The [[* Introduction][Introduction]] :: provides a brief overview of the design and use
14 of Org-Babel including tutorials and examples.
15 - In [[* Getting started][Getting Started]] :: find instructions for installing org-babel
16 into your emacs configuration.
17 - The [[* Tasks][Tasks]] :: section contains current and past tasks roughly ordered
18 by TODO state, then importance or date-completed. This would be
19 a good place to suggest ideas for development.
20 - The [[* Bugs][Bugs]] :: section contains bug reports.
21 - The [[* Tests][Tests]] :: section consists of a large table which can be
22 evaluated to run Org-Babel's functional test suite. This
23 provides a good overview of the current functionality with
24 pointers to example source blocks.
25 - The [[* Sandbox][Sandbox]] :: demonstrates much of the early/basic functionality
26 through commented source-code blocks.
28 Also see the [[file:library-of-babel.org][Library of Babel]], an extensible collection of ready-made
29 and easily-shortcut-callable source-code blocks for handling common
34 Org-Babel enables *communication* between programming languages and
38 - communication between programs :: Data passes seamlessly between
39 different programming languages, Org-Mode constructs (tables,
40 file links, example text) and interactive comint buffers.
41 - communication between people :: Data and calculations are embedded
42 in the same document as notes explanations and reports.
44 ** communication between programs
46 Org-Mode supports embedded blocks of source code (in any language)
47 inside of Org documents. Org-Babel allows these blocks of code to be
48 executed from within Org-Mode with natural handling of their inputs
52 with both scalar, file, and table output
54 *** reading information from tables
56 *** reading information from other source blocks (disk usage in your home directory)
58 This will work for Linux and Mac users, not so sure about shell
59 commands for windows users.
61 To run place the cursor on the =#+begin_src= line of the source block
62 labeled directory-pie and press =\C-c\C-c=.
64 #+srcname: directories
65 #+begin_src bash :results replace
66 cd ~ && du -sc * |grep -v total
69 #+resname: directories
71 | 11882808 | "Documents" |
72 | 8210024 | "Downloads" |
73 | 879800 | "Library" |
76 | 5307664 | "Pictures" |
85 #+srcname: directory-pie
86 #+begin_src R :var dirs = directories :session R-pie-example
87 pie(dirs[,1], labels = dirs[,2])
95 *** operations in/on tables
97 #+tblname: grades-table
98 | student | grade | letter |
99 |---------+-------+--------|
106 #+TBLFM: $2='(sbe random-score-generator)::$3='(sbe assign-grade (score $2))
108 #+srcname: assign-grade
109 #+begin_src ruby :var score=99
120 #+srcname: random-score-generator
125 #+srcname: show-distribution
126 #+begin_src R :var grades=grades-table
131 ** communication between people
132 Quick overview of Org-Mode's exportation abilities, with links to the
133 online Org-Mode documentation, a focus on source-code blocks, and the
134 exportation options provided by Org-Babel.
136 *** Interactive tutorial
137 This would demonstrate applicability to Reproducible Research, and
138 Literate Programming.
140 *** Tests embedded in documentation
141 org-babels own functional tests are contained in a large org-mode
142 table, allowing the test suite to be run be evaluation of the table
143 and the results to be collected in the same table.
145 *** Emacs initialization files stored in Org-Mode buffers
146 Using `org-babel-tangle' it is possible to embed your Emacs
147 initialization into org-mode files. This allows for folding,
148 note-taking, todo's etc... embedded with the source-code of your Emacs
149 initialization, and through org-mode's publishing features aids in
150 sharing your customizations with others.
152 It may be worthwhile to create a fork of Phil Hagelberg's
153 [[http://github.com/technomancy/emacs-starter-kit/tree/master][emacs-starter-kit]] which uses literate org-mode files for all of the
154 actual elisp customization. These org-mode files could then be
155 exported to html and used to populate the repositories wiki on [[http://github.com/][github]].
160 *** code evaluation (comint buffer sessions and external processes)
161 There are two main ways to evaluate source blocks with org-babel.
163 - external :: By default (if the =:session= header argument is not
164 present) all source code blocks are evaluated in
165 external processes. In these cases an external process
166 is used to evaluate the source-code blocks.
167 - session :: Session based evaluation uses persistent sessions in
168 comint buffers. Sessions can be used across multiple
169 source blocks setting and accessing variables in the
172 Evaluating source blocks in sessions also allows for
173 interaction with the code. To jump to the session of a
174 source block use the `org-babel-pop-to-session' command
175 or press =M-[down]= while inside of a source code block.
176 When called with a prefix argument
177 `org-babel-pop-to-session' will evaluate all header
178 arguments before jumping to the source-code block.
180 *** results (values and outputs)
181 Either the *value* or the *output* of source code blocks can be
182 collected after evaluation.
184 - value :: The default way to collect results from a source-code block
185 is to return the value of the last statement in the block.
186 This can be thought of as the return value of the block.
187 In this case any printed output of the block is ignored.
188 This can be though of a similar to a "functional" value of
190 - output :: Another way of generating results from a source-code block
191 is to collect the output generated by the execution of the
192 block. In this case all printed output is collected
193 throughout the execution of the block. This can be
194 thought of as similar to a "script" style of evaluation.
198 Add the following lines to your .emacs, replacing the path as
199 appropriate. A good place to check that things are up and running
200 would then be [[#sandbox][the sandbox]].
201 #+begin_src emacs-lisp
202 (add-to-list 'load-path "/path/to/org-babel/lisp")
203 (require 'org-babel-init)
208 ** PROPOSED optional timestamp for output
209 Add option to place an (inactive) timestamp at the #+resname, to
210 record when that output was generated.
212 *** source code block timestamps (optional addition)
213 If we did this would we then want to place a timestamp on the
214 source-code block, so that we would know if the results are
215 current or out of date? This would have the effect of caching the
216 results of calculations and then only re-running if the
217 source-code has changed. For the caching to work we would need to
218 check not only the timestamp on a source-code block, but also the
219 timestamps of any tables or source-code blocks referenced by the
220 original source-code block.
222 **** maintaining source-code block timestamps
223 It may make sense to add a hook to `org-edit-special' which could
224 update the source-code blocks timestamp. If the user edits the
225 contents of a source-code block directly I can think of no
226 efficient way of maintaining the timestamp.
228 ** TODO use example block for large amounts of stdout output?
229 We're currently `examplizing' with : at the beginning of the line,
230 but should larger amounts of output be in a
231 \#+begin_example...\#+end_example block? What's the cutoff? > 1
232 line? This would be nice as it would allow folding of lengthy
233 output. Sometimes one will want to see stdout just to check
234 everything looks OK, and then fold it away.
236 I'm addressing this in branch 'examplizing-output'.
237 Yea, that makes sense. (either that or allow folding of large
238 blocks escaped with =:=).
240 Proposed cutoff of 10 lines, we can save this value in a user
241 customizable variable.
243 ** TODO make tangle files read-only?
244 With a file-local variable setting, yea that makes sense. Maybe
245 the header should reference the related org-mode file.
246 ** TODO take default values for header args from properties
247 Use file-wide and subtree wide properties to set default values for
249 ** TODO support for working with =*Org Edit Src Example*= buffers [2/4]
250 *** TODO optionally evaluate header references when we switch to =*Org Edit Src*= buffer
251 That seems to imply that the header references need to be evaluated
252 and transformed into the target language object when we hit C-c ' to
253 enter the *Org Edit Src* buffer [DED]
255 Good point, I heartily agree that this should be supported [Eric]
257 (or at least before the first time we attempt to evaluate code in that
258 buffer -- I suppose there might be an argument for lazy evaluation, in
259 case someone hits C-c ' but is "just looking" and not actually
260 evaluating anything.) Of course if evaluating the reference is
261 computationally intensive then the user might have to wait before they
262 get the *Org Edit Src* buffer. [DED]
264 I fear that it may be hard to anticipate when the references will be
265 needed, some major-modes do on-the-fly evaluation while the buffer is
266 being edited. I think that we should either do this before the buffer
267 is opened or not at all, specifically I think we should resolve
268 references if the user calls C-c ' with a prefix argument. Does that
269 sound reasonable? [Eric]
273 [Dan] So now that we have org-src-mode and org-src-mode-hook, I guess
274 org-babel should do this by using the hook to make sure that, when C-c
275 C-' is issued on a source block, any references are resolved and
276 assignments are made in the appropriate session.
277 *** TODO set buffer-local-process variables appropriately [DED]
278 I think something like this would be great. You've probably
279 already thought of this, but just to note it down: it would be really
280 nice if org-babel's notion of a buffer's 'session/process' played
281 nicely with ESS's notion of the buffer's session/process. ESS keeps
282 the current process name for a buffer in a buffer-local variable
283 ess-local-process-name. So one thing we will probably want to do is
284 make sure that the *Org Edit Src Example* buffer sets that variable
287 I had not thought of that, but I agree whole heartedly. [Eric]
289 Once this is done every variable should be able to dump regions into
290 their inferior-process buffer using major-mode functions.
291 *** DEFERRED send code to inferior process
292 Another thought on this topic: I think we will want users to send
293 chunks of code to the interpreter from within the *Org Edit Src*
294 buffer, and I think that's what you have in mind already. In ESS that
295 is done using the ess-eval-* functions. [DED]
297 I think we can leave this up to the major-mode in the source code
298 buffer, as almost every source-code major mode will have functions for
299 doing things like sending regions to the inferior process. If
300 anything we might need to set the value of the buffer local inferior
301 process variable. [Eric]
303 *** DONE some possible requests/proposed changes for Carsten [4/4]
304 While I remember, some possible requests/proposed changes for Carsten
305 come to mind in that regard:
307 **** DONE Remap C-x C-s to save the source to the org buffer?
308 I've done this personally and I find it essential. I'm using
309 #+begin_src emacs-lisp
310 (defun org-edit-src-save ()
311 "Update the parent org buffer with the edited source code, save
312 the parent org-buffer, and return to the source code edit
321 (define-key org-exit-edit-mode-map "\C-x\C-s" 'org-edit-src-save)
325 I think this is great, but I think it should be implemented in the
328 **** DEFERRED Rename buffer and minor mode?
329 Something shorter than *Org Edit Src Example* for the buffer
330 name. org-babel is bringing org's source code interaction to a
331 level of maturity where the 'example' is no longer
332 appropriate. And if further keybindings are going to be added to
333 the minor mode then maybe org-edit-src-mode is a better name than
336 Maybe we should name the buffer with a combination of the source
337 code and the session. I think that makes sense.
339 [ES] Are you also suggesting a new org-edit-src minor mode?
340 [DED] org-exit-edit-mode is a minor mode that already exists:
342 Minor mode installing a single key binding, "C-c '" to exit special edit.
344 org-edit-src-save now has a binding in that mode, so I guess all
345 I'm saying at this stage is that it's a bit of a misnomer. But
346 perhaps we will also have more functionality to add to that minor
347 mode, making it even more of a misnomer. Perhaps something like
348 org-src-mode would be better.
349 **** DONE Changed minor mode name and added hooks
351 **** DEFERRED a hook called when the src edit buffer is created
352 This should be implemented in the org-mode core
354 ** TODO resolve references to other org buffers/files
355 This would allow source blocks to call upon tables, source-blocks,
356 and results in other org buffers/files.
359 - [[file:lisp/org-babel-ref.el::TODO%20allow%20searching%20for%20names%20in%20other%20buffers][org-babel-ref.el:searching-in-other-buffers]]
360 - [[file:lisp/org-babel.el::defun%20org-babel%20find%20named%20result%20name][org-babel.el#org-babel-find-named-result]]
361 ** TODO resolve references to other non-org files
362 - tabular data in .csv, .tsv etc format
363 - files of interpreted code: anything stopping us giving such files
364 similar status to a source code block?
365 - Would be nice to allow org and non-org files to be remote
366 ** TODO figure out how to handle errors during evaluation
367 R has a try function, with error handling, along the lines of
368 python. I bet ruby does too. Maybe more of an issue for functional
369 style; in my proposed scripting style the error just gets dumped to
370 the org buffer and the user is thus alerted.
371 ** TODO figure out how to handle graphic output
372 This is listed under [[* graphical output][graphical output]] in out objectives.
374 This should take advantage of the =:results file= option, and
375 languages which almost always produce graphical output should set
376 =:results file= to true by default. That would handle placing these
377 results in the buffer. Then if there is a combination of =silent= and
378 =file= =:results= headers we could drop the results to a temp buffer
379 and pop open that buffer...
380 ** TODO =\C-c \C-o= to open results of source block
381 by adding a =defadvice= to =org-open-at-point= we can use the common
382 =\C-c \C-o= keybinding to open the results of a source-code block.
383 This would be especially useful for source-code blocks which generate
384 graphical results and insert a file link as the results in the
385 org-mode buffer. (see [[* figure out how to handle graphic output][TODO figure out how to handle graphic output]]).
386 This could also act reasonably with other results types...
388 - file :: use org-open-at-point to open the file
389 - scalar :: open results unquoted in a new buffer
390 - tabular :: export the table to a new buffer and open that buffer
392 ** TODO Finalise behaviour regarding vector/scalar output
393 *** DONE Stop spaces causing vector output
394 This simple example of multilingual chaining produces vector output if
395 there are spaces in the message and scalar otherwise.
399 #+begin_src R :var msg=msg-from-python
400 paste(msg, "und R", sep=" ")
404 : org-babel speaks elisp y python und R
406 #+srcname: msg-from-python
407 #+begin_src python :var msg=msg-from-elisp
411 #+srcname: msg-from-elisp
412 #+begin_src emacs-lisp :var msg="org-babel speaks"
413 (concat msg " elisp")
415 ** STARTED share org-babel [1/4]
416 how should we share org-babel?
418 *** DONE post to org-mode
419 *** TODO post to ess mailing list
420 *** TODO create a org-babel page on worg
421 *** TODO create a short screencast demonstrating org-babel in action
424 we need to think up some good examples
426 **** interactive tutorials
427 This could be a place to use [[* org-babel assertions][org-babel assertions]].
429 for example the first step of a tutorial could assert that the version
430 of the software-package (or whatever) is equal to some value, then
431 source-code blocks could be used with confidence (and executed
432 directly from) the rest of the tutorial.
434 **** answering a text-book question w/code example
435 org-babel is an ideal environment enabling both the development and
436 demonstrationg of the code snippets required as answers to many
439 **** something using tables
440 maybe something along the lines of calculations from collected grades
443 Maybe something like the following which outputs sizes of directories
444 under the home directory, and then instead of the trivial =emacs-lisp=
445 block we could use an R block to create a nice pie chart of the
449 #+begin_src bash :results replace
453 #+begin_src emacs-lisp :var sizes=sizes :results replace
457 ** TODO command line execution
458 Allow source code blocks to be called form the command line. This
459 will be easy using the =sbe= function in [[file:lisp/org-babel-table.el][org-babel-table.el]].
461 This will rely upon [[* resolve references to other buffers][resolve references to other buffers]].
463 ** TODO inline source code blocks [3/5]
464 Like the =\R{ code }= blocks
466 not sure what the format should be, maybe just something simple
467 like =src_lang[]{}= where lang is the name of the source code
468 language to be evaluated, =[]= is optional and contains any header
469 arguments and ={}= contains the code.
471 (see [[* (sandbox) inline source blocks][the-sandbox]])
473 *** DONE evaluation with \C-c\C-c
474 Putting aside the header argument issue for now we can just run these
475 with the following default header arguments
476 - =:results= :: silent
477 - =:exports= :: results
479 *** DONE inline exportation
480 Need to add an interblock hook (or some such) through org-exp-blocks
481 *** DONE header arguments
482 We should make it possible to use header arguments.
484 *** TODO fontification
485 we should color these blocks differently
487 *** TODO refine html exportation
488 should use a span class, and should show original source in tool-tip
490 ** TODO formulate general rules for handling vectors and tables / matrices with names
491 This is non-trivial, but may be worth doing, in particular to
492 develop a nice framework for sending data to/from R.
494 In R, indexing vector elements, and rows and columns, using
495 strings rather than integers is an important part of the
497 - elements of a vector may have names
498 - matrices and data.frames may have "column names" and "row names"
499 which can be used for indexing
500 - In a data frame, row names *must* be unique
508 > mat <- matrix(1:4, nrow=2, ncol=2, dimnames=list(c("r1","r2"), c("c1","c2")))
513 > # The names are separate from the data: they do not interfere with operations on the data
520 > df <- data.frame(var1=1:26, var2=26:1, row.names=letters)
522 [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
528 So it's tempting to try to provide support for this in org-babel. For example
529 - allow R to refer to columns of a :var reference by their names
530 - When appropriate, results from R appear in the org buffer with "named
533 However none (?) of the other languages we are currently supporting
534 really have a native matrix type, let alone "column names" or "row
535 names". Names are used in e.g. python and perl to refer to entries
538 It currently seems to me that support for this in org-babel would
539 require setting rules about when org tables are considered to have
540 named columns/fields, and ensuring that (a) languages with a notion
541 of named columns/fields use them appropriately and (b) languages
542 with no such notion do not treat then as data.
544 - Org allows something that *looks* like column names to be separated
546 - Org also allows a row to *function* as column names when special
547 markers are placed in the first column. An hline is unnecessary
548 (indeed hlines are purely cosmetic in org [correct?]
549 - Org does not have a notion of "row names" [correct?]
551 The full org table functionality exeplified [[http://orgmode.org/manual/Advanced-features.html#Advanced-features][here]] has features that
552 we would not support in e.g. R (like names for the row below).
554 *** Initial statement: allow tables with hline to be passed as args into R
555 This doesn't seem to work at the moment (example below). It would
556 also be nice to have a natural way for the column names of the org
557 table to become the column names of the R data frame, and to have
558 the option to specify that the first column is to be used as row
559 names in R (these must be unique). But this might require a bit of
564 | col1 | col2 | col3 |
565 |------+---------+------|
573 #+begin_src R var tabel=egtable
578 | "col1" | "col2" | "col3" |
579 |--------+-----------+--------|
581 | 4 | "schulte" | 6 |
584 Another example is in the [[*operations%20in%20on%20tables][grades example]].
586 ** TODO re-implement helper functions from org-R
587 *** Initial statement [Eric]
588 Much of the power of org-R seems to be in it's helper functions for
589 the quick graphing of tables. Should we try to re-implement these
590 functions on top of org-babel?
592 I'm thinking this may be useful both to add features to org-babel-R and
593 also to potentially suggest extensions of the framework. For example
594 one that comes to mind is the ability to treat a source-code block
595 like a function which accepts arguments and returns results. Actually
596 this can be it's own TODO (see [[* source blocks as functions][source blocks as functions]]).
598 - We want to provide convenient off-the-shelf actions
599 (e.g. plotting data) that make use of our new code evaluation
600 environment but do not require any actual coding.
601 *** Initial Design proposal [Dan]
602 - *Input data* will be specified using the same mechanism as :var
603 references, thus the input data may come from a table, or
604 another source block, and it is initially available as an elisp
606 - We introduce a new #+ line, e.g. #+BABELDO. C-c C-c on that
607 line will apply an *action* to the referenced data.
608 - *Actions correspond to source blocks*: our library of available
609 actions will be a library of org-babel source blocks. Thus the
610 code for executing an action, and the code for dealing with the
611 output of the action will be the same code as for executing
612 source blocks in general
613 - Optionally, the user can have the relevant source block inserted
614 into the org buffer after the (say) #+BABELDO line. This will
615 allow the user to fine tune the action by modifying the code
616 (especially useful for plots).
617 - So maybe a #+BABELDO line will have header args
618 - :data (a reference to a table or source code block)
619 - :action (or should that be :srcname?) which will be something
620 like :action pie-chart, referring to a source block which will
621 be executed with the :data referent passed in using a :var arg.
622 - :showcode or something controlling whether to show the code
624 *** Modification to design
625 I'm implementing this, at least initially, as a new interpreter
626 named 'babel', which has an empty body. 'babel' blocks take
627 a :srcname header arg, and look for the source-code block with
628 that name. They then execute the referenced block, after first
629 appending their own header args on to the target block's header
632 If the target block is in the library of babel (a.o.t. e.g. the
633 current buffer), then the code in the block will refer to the
634 input data with a name dictated by convention (e.g. __data__
635 (something which is syntactically legal in all languages...). Thus
636 the babel block will use a :var __data__ = whatever header arg to
637 reference the data to be plotted.
640 This is covered by the [[file:library-of-babel.org][Library of Babel]], which will contain
641 ready-made source blocks designed to carry out useful common tasks.
642 ** PROPOSED Are we happy with current behaviour regarding vector/scalar output?
643 This simple example of multilingual chaining produces vector output if
644 there are spaces in the message and scalar otherwise.
646 #+begin_src R :var msg=msg-from-python
647 paste(msg, "und_R", sep="_")
650 #+srcname: msg-from-python
651 #+begin_src python :var msg=msg-from-elisp
655 #+srcname: msg-from-elisp
656 #+begin_src emacs-lisp :var msg="org-babel_speaks"
657 (concat msg "_elisp")
660 ** PROPOSED conversion between org-babel and noweb (e.g. .Rnw) format
661 I haven't thought about this properly. Just noting it down. What
662 Sweave uses is called "R noweb" (.Rnw).
664 I found a good description of noweb in the following article (see
665 the [[http://www.cs.tufts.edu/~nr/pubs/lpsimp.pdf][pdf]]).
667 I think there are two parts to noweb, the construction of
668 documentation and the extraction of source-code (with notangle).
670 *documentation*: org-mode handles all of our documentation needs in
671 a manner that I believe is superior to noweb.
673 *source extraction* At this point I don't see anyone writing large
674 applications with 100% of the source code contained in org-babel
675 files, rather I see org-babel files containing things like
676 - notes with active code chunks
677 - interactive tutorials
678 - requirements documents with code running test suites
679 - and of course experimental reports with the code to run the
680 experiment, and perform analysis
682 Basically I think the scope of the programs written in org-babel
683 (at least initially) will be small enough that it wont require the
684 addition of a tangle type program to extract all of the source code
685 into a running application.
687 On the other hand, since we already have named blocks of source
688 code which reference other blocks on which they rely, this
689 shouldn't be too hard to implement either on our own, or possibly
690 relying on something like noweb/notangle.
692 ** PROPOSED support for passing paths to files between source blocks
693 Maybe this should be it's own result type (in addition to scalars and
694 vectors). The reason being that some source-code blocks (for example
695 ditaa or anything that results in the creation of a file) may want to
696 pass a file path back to org-mode which could then be inserted into
697 the org-mode buffer as a link to the file...
699 This would allow for display of images upon export providing
700 functionality similar to =org-exp-blocks= only in a more general
703 ** DEFERRED use textConnection to pass tsv to R?
704 When passing args from the org buffer to R, the following route is
705 used: arg in buffer -> elisp -> tsv on file -> data frame in R. I
706 think it would be possible to avoid having to write to file by
707 constructing an R expression in org-babel-R-assign-elisp, something
710 #+begin_src emacs-lisp
711 (org-babel-R-input-command
712 (format "%s <- read.table(textConnection(\"%s\"), sep=\"\\t\", as.is=TRUE)"
713 name (orgtbl-to-tsv value '(:sep "\t" :fmt org-babel-R-quote-tsv-field))))
716 I haven't tried to implement this yet as it's basically just
717 fiddling with something that works. The only reason for it I can
718 think of would be efficiency and I haven't tested that.
720 This Didn't work after an initial test. I still think this is a
721 good idea (I also think we should try to do something similar when
722 writing out results frmo R to elisp) however as it wouldn't result
723 in any functional changes I'm bumping it down to deferred for
728 #+tblname: quick-test
731 #+srcname: quick-test-src-blk
732 #+begin_src R :var vec=quick-test
738 ** DEFERRED re-implement R evaluation using ess-command or ess-execute
739 I don't have any complaints with the current R evaluation code or
740 behaviour, but I think it would be good to use the ESS functions
741 from a political point of view. Plus of course it has the normal
742 benefits of an API (insulates us from any underlying changes etc). [DED]
744 I'll look into this. I believe that I looked at and rejected these
745 functions initially but now I can't remember why. I agree with
746 your overall point about using API's where available. I will take
747 a look back at these and either switch to using the ess commands,
748 or at least articulate under this TODO the reasons for using our
749 custom R-interaction commands. [Eric]
753 Lets just replace =org-babel-R-input-command= with =ess-execute=.
755 I tried this, and although it works in some situations, I find that
756 =ess-command= will often just hang indefinitely without returning
757 results. Also =ess-execute= will occasionally hang, and pops up
758 the buffer containing the results of the command's execution, which
759 is undesirable. For now these functions can not be used. Maybe
760 someone more familiar with the ESS code can recommend proper usage
761 of =ess-command= or some other lower-level function which could be
762 used in place of [[file:lisp/org-babel-R.el::defun%20org-babel%20R%20input%20command%20command][org-babel-R-input-command]].
766 #+begin_quote ess-command
767 (ess-command COM &optional BUF SLEEP NO-PROMPT-CHECK)
769 Send the ESS process command COM and delete the output
770 from the ESS process buffer. If an optional second argument BUF exists
771 save the output in that buffer. BUF is erased before use.
772 COM should have a terminating newline.
773 Guarantees that the value of .Last.value will be preserved.
774 When optional third arg SLEEP is non-nil, `(sleep-for (* a SLEEP))'
775 will be used in a few places where `a' is proportional to `ess-cmd-delay'.
778 #+begin_quote ess-execute
779 (ess-execute COMMAND &optional INVERT BUFF MESSAGE)
781 Send a command to the ESS process.
782 A newline is automatically added to COMMAND. Prefix arg (or second arg
783 INVERT) means invert the meaning of
784 `ess-execute-in-process-buffer'. If INVERT is 'buffer, output is
785 forced to go to the process buffer. If the output is going to a
786 buffer, name it *BUFF*. This buffer is erased before use. Optional
787 fourth arg MESSAGE is text to print at the top of the buffer (defaults
788 to the command if BUFF is not given.)
791 *** out current setup
793 1) The body of the R source code block is wrapped in a function
794 2) The function is called inside of a =write.table= function call
795 writing the results to a table
796 3) The table is read using =org-table-import=
798 ** DEFERRED Rework Interaction with Running Processes [2/5]
799 *** DONE robust to errors interrupting execution
801 #+srcname: long-runner-ruby
802 #+begin_src ruby :results silent
807 *** DEFERRED use =C-g= keyboard-quit to push processing into the background
808 This may be possible using the `run-with-timer' command.
810 I have no idea how this could work...
812 #+srcname: long-runner-ruby
813 #+begin_src ruby :results silent
818 *** TODO ability to select which of multiple sessions is being used
819 Increasingly it is looking like we're going to want to run all
820 source code blocks in comint buffer (sessions). Which will have
822 1) allowing background execution
823 2) maintaining state between source-blocks
824 - allowing inline blocks w/o header arguments
827 (like ess-switch-process in .R buffers)
829 Maybe this could be packaged into a header argument, something
830 like =:R_session= which could accept either the name of the
831 session to use, or the string =prompt=, in which case we could use
832 the =ess-switch-process= command to select a new process.
834 *** TODO evaluation of shell code as background process?
835 After C-c C-c on an R code block, the process may appear to
836 block, but C-g can be used to reclaim control of the .org buffer,
837 without interrupting the R evalution. However I believe this is not
838 true of bash/sh evaluation. [Haven't tried other languages] Perhaps
839 a solution is just to background the individual shell commands.
841 The other languages (aside from emacs lisp) are run through the
842 shell, so if we find a shell solution it should work for them as
845 Adding an ampersand seems to be a supported way to run commands in
846 the background (see [[http://www.emacswiki.org/emacs/ExecuteExternalCommand#toc4][external-commands]]). Although a more extensible
847 solution may involve the use of the [[elisp:(progn (describe-function 'call-process-region) nil)][call-process-region]] function.
849 Going to try this out in a new file [[file:lisp/org-babel-proc.el][org-babel-proc.el]]. This should
850 contain functions for asynchronously running generic shell commands
851 in the background, and then returning their input.
853 **** partial update of org-mode buffer
854 The sleekest solution to this may be using a comint buffer, and
855 then defining a filter function which would incrementally interpret
856 the results as they are returned, including insertion into the
857 org-mode buffer. This may actually cause more problems than it is
858 worth, what with the complexities of identifying the types of
859 incrementally returned results, and the need for maintenance of a
860 process marker in the org buffer.
862 **** 'working' spinner
863 It may be nice and not too difficult to place a spinner on/near the
864 evaluating source code block
866 *** TODO conversion of output from interactive shell, R (and python) sessions to org-babel buffers
867 [DED] This would be a nice feature I think. Although an org-babel
868 purist would say that it's working the wrong way round... After
869 some interactive work in a *R* buffer, you save the buffer, maybe
870 edit out some lines, and then convert it to org-babel format for
871 posterity. Same for a shell session either in a *shell* buffer, or
872 pasted from another terminal emulator. And python of course.
874 ** DEFERRED improve the source-block snippet
875 any real improvement seems somewhat beyond the ability of yasnippet
878 [[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]]
880 ,#name : Chapter title
883 ${1:$(make-string (string-width text) ?\=)}
888 [[file:snippets/org-mode/sb][sb -- snippet]]
890 waiting for guidance from those more familiar with yasnippets
891 ** DONE exclusive =exports= params
893 #+srcname: implement-export-exclusivity
900 ** DONE LoB: allow output in buffer
901 ** DONE allow default header arguments by language
902 org-babel-default-header-args:lang-name
904 An example of when this is useful is for languages which always return
905 files as their results (e.g. [[*** ditaa][ditaa]], and [[*** gnuplot][gnuplot]]).
906 ** DONE singe-function tangling and loading elisp from literate org-mode file [3/3]
908 This function should tangle the org-mode file for elisp, and then call
909 `load-file' on the resulting tangled file.
911 #+srcname: test-loading-embedded-emacs-lisp
912 #+begin_src emacs-lisp :results replace
913 (setq test-tangle-advert nil)
914 (setq test-tangle-loading nil)
915 (setq results (list :before test-tangle-loading test-tangle-advert))
916 (org-babel-load-file "test-tangle.org")
917 (setq results (list (list :after test-tangle-loading test-tangle-advert) results))
918 (delete-file "test-tangle.el")
922 #+resname: test-loading-embedded-emacs-lisp
923 | :before | nil | nil |
924 | :after | "org-babel tangles" | "use org-babel-tangle for all your emacs initialization files!!" |
926 *** DONE add optional language limiter to org-babel-tangle
927 This should check to see if there is any need to re-export
929 *** DONE ensure that org-babel-tangle returns the path to the tangled file(s)
931 #+srcname: test-return-value-of-org-babel-tangle
932 #+begin_src emacs-lisp :results replace
933 (mapcar #'file-name-nondirectory (org-babel-tangle-file "test-tangle.org" "emacs-lisp"))
939 *** DONE only tangle the file if it's actually necessary
940 ** DONE add a function to jump to a source-block by name
941 I've had an initial stab at that in org-babel-find-named-block
942 (library-of-babel branch).
944 At the same time I introduced org-babel-named-src-block-regexp, to
945 match src-blocks with srcname.
947 This is now working with the command
948 `org-babel-goto-named-source-block', all we need is a good key
951 ** DONE add =:none= session argument (for purely functional execution) [4/4]
952 This would allow source blocks to be run in their own new process
954 - These blocks could then also be run in the background (since we can
955 detach and just wait for the process to signal that it has terminated)
956 - We wouldn't be drowning in session buffers after running the tests
957 - we can re-use much of the session code to run in a more /functional/
960 While session provide a lot of cool features, like persistent
961 environments, [[* DONE function to bring up inferior-process buffer][pop-to-session]], and hints at exportation for
962 org-babel-tangle, they also have some down sides and I'm thinking that
963 session-based execution maybe shouldn't be the default behavior.
965 Down-sides to sessions
966 - *much* more complicated than functional evaluation
967 - maintaining the state of the session has weird issues
968 - waiting for evaluation to finish
969 - prompt issues like [[* TODO weird escaped characters in shell prompt break shell evaluation][shell-prompt-escapes-bug]]
970 - can't run in background
971 - litter emacs with session buffers
975 #+srcname: ruby-task-no-session
976 #+begin_src ruby :results replace output
982 #+resname: ruby-task-no-session
987 #+srcname: task-python-none-session
988 #+begin_src python :session none :results replace value
994 #+resname: task-python-none-session
999 #+srcname: task-session-none-sh
1000 #+begin_src sh :results replace
1005 #+resname: task-session-none-sh
1011 #+srcname: task-no-session-R
1012 #+begin_src R :results replace output
1019 #+resname: task-no-session-R
1023 ** DONE fully purge org-babel-R of direct comint interaction
1024 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
1026 ** DONE Create objects in top level (global) environment [5/5]
1029 *** initial requirement statement [DED]
1030 At the moment, objects created by computations performed in the
1031 code block are evaluated in the scope of the
1032 code-block-function-body and therefore disappear when the code
1033 block is evaluated {unless you employ some extra trickery like
1034 assign('name', object, env=globalenv()) }. I think it will be
1035 desirable to also allow for a style wherein objects that are
1036 created in one code block persist in the R global environment and
1037 can be re-used in a separate block.
1039 This is what Sweave does, and while I'm not saying we have to be
1040 the same as Sweave, it wouldn't be hard for us to provide the same
1041 behaviour in this case; if we don't, we risk undeservedly being
1042 written off as an oddity by some.
1044 IOW one aspect of org-babel is that of a sort of functional
1045 meta-programming language. This is crazy, in a very good
1046 way. Nevertheless, wrt R I think there's going to be a lot of value
1047 in providing for a working style in which the objects are stored in
1048 the R session, rather than elisp/org buffer. This will be a very
1049 familiar working style to lots of people.
1051 There are no doubt a number of different ways of accomplishing
1052 this, the simplest being a hack like adding
1055 for(objname in ls())
1056 assign(objname, get(objname), envir=globalenv())
1059 to the source code block function body. (Maybe wrap it in an on.exit() call).
1061 However this may deserve to be thought about more carefully, perhaps
1062 with a view to having a uniform approach across languages. E.g. shell
1063 code blocks have the same semantics at the moment (no persistence of
1064 variables across code blocks), because the body is evaluated in a new
1065 bash shell process rather than a running shell. And I guess the same
1066 is true for python. However, in both these cases, you could imagine
1067 implementing the alternative in which the body is evaluated in a
1068 persistent interactive session. It's just that it's particularly
1069 natural for R, seeing as both ESS and org-babel evaluate commands in a
1070 single persistent R session.
1074 Thanks for bringing this up. I think you are absolutely correct that we
1075 should provide support for a persistent environment (maybe called a
1076 *session*) in which to evaluate code blocks. I think the current setup
1077 demonstrates my personal bias for a functional style of programming
1078 which is certainly not ideal in all contexts.
1080 While the R function you mention does look like an elegant solution, I
1081 think we should choose an implementation that would be the same across
1082 all source code types. Specifically I think we should allow the user to
1083 specify an optional *session* as a header variable (when not present we
1084 assume a default session for each language). The session name could be
1085 used to name a comint buffer (like the *R* buffer) in which all
1086 evaluation would take place (within which variables would retain their
1087 values --at least once I remove some of the functional method wrappings
1088 currently in place-- ).
1090 This would allow multiple environments to be used in the same buffer,
1091 and once this setup was implemented we should be able to fairly easily
1092 implement commands for jumping between source code blocks and the
1093 related session buffers, as well as for dumping the last N commands from
1094 a session into a new or existing source code block.
1096 Please let me know if you foresee any problems with this proposed setup,
1097 or if you think any parts might be confusing for people coming from
1098 Sweave. I'll hopefully find some time to work on this later in the
1101 *** can functional and interpreted/interactive models coexist?
1103 Even though both of these use the same =*R*= buffer the value of =a=
1104 is not preserved because it is assigned inside of a functional
1107 #+srcname: task-R-sessions
1114 #+srcname: task-R-same-session
1119 This functional wrapper was implemented in order to efficiently return
1120 the results of the execution of the entire source code block. However
1121 it inhibits the evaluation of source code blocks in the top level,
1122 which would allow for persistence of variable assignment across
1123 evaluations. How can we allow *both* evaluation in the top level, and
1124 efficient capture of the return value of an entire source code block
1125 in a language independent manner?
1127 Possible solutions...
1128 1) we can't so we will have to implement two types of evaluation
1129 depending on which is appropriate (functional or imperative)
1130 2) we remove the functional wrapper and parse the source code block
1131 into it's top level statements (most often but not always on line
1132 breaks) so that we can isolate the final segment which is our
1134 3) we add some sort of "#+return" line to the code block
1135 4) we take advantage of each languages support for meta-programming
1136 through =eval= type functions, and use said to evaluate the entire
1137 blocks in such a way that their environment can be combined with the
1138 global environment, and their results are still captured.
1139 5) I believe that most modern languages which support interactive
1140 sessions have support for a =last_result= type function, which
1141 returns the result of the last input without re-calculation. If
1142 widely enough present this would be the ideal solution to a
1143 combination of functional and imperative styles.
1145 None of these solutions seem very desirable, but for now I don't see
1146 what else would be possible.
1148 Of these options I was leaning towards (1) and (4) but now believe
1149 that if it is possible option (5) will be ideal.
1151 **** (1) both functional and imperative evaluation
1153 - can take advantage of built in functions for sending regions to the
1155 - retains the proven tested and working functional wrappers
1158 - introduces the complication of keeping track of which type of
1159 evaluation is best suited to a particular context
1160 - the current functional wrappers may require some changes in order to
1161 include the existing global context
1163 **** (4) exploit language meta-programming constructs to explicitly evaluate code
1165 - only one type of evaluation
1168 - some languages may not have sufficient meta-programming constructs
1170 **** (5) exploit some =last_value= functionality if present
1172 Need to ensure that most languages have such a function, those without
1173 will simply have to implement their own similar solution...
1175 | language | =last_value= function |
1176 |------------+-----------------------------|
1180 | shell | see [[* last command for shells][last command for shells]] |
1181 | emacs-lisp | see [[* emacs-lisp will be a special case][special-case]] |
1183 #+srcname: task-last-value
1188 ***** last command for shells
1189 Do this using the =tee= shell command, and continually pipe the output
1192 Got this idea from the following [[http://linux.derkeiler.com/Mailing-Lists/Fedora/2004-01/0898.html][email-thread]].
1194 suggested from mailing list
1196 #+srcname: bash-save-last-output-to-file
1200 bash -c "$line" | tee /tmp/last.out1
1201 mv /tmp/last.out1 /tmp/last.out
1205 another proposed solution from the above thread
1207 #+srcname: bash-save-in-variable
1210 # so - Save Output. Saves output of command in OUT shell variable.
1220 "^M": " | tee /tmp/h_lastcmd.out ^[k"
1223 export __=/tmp/h_lastcmd.out
1225 If you try it, Alt-k will stand for the old Enter; use "command $__" to
1226 access the last output.
1232 Herculano de Lima Einloft Neto
1235 ***** emacs-lisp will be a special case
1236 While it is possible for emacs-lisp to be run in a console type
1237 environment (see the =elim= function) it is *not* possible to run
1238 emacs-lisp in a different *session*. Meaning any variable set top
1239 level of the console environment will be set *everywhere* inside
1240 emacs. For this reason I think that it doesn't make any sense to
1241 worry about session support for emacs-lisp.
1243 *** Further thoughts on 'scripting' vs. functional approaches
1245 These are just thoughts, I don't know how sure I am about this.
1246 And again, perhaps I'm not saying anything very radical, just that
1247 it would be nice to have some options supporting things like
1248 receiving text output in the org buffer.
1250 I can see that you've already gone some way down the road towards
1251 the 'last value' approach, so sorry if my comments come rather
1252 late. I am concerned that we are not giving sufficient attention
1253 to stdout / the text that is returned by the interpreters. In
1254 contrast, many of our potential users will be accustomed to a
1255 'scripting' approach, where they are outputting text at various
1256 points in the code block, not just at the end. I am leaning
1257 towards thinking that we should have 2 modes of evaluation:
1258 'script' mode, and 'functional' mode.
1260 In script mode, evaluation of a code block would result in *all*
1261 text output from that code block appearing as output in the org
1262 buffer, presumably as an #+begin_example...#+end_example. There
1263 could be an :echo option controlling whether the input commands
1264 also appear in the output. [This is like Sweave].
1266 In functional mode, the *result* of the code block is available as
1267 an elisp object, and may appear in the org buffer as an org
1268 table/string, via the mechanisms you have developed already.
1270 One thing I'm wondering about is whether, in script mode, there
1271 simply should not be a return value. Perhaps this is not so
1272 different from what exists: script mode would be new, and what
1273 exists currently would be functional mode.
1275 I think it's likely that, while code evaluation will be exciting
1276 to people, a large majority of our users in a large majority of
1277 their usage will not attempt to actually use the return value from
1278 a source code block in any meaningful way. In that case, it seems
1279 rather restrictive to only allow them to see output from the end
1282 Instead I think the most accessible way to introduce org-babel to
1283 people, at least while they are learning it, is as an immensely
1284 powerful environment in which to embed their 'scripts', which now
1285 also allows them to 'run' their 'scripts'. Especially as such
1286 people are likely to be the least capable of the user-base, a
1287 possible design-rule would be to make the scripting style of usage
1288 easy (default?), perhaps requiring a special option to enable a
1289 functional style. Those who will use the functional style won't
1290 have a problem understanding what's going on, whereas the 'skript
1291 kiddies' might not even know the syntax for defining a function in
1292 their language of choice. And of course we can allow the user to
1293 set a variable in their .emacs controlling the preference, so that
1294 functional users are not inconveniennced by having to provide
1295 header args the whole time.
1297 Please don't get the impression that I am down-valuing the
1298 functional style of org-babel. I am constantly horrified at the
1299 messy 'scripts' that my colleagues produce in perl or R or
1300 whatever! Nevertheless that seems to be how a lot of people work.
1302 I think you were leaning towards the last-value approach because
1303 it offered the possibility of unified code supporting both the
1304 single evaluation environment and the functional style. If you
1305 agree with any of the above then perhaps it will impact upon this
1306 and mean that the code in the two branches has to differ a bit. In
1307 that case, functional mode could perhaps after all evaluate each
1308 code block in its own environment, thus (re)approaching 'true'
1309 functional programming (side-effects are hard to achieve).
1313 echo "There are `wc -l files` files in this directory"
1317 *** even more thoughts on evaluation, results, models and options
1319 Thanks Dan, These comments are invaluable.
1321 What do you think about this as a new list of priorities/requirements
1322 for the execution of source-code blocks.
1325 1) we want the evaluation of the source code block to take place in a
1326 session which can persist state (variables, current directory,
1328 2) source code blocks can specify their session with a header argument
1329 3) each session should correspond to an Emacs comint buffer so that the
1330 user can drop into the session and experiment with live code
1333 1) each source-code block generates some form of results which (as
1334 we have already implemented) is transfered into emacs-lisp
1335 after which it can be inserted into the org-mode buffer, or
1336 used by other source-code blocks
1337 2) when the results are translated into emacs-lisp, forced to be
1338 interpreted as a scalar (dumping their raw values into the
1339 org-mode buffer), as a vector (which is often desirable with R
1340 code blocks), or interpreted on the fly (the default option).
1341 Note that this is very nearly currently implemented through the
1342 [[* DONE results-type header (vector/file)][results-type-header]].
1343 3) there should be *two* means of collecting results from the
1344 execution of a source code block. *Either* the value of the
1345 last statement of the source code block, or the collection of
1346 all that has been passed to STDOUT during the evaluation.
1348 **** header argument or return line (*header argument*)
1350 Rather than using a header argument to specify how the return value
1351 should be passed back, I'm leaning towards the use of a =#+RETURN=
1352 line inside the block. If such a line *is not present* then we
1353 default to using STDOUT to collect results, but if such a line *is
1354 present* then we use it's value as the results of the block. I
1355 think this will allow for the most elegant specification between
1356 functional and script execution. This also cleans up some issues
1357 of implementation and finding which statement is the last
1360 Having given this more thought, I think a header argument is
1361 preferable. The =#+return:= line adds new complicating syntax for
1362 something that does little more than we would accomplish through
1363 the addition of a header argument. The only benefit being that we
1364 know where the final statement starts, which is not an issue in
1365 those languages which contain 'last value' operators.
1367 new header =:results= arguments
1368 - script :: explicitly states that we want to use STDOUT to
1369 initialize our results
1370 - return_last :: stdout is ignored instead the *value* of the final
1371 statement in the block is returned
1372 - echo :: means echo the contents of the source-code block along
1373 with the results (this implies the *script* =:results=
1376 *** DONE rework evaluation lang-by-lang [4/4]
1378 This should include...
1379 - functional results working with the comint buffer
1381 - script :: return the output of STDOUT
1382 - write a macro which runs the first redirection, executes the
1383 body, then runs the second redirection
1384 - last :: return the value of the last statement
1387 - sessions in comint buffers
1389 **** DONE Ruby [4/4]
1390 - [X] functional results working with comint
1391 - [X] script results
1392 - [X] ensure scalar/vector results args are taken into consideration
1393 - [X] ensure callable by other source block
1395 #+srcname: ruby-use-last-output
1396 #+begin_src ruby :results replace
1403 #+resname: ruby-use-last-output
1406 #+srcname: task-call-use-last-output
1407 #+begin_src ruby :var last=ruby-use-last-output :results replace
1408 last.flatten.size + 1
1411 #+resname: task-call-use-last-output
1416 #+srcname: first-ruby-session-task
1417 #+begin_src ruby :session schulte :results silent
1421 #+srcname: second-ruby-session-task
1422 #+begin_src ruby :session schulte :results silent
1426 #+srcname: without-the-right-session
1427 #+begin_src ruby :results silent
1433 - [X] functional results working with comint
1434 - [X] script results
1435 - [X] ensure scalar/vector results args are taken into consideration
1436 - [X] ensure callable by other source block
1438 To redirect output to a file, you can use the =sink()= command.
1441 #+begin_src R :results value vector silent
1448 #+srcname: task-R-use-other-output
1449 #+begin_src R :var twoentyseven=task_R_B() :results replace value
1454 #+resname: task-R-use-other-output
1457 **** DONE Python [4/4]
1458 - [X] functional results working with comint
1459 - [X] script results
1460 - [X] ensure scalar/vector results args are taken into consideration
1461 - [X] ensure callable by other source block
1463 #+srcname: task-new-eval-for-python
1464 #+begin_src python :results silent output scalar
1470 #+srcname: task-use-new-eval
1471 #+begin_src python :var tasking=task-new-eval-for-python() :results replace
1475 #+resname: task-use-new-eval
1478 **** DONE Shells [4/4]
1479 - [X] functional results working with comint
1480 - [X] script results
1481 - [X] ensure scalar/vector results args are taken into consideration
1482 - [X] ensure callable by other source block
1484 #+srcname: task-shell-new-evaluation
1485 #+begin_src sh :results silent value scalar
1490 #+srcname: task-call-other-shell
1491 #+begin_src sh :var other=task-shell-new-evaluation() :results replace scalar
1492 echo $other ' is the old date'
1495 #+resname: task-call-other-shell
1496 : $ Fri Jun 12 13:08:37 PDT 2009 is the old date
1498 *** DONE implement a *session* header argument [4/4]
1499 =:session= header argument to override the default *session* buffer
1503 #+srcname: task-ruby-named-session
1504 #+begin_src ruby :session schulte :results replace
1505 schulte = :in_schulte
1508 #+resname: task-ruby-named-session
1511 #+srcname: another-in-schulte
1512 #+begin_src ruby :session schulte
1516 #+resname: another-in-schulte
1523 #+srcname: python-session-task
1524 #+begin_src python :session what :results silent
1528 #+srcname: python-get-from-session
1529 #+begin_src python :session what :results replace
1533 #+resname: python-get-from-session
1538 #+srcname: task-shell-sessions
1539 #+begin_src sh :session what
1543 #+srcname: task-shell-sessions-what
1544 #+begin_src sh :session what :results replace
1548 #+resname: task-shell-sessions-what
1553 #+srcname: task-R-session
1554 #+begin_src R :session what :results replace
1560 #+resname: task-R-session
1563 #+srcname: another-task-R-session
1564 #+begin_src R :session what :results replace
1568 *** DONE function to bring up inferior-process buffer [4/4]
1570 This should be callable from inside of a source-code block in an
1571 org-mode buffer. It should evaluate the header arguments, then bring
1572 up the inf-proc buffer using =pop-to-buffer=.
1574 For lack of a better place, lets add this to the `org-metadown-hook'
1577 To give this a try, place the cursor on a source block with variables,
1578 (optionally git a prefix argument) then hold meta and press down.
1582 #+srcname: task-ruby-pop-to-session
1583 #+begin_src ruby :var num=9 :var another="something else"
1584 num.times{|n| puts another}
1589 #+srcname: task-python-pop-to-session
1590 #+begin_src python :var num=9 :var another="something else"
1595 #+srcname: task-R-pop-to-session
1596 #+begin_src R :var a=9 :var b=8
1602 #+srcname: task-shell-pop-sessions
1603 #+begin_src sh :var NAME="eric"
1607 *** DEFERRED function to dump last N lines from inf-proc buffer into the current source block
1609 Callable with a prefix argument to specify how many lines should be
1610 dumped into the source-code buffer.
1612 *** REJECTED comint notes
1614 Implementing comint integration in [[file:lisp/org-babel-comint.el][org-babel-comint.el]].
1617 - handling of outputs
1618 - split raw output from process by prompts
1619 - a ring of the outputs, buffer-local, `org-babel-comint-output-ring'
1620 - a switch for dumping all outputs to a buffer
1621 - inputting commands
1623 Lets drop all this language specific stuff, and just use
1624 org-babel-comint to split up our outputs, and return either the last
1625 value of an execution or the combination of values from the
1628 **** comint filter functions
1629 : ;; comint-input-filter-functions hook process-in-a-buffer
1630 : ;; comint-output-filter-functions hook function modes.
1631 : ;; comint-preoutput-filter-functions hook
1632 : ;; comint-input-filter function ...
1634 #+srcname: obc-filter-ruby
1635 #+begin_src ruby :results last
1643 ** DONE Remove protective commas from # comments before evaluating
1644 org inserts protective commas in front of ## comments in language
1645 modes that use them. We need to remove them prior to sending code
1648 #+srcname: testing-removal-of-protective-comas
1650 ,# this one might break it??
1654 ** DONE pass multiple reference arguments into R
1655 Can we do this? I wasn't sure how to supply multiple 'var' header
1656 args. Just delete this if I'm being dense.
1658 This should be working, see the following example...
1660 #+srcname: two-arg-example
1661 #+begin_src R :var n=2 :var m=8
1665 #+resname: two-arg-example
1668 ** DONE ensure that table ranges work
1669 when a table range is passed to org-babel as an argument, it should be
1670 interpreted as a vector.
1673 | 2 | 3 | Fixnum:1 |
1674 | 3 | 4 | Array:123456 |
1678 #+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)))
1680 #+srcname: simple-sbe-example
1681 #+begin_src emacs-lisp
1685 #+srcname: task-table-range
1686 #+begin_src ruby :var n=simple-sbe-example
1690 #+srcname: simple-results
1691 #+begin_src emacs-lisp :var n=task-table-range(n=(1 2 3))
1695 #+resname: simple-results
1698 #+srcname: task-arr-referent
1699 #+begin_src ruby :var ar=(1 2 3)
1703 #+resname: task-arr-referent
1706 ** DONE global variable indicating default to vector output
1707 how about an alist... =org-babel-default-header-args= this may already
1708 exist... just execute the following and all source blocks will default
1711 #+begin_src emacs-lisp
1712 (setq org-babel-default-header-args '((:results . "vector")))
1715 ** DONE name named results if source block is named
1716 currently this isn't happening although it should be
1718 #+srcname: test-naming-named-source-blocks
1719 #+begin_src emacs-lisp
1723 #+resname: test-naming-named-source-blocks
1725 ** DONE (simple caching) check for named results before source blocks
1726 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]]
1727 ** DONE set =:results silent= when eval with prefix argument
1729 #+begin_src emacs-lisp
1732 ** DONE results-type header (vector/file) [3/3]
1733 In response to a point in Dan's email. We should allow the user to
1734 force scalar or vector results. This could be done with a header
1735 argument, and the default behavior could be controlled through a
1736 configuration variable.
1738 #+srcname: task-trivial-vector
1739 #+begin_src ruby :results replace vector
1746 since it doesn't make sense to turn a vector into a scalar, lets
1747 just add a two values...
1749 - vector :: forces the results to be a vector (potentially 1 dimensional)
1750 - file :: this throws an error if the result isn't a string, and
1751 tries to treat it as a path to a file.
1753 I'm just going to cram all of these into the =:results= header
1754 argument. Then if we allow multiple header arguments it should
1755 work out, for example one possible header argument string could be
1756 =:results replace vector file=, which would *replace* any existing
1757 results forcing the results into an org-mode table, and
1758 interpreting any strings as file paths.
1760 *** DONE multiple =:results= headers
1762 #+srcname: multiple-result-headers
1763 #+begin_src ruby :results replace silent
1769 *** DONE file result types
1770 When inserting into an org-mode buffer create a link with the path
1771 being the value, and optionally the display being the
1772 =file-name-nondirectory= if it exists.
1774 #+srcname: task-file-result
1775 #+begin_src python :results replace file
1780 [[something][something]]
1783 This will be useful because blocks like =ditaa= and =dot= can return
1784 the string path of their files, and can add =file= to their results
1787 *** DONE vector result types
1789 #+srcname: task-force-results
1790 #+begin_src emacs-lisp :results vector
1797 ** DONE results name
1798 In order to do this we will need to start naming our results.
1799 Since the source blocks are named with =#+srcname:= lines we can
1800 name results with =#+resname:= lines (if the source block has no
1801 name then no name is given to the =#+resname:= line on creation,
1802 otherwise the name of the source block is used).
1804 This will have the additional benefit of allowing results and
1805 source blocks to be located in different places in a buffer (and
1806 eventually in different buffers entirely).
1808 #+srcname: developing-resnames
1809 #+begin_src emacs-lisp :results silent
1813 Once source blocks are able to find their own =#+resname:= lines
1816 #+srcname: sbe-w-new-results
1817 #+begin_src emacs-lisp :results replace
1818 (sbe "developing-resnames")
1824 *** TODO change the results insertion functions to use these lines
1826 *** TODO teach references to resolve =#+resname= lines.
1828 ** DONE org-babel tests org-babel [1/1]
1829 since we are accumulating this nice collection of source-code blocks
1830 in the sandbox section we should make use of them as unit tests.
1831 What's more, we should be able to actually use org-babel to run these
1834 We would just need to cycle over every source code block under the
1835 sandbox, run it, and assert that the return value is equal to what we
1838 I have the feeling that this should be possible using only org-babel
1839 functions with minimal or no additional elisp. It would be very cool
1840 for org-babel to be able to test itself.
1842 This is now done, see [[* Tests]].
1844 *** DEFERRED org-babel assertions (may not be necessary)
1845 These could be used to make assertions about the results of a
1846 source-code block. If the assertion fails then the point could be
1847 moved to the block, and error messages and highlighting etc... could
1850 ** DONE make C-c C-c work anywhere within source code block?
1851 This seems like it would be nice to me, but perhaps it would be
1852 inefficient or ugly in implementation? I suppose you could search
1853 forward, and if you find #+end_src before you find #+begin_src,
1854 then you're inside one. [DED]
1856 Agreed, I think inside of the =#+srcname: line= would be useful as
1859 #+srcname: testing-out-cc
1860 #+begin_src emacs-lisp
1864 ** DONE integration with org tables
1865 We should make it easy to call org-babel source blocks from org-mode
1866 table formulas. This is practical now that it is possible to pass
1867 arguments to org-babel source blocks.
1869 See the related [[* (sandbox) integration w/org tables][sandbox]] header for tests/examples.
1871 *** digging in org-table.el
1872 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.
1874 Should be a hook in [[file:~/src/org/lisp/org-table.el::defun%20org%20table%20eval%20formula%20optional%20arg%20equation][org-table-eval-formula]].
1876 Looks like I need to change this [[file:~/src/org/lisp/org-table.el::if%20lispp][if statement]] (line 2239) into a cond
1879 ** DONE source blocks as functions
1881 Allow source code blocks to be called like functions, with arguments
1882 specified. We are already able to call a source-code block and assign
1883 it's return result to a variable. This would just add the ability to
1884 specify the values of the arguments to the source code block assuming
1885 any exist. For an example see
1887 When a variable appears in a header argument, how do we differentiate
1888 between it's value being a reference or a literal value? I guess this
1889 could work just like a programming language. If it's escaped or in
1890 quotes, then we count it as a literal, otherwise we try to look it up
1893 ** DONE folding of code blocks? [2/2]
1894 [DED] In similar way to using outline-minor-mode for folding function
1895 bodies, can we fold code blocks? #+begin whatever statements are
1896 pretty ugly, and in any case when you're thinking about the overall
1897 game plan you don't necessarily want to see the code for each Step.
1899 *** DONE folding of source code block
1900 Sounds good, and wasn't too hard to implement. Code blocks should
1901 now be fold-able in the same manner as headlines (by pressing TAB
1904 *** REJECTED folding of results
1905 So, lets do a three-stage tab cycle... First fold the src block,
1906 then fold the results, then unfold.
1908 There's no way to tell if the results are a table or not w/o
1909 actually executing the block which would be too expensive of an
1912 ** DONE selective export of text, code, figures
1913 [DED] The org-babel buffer contains everything (code, headings and
1914 notes/prose describing what you're up to, textual/numeric/graphical
1915 code output, etc). However on export to html / LaTeX one might want
1916 to include only a subset of that content. For example you might
1917 want to create a presentation of what you've done which omits the
1920 [EMS] So I think this should be implemented as a property which can
1921 be set globally or on the outline header level (I need to review
1922 the mechanics of org-mode properties). And then as a source block
1923 header argument which will apply only to a specific source code
1924 block. A header argument of =:export= with values of
1926 - =code= :: just show the code in the source code block
1927 - =none= :: don't show the code or the results of the evaluation
1928 - =results= :: just show the results of the code evaluation (don't
1929 show the actual code)
1930 - =both= :: show both the source code, and the results
1932 this will be done in [[* (sandbox) selective export][(sandbox) selective export]].
1934 ** DONE a header argument specifying silent evaluation (no output)
1935 This would be useful across all types of source block. Currently
1936 there is a =:replace t= option to control output, this could be
1937 generalized to an =:output= option which could take the following
1938 options (maybe more)
1940 - =t= :: this would be the default, and would simply insert the
1941 results after the source block
1942 - =replace= :: to replace any results which may already be there
1943 - =silent= :: this would inhibit any insertion of the results
1945 This is now implemented see the example in the [[* silent evaluation][sandbox]]
1947 ** DONE assign variables from tables in R
1948 This is now working (see [[* (sandbox table) R][(sandbox-table)-R]]). Although it's not that
1949 impressive until we are able to print table results from R.
1951 ** DONE insert 2-D R results as tables
1952 everything is working but R and shell
1958 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
1959 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]]
1960 as a means of converting multidimensional R objects to emacs lisp.
1962 It may be as simple as first checking if the data is multidimensional,
1963 and then, if so using =write= to write the data out to a temporary
1964 file from which emacs can read the data in using =org-table-import=.
1966 Looking into this further, is seems that there is no such thing as a
1967 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
1968 deal with trivial vectors (scalars) in R. I'm tempted to just treat
1969 them as vectors, but then that would lead to a proliferation of
1970 trivial 1-cell tables...
1972 ** DONE allow variable initialization from source blocks
1973 Currently it is possible to initialize a variable from an org-mode
1974 table with a block argument like =table=sandbox= (note that the
1975 variable doesn't have to named =table=) as in the following example
1981 #+begin_src emacs-lisp :var table=sandbox :results replace
1982 (message (format "table = %S" table))
1985 : "table = ((1 2 3) (4 \"schulte\" 6))"
1987 It would be good to allow initialization of variables from the results
1988 of other source blocks in the same manner. This would probably
1989 require the addition of =#+SRCNAME: example= lines for the naming of
1990 source blocks, also the =table=sandbox= syntax may have to be expanded
1991 to specify whether the target is a source code block or a table
1992 (alternately we could just match the first one with the given name
1993 whether it's a table or a source code block).
1995 At least initially I'll try to implement this so that there is no need
1996 to specify whether the reference is to a table or a source-code block.
1997 That seems to be simpler both in terms of use and implementation.
1999 This is now working for emacs-lisp, ruby and python (and mixtures of
2000 the three) source blocks. See the examples in the [[* (sandbox) referencing other source blocks][sandbox]].
2002 This is currently working only with emacs lisp as in the following
2003 example in the [[* emacs lisp source reference][emacs lisp source reference]].
2006 ** TODO Add languages [1/6]
2007 I'm sure there are many more that aren't listed here. Please add
2008 them, and bubble any that you particularly care about up to the top.
2010 Any new language should be implemented in a org-babel-lang.el file.
2011 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
2012 [[file:lisp/org-babel-R.el][org-babel-R.el]].
2015 This could probably be added to [[file:lisp/org-babel-script.el][org-babel-script.el]]
2020 (see [[* file result types][file result types]])
2022 #+srcname: implementing-ditaa
2023 #+begin_src ditaa :results replace :file blue.png :cmdline -r
2033 #+resname: implementing-ditaa
2034 [[file:blue.png][blue.png]]
2037 (see [[* file result types][file result types]])
2039 - a =file= header argument
2040 - a =cmdline= header argument
2042 - scalar variables should be replaced in the body of the gnuplot code
2043 - vector variables should be exported to tab-separated files, and
2044 the variable names should be replaced with the path to the files
2046 #+PLOT: title:"Citas" ind:1 deps:(3) type:2d with:histograms set:"yrange [0:]"
2047 #+TBLNAME: gnuplot-data
2048 | independent var | first dependent var | second dependent var |
2049 |-----------------+---------------------+----------------------|
2050 | 0.1 | 0.425 | 0.375 |
2051 | 0.2 | 0.3125 | 0.3375 |
2052 | 0.3 | 0.24999993 | 0.28333338 |
2053 | 0.4 | 0.275 | 0.28125 |
2054 | 0.5 | 0.26 | 0.27 |
2055 | 0.6 | 0.25833338 | 0.24999993 |
2056 | 0.7 | 0.24642845 | 0.23928553 |
2057 | 0.8 | 0.23125 | 0.2375 |
2058 | 0.9 | 0.23333323 | 0.2333332 |
2059 | 1 | 0.2225 | 0.22 |
2060 | 1.1 | 0.20909075 | 0.22272708 |
2061 | 1.2 | 0.19999998 | 0.21458333 |
2062 | 1.3 | 0.19615368 | 0.21730748 |
2064 #+srcname: implementing-gnuplot
2065 #+begin_src gnuplot :var data=gnuplot-data
2066 set title "Implementing Gnuplot"
2067 plot "data" using 1:2 with lines
2071 (see [[* file result types][file result types]])
2074 (see [[* file result types][file result types]])
2078 ** TODO avoid stripping whitespace from output when :results output
2079 ** TODO prompt characters appearing in output with R
2080 #+begin_src R :session *R* :results output
2089 ** TODO o-b-execute-subtree overwrites heading when subtree is folded
2091 Try M-x org-babel-execute-subtree with the subtree folded and
2092 point at the beginning of the heading line.
2097 ** TODO Allow source blocks to be recognised when #+ are not first characters on the line
2098 I think Carsten has recently altered the core so that #+ can have
2099 preceding whitespace, at least for literal/code examples. org-babel
2100 should support this.
2102 ** TODO non-orgtbl formatted lists
2105 #+srcname: this-doesn't-match-orgtbl
2106 #+begin_src emacs-lisp :results replace
2107 '((:results . "replace"))
2110 #+resname: this-doesn't-match-orgtbl
2112 ** TODO collapsing consecutive newlines in string output
2114 #+srcname: multi-line-string-output
2115 #+begin_src ruby :results output
2116 "the first line ends here
2119 and this is the second one
2124 #+resname: multi-line-string-output
2126 ** PROPOSED external shell execution can't isolate return values
2127 I have no idea how to do this as of yet. The result is that when
2128 shell functions are run w/o a session there is no difference between
2129 the =output= and =value= result arguments.
2131 Yea, I don't know how to do this either. I searched extensively on
2132 how to isolate the *last* output of a series of shell commands (see
2133 [[* last command for
2134 shells][last command for shells]]). The results of the search were basically
2135 that it was not possible (or at least not accomplish-able with a
2136 reasonable amount of effort).
2138 That fact combined with the tenancy to all ways use standard out in
2139 shell scripts led me to treat these two options (=output= and =value=)
2140 as identical in shell evaluation. Not ideal but maybe good enough for
2143 In the `results' branch I've changed this so that they're not quite
2144 identical: output results in raw stdout contents, whereas value
2145 converts it to elisp, perhaps to a table if it looks tabular. This is
2146 the same for the other languages. [Dan]
2148 ** TODO are the org-babel-trim s necessary?
2149 at the end of e.g. org-babel-R-evaluate, org-babel-python-evaluate, but
2150 not org-babel-ruby-evaluate
2151 ** TODO elisp reference fails for literal number
2152 #+srcname: elisp-test(a=4)
2153 #+begin_src emacs-lisp
2156 ** TODO use new merge function [[file:lisp/org-babel-ref.el::t%20nil%20org%20combine%20plists%20args%20nil][here]]?
2157 And at other occurrences of org-combine-plists?
2158 ** TODO LoB: with output to buffer, not working in buffers other than library-of-babel.org
2159 I haven't fixed this yet. org-babel-ref-resolve-reference moves
2160 point around, inside a save-excursion. Somehow when it comes to
2161 inserting the results (after possible further recursive calls to
2162 org-babel-ref-resolve-reference), point hasn't gone back to the
2164 ** TODO creeping blank lines
2165 There's still inappropriate addition of blank lines in some circumstances. E.g.
2171 Compare the results of
2172 #+lob: python-add(a=5, b=17)
2174 #+resname: python-add(a=5, b=17)
2176 --------------------------------
2184 ---------------------
2186 Hmm, it's a bit confusing. I think it's to do with the fact that
2187 LoB removes the entire (#+resname and result) and starts from
2188 scratch, whereas #+begin_src only removes the result. I haven't
2189 worked out what the correct fix is yet.
2190 ** DEFERRED weird escaped characters in shell prompt break shell evaluation
2191 E.g. this doesn't work. Should the shell sessions set a sane prompt
2192 when they start up? Or is it a question of altering
2193 comint-prompt-regexp? Or altering org-babel regexps?
2196 black=30 ; red=31 ; green=32 ; yellow=33 ; blue=34 ; magenta=35 ; cyan=36 ; white=37
2199 export PS1="\[\033[${prompt_col}m\]\w${prompt_char} \[\033[0m\]"
2202 I just pushed a good amount of changes, could you see if your shell
2203 problems still exist?
2205 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
2206 org-babel-sh.el, raw gets the value
2208 ("" "
\e[0m Sun Jun 14 19:26:24 EDT 2009\n" "
\e[0m org_babel_sh_eoe\n" "
\e[0m ")
2210 and therefore (member org-babel-sh-eoe-output ...) fails
2212 I think that `comint-prompt-regexp' needs to be altered to match
2213 the shell prompt. This shouldn't be too difficult to do by hand,
2214 using the `regexp-builder' command and should probably be part of
2215 the user's regular emacs init. I can't think of a way for us to
2216 set this automatically, and we are SOL without a regexp to match
2219 ** DONE cursor movement when evaluating source blocks
2220 E.g. the pie chart example. Despite the save-window-excursion in
2221 org-babel-execute:R. (I never learned how to do this properly: org-R
2222 jumps all over the place...)
2224 I don't see this now [ded]
2226 ** DONE LoB: calls fail if reference has single character name
2227 commit 21d058869df1ff23f4f8cc26f63045ac9c0190e2
2228 **** This doesn't work
2229 #+lob: R-plot(data=X)
2248 #+lob: R-plot(data=XX)
2250 ** DONE make :results replace the default?
2251 I'm tending to think that appending results to pre-existing results
2252 creates mess, and that the cleaner `replace' option should be the
2253 default. E.g. when a source block creates an image, we would want
2254 that to be updated, rather than have a new one be added.
2258 ** DONE ruby evaluation not working under ubuntu emacs 23
2259 With emacs 23.0.91.1 on ubuntu, for C-h f run-ruby I have the
2260 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.
2263 run-ruby is an interactive compiled Lisp function.
2267 Run an inferior Ruby process, input and output via buffer *ruby*.
2268 If there is a process already running in `*ruby*', switch to that buffer.
2269 With argument, allows you to edit the command line (default is value
2270 of `ruby-program-name'). Runs the hooks `inferior-ruby-mode-hook'
2271 (after the `comint-mode-hook' is run).
2272 (Type C-h m in the process buffer for a list of commands.)
2275 So, I may have a non-standard inf-ruby.el. Here's my version of
2279 run-ruby is an interactive Lisp function in `inf-ruby.el'.
2281 (run-ruby &optional COMMAND NAME)
2283 Run an inferior Ruby process, input and output via buffer *ruby*.
2284 If there is a process already running in `*ruby*', switch to that buffer.
2285 With argument, allows you to edit the command line (default is value
2286 of `ruby-program-name'). Runs the hooks `inferior-ruby-mode-hook'
2287 (after the `comint-mode-hook' is run).
2288 (Type C-h m in the process buffer for a list of commands.)
2291 It seems we could either bundle my version of inf-ruby.el (as it's
2292 the newest). Or we could change the use of `run-ruby' so that it
2293 is robust across multiple distributions. I think I'd prefer the
2294 former, unless the older version of inf-ruby is actually bundled
2295 with emacs, in which case maybe we should go out of our way to
2296 support it. Thoughts?
2298 I think for now I'll just include the latest [[file:util/inf-ruby.el][inf-ruby.el]] in the
2299 newly created utility directory. I doubt anyone would have a
2300 problem using the latest version of this file.
2301 ** DONE test failing forcing vector results with =test-forced-vector-results= ruby code block
2302 Note that this only seems to happen the *second* time the test table
2305 #+srcname: bug-trivial-vector
2306 #+begin_src emacs-lisp :results vector silent
2310 #+srcname: bug-forced-vector-results
2311 #+begin_src ruby :var triv=test-trivial-vector :results silent
2315 mysteriously this seems to be fixed...
2316 ** DONE defunct R sessions
2317 Sometimes an old R session will turn defunct, and newly inserted code
2318 will not be evaluated (leading to a hang).
2320 This seems to be fixed by using `inferior-ess-send-input' rather than `comint-send-input'.
2321 ** DONE ruby fails on first call to non-default session
2323 #+srcname: bug-new-session
2324 #+begin_src ruby :session is-new
2328 ** DONE when reading results from =#+resname= line
2330 Errors when trying to read from resname lines.
2332 #+resname: bug-in-resname
2335 #+srcname: bug-in-resname-reader
2336 #+begin_src emacs-lisp :var buggy=bug-in-resname() :results silent
2340 ** DONE R-code broke on "org-babel" rename
2342 #+srcname: bug-R-babels
2347 ** DONE error on trivial R results
2349 So I know it's generally not a good idea to squash error without
2350 handling them, but in this case the error almost always means that
2351 there was no file contents to be read by =org-table-import=, so I
2354 #+srcname: bug-trivial-r1
2355 #+begin_src R :results replace
2356 pie(c(1, 2, 3), labels = c(1, 2, 3))
2359 #+srcname: bug-trivial-r2
2360 #+begin_src R :results replace
2364 #+resname: bug-trivial-r2
2367 #+srcname: bug-trivial-r3
2368 #+begin_src R :results replace
2372 #+resname: bug-trivial-r3
2377 ** DONE ruby new variable creation (multi-line ruby blocks)
2378 Actually it looks like we were dropping all but the last line.
2380 #+srcname: multi-line-ruby-test
2381 #+begin_src ruby :var table=bug-numerical-table :results replace
2383 table.each{|n| total += n}
2390 ** DONE R code execution seems to choke on certain inputs
2391 Currently the R code seems to work on vertical (but not landscape)
2394 #+srcname: little-fake
2395 #+begin_src emacs-lisp
2400 #+begin_src R :var num=little-fake
2407 #+srcname: set-debug-on-error
2408 #+begin_src emacs-lisp :results silent
2409 (setq debug-on-error t)
2412 #+srcname: bug-numerical-table
2413 #+begin_src emacs-lisp :results silent
2420 #+srcname: bug-R-number-evaluation
2421 #+begin_src R :var table=bug-numerical-table
2430 #+tblname: bug-vert-table
2435 #+srcname: bug-R-vertical-table
2436 #+begin_src R :var table=bug-vert-table :results silent
2440 ** DONE org bug/request: prevent certain org behaviour within code blocks
2441 E.g. [[]] gets recognised as a link (when there's text inside the
2442 brackets). This is bad for R code at least, and more generally
2443 could be argued to be inappropriate. Is it difficult to get org to
2444 ignore text in code blocks? [DED]
2446 I believe Carsten addressed this recently on the mailing list with
2447 the comment that it was indeed a difficult issue. I believe this
2448 may be one area where we could wait for an upstream (org-mode) fix.
2450 [Dan] Carsten has fixed this now in the core.
2452 ** DONE with :results replace, non-table output doesn't replace table output
2453 And vice versa. E.g. Try this first with table and then with len(table) [DED]
2454 #+begin_src python :var table=sandbox :results replace
2459 | 4 | "schulte" | 6 |
2462 Yes, this is certainly a problem. I fear that if we begin replacing
2463 anything immediately following a source block (regardless of whether
2464 it matches the type of our current results) we may accidentally delete
2465 hand written portions of the user's org-mode buffer.
2467 I think that the best solution here would be to actually start
2468 labeling results with a line that looks something like...
2472 This would have a couple of benefits...
2473 1) we wouldn't have to worry about possibly deleting non-results
2474 (which is currently an issue)
2475 2) we could reliably replace results even if there are different types
2476 3) we could reference the results of a source-code block in variable
2477 definitions, which would be useful if for example we don't wish to
2478 re-run a source-block every time because it is long-running.
2480 Thoughts? If no-one objects, I believe I will implement the labeling
2483 ** DONE extra quotes for nested string
2484 Well R appears to be reading the tables without issue...
2486 these *should* be quoted
2488 #+begin_src sh :results replace
2493 | "README.markdown" |
2496 | "existing_tools" |
2500 | "test-export.html" |
2501 | "test-export.org" |
2503 #+srcname: test-quotes
2504 #+begin_src ruby :var tab=ls
2510 #+srcname: test-quotes
2511 #+begin_src R :var tab=ls
2517 ** DONE simple ruby arrays not working
2519 As an example eval the following. Adding a line to test
2521 #+tblname: simple-ruby-array
2524 #+srcname: ruby-array-test
2525 #+begin_src ruby :var ar = simple-ruby-array :results silent
2529 ** DONE space trailing language name
2530 fix regexp so it works when there's a space trailing the language name
2532 #+srcname: test-trailing-space
2537 ** DONE Args out of range error
2539 The following block resulted in the error below [DED]. It ran without
2540 error directly in the shell.
2543 for platf in ill aff ; do
2544 for pop in CEU YRI ASI ; do
2545 rm -f $platf/hapmap-genos-$pop-all $platf/hapmap-rs-all
2546 cat $platf/hapmap-genos-$pop-* > $platf/hapmap-genos-$pop-all
2547 cat $platf/hapmap-rs-* > $platf/hapmap-rs-all
2552 executing source block with sh...
2553 finished executing source block
2554 string-equal: Args out of range: "", -1, 0
2556 the error =string-equal: Args out of range: "", -1, 0= looks like what
2557 used to be output when the block returned an empty results string.
2558 This should be fixed in the current version, you should now see the
2559 following message =no result returned by source block=.
2561 ** DONE ruby arrays not recognized as such
2563 Something is wrong in [[file:lisp/org-babel-script.el]] related to the
2564 recognition of ruby arrays as such.
2566 #+begin_src ruby :results replace
2572 #+begin_src python :results replace
2584 Evaluate all the cells in this table for a comprehensive test of the
2585 org-babel functionality.
2587 *Note*: if you have customized =org-babel-default-header-args= then some
2588 of these tests may fail.
2590 #+TBLNAME: org-babel-tests
2591 | functionality | block | arg | expected | results | pass |
2592 |-------------------------+----------------------------+-----+-------------+-------------+------|
2593 | basic evaluation | | | | | pass |
2594 |-------------------------+----------------------------+-----+-------------+-------------+------|
2595 | emacs lisp | basic-elisp | | 5 | 5 | pass |
2596 | shell | basic-shell | | 6 | 6 | pass |
2597 | ruby | basic-ruby | | org-babel | org-babel | pass |
2598 | python | basic-python | | hello world | hello world | pass |
2599 | R | basic-R | | 13 | 13 | pass |
2600 |-------------------------+----------------------------+-----+-------------+-------------+------|
2601 | tables | | | | | pass |
2602 |-------------------------+----------------------------+-----+-------------+-------------+------|
2603 | emacs lisp | table-elisp | | 3 | 3 | pass |
2604 | ruby | table-ruby | | 1-2-3 | 1-2-3 | pass |
2605 | python | table-python | | 5 | 5 | pass |
2606 | R | table-R | | 3.5 | 3.5 | pass |
2607 |-------------------------+----------------------------+-----+-------------+-------------+------|
2608 | source block references | | | | | pass |
2609 |-------------------------+----------------------------+-----+-------------+-------------+------|
2610 | all languages | chained-ref-last | | Array | Array | pass |
2611 |-------------------------+----------------------------+-----+-------------+-------------+------|
2612 | source block functions | | | | | pass |
2613 |-------------------------+----------------------------+-----+-------------+-------------+------|
2614 | emacs lisp | defun-fibb | | fibbd | fibbd | pass |
2615 | run over | Fibonacci | 0 | 1 | 1 | pass |
2616 | a | Fibonacci | 1 | 1 | 1 | pass |
2617 | variety | Fibonacci | 2 | 2 | 2 | pass |
2618 | of | Fibonacci | 3 | 3 | 3 | pass |
2619 | different | Fibonacci | 4 | 5 | 5 | pass |
2620 | arguments | Fibonacci | 5 | 8 | 8 | pass |
2621 |-------------------------+----------------------------+-----+-------------+-------------+------|
2622 | bugs and tasks | | | | | pass |
2623 |-------------------------+----------------------------+-----+-------------+-------------+------|
2624 | simple ruby arrays | ruby-array-test | | 3 | 3 | pass |
2625 | R number evaluation | bug-R-number-evaluation | | 2 | 2 | pass |
2626 | multi-line ruby blocks | multi-line-ruby-test | | 2 | 2 | pass |
2627 | forcing vector results | test-forced-vector-results | | Array | Array | pass |
2628 |-------------------------+----------------------------+-----+-------------+-------------+------|
2629 | sessions | | | | | pass |
2630 |-------------------------+----------------------------+-----+-------------+-------------+------|
2631 | set ruby session | set-ruby-session-var | | :set | :set | pass |
2632 | get from ruby session | get-ruby-session-var | | 3 | 3 | pass |
2633 | set python session | set-python-session-var | | set | set | pass |
2634 | get from python session | get-python-session-var | | 4 | 4 | pass |
2635 | set R session | set-R-session-var | | set | set | pass |
2636 | get from R session | get-R-session-var | | 5 | 5 | pass |
2637 #+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))
2638 #+TBLFM: $5=""::$6=""
2640 The second line (followed by replacing '[]' with '') can be used to blank out the table results, in the absence of a better method.
2644 #+srcname: basic-elisp
2645 #+begin_src emacs-lisp :results silent
2650 #+srcname: basic-shell
2651 #+begin_src sh :results silent
2656 #+srcname: date-simple
2657 #+begin_src sh :results silent
2661 #+srcname: basic-ruby
2662 #+begin_src ruby :results silent
2667 #+srcname: basic-python
2668 #+begin_src python :results silent
2674 #+begin_src R :results silent
2682 #+tblname: test-table
2686 #+srcname: table-elisp
2687 #+begin_src emacs-lisp :results silent :var table=test-table
2688 (length (car table))
2692 #+srcname: table-ruby
2693 #+begin_src ruby :results silent :var table=test-table
2694 table.first.join("-")
2698 #+srcname: table-python
2699 #+begin_src python :var table=test-table
2705 #+begin_src R :var table=test-table
2712 Lets pass a references through all of our languages...
2714 Lets start by reversing the table from the previous examples
2716 #+srcname: chained-ref-first
2717 #+begin_src python :var table = test-table
2722 #+resname: chained-ref-first
2726 Take the first part of the list
2728 #+srcname: chained-ref-second
2729 #+begin_src R :var table = chained-ref-first
2733 #+resname: chained-ref-second
2737 Turn the numbers into string
2739 #+srcname: chained-ref-third
2740 #+begin_src emacs-lisp :var table = chained-ref-second
2741 (mapcar (lambda (el) (format "%S" el)) table)
2744 #+resname: chained-ref-third
2747 and Check that it is still a list
2749 #+srcname: chained-ref-last
2750 #+begin_src ruby :var table=chained-ref-third
2755 ** source blocks as functions
2757 #+srcname: defun-fibb
2758 #+begin_src emacs-lisp :results silent
2759 (defun fibbd (n) (if (< n 2) 1 (+ (fibbd (- n 1)) (fibbd (- n 2)))))
2763 #+srcname: fibonacci
2764 #+begin_src emacs-lisp :results silent :var n=7
2774 ** sbe tests (these don't seem to be working...)
2775 Testing the insertion of results into org-mode tables.
2777 #+srcname: multi-line-output
2778 #+begin_src ruby :results replace
2779 "the first line ends here
2782 and this is the second one
2788 : the first line ends here\n\n\n and this is the second one\n\neven a third
2790 #+srcname: multi-line-error
2791 #+begin_src ruby :results replace
2792 raise "oh nooooooooooo"
2798 | the first line ends here... | -:5: warning: parenthesize argument(s) for future version... |
2799 #+TBLFM: $1='(sbe "multi-line-output")::$2='(sbe "multi-line-error")
2801 ** forcing results types tests
2803 #+srcname: test-trivial-vector
2804 #+begin_src emacs-lisp :results vector silent
2808 #+srcname: test-forced-vector-results
2809 #+begin_src ruby :var triv=test-trivial-vector :results silent
2815 #+srcname: set-ruby-session-var
2816 #+begin_src ruby :session :results silent
2821 #+srcname: get-ruby-session-var
2822 #+begin_src ruby :session :results silent
2826 #+srcname: set-python-session-var
2827 #+begin_src python :session
2832 #+srcname: get-python-session-var
2833 #+begin_src python :session
2837 #+srcname: set-R-session-var
2838 #+begin_src R :session
2843 #+srcname: get-R-session-var
2844 #+begin_src R :session
2853 To run these examples evaluate [[file:lisp/org-babel-init.el][org-babel-init.el]]
2855 ** org-babel.el beginning functionality
2857 #+begin_src sh :results replace
2862 : Sun Jul 5 18:54:39 EDT 2009
2869 : Sun Jul 05 18:54:35 -0400 2009
2881 #+begin_src R :results replace
2891 hist(rgamma(20,3,3))
2896 ** org-babel plays with tables
2897 Alright, this should demonstrate both the ability of org-babel to read
2898 tables into a lisp source code block, and to then convert the results
2899 of the source code block into an org table. It's using the classic
2900 "lisp is elegant" demonstration transpose function. To try this
2903 1. evaluate [[file:lisp/org-babel-init.el]] to load org-babel and friends
2904 2. evaluate the transpose definition =\C-c\\C-c= on the beginning of
2906 3. evaluate the next source code block, this should read in the table
2907 because of the =:var table=previous=, then transpose the table, and
2908 finally it should insert the transposed table into the buffer
2909 immediately following the block
2913 #+begin_src emacs-lisp :results silent
2914 (defun transpose (table)
2915 (apply #'mapcar* #'list table))
2923 #+begin_src emacs-lisp :var table=sandbox :results replace
2928 #+begin_src emacs-lisp
2933 | 1 | 2 | 3 | 4 | 5 |
2937 #+begin_src ruby :var table=sandbox :results replace
2938 table.first.join(" - ")
2944 #+begin_src python :var table=sandbox
2949 #+begin_src ruby :var table=sandbox :results replace
2954 : [[1, 2, 3], [4, "schulte", 6]]
2958 | 4 | "schulte" | 6 |
2960 #+begin_src python :var table=sandbox :results replace
2966 | "__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" |
2968 *** (sandbox table) R
2970 #+TBLNAME: sandbox_r
2974 #+begin_src R :results replace
2975 x <- c(rnorm(10, mean=-3, sd=1), rnorm(10, mean=3, sd=1))
2979 | -3.35473133869346 |
2981 | -3.32819924928633 |
2982 | -2.97310212756194 |
2983 | -2.09640758369576 |
2984 | -5.06054014378736 |
2985 | -2.20713700711221 |
2986 | -1.37618039712037 |
2987 | -1.95839385821742 |
2988 | -3.90407396475502 |
2989 | 2.51168071590226 |
2990 | 3.96753011570494 |
2991 | 3.31793212627865 |
2992 | 1.99829753972341 |
2993 | 4.00403686419829 |
2994 | 4.63723764452927 |
2995 | 3.94636744261313 |
2996 | 3.58355906547775 |
2997 | 3.01563442274226 |
3000 #+begin_src R var tabel=sandbox_r :results replace
3005 | 4 | "schulte" | 6 |
3008 Now shell commands are converted to tables using =org-table-import=
3009 and if these tables are non-trivial (i.e. have multiple elements) then
3010 they are imported as org-mode tables...
3012 #+begin_src sh :results replace
3016 | "total" | 208 | "" | "" | "" | "" | "" | "" |
3017 | "-rw-r--r--" | 1 | "dan" | "dan" | 57 | 2009 | 15 | "block" |
3018 | "-rw-r--r--" | 1 | "dan" | "dan" | 35147 | 2009 | 15 | "COPYING" |
3019 | "-rw-r--r--" | 1 | "dan" | "dan" | 722 | 2009 | 18 | "examples.org" |
3020 | "drwxr-xr-x" | 4 | "dan" | "dan" | 4096 | 2009 | 19 | "existing_tools" |
3021 | "-rw-r--r--" | 1 | "dan" | "dan" | 2207 | 2009 | 14 | "intro.org" |
3022 | "drwxr-xr-x" | 2 | "dan" | "dan" | 4096 | 2009 | 18 | "org-babel" |
3023 | "-rw-r--r--" | 1 | "dan" | "dan" | 277 | 2009 | 20 | "README.markdown" |
3024 | "-rw-r--r--" | 1 | "dan" | "dan" | 11837 | 2009 | 18 | "rorg.html" |
3025 | "-rw-r--r--" | 1 | "dan" | "dan" | 61829 | 2009 | 19 | "#rorg.org#" |
3026 | "-rw-r--r--" | 1 | "dan" | "dan" | 60190 | 2009 | 19 | "rorg.org" |
3027 | "-rw-r--r--" | 1 | "dan" | "dan" | 972 | 2009 | 11 | "test-export.org" |
3030 ** silent evaluation
3038 #+begin_src ruby :results silent
3042 #+begin_src ruby :results replace
3049 ** (sandbox) referencing other source blocks
3050 Doing this in emacs-lisp first because it's trivial to convert
3051 emacs-lisp results to and from emacs-lisp.
3053 *** emacs lisp source reference
3054 This first example performs a calculation in the first source block
3055 named =top=, the results of this calculation are then saved into the
3056 variable =first= by the header argument =:var first=top=, and it is
3057 used in the calculations of the second source block.
3060 #+begin_src emacs-lisp
3064 #+begin_src emacs-lisp :var first=top :results replace
3070 This example is the same as the previous only the variable being
3071 passed through is a table rather than a number.
3073 #+begin_src emacs-lisp :results silent
3074 (defun transpose (table)
3075 (apply #'mapcar* #'list table))
3078 #+TBLNAME: top_table
3082 #+SRCNAME: second_src_example
3083 #+begin_src emacs-lisp :var table=top_table
3087 #+begin_src emacs-lisp :var table=second_src_example :results replace
3092 | 4 | "schulte" | 6 |
3094 Now working for ruby
3101 #+begin_src ruby :var other=start :results replace
3107 #+SRCNAME: start_two
3112 #+begin_src python :var another=start_two :results replace
3117 Since all variables are converted into Emacs Lisp it is no problem to
3118 reference variables specified in another language.
3120 #+SRCNAME: ruby-block
3125 #+SRCNAME: lisp_block
3126 #+begin_src emacs-lisp :var ruby-variable=ruby-block
3130 #+begin_src python :var lisp_var=lisp_block
3139 #+begin_src R :results replace
3146 #+begin_src R :var other=first_r :results replace
3153 ** (sandbox) selective export
3155 For exportation tests and examples see (including exportation of
3156 inline source code blocks) [[file:test-export.org]]
3159 ** (sandbox) source blocks as functions
3162 #+begin_src emacs-lisp :results silent
3167 #+begin_src emacs-lisp :var n=default :results replace
3173 #+begin_src emacs-lisp :var result=triple(n=3, m=98) :results replace
3179 The following just demonstrates the ability to assign variables to
3180 literal values, which was not implemented until recently.
3182 #+begin_src ruby :var num="eric" :results replace
3189 ** (sandbox) inline source blocks
3191 This is an inline source code block src_ruby{1 + 6}. And another
3192 source block with text output src_emacs-lisp{"eric"}.
3194 This is an inline source code block with header
3195 arguments. src_ruby[:var n=fibbd( n = 0 )]{n}
3198 ** (sandbox) integration w/org tables
3200 #+begin_src emacs-lisp :results silent
3201 (defun fibbd (n) (if (< n 2) 1 (+ (fibbd (- n 1)) (fibbd (- n 2)))))
3205 #+begin_src emacs-lisp :var n=4 :results silent
3209 #+begin_src emacs-lisp :results silent
3210 (mapcar #'fibbd '(0 1 2 3 4 5 6 7 8))
3213 Something is not working here. The function `sbe ' works fine when
3214 called from outside of the table (see the source block below), but
3215 produces an error when called from inside the table. I think there
3216 must be some narrowing going on during intra-table emacs-lisp
3219 | original | fibbd |
3220 |----------+-------|
3231 #+TBLFM: $2='(sbe "fibbd" (n $1))
3235 #+begin_src emacs-lisp :results silent
3236 (sbe 'fibbd (n "8"))
3241 LocalWords: DBlocks dblocks org-babel el eric fontification