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 LoB: allow output in buffer
892 ** DONE allow default header arguments by language
893 org-babel-default-header-args:lang-name
895 An example of when this is useful is for languages which always return
896 files as their results (e.g. [[*** ditaa][ditaa]], and [[*** gnuplot][gnuplot]]).
897 ** DONE singe-function tangling and loading elisp from literate org-mode file [3/3]
899 This function should tangle the org-mode file for elisp, and then call
900 `load-file' on the resulting tangled file.
902 #+srcname: test-loading-embedded-emacs-lisp
903 #+begin_src emacs-lisp :results replace
904 (setq test-tangle-advert nil)
905 (setq test-tangle-loading nil)
906 (setq results (list :before test-tangle-loading test-tangle-advert))
907 (org-babel-load-file "test-tangle.org")
908 (setq results (list (list :after test-tangle-loading test-tangle-advert) results))
909 (delete-file "test-tangle.el")
913 #+resname: test-loading-embedded-emacs-lisp
914 | :before | nil | nil |
915 | :after | "org-babel tangles" | "use org-babel-tangle for all your emacs initialization files!!" |
917 *** DONE add optional language limiter to org-babel-tangle
918 This should check to see if there is any need to re-export
920 *** DONE ensure that org-babel-tangle returns the path to the tangled file(s)
922 #+srcname: test-return-value-of-org-babel-tangle
923 #+begin_src emacs-lisp :results replace
924 (mapcar #'file-name-nondirectory (org-babel-tangle-file "test-tangle.org" "emacs-lisp"))
930 *** DONE only tangle the file if it's actually necessary
931 ** DONE add a function to jump to a source-block by name
932 I've had an initial stab at that in org-babel-find-named-block
933 (library-of-babel branch).
935 At the same time I introduced org-babel-named-src-block-regexp, to
936 match src-blocks with srcname.
938 This is now working with the command
939 `org-babel-goto-named-source-block', all we need is a good key
942 ** DONE add =:none= session argument (for purely functional execution) [4/4]
943 This would allow source blocks to be run in their own new process
945 - These blocks could then also be run in the background (since we can
946 detach and just wait for the process to signal that it has terminated)
947 - We wouldn't be drowning in session buffers after running the tests
948 - we can re-use much of the session code to run in a more /functional/
951 While session provide a lot of cool features, like persistent
952 environments, [[* DONE function to bring up inferior-process buffer][pop-to-session]], and hints at exportation for
953 org-babel-tangle, they also have some down sides and I'm thinking that
954 session-based execution maybe shouldn't be the default behavior.
956 Down-sides to sessions
957 - *much* more complicated than functional evaluation
958 - maintaining the state of the session has weird issues
959 - waiting for evaluation to finish
960 - prompt issues like [[* TODO weird escaped characters in shell prompt break shell evaluation][shell-prompt-escapes-bug]]
961 - can't run in background
962 - litter emacs with session buffers
966 #+srcname: ruby-task-no-session
967 #+begin_src ruby :results replace output
973 #+resname: ruby-task-no-session
978 #+srcname: task-python-none-session
979 #+begin_src python :session none :results replace value
985 #+resname: task-python-none-session
990 #+srcname: task-session-none-sh
991 #+begin_src sh :results replace
996 #+resname: task-session-none-sh
1002 #+srcname: task-no-session-R
1003 #+begin_src R :results replace output
1010 #+resname: task-no-session-R
1014 ** DONE fully purge org-babel-R of direct comint interaction
1015 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
1017 ** DONE Create objects in top level (global) environment [5/5]
1020 *** initial requirement statement [DED]
1021 At the moment, objects created by computations performed in the
1022 code block are evaluated in the scope of the
1023 code-block-function-body and therefore disappear when the code
1024 block is evaluated {unless you employ some extra trickery like
1025 assign('name', object, env=globalenv()) }. I think it will be
1026 desirable to also allow for a style wherein objects that are
1027 created in one code block persist in the R global environment and
1028 can be re-used in a separate block.
1030 This is what Sweave does, and while I'm not saying we have to be
1031 the same as Sweave, it wouldn't be hard for us to provide the same
1032 behaviour in this case; if we don't, we risk undeservedly being
1033 written off as an oddity by some.
1035 IOW one aspect of org-babel is that of a sort of functional
1036 meta-programming language. This is crazy, in a very good
1037 way. Nevertheless, wrt R I think there's going to be a lot of value
1038 in providing for a working style in which the objects are stored in
1039 the R session, rather than elisp/org buffer. This will be a very
1040 familiar working style to lots of people.
1042 There are no doubt a number of different ways of accomplishing
1043 this, the simplest being a hack like adding
1046 for(objname in ls())
1047 assign(objname, get(objname), envir=globalenv())
1050 to the source code block function body. (Maybe wrap it in an on.exit() call).
1052 However this may deserve to be thought about more carefully, perhaps
1053 with a view to having a uniform approach across languages. E.g. shell
1054 code blocks have the same semantics at the moment (no persistence of
1055 variables across code blocks), because the body is evaluated in a new
1056 bash shell process rather than a running shell. And I guess the same
1057 is true for python. However, in both these cases, you could imagine
1058 implementing the alternative in which the body is evaluated in a
1059 persistent interactive session. It's just that it's particularly
1060 natural for R, seeing as both ESS and org-babel evaluate commands in a
1061 single persistent R session.
1065 Thanks for bringing this up. I think you are absolutely correct that we
1066 should provide support for a persistent environment (maybe called a
1067 *session*) in which to evaluate code blocks. I think the current setup
1068 demonstrates my personal bias for a functional style of programming
1069 which is certainly not ideal in all contexts.
1071 While the R function you mention does look like an elegant solution, I
1072 think we should choose an implementation that would be the same across
1073 all source code types. Specifically I think we should allow the user to
1074 specify an optional *session* as a header variable (when not present we
1075 assume a default session for each language). The session name could be
1076 used to name a comint buffer (like the *R* buffer) in which all
1077 evaluation would take place (within which variables would retain their
1078 values --at least once I remove some of the functional method wrappings
1079 currently in place-- ).
1081 This would allow multiple environments to be used in the same buffer,
1082 and once this setup was implemented we should be able to fairly easily
1083 implement commands for jumping between source code blocks and the
1084 related session buffers, as well as for dumping the last N commands from
1085 a session into a new or existing source code block.
1087 Please let me know if you foresee any problems with this proposed setup,
1088 or if you think any parts might be confusing for people coming from
1089 Sweave. I'll hopefully find some time to work on this later in the
1092 *** can functional and interpreted/interactive models coexist?
1094 Even though both of these use the same =*R*= buffer the value of =a=
1095 is not preserved because it is assigned inside of a functional
1098 #+srcname: task-R-sessions
1105 #+srcname: task-R-same-session
1110 This functional wrapper was implemented in order to efficiently return
1111 the results of the execution of the entire source code block. However
1112 it inhibits the evaluation of source code blocks in the top level,
1113 which would allow for persistence of variable assignment across
1114 evaluations. How can we allow *both* evaluation in the top level, and
1115 efficient capture of the return value of an entire source code block
1116 in a language independent manner?
1118 Possible solutions...
1119 1) we can't so we will have to implement two types of evaluation
1120 depending on which is appropriate (functional or imperative)
1121 2) we remove the functional wrapper and parse the source code block
1122 into it's top level statements (most often but not always on line
1123 breaks) so that we can isolate the final segment which is our
1125 3) we add some sort of "#+return" line to the code block
1126 4) we take advantage of each languages support for meta-programming
1127 through =eval= type functions, and use said to evaluate the entire
1128 blocks in such a way that their environment can be combined with the
1129 global environment, and their results are still captured.
1130 5) I believe that most modern languages which support interactive
1131 sessions have support for a =last_result= type function, which
1132 returns the result of the last input without re-calculation. If
1133 widely enough present this would be the ideal solution to a
1134 combination of functional and imperative styles.
1136 None of these solutions seem very desirable, but for now I don't see
1137 what else would be possible.
1139 Of these options I was leaning towards (1) and (4) but now believe
1140 that if it is possible option (5) will be ideal.
1142 **** (1) both functional and imperative evaluation
1144 - can take advantage of built in functions for sending regions to the
1146 - retains the proven tested and working functional wrappers
1149 - introduces the complication of keeping track of which type of
1150 evaluation is best suited to a particular context
1151 - the current functional wrappers may require some changes in order to
1152 include the existing global context
1154 **** (4) exploit language meta-programming constructs to explicitly evaluate code
1156 - only one type of evaluation
1159 - some languages may not have sufficient meta-programming constructs
1161 **** (5) exploit some =last_value= functionality if present
1163 Need to ensure that most languages have such a function, those without
1164 will simply have to implement their own similar solution...
1166 | language | =last_value= function |
1167 |------------+-----------------------------|
1171 | shell | see [[* last command for shells][last command for shells]] |
1172 | emacs-lisp | see [[* emacs-lisp will be a special case][special-case]] |
1174 #+srcname: task-last-value
1179 ***** last command for shells
1180 Do this using the =tee= shell command, and continually pipe the output
1183 Got this idea from the following [[http://linux.derkeiler.com/Mailing-Lists/Fedora/2004-01/0898.html][email-thread]].
1185 suggested from mailing list
1187 #+srcname: bash-save-last-output-to-file
1191 bash -c "$line" | tee /tmp/last.out1
1192 mv /tmp/last.out1 /tmp/last.out
1196 another proposed solution from the above thread
1198 #+srcname: bash-save-in-variable
1201 # so - Save Output. Saves output of command in OUT shell variable.
1211 "^M": " | tee /tmp/h_lastcmd.out ^[k"
1214 export __=/tmp/h_lastcmd.out
1216 If you try it, Alt-k will stand for the old Enter; use "command $__" to
1217 access the last output.
1223 Herculano de Lima Einloft Neto
1226 ***** emacs-lisp will be a special case
1227 While it is possible for emacs-lisp to be run in a console type
1228 environment (see the =elim= function) it is *not* possible to run
1229 emacs-lisp in a different *session*. Meaning any variable set top
1230 level of the console environment will be set *everywhere* inside
1231 emacs. For this reason I think that it doesn't make any sense to
1232 worry about session support for emacs-lisp.
1234 *** Further thoughts on 'scripting' vs. functional approaches
1236 These are just thoughts, I don't know how sure I am about this.
1237 And again, perhaps I'm not saying anything very radical, just that
1238 it would be nice to have some options supporting things like
1239 receiving text output in the org buffer.
1241 I can see that you've already gone some way down the road towards
1242 the 'last value' approach, so sorry if my comments come rather
1243 late. I am concerned that we are not giving sufficient attention
1244 to stdout / the text that is returned by the interpreters. In
1245 contrast, many of our potential users will be accustomed to a
1246 'scripting' approach, where they are outputting text at various
1247 points in the code block, not just at the end. I am leaning
1248 towards thinking that we should have 2 modes of evaluation:
1249 'script' mode, and 'functional' mode.
1251 In script mode, evaluation of a code block would result in *all*
1252 text output from that code block appearing as output in the org
1253 buffer, presumably as an #+begin_example...#+end_example. There
1254 could be an :echo option controlling whether the input commands
1255 also appear in the output. [This is like Sweave].
1257 In functional mode, the *result* of the code block is available as
1258 an elisp object, and may appear in the org buffer as an org
1259 table/string, via the mechanisms you have developed already.
1261 One thing I'm wondering about is whether, in script mode, there
1262 simply should not be a return value. Perhaps this is not so
1263 different from what exists: script mode would be new, and what
1264 exists currently would be functional mode.
1266 I think it's likely that, while code evaluation will be exciting
1267 to people, a large majority of our users in a large majority of
1268 their usage will not attempt to actually use the return value from
1269 a source code block in any meaningful way. In that case, it seems
1270 rather restrictive to only allow them to see output from the end
1273 Instead I think the most accessible way to introduce org-babel to
1274 people, at least while they are learning it, is as an immensely
1275 powerful environment in which to embed their 'scripts', which now
1276 also allows them to 'run' their 'scripts'. Especially as such
1277 people are likely to be the least capable of the user-base, a
1278 possible design-rule would be to make the scripting style of usage
1279 easy (default?), perhaps requiring a special option to enable a
1280 functional style. Those who will use the functional style won't
1281 have a problem understanding what's going on, whereas the 'skript
1282 kiddies' might not even know the syntax for defining a function in
1283 their language of choice. And of course we can allow the user to
1284 set a variable in their .emacs controlling the preference, so that
1285 functional users are not inconveniennced by having to provide
1286 header args the whole time.
1288 Please don't get the impression that I am down-valuing the
1289 functional style of org-babel. I am constantly horrified at the
1290 messy 'scripts' that my colleagues produce in perl or R or
1291 whatever! Nevertheless that seems to be how a lot of people work.
1293 I think you were leaning towards the last-value approach because
1294 it offered the possibility of unified code supporting both the
1295 single evaluation environment and the functional style. If you
1296 agree with any of the above then perhaps it will impact upon this
1297 and mean that the code in the two branches has to differ a bit. In
1298 that case, functional mode could perhaps after all evaluate each
1299 code block in its own environment, thus (re)approaching 'true'
1300 functional programming (side-effects are hard to achieve).
1304 echo "There are `wc -l files` files in this directory"
1308 *** even more thoughts on evaluation, results, models and options
1310 Thanks Dan, These comments are invaluable.
1312 What do you think about this as a new list of priorities/requirements
1313 for the execution of source-code blocks.
1316 1) we want the evaluation of the source code block to take place in a
1317 session which can persist state (variables, current directory,
1319 2) source code blocks can specify their session with a header argument
1320 3) each session should correspond to an Emacs comint buffer so that the
1321 user can drop into the session and experiment with live code
1324 1) each source-code block generates some form of results which (as
1325 we have already implemented) is transfered into emacs-lisp
1326 after which it can be inserted into the org-mode buffer, or
1327 used by other source-code blocks
1328 2) when the results are translated into emacs-lisp, forced to be
1329 interpreted as a scalar (dumping their raw values into the
1330 org-mode buffer), as a vector (which is often desirable with R
1331 code blocks), or interpreted on the fly (the default option).
1332 Note that this is very nearly currently implemented through the
1333 [[* DONE results-type header (vector/file)][results-type-header]].
1334 3) there should be *two* means of collecting results from the
1335 execution of a source code block. *Either* the value of the
1336 last statement of the source code block, or the collection of
1337 all that has been passed to STDOUT during the evaluation.
1339 **** header argument or return line (*header argument*)
1341 Rather than using a header argument to specify how the return value
1342 should be passed back, I'm leaning towards the use of a =#+RETURN=
1343 line inside the block. If such a line *is not present* then we
1344 default to using STDOUT to collect results, but if such a line *is
1345 present* then we use it's value as the results of the block. I
1346 think this will allow for the most elegant specification between
1347 functional and script execution. This also cleans up some issues
1348 of implementation and finding which statement is the last
1351 Having given this more thought, I think a header argument is
1352 preferable. The =#+return:= line adds new complicating syntax for
1353 something that does little more than we would accomplish through
1354 the addition of a header argument. The only benefit being that we
1355 know where the final statement starts, which is not an issue in
1356 those languages which contain 'last value' operators.
1358 new header =:results= arguments
1359 - script :: explicitly states that we want to use STDOUT to
1360 initialize our results
1361 - return_last :: stdout is ignored instead the *value* of the final
1362 statement in the block is returned
1363 - echo :: means echo the contents of the source-code block along
1364 with the results (this implies the *script* =:results=
1367 *** DONE rework evaluation lang-by-lang [4/4]
1369 This should include...
1370 - functional results working with the comint buffer
1372 - script :: return the output of STDOUT
1373 - write a macro which runs the first redirection, executes the
1374 body, then runs the second redirection
1375 - last :: return the value of the last statement
1378 - sessions in comint buffers
1380 **** DONE Ruby [4/4]
1381 - [X] functional results working with comint
1382 - [X] script results
1383 - [X] ensure scalar/vector results args are taken into consideration
1384 - [X] ensure callable by other source block
1386 #+srcname: ruby-use-last-output
1387 #+begin_src ruby :results replace
1394 #+resname: ruby-use-last-output
1397 #+srcname: task-call-use-last-output
1398 #+begin_src ruby :var last=ruby-use-last-output :results replace
1399 last.flatten.size + 1
1402 #+resname: task-call-use-last-output
1407 #+srcname: first-ruby-session-task
1408 #+begin_src ruby :session schulte :results silent
1412 #+srcname: second-ruby-session-task
1413 #+begin_src ruby :session schulte :results silent
1417 #+srcname: without-the-right-session
1418 #+begin_src ruby :results silent
1424 - [X] functional results working with comint
1425 - [X] script results
1426 - [X] ensure scalar/vector results args are taken into consideration
1427 - [X] ensure callable by other source block
1429 To redirect output to a file, you can use the =sink()= command.
1432 #+begin_src R :results value vector silent
1439 #+srcname: task-R-use-other-output
1440 #+begin_src R :var twoentyseven=task_R_B() :results replace value
1445 #+resname: task-R-use-other-output
1448 **** DONE Python [4/4]
1449 - [X] functional results working with comint
1450 - [X] script results
1451 - [X] ensure scalar/vector results args are taken into consideration
1452 - [X] ensure callable by other source block
1454 #+srcname: task-new-eval-for-python
1455 #+begin_src python :results silent output scalar
1461 #+srcname: task-use-new-eval
1462 #+begin_src python :var tasking=task-new-eval-for-python() :results replace
1466 #+resname: task-use-new-eval
1469 **** DONE Shells [4/4]
1470 - [X] functional results working with comint
1471 - [X] script results
1472 - [X] ensure scalar/vector results args are taken into consideration
1473 - [X] ensure callable by other source block
1475 #+srcname: task-shell-new-evaluation
1476 #+begin_src sh :results silent value scalar
1481 #+srcname: task-call-other-shell
1482 #+begin_src sh :var other=task-shell-new-evaluation() :results replace scalar
1483 echo $other ' is the old date'
1486 #+resname: task-call-other-shell
1487 : $ Fri Jun 12 13:08:37 PDT 2009 is the old date
1489 *** DONE implement a *session* header argument [4/4]
1490 =:session= header argument to override the default *session* buffer
1494 #+srcname: task-ruby-named-session
1495 #+begin_src ruby :session schulte :results replace
1496 schulte = :in_schulte
1499 #+resname: task-ruby-named-session
1502 #+srcname: another-in-schulte
1503 #+begin_src ruby :session schulte
1507 #+resname: another-in-schulte
1514 #+srcname: python-session-task
1515 #+begin_src python :session what :results silent
1519 #+srcname: python-get-from-session
1520 #+begin_src python :session what :results replace
1524 #+resname: python-get-from-session
1529 #+srcname: task-shell-sessions
1530 #+begin_src sh :session what
1534 #+srcname: task-shell-sessions-what
1535 #+begin_src sh :session what :results replace
1539 #+resname: task-shell-sessions-what
1544 #+srcname: task-R-session
1545 #+begin_src R :session what :results replace
1551 #+resname: task-R-session
1554 #+srcname: another-task-R-session
1555 #+begin_src R :session what :results replace
1559 *** DONE function to bring up inferior-process buffer [4/4]
1561 This should be callable from inside of a source-code block in an
1562 org-mode buffer. It should evaluate the header arguments, then bring
1563 up the inf-proc buffer using =pop-to-buffer=.
1565 For lack of a better place, lets add this to the `org-metadown-hook'
1568 To give this a try, place the cursor on a source block with variables,
1569 (optionally git a prefix argument) then hold meta and press down.
1573 #+srcname: task-ruby-pop-to-session
1574 #+begin_src ruby :var num=9 :var another="something else"
1575 num.times{|n| puts another}
1580 #+srcname: task-python-pop-to-session
1581 #+begin_src python :var num=9 :var another="something else"
1586 #+srcname: task-R-pop-to-session
1587 #+begin_src R :var a=9 :var b=8
1593 #+srcname: task-shell-pop-sessions
1594 #+begin_src sh :var NAME="eric"
1598 *** DEFERRED function to dump last N lines from inf-proc buffer into the current source block
1600 Callable with a prefix argument to specify how many lines should be
1601 dumped into the source-code buffer.
1603 *** REJECTED comint notes
1605 Implementing comint integration in [[file:lisp/org-babel-comint.el][org-babel-comint.el]].
1608 - handling of outputs
1609 - split raw output from process by prompts
1610 - a ring of the outputs, buffer-local, `org-babel-comint-output-ring'
1611 - a switch for dumping all outputs to a buffer
1612 - inputting commands
1614 Lets drop all this language specific stuff, and just use
1615 org-babel-comint to split up our outputs, and return either the last
1616 value of an execution or the combination of values from the
1619 **** comint filter functions
1620 : ;; comint-input-filter-functions hook process-in-a-buffer
1621 : ;; comint-output-filter-functions hook function modes.
1622 : ;; comint-preoutput-filter-functions hook
1623 : ;; comint-input-filter function ...
1625 #+srcname: obc-filter-ruby
1626 #+begin_src ruby :results last
1634 ** DONE Remove protective commas from # comments before evaluating
1635 org inserts protective commas in front of ## comments in language
1636 modes that use them. We need to remove them prior to sending code
1639 #+srcname: testing-removal-of-protective-comas
1641 ,# this one might break it??
1645 ** DONE pass multiple reference arguments into R
1646 Can we do this? I wasn't sure how to supply multiple 'var' header
1647 args. Just delete this if I'm being dense.
1649 This should be working, see the following example...
1651 #+srcname: two-arg-example
1652 #+begin_src R :var n=2 :var m=8
1656 #+resname: two-arg-example
1659 ** DONE ensure that table ranges work
1660 when a table range is passed to org-babel as an argument, it should be
1661 interpreted as a vector.
1664 | 2 | 3 | Fixnum:1 |
1665 | 3 | 4 | Array:123456 |
1669 #+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)))
1671 #+srcname: simple-sbe-example
1672 #+begin_src emacs-lisp
1676 #+srcname: task-table-range
1677 #+begin_src ruby :var n=simple-sbe-example
1681 #+srcname: simple-results
1682 #+begin_src emacs-lisp :var n=task-table-range(n=(1 2 3))
1686 #+resname: simple-results
1689 #+srcname: task-arr-referent
1690 #+begin_src ruby :var ar=(1 2 3)
1694 #+resname: task-arr-referent
1697 ** DONE global variable indicating default to vector output
1698 how about an alist... =org-babel-default-header-args= this may already
1699 exist... just execute the following and all source blocks will default
1702 #+begin_src emacs-lisp
1703 (setq org-babel-default-header-args '((:results . "vector")))
1706 ** DONE name named results if source block is named
1707 currently this isn't happening although it should be
1709 #+srcname: test-naming-named-source-blocks
1710 #+begin_src emacs-lisp
1714 #+resname: test-naming-named-source-blocks
1716 ** DONE (simple caching) check for named results before source blocks
1717 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]]
1718 ** DONE set =:results silent= when eval with prefix argument
1720 #+begin_src emacs-lisp
1723 ** DONE results-type header (vector/file) [3/3]
1724 In response to a point in Dan's email. We should allow the user to
1725 force scalar or vector results. This could be done with a header
1726 argument, and the default behavior could be controlled through a
1727 configuration variable.
1729 #+srcname: task-trivial-vector
1730 #+begin_src ruby :results replace vector
1737 since it doesn't make sense to turn a vector into a scalar, lets
1738 just add a two values...
1740 - vector :: forces the results to be a vector (potentially 1 dimensional)
1741 - file :: this throws an error if the result isn't a string, and
1742 tries to treat it as a path to a file.
1744 I'm just going to cram all of these into the =:results= header
1745 argument. Then if we allow multiple header arguments it should
1746 work out, for example one possible header argument string could be
1747 =:results replace vector file=, which would *replace* any existing
1748 results forcing the results into an org-mode table, and
1749 interpreting any strings as file paths.
1751 *** DONE multiple =:results= headers
1753 #+srcname: multiple-result-headers
1754 #+begin_src ruby :results replace silent
1760 *** DONE file result types
1761 When inserting into an org-mode buffer create a link with the path
1762 being the value, and optionally the display being the
1763 =file-name-nondirectory= if it exists.
1765 #+srcname: task-file-result
1766 #+begin_src python :results replace file
1771 [[something][something]]
1774 This will be useful because blocks like =ditaa= and =dot= can return
1775 the string path of their files, and can add =file= to their results
1778 *** DONE vector result types
1780 #+srcname: task-force-results
1781 #+begin_src emacs-lisp :results vector
1788 ** DONE results name
1789 In order to do this we will need to start naming our results.
1790 Since the source blocks are named with =#+srcname:= lines we can
1791 name results with =#+resname:= lines (if the source block has no
1792 name then no name is given to the =#+resname:= line on creation,
1793 otherwise the name of the source block is used).
1795 This will have the additional benefit of allowing results and
1796 source blocks to be located in different places in a buffer (and
1797 eventually in different buffers entirely).
1799 #+srcname: developing-resnames
1800 #+begin_src emacs-lisp :results silent
1804 Once source blocks are able to find their own =#+resname:= lines
1807 #+srcname: sbe-w-new-results
1808 #+begin_src emacs-lisp :results replace
1809 (sbe "developing-resnames")
1815 *** TODO change the results insertion functions to use these lines
1817 *** TODO teach references to resolve =#+resname= lines.
1819 ** DONE org-babel tests org-babel [1/1]
1820 since we are accumulating this nice collection of source-code blocks
1821 in the sandbox section we should make use of them as unit tests.
1822 What's more, we should be able to actually use org-babel to run these
1825 We would just need to cycle over every source code block under the
1826 sandbox, run it, and assert that the return value is equal to what we
1829 I have the feeling that this should be possible using only org-babel
1830 functions with minimal or no additional elisp. It would be very cool
1831 for org-babel to be able to test itself.
1833 This is now done, see [[* Tests]].
1835 *** DEFERRED org-babel assertions (may not be necessary)
1836 These could be used to make assertions about the results of a
1837 source-code block. If the assertion fails then the point could be
1838 moved to the block, and error messages and highlighting etc... could
1841 ** DONE make C-c C-c work anywhere within source code block?
1842 This seems like it would be nice to me, but perhaps it would be
1843 inefficient or ugly in implementation? I suppose you could search
1844 forward, and if you find #+end_src before you find #+begin_src,
1845 then you're inside one. [DED]
1847 Agreed, I think inside of the =#+srcname: line= would be useful as
1850 #+srcname: testing-out-cc
1851 #+begin_src emacs-lisp
1855 ** DONE integration with org tables
1856 We should make it easy to call org-babel source blocks from org-mode
1857 table formulas. This is practical now that it is possible to pass
1858 arguments to org-babel source blocks.
1860 See the related [[* (sandbox) integration w/org tables][sandbox]] header for tests/examples.
1862 *** digging in org-table.el
1863 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.
1865 Should be a hook in [[file:~/src/org/lisp/org-table.el::defun%20org%20table%20eval%20formula%20optional%20arg%20equation][org-table-eval-formula]].
1867 Looks like I need to change this [[file:~/src/org/lisp/org-table.el::if%20lispp][if statement]] (line 2239) into a cond
1870 ** DONE source blocks as functions
1872 Allow source code blocks to be called like functions, with arguments
1873 specified. We are already able to call a source-code block and assign
1874 it's return result to a variable. This would just add the ability to
1875 specify the values of the arguments to the source code block assuming
1876 any exist. For an example see
1878 When a variable appears in a header argument, how do we differentiate
1879 between it's value being a reference or a literal value? I guess this
1880 could work just like a programming language. If it's escaped or in
1881 quotes, then we count it as a literal, otherwise we try to look it up
1884 ** DONE folding of code blocks? [2/2]
1885 [DED] In similar way to using outline-minor-mode for folding function
1886 bodies, can we fold code blocks? #+begin whatever statements are
1887 pretty ugly, and in any case when you're thinking about the overall
1888 game plan you don't necessarily want to see the code for each Step.
1890 *** DONE folding of source code block
1891 Sounds good, and wasn't too hard to implement. Code blocks should
1892 now be fold-able in the same manner as headlines (by pressing TAB
1895 *** REJECTED folding of results
1896 So, lets do a three-stage tab cycle... First fold the src block,
1897 then fold the results, then unfold.
1899 There's no way to tell if the results are a table or not w/o
1900 actually executing the block which would be too expensive of an
1903 ** DONE selective export of text, code, figures
1904 [DED] The org-babel buffer contains everything (code, headings and
1905 notes/prose describing what you're up to, textual/numeric/graphical
1906 code output, etc). However on export to html / LaTeX one might want
1907 to include only a subset of that content. For example you might
1908 want to create a presentation of what you've done which omits the
1911 [EMS] So I think this should be implemented as a property which can
1912 be set globally or on the outline header level (I need to review
1913 the mechanics of org-mode properties). And then as a source block
1914 header argument which will apply only to a specific source code
1915 block. A header argument of =:export= with values of
1917 - =code= :: just show the code in the source code block
1918 - =none= :: don't show the code or the results of the evaluation
1919 - =results= :: just show the results of the code evaluation (don't
1920 show the actual code)
1921 - =both= :: show both the source code, and the results
1923 this will be done in [[* (sandbox) selective export][(sandbox) selective export]].
1925 ** DONE a header argument specifying silent evaluation (no output)
1926 This would be useful across all types of source block. Currently
1927 there is a =:replace t= option to control output, this could be
1928 generalized to an =:output= option which could take the following
1929 options (maybe more)
1931 - =t= :: this would be the default, and would simply insert the
1932 results after the source block
1933 - =replace= :: to replace any results which may already be there
1934 - =silent= :: this would inhibit any insertion of the results
1936 This is now implemented see the example in the [[* silent evaluation][sandbox]]
1938 ** DONE assign variables from tables in R
1939 This is now working (see [[* (sandbox table) R][(sandbox-table)-R]]). Although it's not that
1940 impressive until we are able to print table results from R.
1942 ** DONE insert 2-D R results as tables
1943 everything is working but R and shell
1949 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
1950 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]]
1951 as a means of converting multidimensional R objects to emacs lisp.
1953 It may be as simple as first checking if the data is multidimensional,
1954 and then, if so using =write= to write the data out to a temporary
1955 file from which emacs can read the data in using =org-table-import=.
1957 Looking into this further, is seems that there is no such thing as a
1958 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
1959 deal with trivial vectors (scalars) in R. I'm tempted to just treat
1960 them as vectors, but then that would lead to a proliferation of
1961 trivial 1-cell tables...
1963 ** DONE allow variable initialization from source blocks
1964 Currently it is possible to initialize a variable from an org-mode
1965 table with a block argument like =table=sandbox= (note that the
1966 variable doesn't have to named =table=) as in the following example
1972 #+begin_src emacs-lisp :var table=sandbox :results replace
1973 (message (format "table = %S" table))
1976 : "table = ((1 2 3) (4 \"schulte\" 6))"
1978 It would be good to allow initialization of variables from the results
1979 of other source blocks in the same manner. This would probably
1980 require the addition of =#+SRCNAME: example= lines for the naming of
1981 source blocks, also the =table=sandbox= syntax may have to be expanded
1982 to specify whether the target is a source code block or a table
1983 (alternately we could just match the first one with the given name
1984 whether it's a table or a source code block).
1986 At least initially I'll try to implement this so that there is no need
1987 to specify whether the reference is to a table or a source-code block.
1988 That seems to be simpler both in terms of use and implementation.
1990 This is now working for emacs-lisp, ruby and python (and mixtures of
1991 the three) source blocks. See the examples in the [[* (sandbox) referencing other source blocks][sandbox]].
1993 This is currently working only with emacs lisp as in the following
1994 example in the [[* emacs lisp source reference][emacs lisp source reference]].
1997 ** TODO Add languages [1/6]
1998 I'm sure there are many more that aren't listed here. Please add
1999 them, and bubble any that you particularly care about up to the top.
2001 Any new language should be implemented in a org-babel-lang.el file.
2002 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
2003 [[file:lisp/org-babel-R.el][org-babel-R.el]].
2006 This could probably be added to [[file:lisp/org-babel-script.el][org-babel-script.el]]
2011 (see [[* file result types][file result types]])
2013 #+srcname: implementing-ditaa
2014 #+begin_src ditaa :results replace :file blue.png :cmdline -r
2024 #+resname: implementing-ditaa
2025 [[file:blue.png][blue.png]]
2028 (see [[* file result types][file result types]])
2030 - a =file= header argument
2031 - a =cmdline= header argument
2033 - scalar variables should be replaced in the body of the gnuplot code
2034 - vector variables should be exported to tab-separated files, and
2035 the variable names should be replaced with the path to the files
2037 #+PLOT: title:"Citas" ind:1 deps:(3) type:2d with:histograms set:"yrange [0:]"
2038 #+TBLNAME: gnuplot-data
2039 | independent var | first dependent var | second dependent var |
2040 |-----------------+---------------------+----------------------|
2041 | 0.1 | 0.425 | 0.375 |
2042 | 0.2 | 0.3125 | 0.3375 |
2043 | 0.3 | 0.24999993 | 0.28333338 |
2044 | 0.4 | 0.275 | 0.28125 |
2045 | 0.5 | 0.26 | 0.27 |
2046 | 0.6 | 0.25833338 | 0.24999993 |
2047 | 0.7 | 0.24642845 | 0.23928553 |
2048 | 0.8 | 0.23125 | 0.2375 |
2049 | 0.9 | 0.23333323 | 0.2333332 |
2050 | 1 | 0.2225 | 0.22 |
2051 | 1.1 | 0.20909075 | 0.22272708 |
2052 | 1.2 | 0.19999998 | 0.21458333 |
2053 | 1.3 | 0.19615368 | 0.21730748 |
2055 #+srcname: implementing-gnuplot
2056 #+begin_src gnuplot :var data=gnuplot-data
2057 set title "Implementing Gnuplot"
2058 plot "data" using 1:2 with lines
2062 (see [[* file result types][file result types]])
2065 (see [[* file result types][file result types]])
2069 ** TODO prompt characters appearing in output with R
2070 #+begin_src R :session *R* :results output
2079 ** TODO o-b-execute-subtree overwrites heading when subtree is folded
2081 Try M-x org-babel-execute-subtree with the subtree folded and
2082 point at the beginning of the heading line.
2087 ** TODO Allow source blocks to be recognised when #+ are not first characters on the line
2088 I think Carsten has recently altered the core so that #+ can have
2089 preceding whitespace, at least for literal/code examples. org-babel
2090 should support this.
2092 ** TODO non-orgtbl formatted lists
2095 #+srcname: this-doesn't-match-orgtbl
2096 #+begin_src emacs-lisp :results replace
2097 '((:results . "replace"))
2100 #+resname: this-doesn't-match-orgtbl
2102 ** TODO collapsing consecutive newlines in string output
2104 #+srcname: multi-line-string-output
2105 #+begin_src ruby :results output
2106 "the first line ends here
2109 and this is the second one
2114 #+resname: multi-line-string-output
2116 ** PROPOSED external shell execution can't isolate return values
2117 I have no idea how to do this as of yet. The result is that when
2118 shell functions are run w/o a session there is no difference between
2119 the =output= and =value= result arguments.
2121 Yea, I don't know how to do this either. I searched extensively on
2122 how to isolate the *last* output of a series of shell commands (see
2123 [[* last command for
2124 shells][last command for shells]]). The results of the search were basically
2125 that it was not possible (or at least not accomplish-able with a
2126 reasonable amount of effort).
2128 That fact combined with the tenancy to all ways use standard out in
2129 shell scripts led me to treat these two options (=output= and =value=)
2130 as identical in shell evaluation. Not ideal but maybe good enough for
2133 In the `results' branch I've changed this so that they're not quite
2134 identical: output results in raw stdout contents, whereas value
2135 converts it to elisp, perhaps to a table if it looks tabular. This is
2136 the same for the other languages. [Dan]
2138 ** TODO are the org-babel-trim s necessary?
2139 at the end of e.g. org-babel-R-evaluate, org-babel-python-evaluate, but
2140 not org-babel-ruby-evaluate
2141 ** TODO elisp reference fails for literal number
2142 #+srcname: elisp-test(a=4)
2143 #+begin_src emacs-lisp
2146 ** TODO use new merge function [[file:lisp/org-babel-ref.el::t%20nil%20org%20combine%20plists%20args%20nil][here]]?
2147 And at other occurrences of org-combine-plists?
2148 ** TODO LoB: with output to buffer, not working in buffers other than library-of-babel.org
2149 I haven't fixed this yet. org-babel-ref-resolve-reference moves
2150 point around, inside a save-excursion. Somehow when it comes to
2151 inserting the results (after possible further recursive calls to
2152 org-babel-ref-resolve-reference), point hasn't gone back to the
2154 ** TODO creeping blank lines
2155 There's still inappropriate addition of blank lines in some circumstances. E.g.
2161 Compare the results of
2162 #+lob: python-add(a=5, b=17)
2164 #+resname: python-add(a=5, b=17)
2166 --------------------------------
2174 ---------------------
2176 Hmm, it's a bit confusing. I think it's to do with the fact that
2177 LoB removes the entire (#+resname and result) and starts from
2178 scratch, whereas #+begin_src only removes the result. I haven't
2179 worked out what the correct fix is yet.
2180 ** DEFERRED weird escaped characters in shell prompt break shell evaluation
2181 E.g. this doesn't work. Should the shell sessions set a sane prompt
2182 when they start up? Or is it a question of altering
2183 comint-prompt-regexp? Or altering org-babel regexps?
2186 black=30 ; red=31 ; green=32 ; yellow=33 ; blue=34 ; magenta=35 ; cyan=36 ; white=37
2189 export PS1="\[\033[${prompt_col}m\]\w${prompt_char} \[\033[0m\]"
2192 I just pushed a good amount of changes, could you see if your shell
2193 problems still exist?
2195 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
2196 org-babel-sh.el, raw gets the value
2198 ("" "
\e[0m Sun Jun 14 19:26:24 EDT 2009\n" "
\e[0m org_babel_sh_eoe\n" "
\e[0m ")
2200 and therefore (member org-babel-sh-eoe-output ...) fails
2202 I think that `comint-prompt-regexp' needs to be altered to match
2203 the shell prompt. This shouldn't be too difficult to do by hand,
2204 using the `regexp-builder' command and should probably be part of
2205 the user's regular emacs init. I can't think of a way for us to
2206 set this automatically, and we are SOL without a regexp to match
2209 ** DONE cursor movement when evaluating source blocks
2210 E.g. the pie chart example. Despite the save-window-excursion in
2211 org-babel-execute:R. (I never learned how to do this properly: org-R
2212 jumps all over the place...)
2214 I don't see this now [ded]
2216 ** DONE LoB: calls fail if reference has single character name
2217 commit 21d058869df1ff23f4f8cc26f63045ac9c0190e2
2218 **** This doesn't work
2219 #+lob: R-plot(data=X)
2238 #+lob: R-plot(data=XX)
2240 ** DONE make :results replace the default?
2241 I'm tending to think that appending results to pre-existing results
2242 creates mess, and that the cleaner `replace' option should be the
2243 default. E.g. when a source block creates an image, we would want
2244 that to be updated, rather than have a new one be added.
2248 ** DONE ruby evaluation not working under ubuntu emacs 23
2249 With emacs 23.0.91.1 on ubuntu, for C-h f run-ruby I have the
2250 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.
2253 run-ruby is an interactive compiled Lisp function.
2257 Run an inferior Ruby process, input and output via buffer *ruby*.
2258 If there is a process already running in `*ruby*', switch to that buffer.
2259 With argument, allows you to edit the command line (default is value
2260 of `ruby-program-name'). Runs the hooks `inferior-ruby-mode-hook'
2261 (after the `comint-mode-hook' is run).
2262 (Type C-h m in the process buffer for a list of commands.)
2265 So, I may have a non-standard inf-ruby.el. Here's my version of
2269 run-ruby is an interactive Lisp function in `inf-ruby.el'.
2271 (run-ruby &optional COMMAND NAME)
2273 Run an inferior Ruby process, input and output via buffer *ruby*.
2274 If there is a process already running in `*ruby*', switch to that buffer.
2275 With argument, allows you to edit the command line (default is value
2276 of `ruby-program-name'). Runs the hooks `inferior-ruby-mode-hook'
2277 (after the `comint-mode-hook' is run).
2278 (Type C-h m in the process buffer for a list of commands.)
2281 It seems we could either bundle my version of inf-ruby.el (as it's
2282 the newest). Or we could change the use of `run-ruby' so that it
2283 is robust across multiple distributions. I think I'd prefer the
2284 former, unless the older version of inf-ruby is actually bundled
2285 with emacs, in which case maybe we should go out of our way to
2286 support it. Thoughts?
2288 I think for now I'll just include the latest [[file:util/inf-ruby.el][inf-ruby.el]] in the
2289 newly created utility directory. I doubt anyone would have a
2290 problem using the latest version of this file.
2291 ** DONE test failing forcing vector results with =test-forced-vector-results= ruby code block
2292 Note that this only seems to happen the *second* time the test table
2295 #+srcname: bug-trivial-vector
2296 #+begin_src emacs-lisp :results vector silent
2300 #+srcname: bug-forced-vector-results
2301 #+begin_src ruby :var triv=test-trivial-vector :results silent
2305 mysteriously this seems to be fixed...
2306 ** DONE defunct R sessions
2307 Sometimes an old R session will turn defunct, and newly inserted code
2308 will not be evaluated (leading to a hang).
2310 This seems to be fixed by using `inferior-ess-send-input' rather than `comint-send-input'.
2311 ** DONE ruby fails on first call to non-default session
2313 #+srcname: bug-new-session
2314 #+begin_src ruby :session is-new
2318 ** DONE when reading results from =#+resname= line
2320 Errors when trying to read from resname lines.
2322 #+resname: bug-in-resname
2325 #+srcname: bug-in-resname-reader
2326 #+begin_src emacs-lisp :var buggy=bug-in-resname() :results silent
2330 ** DONE R-code broke on "org-babel" rename
2332 #+srcname: bug-R-babels
2337 ** DONE error on trivial R results
2339 So I know it's generally not a good idea to squash error without
2340 handling them, but in this case the error almost always means that
2341 there was no file contents to be read by =org-table-import=, so I
2344 #+srcname: bug-trivial-r1
2345 #+begin_src R :results replace
2346 pie(c(1, 2, 3), labels = c(1, 2, 3))
2349 #+srcname: bug-trivial-r2
2350 #+begin_src R :results replace
2354 #+resname: bug-trivial-r2
2357 #+srcname: bug-trivial-r3
2358 #+begin_src R :results replace
2362 #+resname: bug-trivial-r3
2367 ** DONE ruby new variable creation (multi-line ruby blocks)
2368 Actually it looks like we were dropping all but the last line.
2370 #+srcname: multi-line-ruby-test
2371 #+begin_src ruby :var table=bug-numerical-table :results replace
2373 table.each{|n| total += n}
2380 ** DONE R code execution seems to choke on certain inputs
2381 Currently the R code seems to work on vertical (but not landscape)
2384 #+srcname: little-fake
2385 #+begin_src emacs-lisp
2390 #+begin_src R :var num=little-fake
2397 #+srcname: set-debug-on-error
2398 #+begin_src emacs-lisp :results silent
2399 (setq debug-on-error t)
2402 #+srcname: bug-numerical-table
2403 #+begin_src emacs-lisp :results silent
2410 #+srcname: bug-R-number-evaluation
2411 #+begin_src R :var table=bug-numerical-table
2420 #+tblname: bug-vert-table
2425 #+srcname: bug-R-vertical-table
2426 #+begin_src R :var table=bug-vert-table :results silent
2430 ** DONE org bug/request: prevent certain org behaviour within code blocks
2431 E.g. [[]] gets recognised as a link (when there's text inside the
2432 brackets). This is bad for R code at least, and more generally
2433 could be argued to be inappropriate. Is it difficult to get org to
2434 ignore text in code blocks? [DED]
2436 I believe Carsten addressed this recently on the mailing list with
2437 the comment that it was indeed a difficult issue. I believe this
2438 may be one area where we could wait for an upstream (org-mode) fix.
2440 [Dan] Carsten has fixed this now in the core.
2442 ** DONE with :results replace, non-table output doesn't replace table output
2443 And vice versa. E.g. Try this first with table and then with len(table) [DED]
2444 #+begin_src python :var table=sandbox :results replace
2449 | 4 | "schulte" | 6 |
2452 Yes, this is certainly a problem. I fear that if we begin replacing
2453 anything immediately following a source block (regardless of whether
2454 it matches the type of our current results) we may accidentally delete
2455 hand written portions of the user's org-mode buffer.
2457 I think that the best solution here would be to actually start
2458 labeling results with a line that looks something like...
2462 This would have a couple of benefits...
2463 1) we wouldn't have to worry about possibly deleting non-results
2464 (which is currently an issue)
2465 2) we could reliably replace results even if there are different types
2466 3) we could reference the results of a source-code block in variable
2467 definitions, which would be useful if for example we don't wish to
2468 re-run a source-block every time because it is long-running.
2470 Thoughts? If no-one objects, I believe I will implement the labeling
2473 ** DONE extra quotes for nested string
2474 Well R appears to be reading the tables without issue...
2476 these *should* be quoted
2478 #+begin_src sh :results replace
2483 | "README.markdown" |
2486 | "existing_tools" |
2490 | "test-export.html" |
2491 | "test-export.org" |
2493 #+srcname: test-quotes
2494 #+begin_src ruby :var tab=ls
2500 #+srcname: test-quotes
2501 #+begin_src R :var tab=ls
2507 ** DONE simple ruby arrays not working
2509 As an example eval the following. Adding a line to test
2511 #+tblname: simple-ruby-array
2514 #+srcname: ruby-array-test
2515 #+begin_src ruby :var ar = simple-ruby-array :results silent
2519 ** DONE space trailing language name
2520 fix regexp so it works when there's a space trailing the language name
2522 #+srcname: test-trailing-space
2527 ** DONE Args out of range error
2529 The following block resulted in the error below [DED]. It ran without
2530 error directly in the shell.
2533 for platf in ill aff ; do
2534 for pop in CEU YRI ASI ; do
2535 rm -f $platf/hapmap-genos-$pop-all $platf/hapmap-rs-all
2536 cat $platf/hapmap-genos-$pop-* > $platf/hapmap-genos-$pop-all
2537 cat $platf/hapmap-rs-* > $platf/hapmap-rs-all
2542 executing source block with sh...
2543 finished executing source block
2544 string-equal: Args out of range: "", -1, 0
2546 the error =string-equal: Args out of range: "", -1, 0= looks like what
2547 used to be output when the block returned an empty results string.
2548 This should be fixed in the current version, you should now see the
2549 following message =no result returned by source block=.
2551 ** DONE ruby arrays not recognized as such
2553 Something is wrong in [[file:lisp/org-babel-script.el]] related to the
2554 recognition of ruby arrays as such.
2556 #+begin_src ruby :results replace
2562 #+begin_src python :results replace
2574 Evaluate all the cells in this table for a comprehensive test of the
2575 org-babel functionality.
2577 *Note*: if you have customized =org-babel-default-header-args= then some
2578 of these tests may fail.
2580 #+TBLNAME: org-babel-tests
2581 | functionality | block | arg | expected | results | pass |
2582 |-------------------------+----------------------------+-----+-------------+-------------+------|
2583 | basic evaluation | | | | | pass |
2584 |-------------------------+----------------------------+-----+-------------+-------------+------|
2585 | emacs lisp | basic-elisp | | 5 | 5 | pass |
2586 | shell | basic-shell | | 6 | 6 | pass |
2587 | ruby | basic-ruby | | org-babel | org-babel | pass |
2588 | python | basic-python | | hello world | hello world | pass |
2589 | R | basic-R | | 13 | 13 | pass |
2590 |-------------------------+----------------------------+-----+-------------+-------------+------|
2591 | tables | | | | | pass |
2592 |-------------------------+----------------------------+-----+-------------+-------------+------|
2593 | emacs lisp | table-elisp | | 3 | 3 | pass |
2594 | ruby | table-ruby | | 1-2-3 | 1-2-3 | pass |
2595 | python | table-python | | 5 | 5 | pass |
2596 | R | table-R | | 3.5 | 3.5 | pass |
2597 |-------------------------+----------------------------+-----+-------------+-------------+------|
2598 | source block references | | | | | pass |
2599 |-------------------------+----------------------------+-----+-------------+-------------+------|
2600 | all languages | chained-ref-last | | Array | Array | pass |
2601 |-------------------------+----------------------------+-----+-------------+-------------+------|
2602 | source block functions | | | | | pass |
2603 |-------------------------+----------------------------+-----+-------------+-------------+------|
2604 | emacs lisp | defun-fibb | | fibbd | fibbd | pass |
2605 | run over | Fibonacci | 0 | 1 | 1 | pass |
2606 | a | Fibonacci | 1 | 1 | 1 | pass |
2607 | variety | Fibonacci | 2 | 2 | 2 | pass |
2608 | of | Fibonacci | 3 | 3 | 3 | pass |
2609 | different | Fibonacci | 4 | 5 | 5 | pass |
2610 | arguments | Fibonacci | 5 | 8 | 8 | pass |
2611 |-------------------------+----------------------------+-----+-------------+-------------+------|
2612 | bugs and tasks | | | | | pass |
2613 |-------------------------+----------------------------+-----+-------------+-------------+------|
2614 | simple ruby arrays | ruby-array-test | | 3 | 3 | pass |
2615 | R number evaluation | bug-R-number-evaluation | | 2 | 2 | pass |
2616 | multi-line ruby blocks | multi-line-ruby-test | | 2 | 2 | pass |
2617 | forcing vector results | test-forced-vector-results | | Array | Array | pass |
2618 |-------------------------+----------------------------+-----+-------------+-------------+------|
2619 | sessions | | | | | pass |
2620 |-------------------------+----------------------------+-----+-------------+-------------+------|
2621 | set ruby session | set-ruby-session-var | | :set | :set | pass |
2622 | get from ruby session | get-ruby-session-var | | 3 | 3 | pass |
2623 | set python session | set-python-session-var | | set | set | pass |
2624 | get from python session | get-python-session-var | | 4 | 4 | pass |
2625 | set R session | set-R-session-var | | set | set | pass |
2626 | get from R session | get-R-session-var | | 5 | 5 | pass |
2627 #+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))
2628 #+TBLFM: $5=""::$6=""
2630 The second line (followed by replacing '[]' with '') can be used to blank out the table results, in the absence of a better method.
2634 #+srcname: basic-elisp
2635 #+begin_src emacs-lisp :results silent
2640 #+srcname: basic-shell
2641 #+begin_src sh :results silent
2646 #+srcname: date-simple
2647 #+begin_src sh :results silent
2651 #+srcname: basic-ruby
2652 #+begin_src ruby :results silent
2657 #+srcname: basic-python
2658 #+begin_src python :results silent
2664 #+begin_src R :results silent
2672 #+tblname: test-table
2676 #+srcname: table-elisp
2677 #+begin_src emacs-lisp :results silent :var table=test-table
2678 (length (car table))
2682 #+srcname: table-ruby
2683 #+begin_src ruby :results silent :var table=test-table
2684 table.first.join("-")
2688 #+srcname: table-python
2689 #+begin_src python :var table=test-table
2695 #+begin_src R :var table=test-table
2702 Lets pass a references through all of our languages...
2704 Lets start by reversing the table from the previous examples
2706 #+srcname: chained-ref-first
2707 #+begin_src python :var table = test-table
2712 #+resname: chained-ref-first
2716 Take the first part of the list
2718 #+srcname: chained-ref-second
2719 #+begin_src R :var table = chained-ref-first
2723 #+resname: chained-ref-second
2727 Turn the numbers into string
2729 #+srcname: chained-ref-third
2730 #+begin_src emacs-lisp :var table = chained-ref-second
2731 (mapcar (lambda (el) (format "%S" el)) table)
2734 #+resname: chained-ref-third
2737 and Check that it is still a list
2739 #+srcname: chained-ref-last
2740 #+begin_src ruby :var table=chained-ref-third
2745 ** source blocks as functions
2747 #+srcname: defun-fibb
2748 #+begin_src emacs-lisp :results silent
2749 (defun fibbd (n) (if (< n 2) 1 (+ (fibbd (- n 1)) (fibbd (- n 2)))))
2753 #+srcname: fibonacci
2754 #+begin_src emacs-lisp :results silent :var n=7
2764 ** sbe tests (these don't seem to be working...)
2765 Testing the insertion of results into org-mode tables.
2767 #+srcname: multi-line-output
2768 #+begin_src ruby :results replace
2769 "the first line ends here
2772 and this is the second one
2778 : the first line ends here\n\n\n and this is the second one\n\neven a third
2780 #+srcname: multi-line-error
2781 #+begin_src ruby :results replace
2782 raise "oh nooooooooooo"
2788 | the first line ends here... | -:5: warning: parenthesize argument(s) for future version... |
2789 #+TBLFM: $1='(sbe "multi-line-output")::$2='(sbe "multi-line-error")
2791 ** forcing results types tests
2793 #+srcname: test-trivial-vector
2794 #+begin_src emacs-lisp :results vector silent
2798 #+srcname: test-forced-vector-results
2799 #+begin_src ruby :var triv=test-trivial-vector :results silent
2805 #+srcname: set-ruby-session-var
2806 #+begin_src ruby :session :results silent
2811 #+srcname: get-ruby-session-var
2812 #+begin_src ruby :session :results silent
2816 #+srcname: set-python-session-var
2817 #+begin_src python :session
2822 #+srcname: get-python-session-var
2823 #+begin_src python :session
2827 #+srcname: set-R-session-var
2828 #+begin_src R :session
2833 #+srcname: get-R-session-var
2834 #+begin_src R :session
2843 To run these examples evaluate [[file:lisp/org-babel-init.el][org-babel-init.el]]
2845 ** org-babel.el beginning functionality
2847 #+begin_src sh :results replace
2852 : Sun Jul 5 18:54:39 EDT 2009
2859 : Sun Jul 05 18:54:35 -0400 2009
2871 #+begin_src R :results replace
2881 hist(rgamma(20,3,3))
2886 ** org-babel plays with tables
2887 Alright, this should demonstrate both the ability of org-babel to read
2888 tables into a lisp source code block, and to then convert the results
2889 of the source code block into an org table. It's using the classic
2890 "lisp is elegant" demonstration transpose function. To try this
2893 1. evaluate [[file:lisp/org-babel-init.el]] to load org-babel and friends
2894 2. evaluate the transpose definition =\C-c\\C-c= on the beginning of
2896 3. evaluate the next source code block, this should read in the table
2897 because of the =:var table=previous=, then transpose the table, and
2898 finally it should insert the transposed table into the buffer
2899 immediately following the block
2903 #+begin_src emacs-lisp :results silent
2904 (defun transpose (table)
2905 (apply #'mapcar* #'list table))
2913 #+begin_src emacs-lisp :var table=sandbox :results replace
2918 #+begin_src emacs-lisp
2923 | 1 | 2 | 3 | 4 | 5 |
2927 #+begin_src ruby :var table=sandbox :results replace
2928 table.first.join(" - ")
2934 #+begin_src python :var table=sandbox
2939 #+begin_src ruby :var table=sandbox :results replace
2944 : [[1, 2, 3], [4, "schulte", 6]]
2948 | 4 | "schulte" | 6 |
2950 #+begin_src python :var table=sandbox :results replace
2956 | "__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" |
2958 *** (sandbox table) R
2960 #+TBLNAME: sandbox_r
2964 #+begin_src R :results replace
2965 x <- c(rnorm(10, mean=-3, sd=1), rnorm(10, mean=3, sd=1))
2969 | -3.35473133869346 |
2971 | -3.32819924928633 |
2972 | -2.97310212756194 |
2973 | -2.09640758369576 |
2974 | -5.06054014378736 |
2975 | -2.20713700711221 |
2976 | -1.37618039712037 |
2977 | -1.95839385821742 |
2978 | -3.90407396475502 |
2979 | 2.51168071590226 |
2980 | 3.96753011570494 |
2981 | 3.31793212627865 |
2982 | 1.99829753972341 |
2983 | 4.00403686419829 |
2984 | 4.63723764452927 |
2985 | 3.94636744261313 |
2986 | 3.58355906547775 |
2987 | 3.01563442274226 |
2990 #+begin_src R var tabel=sandbox_r :results replace
2995 | 4 | "schulte" | 6 |
2998 Now shell commands are converted to tables using =org-table-import=
2999 and if these tables are non-trivial (i.e. have multiple elements) then
3000 they are imported as org-mode tables...
3002 #+begin_src sh :results replace
3006 | "total" | 208 | "" | "" | "" | "" | "" | "" |
3007 | "-rw-r--r--" | 1 | "dan" | "dan" | 57 | 2009 | 15 | "block" |
3008 | "-rw-r--r--" | 1 | "dan" | "dan" | 35147 | 2009 | 15 | "COPYING" |
3009 | "-rw-r--r--" | 1 | "dan" | "dan" | 722 | 2009 | 18 | "examples.org" |
3010 | "drwxr-xr-x" | 4 | "dan" | "dan" | 4096 | 2009 | 19 | "existing_tools" |
3011 | "-rw-r--r--" | 1 | "dan" | "dan" | 2207 | 2009 | 14 | "intro.org" |
3012 | "drwxr-xr-x" | 2 | "dan" | "dan" | 4096 | 2009 | 18 | "org-babel" |
3013 | "-rw-r--r--" | 1 | "dan" | "dan" | 277 | 2009 | 20 | "README.markdown" |
3014 | "-rw-r--r--" | 1 | "dan" | "dan" | 11837 | 2009 | 18 | "rorg.html" |
3015 | "-rw-r--r--" | 1 | "dan" | "dan" | 61829 | 2009 | 19 | "#rorg.org#" |
3016 | "-rw-r--r--" | 1 | "dan" | "dan" | 60190 | 2009 | 19 | "rorg.org" |
3017 | "-rw-r--r--" | 1 | "dan" | "dan" | 972 | 2009 | 11 | "test-export.org" |
3020 ** silent evaluation
3028 #+begin_src ruby :results silent
3032 #+begin_src ruby :results replace
3039 ** (sandbox) referencing other source blocks
3040 Doing this in emacs-lisp first because it's trivial to convert
3041 emacs-lisp results to and from emacs-lisp.
3043 *** emacs lisp source reference
3044 This first example performs a calculation in the first source block
3045 named =top=, the results of this calculation are then saved into the
3046 variable =first= by the header argument =:var first=top=, and it is
3047 used in the calculations of the second source block.
3050 #+begin_src emacs-lisp
3054 #+begin_src emacs-lisp :var first=top :results replace
3060 This example is the same as the previous only the variable being
3061 passed through is a table rather than a number.
3063 #+begin_src emacs-lisp :results silent
3064 (defun transpose (table)
3065 (apply #'mapcar* #'list table))
3068 #+TBLNAME: top_table
3072 #+SRCNAME: second_src_example
3073 #+begin_src emacs-lisp :var table=top_table
3077 #+begin_src emacs-lisp :var table=second_src_example :results replace
3082 | 4 | "schulte" | 6 |
3084 Now working for ruby
3091 #+begin_src ruby :var other=start :results replace
3097 #+SRCNAME: start_two
3102 #+begin_src python :var another=start_two :results replace
3107 Since all variables are converted into Emacs Lisp it is no problem to
3108 reference variables specified in another language.
3110 #+SRCNAME: ruby-block
3115 #+SRCNAME: lisp_block
3116 #+begin_src emacs-lisp :var ruby-variable=ruby-block
3120 #+begin_src python :var lisp_var=lisp_block
3129 #+begin_src R :results replace
3136 #+begin_src R :var other=first_r :results replace
3143 ** (sandbox) selective export
3145 For exportation tests and examples see (including exportation of
3146 inline source code blocks) [[file:test-export.org]]
3149 ** (sandbox) source blocks as functions
3152 #+begin_src emacs-lisp :results silent
3157 #+begin_src emacs-lisp :var n=default :results replace
3163 #+begin_src emacs-lisp :var result=triple(n=3, m=98) :results replace
3169 The following just demonstrates the ability to assign variables to
3170 literal values, which was not implemented until recently.
3172 #+begin_src ruby :var num="eric" :results replace
3179 ** (sandbox) inline source blocks
3181 This is an inline source code block src_ruby{1 + 6}. And another
3182 source block with text output src_emacs-lisp{"eric"}.
3184 This is an inline source code block with header
3185 arguments. src_ruby[:var n=fibbd( n = 0 )]{n}
3188 ** (sandbox) integration w/org tables
3190 #+begin_src emacs-lisp :results silent
3191 (defun fibbd (n) (if (< n 2) 1 (+ (fibbd (- n 1)) (fibbd (- n 2)))))
3195 #+begin_src emacs-lisp :var n=4 :results silent
3199 #+begin_src emacs-lisp :results silent
3200 (mapcar #'fibbd '(0 1 2 3 4 5 6 7 8))
3203 Something is not working here. The function `sbe ' works fine when
3204 called from outside of the table (see the source block below), but
3205 produces an error when called from inside the table. I think there
3206 must be some narrowing going on during intra-table emacs-lisp
3209 | original | fibbd |
3210 |----------+-------|
3221 #+TBLFM: $2='(sbe "fibbd" (n $1))
3225 #+begin_src emacs-lisp :results silent
3226 (sbe 'fibbd (n "8"))
3231 LocalWords: DBlocks dblocks org-babel el eric fontification