1 ;;; org-feed.el --- Add RSS feed items to Org files
3 ;; Copyright (C) 2009, 2010 Free Software Foundation, Inc.
5 ;; Author: Carsten Dominik <carsten at orgmode dot org>
6 ;; Keywords: outlines, hypermedia, calendar, wp
7 ;; Homepage: http://orgmode.org
10 ;; This file is part of GNU Emacs.
12 ;; GNU Emacs is free software: you can redistribute it and/or modify
13 ;; it under the terms of the GNU General Public License as published by
14 ;; the Free Software Foundation, either version 3 of the License, or
15 ;; (at your option) any later version.
17 ;; GNU Emacs is distributed in the hope that it will be useful,
18 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
19 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
20 ;; GNU General Public License for more details.
22 ;; You should have received a copy of the GNU General Public License
23 ;; along with GNU Emacs. If not, see <http://www.gnu.org/licenses/>.
24 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
28 ;; This module allows to create and change entries in an Org-mode
29 ;; file triggered by items in an RSS feed. The basic functionality is
30 ;; geared toward simply adding new items found in a feed as outline nodes
31 ;; to an Org file. Using hooks, arbitrary actions can be triggered for
32 ;; new or changed items.
34 ;; Selecting feeds and target locations
35 ;; ------------------------------------
37 ;; This module is configured through a single variable, `org-feed-alist'.
38 ;; Here is an example, using a notes/tasks feed from reQall.com.
40 ;; (setq org-feed-alist
42 ;; "http://www.reqall.com/user/feeds/rss/a1b2c3....."
43 ;; "~/org/feeds.org" "ReQall Entries")
45 ;; With this setup, the command `M-x org-feed-update-all' will
46 ;; collect new entries in the feed at the given URL and create
47 ;; entries as subheadings under the "ReQall Entries" heading in the
48 ;; file "~/org/feeds.org". Each feed should normally have its own
49 ;; heading - however see the `:drawer' parameter.
51 ;; Besides these standard elements that need to be specified for each
52 ;; feed, keyword-value pairs can set additional options. For example,
53 ;; to de-select transitional entries with a title containing
55 ;; "reQall is typing what you said",
57 ;; you could use the `:filter' argument:
59 ;; (setq org-feed-alist
61 ;; "http://www.reqall.com/user/feeds/rss/a1b2c3....."
62 ;; "~/org/feeds.org" "ReQall Entries"
63 ;; :filter my-reqall-filter)))
65 ;; (defun my-reqall-filter (e)
66 ;; (if (string-match "reQall is typing what you said"
67 ;; (plist-get e :title))
71 ;; See the docstring for `org-feed-alist' for more details.
74 ;; Keeping track of previously added entries
75 ;; -----------------------------------------
77 ;; Since Org allows you to delete, archive, or move outline nodes,
78 ;; org-feed.el needs to keep track of which feed items have been handled
79 ;; before, so that they will not be handled again. For this, org-feed.el
80 ;; stores information in a special drawer, FEEDSTATUS, under the heading
81 ;; that received the input of the feed. You should add FEEDSTATUS
82 ;; to your list of drawers in the files that receive feed input:
84 ;; #+DRAWERS: PROPERTIES LOGBOOK FEEDSTATUS
89 ;; org-feed.el is based on ideas by Brad Bozarth who implemented a
90 ;; similar mechanism using shell and awk scripts.
97 (declare-function url-retrieve-synchronously
"url" (url))
98 (declare-function xml-node-children
"xml" (node))
99 (declare-function xml-get-children
"xml" (node child-name
))
100 (declare-function xml-get-attribute
"xml" (node attribute
))
101 (declare-function xml-get-attribute-or-nil
"xml" (node attribute
))
102 (declare-function xml-substitute-special
"xml" (string))
104 (defgroup org-feed nil
105 "Options concerning RSS feeds as inputs for Org files."
109 (defcustom org-feed-alist nil
110 "Alist specifying RSS feeds that should create inputs for Org.
111 Each entry in this list specified an RSS feed tat should be queried
112 to create inbox items in Org. Each entry is a list with the following items:
114 name a custom name for this feed
116 file the target Org file where entries should be listed
117 headline the headline under which entries should be listed
119 Additional arguments can be given using keyword-value pairs. Many of these
120 specify functions that receive one or a list of \"entries\" as their single
121 argument. An entry is a property list that describes a feed item. The
122 property list has properties for each field in the item, for example `:title'
123 for the `<title>' field and `:pubDate' for the publication date. In addition,
124 it contains the following properties:
126 `:item-full-text' the full text in the <item> tag
127 `:guid-permalink' t when the guid property is a permalink
129 Here are the keyword-value pair allows in `org-feed-alist'.
132 The name of the drawer for storing feed information. The default is
133 \"FEEDSTATUS\". Using different drawers for different feeds allows
134 several feeds to target the same inbox heading.
136 :filter filter-function
137 A function to select interesting entries in the feed. It gets a single
138 entry as parameter. It should return the entry if it is relevant, or
141 :template template-string
142 The default action on new items in the feed is to add them as children
143 under the headline for the feed. The template describes how the entry
144 should be formatted. If not given, it defaults to
145 `org-feed-default-template'.
147 :formatter formatter-function
148 Instead of relying on a template, you may specify a function to format
149 the outline node to be inserted as a child. This function gets passed
150 a property list describing a single feed item, and it should return a
151 string that is a properly formatted Org outline node of level 1.
153 :new-handler function
154 If adding new items as children to the outline is not what you want
155 to do with new items, define a handler function that is called with
156 a list of all new items in the feed, each one represented as a property
157 list. The handler should do what needs to be done, and org-feed will
158 mark all items given to this handler as \"handled\", i.e. they will not
159 be passed to this handler again in future readings of the feed.
160 When the handler is called, point will be at the feed headline.
162 :changed-handler function
163 This function gets passed a list of all entries that have been
164 handled before, but are now still in the feed and have *changed*
165 since last handled (as evidenced by a different sha1 hash).
166 When the handler is called, point will be at the feed headline.
169 This function gets passed a buffer, and should return a list
170 of entries, each being a property list containing the
171 `:guid' and `:item-full-text' keys. The default is
172 `org-feed-parse-rss-feed'; `org-feed-parse-atom-feed' is an
175 :parse-entry function
176 This function gets passed an entry as returned by the parse-feed
177 function, and should return the entry with interesting properties added.
178 The default is `org-feed-parse-rss-entry'; `org-feed-parse-atom-entry'
182 (list :value
("" "http://" "" "")
184 (string :tag
"Feed URL")
185 (file :tag
"File for inbox")
186 (string :tag
"Headline for inbox")
189 (list :inline t
:tag
"Filter"
191 (symbol :tag
"Filter Function"))
192 (list :inline t
:tag
"Template"
194 (string :tag
"Template"))
195 (list :inline t
:tag
"Formatter"
197 (symbol :tag
"Formatter Function"))
198 (list :inline t
:tag
"New items handler"
200 (symbol :tag
"Handler Function"))
201 (list :inline t
:tag
"Changed items"
202 (const :changed-handler
)
203 (symbol :tag
"Handler Function"))
204 (list :inline t
:tag
"Parse Feed"
206 (symbol :tag
"Parse Feed Function"))
207 (list :inline t
:tag
"Parse Entry"
209 (symbol :tag
"Parse Entry Function"))
212 (defcustom org-feed-drawer
"FEEDSTATUS"
213 "The name of the drawer for feed status information.
214 Each feed may also specify its own drawer name using the `:drawer'
215 parameter in `org-feed-alist'.
216 Note that in order to make these drawers behave like drawers, they must
217 be added to the variable `org-drawers' or configured with a #+DRAWERS
220 :type
'(string :tag
"Drawer Name"))
222 (defcustom org-feed-default-template
"\n* %h\n %U\n %description\n %a\n"
223 "Template for the Org node created from RSS feed items.
224 This is just the default, each feed can specify its own.
225 Any fields from the feed item can be interpolated into the template with
226 %name, for example %title, %description, %pubDate etc. In addition, the
227 following special escapes are valid as well:
229 %h the title, or the first line of the description
230 %t the date as a stamp, either from <pubDate> (if present), or
233 %u,%U like %t,%T, but inactive time stamps
234 %a A link, from <guid> if that is a permalink, else from <link>"
236 :type
'(string :tag
"Template"))
238 (defcustom org-feed-save-after-adding t
239 "Non-nil means save buffer after adding new feed items."
243 (defcustom org-feed-retrieve-method
'url-retrieve-synchronously
244 "The method to be used to retrieve a feed URL.
245 This can be `curl' or `wget' to call these external programs, or it can be
246 an Emacs Lisp function that will return a buffer containing the content
247 of the file pointed to by the URL."
250 (const :tag
"Internally with url.el" url-retrieve-synchronously
)
251 (const :tag
"Externally with curl" curl
)
252 (const :tag
"Externally with wget" wget
)
253 (function :tag
"Function")))
255 (defcustom org-feed-before-adding-hook nil
256 "Hook that is run before adding new feed items to a file.
257 You might want to commit the file in its current state to version control,
262 (defcustom org-feed-after-adding-hook nil
263 "Hook that is run after new items have been added to a file.
264 Depending on `org-feed-save-after-adding', the buffer will already
269 (defvar org-feed-buffer
"*Org feed*"
270 "The buffer used to retrieve a feed.")
273 (defun org-feed-update-all ()
274 "Get inbox items from all feeds in `org-feed-alist'."
276 (let ((nfeeds (length org-feed-alist
))
277 (nnew (apply '+ (mapcar 'org-feed-update org-feed-alist
))))
278 (message "%s from %d %s"
279 (cond ((= nnew
0) "No new entries")
280 ((= nnew
1) "1 new entry")
281 (t (format "%d new entries" nnew
)))
283 (if (= nfeeds
1) "feed" "feeds"))))
286 (defun org-feed-update (feed &optional retrieve-only
)
287 "Get inbox items from FEED.
288 FEED can be a string with an association in `org-feed-alist', or
289 it can be a list structured like an entry in `org-feed-alist'."
290 (interactive (list (org-completing-read "Feed name: " org-feed-alist
)))
291 (if (stringp feed
) (setq feed
(assoc feed org-feed-alist
)))
293 (error "No such feed in `org-feed-alist"))
295 (let ((name (car feed
))
298 (headline (nth 3 feed
))
299 (filter (nth 1 (memq :filter feed
)))
300 (formatter (nth 1 (memq :formatter feed
)))
301 (new-handler (nth 1 (memq :new-handler feed
)))
302 (changed-handler (nth 1 (memq :changed-handler feed
)))
303 (template (or (nth 1 (memq :template feed
))
304 org-feed-default-template
))
305 (drawer (or (nth 1 (memq :drawer feed
))
307 (parse-feed (or (nth 1 (memq :parse-feed feed
))
308 'org-feed-parse-rss-feed
))
309 (parse-entry (or (nth 1 (memq :parse-entry feed
))
310 'org-feed-parse-rss-entry
))
311 feed-buffer inbox-pos new-formatted
312 entries old-status status new changed guid-alist e guid olds
)
313 (setq feed-buffer
(org-feed-get-feed url
))
314 (unless (and feed-buffer
(bufferp (get-buffer feed-buffer
)))
315 (error "Cannot get feed %s" name
))
317 (throw 'exit feed-buffer
))
318 (setq entries
(funcall parse-feed feed-buffer
))
319 (ignore-errors (kill-buffer feed-buffer
))
321 (save-window-excursion
322 (setq inbox-pos
(org-feed-goto-inbox-internal file headline
))
323 (setq old-status
(org-feed-read-previous-status inbox-pos drawer
))
324 ;; Add the "handled" status to the appropriate entries
325 (setq entries
(mapcar (lambda (e)
327 (plist-put e
:handled
332 ;; Find out which entries are new and which are changed
334 (if (not (plist-get e
:handled
))
336 (setq olds
(nth 2 (assoc (plist-get e
:guid
) old-status
)))
339 (plist-get e
:item-full-text
))
343 ;; Parse the relevant entries fully
344 (setq new
(mapcar parse-entry new
)
345 changed
(mapcar parse-entry changed
))
349 (setq new
(delq nil
(mapcar filter new
))
350 changed
(delq nil
(mapcar filter new
))))
352 (when (not (or new changed
))
353 (message "No new items in feed %s" name
)
356 ;; Get alist based on guid, to look up entries
359 (mapcar (lambda (e) (list (plist-get e
:guid
) e
)) new
)
360 (mapcar (lambda (e) (list (plist-get e
:guid
) e
)) changed
)))
362 ;; Construct the new status
366 (setq guid
(plist-get e
:guid
))
368 ;; things count as handled if we handle them now,
369 ;; or if they were handled previously
370 (if (assoc guid guid-alist
) t
(plist-get e
:handled
))
371 ;; A hash, to detect changes
372 (sha1 (plist-get e
:item-full-text
))))
375 ;; Handle new items in the feed
379 (goto-char inbox-pos
)
380 (funcall new-handler new
))
381 ;; No custom handler, do the default adding
382 ;; Format the new entries into an alist with GUIDs in the car
385 (lambda (e) (org-feed-format-entry e template formatter
))
388 ;; Insert the new items
389 (org-feed-add-items inbox-pos new-formatted
))
391 ;; Handle changed items in the feed
392 (when (and changed-handler changed
)
393 (goto-char inbox-pos
)
394 (funcall changed-handler changed
))
396 ;; Write the new status
397 ;; We do this only now, in case something goes wrong above, so
398 ;; that would would end up with a status that does not reflect
399 ;; which items truely have been handled
400 (org-feed-write-status inbox-pos drawer status
)
402 ;; Normalize the visibility of the inbox tree
403 (goto-char inbox-pos
)
406 (org-cycle-hide-drawers 'children
)
408 ;; Hooks and messages
409 (when org-feed-save-after-adding
(save-buffer))
410 (message "Added %d new item%s from feed %s to file %s, heading %s"
411 (length new
) (if (> (length new
) 1) "s" "")
413 (file-name-nondirectory file
) headline
)
414 (run-hooks 'org-feed-after-adding-hook
)
418 (defun org-feed-goto-inbox (feed)
419 "Go to the inbox that captures the feed named FEED."
421 (list (if (= (length org-feed-alist
) 1)
423 (org-completing-read "Feed name: " org-feed-alist
))))
424 (if (stringp feed
) (setq feed
(assoc feed org-feed-alist
)))
426 (error "No such feed in `org-feed-alist"))
427 (org-feed-goto-inbox-internal (nth 2 feed
) (nth 3 feed
)))
430 (defun org-feed-show-raw-feed (feed)
431 "Show the raw feed buffer of a feed."
433 (list (if (= (length org-feed-alist
) 1)
435 (org-completing-read "Feed name: " org-feed-alist
))))
436 (if (stringp feed
) (setq feed
(assoc feed org-feed-alist
)))
438 (error "No such feed in `org-feed-alist"))
440 (org-feed-update feed
'retrieve-only
))
441 (goto-char (point-min)))
443 (defun org-feed-goto-inbox-internal (file heading
)
444 "Find or create HEADING in FILE.
445 Switch to that buffer, and return the position of that headline."
448 (goto-char (point-min))
449 (if (re-search-forward
450 (concat "^\\*+[ \t]+" heading
"[ \t]*\\(:.*?:[ \t]*\\)?$")
452 (goto-char (match-beginning 0))
453 (goto-char (point-max))
454 (insert "\n\n* " heading
"\n\n")
455 (org-back-to-heading t
))
458 (defun org-feed-read-previous-status (pos drawer
)
459 "Get the alist of old GUIDs from the entry at POS.
460 This will find DRAWER and extract the alist."
463 (let ((end (save-excursion (org-end-of-subtree t t
))))
464 (if (re-search-forward
465 (concat "^[ \t]*:" drawer
":[ \t]*\n\\([^\000]*?\\)\n[ \t]*:END:")
467 (read (match-string 1))
470 (defun org-feed-write-status (pos drawer status
)
471 "Write the feed STATUS to DRAWER in entry at POS."
474 (let ((end (save-excursion (org-end-of-subtree t t
)))
476 (if (re-search-forward (concat "^[ \t]*:" drawer
":[ \t]*\n")
479 (goto-char (match-end 0))
480 (delete-region (point)
482 (and (re-search-forward "^[ \t]*:END:" nil t
)
483 (match-beginning 0)))))
484 (outline-next-heading)
485 (insert " :" drawer
":\n :END:\n")
486 (beginning-of-line 0))
487 (insert (pp-to-string status
)))))
489 (defun org-feed-add-items (pos entries
)
490 "Add the formatted items to the headline as POS."
494 (unless (looking-at org-complex-heading-regexp
)
495 (error "Wrong position"))
496 (setq level
(org-get-valid-level (length (match-string 1)) 1))
497 (org-end-of-subtree t t
)
498 (skip-chars-backward " \t\n")
499 (beginning-of-line 2)
501 (while (setq entry
(pop entries
))
502 (org-paste-subtree level entry
'yank
))
503 (org-mark-ring-push pos
))))
505 (defun org-feed-format-entry (entry template formatter
)
506 "Format ENTRY so that it can be inserted into an Org file.
507 ENTRY is a property list. This function adds a `:formatted-for-org' property
508 and returns the full property list.
509 If that property is already present, nothing changes."
511 (funcall formatter entry
)
512 (let (dlines fmt tmp indent time name
513 v-h v-t v-T v-u v-U v-a
)
514 (setq dlines
(org-split-string (or (plist-get entry
:description
) "???")
516 v-h
(or (plist-get entry
:title
) (car dlines
) "???")
517 time
(or (if (plist-get entry
:pubDate
)
518 (org-read-date t t
(plist-get entry
:pubDate
)))
520 v-t
(format-time-string (org-time-stamp-format nil nil
) time
)
521 v-T
(format-time-string (org-time-stamp-format t nil
) time
)
522 v-u
(format-time-string (org-time-stamp-format nil t
) time
)
523 v-U
(format-time-string (org-time-stamp-format t t
) time
)
524 v-a
(if (setq tmp
(or (and (plist-get entry
:guid-permalink
)
525 (plist-get entry
:guid
))
526 (plist-get entry
:link
)))
527 (concat "[[" tmp
"]]\n")
531 (goto-char (point-min))
532 (while (re-search-forward "%\\([a-zA-Z]+\\)" nil t
)
533 (setq name
(match-string 1))
535 ((member name
'("h" "t" "T" "u" "U" "a"))
536 (replace-match (symbol-value (intern (concat "v-" name
))) t t
))
537 ((setq tmp
(plist-get entry
(intern (concat ":" name
))))
540 (beginning-of-line 1)
541 (when (looking-at (concat "^\\([ \t]*\\)%" name
"[ \t]*$"))
542 (setq tmp
(org-feed-make-indented-block
543 tmp
(org-get-indentation))))))
544 (replace-match tmp t t
))))
545 (decode-coding-string
546 (buffer-string) (detect-coding-region (point-min) (point-max) t
))))))
548 (defun org-feed-make-indented-block (s n
)
549 "Add indentation of N spaces to a multiline string S."
550 (if (not (string-match "\n" s
))
553 (org-split-string s
"\n")
554 (concat "\n" (make-string n ?\
)))))
556 (defun org-feed-skip-http-headers (buffer)
557 "Remove HTTP headers from BUFFER, and return it.
558 Assumes headers are indeed present!"
559 (with-current-buffer buffer
561 (goto-char (point-min))
562 (search-forward "\n\n")
563 (delete-region (point-min) (point))
566 (defun org-feed-get-feed (url)
567 "Get the RSS feed file at URL and return the buffer."
569 ((eq org-feed-retrieve-method
'url-retrieve-synchronously
)
570 (org-feed-skip-http-headers (url-retrieve-synchronously url
)))
571 ((eq org-feed-retrieve-method
'curl
)
572 (ignore-errors (kill-buffer org-feed-buffer
))
573 (call-process "curl" nil org-feed-buffer nil
"--silent" url
)
575 ((eq org-feed-retrieve-method
'wget
)
576 (ignore-errors (kill-buffer org-feed-buffer
))
577 (call-process "wget" nil org-feed-buffer nil
"-q" "-O" "-" url
)
579 ((functionp org-feed-retrieve-method
)
580 (funcall org-feed-retrieve-method url
))))
582 (defun org-feed-parse-rss-feed (buffer)
583 "Parse BUFFER for RSS feed entries.
584 Returns a list of entries, with each entry a property list,
585 containing the properties `:guid' and `:item-full-text'."
586 (let ((case-fold-search t
)
587 entries beg end item guid entry
)
588 (with-current-buffer buffer
590 (goto-char (point-min))
591 (while (re-search-forward "<item\\>.*?>" nil t
)
593 end
(and (re-search-forward "</item>" nil t
)
594 (match-beginning 0)))
595 (setq item
(buffer-substring beg end
)
596 guid
(if (string-match "<guid\\>.*?>\\(.*?\\)</guid>" item
)
597 (org-match-string-no-properties 1 item
)))
598 (setq entry
(list :guid guid
:item-full-text item
))
602 (nreverse entries
))))
604 (defun org-feed-parse-rss-entry (entry)
605 "Parse the `:item-full-text' field for xml tags and create new properties."
608 (insert (plist-get entry
:item-full-text
))
609 (goto-char (point-min))
610 (while (re-search-forward "<\\([a-zA-Z]+\\>\\).*?>\\([^\000]*?\\)</\\1>"
612 (setq entry
(plist-put entry
613 (intern (concat ":" (match-string 1)))
614 (xml-substitute-special (match-string 2)))))
615 (goto-char (point-min))
616 (unless (re-search-forward "isPermaLink[ \t]*=[ \t]*\"false\"" nil t
)
617 (setq entry
(plist-put entry
:guid-permalink t
))))
620 (defun org-feed-parse-atom-feed (buffer)
621 "Parse BUFFER for Atom feed entries.
622 Returns a list of entries, with each entry a property list,
623 containing the properties `:guid' and `:item-full-text'.
625 The `:item-full-text' property actually contains the sexp
626 formatted as a string, not the original XML data."
628 (with-current-buffer buffer
630 (let ((feed (car (xml-parse-region (point-min) (point-max)))))
634 :guid
(car (xml-node-children (car (xml-get-children entry
'id
))))
635 :item-full-text
(prin1-to-string entry
)))
636 (xml-get-children feed
'entry
)))))
638 (defun org-feed-parse-atom-entry (entry)
639 "Parse the `:item-full-text' as a sexp and create new properties."
640 (let ((xml (car (read-from-string (plist-get entry
:item-full-text
)))))
641 ;; Get first <link href='foo'/>.
642 (setq entry
(plist-put entry
:link
644 (car (xml-get-children xml
'link
))
646 ;; Add <title/> as :title.
647 (setq entry
(plist-put entry
:title
648 (xml-substitute-special
649 (car (xml-node-children
650 (car (xml-get-children xml
'title
)))))))
651 (let* ((content (car (xml-get-children xml
'content
)))
652 (type (xml-get-attribute-or-nil content
'type
)))
655 ((string= type
"text")
656 ;; We like plain text.
657 (setq entry
(plist-put entry
:description
658 (xml-substitute-special
659 (car (xml-node-children content
))))))
660 ((string= type
"html")
661 ;; TODO: convert HTML to Org markup.
662 (setq entry
(plist-put entry
:description
663 (xml-substitute-special
664 (car (xml-node-children content
))))))
665 ((string= type
"xhtml")
666 ;; TODO: convert XHTML to Org markup.
667 (setq entry
(plist-put entry
:description
669 (xml-node-children content
)))))
671 (setq entry
(plist-put entry
:description
672 (format "Unknown '%s' content." type
)))))))
677 ;; arch-tag: 0929b557-9bc4-47f4-9633-30a12dbb5ae2
678 ;;; org-feed.el ends here