X-Git-Url: http://gitweb.fperrin.net/?p=atom.el.git;a=blobdiff_plain;f=atom.el;h=d19bb3c82edc1e5cc06b0b957fff662f54ca42fd;hp=7b43065babb1527f1d63b195d349375ab44e9ad2;hb=4da2ee38a39546325169b6e1ec259b26e2cdf205;hpb=1e75e274945b36bc74eb4e605b023ee85b8a86c5 diff --git a/atom.el b/atom.el index 7b43065..d19bb3c 100644 --- a/atom.el +++ b/atom.el @@ -20,51 +20,90 @@ ;;; Commentary: -;; This is a library for creating an Atom file from a Lisp program. +;; This is a library for creating an Atom feed from a Lisp program. +;; The normal usage is to create a feed with `atom-create', giving it +;; a title and a Web address. Once the feed has been created, entries +;; may be added to the feed, by specifying (at the minimum) a title, a +;; permanent link and the content of the entry. Text-only, HTML and +;; XHTML entries are supported. + +;; It is possible to produce both Atom and RSS feeds. + +;; A typical usage would look like this: + +;; (let ((my-atom-feed (atom-create "My feed" "http://example.org"))) +;; ; A simple, text-only entry +;; (atom-add-text-entry +;; my-atom-feed +;; "Hello world" +;; "http://example.org/hello" +;; "Hello the world!") +;; +;; (atom-add-xhtml-entry +;; my-atom-feed +;; "An XHTML example" +;; "http://example.org/html-example" +;; "

One can also use XHTML in the entries.

") +;; +;; (atom-print my-atom-feed) +;; ;; If you prefer RSS feeds: +;; (atom-to-rss-print my-atom-feed)) + +;; Full documentation is available at . ;;; Code: -(defun atom-create (title author link &optional updated id) +(require 'xml) +(require 'url-parse) +(require 'cl) ; for setf in url-canonalize + +(defun atom-create (title link &optional subtitle self id author updated) "Create a new atom structure. TITLE is the title for the feed, a short, text-only, human readable string. -AUTHOR is the author of the feed. See `atom-massage-author' for -the possible ways to specify it. - LINK is the URL of a page responible for the content of this feed. -UPDATED is the date the feed was last updated. If not given, -`(current-time)' is used. +SUBTITLE is a subtitle for the feed; it can be a bit longer than +TITLE, maybe a paragraph long. + +SELF is the canonical URL to this feed. ID is a unique identifier for this feed. If not given, it -defaults to LINK." - `((title nil ,title) - (link ((href . ,link))) - ,(atom-massage-author author) - (updated nil ,(atom-format-time updated)) - (id nil ,(or id link)))) +defaults to SELF. + +AUTHOR is the author of the feed. See `atom-massage-author' for +the possible ways to specify it. In particular, `nil' uses +`user-full-name' and `user-mail-address'. + +UPDATED is the date the feed was last updated. If not given, +`(current-time)' is used." + (let ((atom-feed (list (list 'title nil title)))) + (atom-modify-entry atom-feed 'link `(((href . ,link)))) + (atom-modify-entry atom-feed 'author (atom-massage-author author)) + (if subtitle (atom-modify-entry atom-feed 'subtitle subtitle)) + (if self (atom-modify-entry atom-feed 'link + `(((href . ,self) (rel . "self") + (type . "application/atom+xml"))))) + (atom-modify-entry atom-feed 'updated (atom-format-time updated)) + (atom-modify-entry atom-feed 'id (or id self link)) + atom-feed)) (defun atom-push-entry (atom entry) "Add the entry ENTRY to the feed ATOM." - (nconc atom (list `(entry nil . ,entry)))) + (nconc atom (list `(entry nil ,@entry)))) (defun atom-modify-entry (entry name val) "Set the NAME element of ENTRY to VAL." - (let ((elem (assoc name entry))) - (if elem - (if (stringp val) - (setcar (cddr elem) val) - (setcdr elem val)) - (setq elem (if (stringp val) - (list name nil val) - (cons name val))) - (nconc entry (list elem))))) + (let ((elem (if (stringp val) + (list name nil val) + (cons name val)))) + (nconc entry (list elem)))) (defun atom-add-entry (atom title link content - &optional summary updated id) + &optional updated id summary) "Add an entry to the atom flux ATOM. Return the newly added entry. @@ -81,9 +120,10 @@ If SUMMARY is not given, the entry will not contain any summary. UPDATED defaults to `(current-time)' if omitted, which is probably not a very good default. -ID defaults to LINK, which is not optimal; TODO give a way to -easily generate IDs. For a given entry, it must not change -between successive generations of the atom feed." +ID defaults to LINK, which is not optimal; see `atom-generate-id' +for a way to create good identifiers. For a given entry, it must +not change between successive generations of the atom feed, even +when the content of the entry ." (let ((entry (list (list 'title nil title)))) (atom-modify-entry entry 'link (list (list (cons 'href link)))) (atom-modify-entry entry 'id (or id link)) @@ -98,43 +138,136 @@ between successive generations of the atom feed." `atom-add-entry' for details.") (defun atom-add-html-entry (atom title link content - &optional summary updated id) + &optional updated id summary) "Add an entry to ATOM, with some HTML content. CONTENT should be a string enconding a valid HTML fragment. See `atom-add-entry' for additional details." (atom-add-entry atom - title link - (atom-massage-html content) - (and summary (atom-massage-html summary)) - updated id)) + title link (atom-massage-html content) + updated id (and summary (atom-massage-html summary)))) (defun atom-add-xhtml-entry (atom title link content - &optional summary updated id) + &optional updated id summary noconvert) "Add an entry to ATOM, with some XHTML content. CONTENT may be given either as a string, or as an XML tree, of a valid XHTML -fragment. See `atom-add-entry' for additional details." - (atom-add-entry atom - title link - (atom-massage-xhtml content) - (and summary (atom-massage-xhtml summary)) - updated id)) +fragment. See `atom-add-entry' for additional details. + +If NOCONVERT is nil, translate all links in CONTENT so that they +are no longer relative to LINK." + (let ((xhtml-content (atom-massage-xhtml content))) + (unless noconvert + (atom-xhtml-convert-links (cadr xhtml-content) link)) + (atom-add-entry atom + title link xhtml-content + updated id (and summary (atom-massage-xhtml summary))))) (defun atom-print (atom) "Print the Atom feed ATOM in the current buffer." - (insert "\n") + (insert atom-xml-declaration) (insert "\n") (xml-print atom) (insert "\n")) +(defun atom-write-file (atom filename) + "Writes the feed ATOM to FILENAME." + (with-temp-buffer + (atom-print atom) + (write-region (point-min) (point-max) filename))) + + +(defun atom-to-rss (atom) + "Translate an Atom feed into an RSS one, returning the translation. + +Some information may be lost or approximated." + (let ((rss (list (assoc 'title atom)))) + (atom-to-rss-translator atom rss '((subtitle . description) + (updated . pubDate) + (link . link))) + (atom-to-rss-modify-time rss) + (atom-to-rss-modify-link rss) + (dolist (entry (xml-get-children atom 'entry)) + (push (atom-to-rss-item entry) rss)) + (reverse rss))) + +(defun atom-to-rss-item (entry) + "Translates an Atom entry into an RSS item." + (let ((item (list (assoc 'title entry)))) + (atom-to-rss-translator + (xml-node-children entry) item + '((id . guid) (content . description) (updated . pubDate) (link . link))) + (atom-to-rss-modify-time item) + (atom-to-rss-modify-link item) + (let ((guid (assoc 'guid item)) + (descr (assoc 'description item))) + (if guid + (setcar (cdr guid) (list (cons 'isPermaLink "false")))) + (if (and descr + (equal (xml-get-attribute descr 'type) "xhtml")) + (setcar (cddr descr) (xml-node-as-text descr)))) + `(item nil ,@item))) + +(defun atom-to-rss-translator (source target translations) + (dolist (translation translations) + (let* ((from (car translation)) + (to (cdr translation)) + (data (copy-tree (cdr (assoc from source))))) + (when data + (atom-modify-entry target to data))))) + +(defun atom-to-rss-modify-link (entry) + (let* ((link (assoc 'link entry)) + (link-addr (xml-get-attribute-or-nil link 'href))) + (when link + (setcar (cdr link) nil) + (setcdr (cdr link) (cons link-addr nil))))) + +(defun atom-print-as-rss (atom) + (let ((rss (atom-to-rss atom))) + (insert atom-xml-declaration) + (insert "\n") + (insert " \n") + (xml-print rss " ") + (insert "\n \n") + (insert ""))) + +(defun atom-to-rss-time (time) + "Translates a string from the format used by Atom into the +format used by RSS." + (let ((system-time-locale "C")) + (format-time-string "%a, %d %b %Y %T %z" (atom-parse-time time)))) + +(defun atom-to-rss-modify-time (entry) + "Modify ENTRY, changing the format of the `pubDate' in it." + (let ((pubDate (assoc 'pubDate entry))) + (setcar (cddr pubDate) + (atom-to-rss-time (car (xml-node-children pubDate)))))) + +(defun atom-to-rss-write-file (atom filename) + "Saves ATOM as a RSS feed into FILENAME." + (with-temp-buffer + (atom-print-as-rss atom) + (write-region nil nil filename))) + +(defvar atom-time-format-string "%Y-%m-%dT%T%z" + "The format for string representation of dates.") + +(defvar atom-xhtml-namespace "http://www.w3.org/1999/xhtml") + +(defvar atom-xml-declaration "\n") + (defun atom-format-time (&optional time) "Format a time according to RFC3339." ;; The time zone must be specified in numeric form, but with a colon between ;; the hour and minute parts. (replace-regexp-in-string - "\\(..\\)$" - ":\\1" - (format-time-string "%Y-%m-%dT%T%z" time))) + "\\(..\\)$" ":\\1" + (format-time-string atom-time-format-string time))) + +(defun atom-parse-time (&optional time) + "Parse a time as specified in RFC3339 into Emacs's native format." + ;; Same remark as in `atom-format-time' + (date-to-time (replace-regexp-in-string ":\\(..\\)$" "\\1" time))) (defun atom-massage-html (content) "Massage CONTENT so it can be used as an HTML fragment in an @@ -144,35 +277,81 @@ Atom feed. CONTENT must be a string." (defun atom-string-to-xml (string) "Convert STRING into a Lisp structure as used by `xml.el'." (with-temp-buffer + (insert "
") (insert string) - (xml-parse-region (point-min) (point-max)))) + (insert "
") + ;; `xml-parse-region' doesn't require that the XML parsed be enclosed in a + ;; root node, and accordingly, returns a list of elements. We are only + ;; interested in the first one, the DIV we just inserted. + (car (xml-parse-region (point-min) (point-max))))) (defun atom-massage-xhtml (content) "Massage CONTENT so it can be used as an XHTML fragment in an Atom feed." (list '((type . "xhtml")) - `(div ((xmlns . "http://www.w3.org/1999/xhtml")) - . ,(or (and (stringp content) - (atom-string-to-xml content)) - content)))) + (or (and (stringp content) + (atom-string-to-xml content)) + `(div ((xmlns . ,atom-xhtml-namespace)) ,@content)))) (defun atom-massage-author (author) "Return an XML node representing the author. AUTHOR can be: - nil, in which case `user-full-name' and `user-mail-address' are used; -- a single string, the full name of the author; +- a single string, the full name of the author; no email address + will be included; - a list with two elements, the full name and the email address of the author; - something else, assumed to be a complete `atomPersonConstruct'." - (cond - ((null author) `(author nil - (name nil ,user-full-name) - (email nil ,user-mail-address))) - ((stringp author) `(author nil - (name nil ,user-full-name))) - ((= 2 (length author)) `(author nil (name nil ,(car author)) - (email nil ,(cadr author)))) - (t `(author nil ,author)))) + `(nil ,@(cond + ((null author) `((name nil ,user-full-name) + (email nil ,user-mail-address))) + ((stringp author) `((name nil ,author))) + ((= 2 (length author)) `((name nil ,(car author)) + (email nil ,(cadr author)))) + (t `(author nil ,author))))) + +(defun atom-xhtml-convert-links (node base) + "Make all links in NODE (a fragment of an XHTML document) +absolute, in the context of BASE, an URL." + (dolist (attr-name (list 'href 'src)) + (let ((attr (assoc attr-name (xml-node-attributes node)))) + (when attr (setcdr attr (url-canonalize (cdr attr) base))))) + (dolist (child (xml-node-children node)) + (when (listp child) (atom-xhtml-convert-links child base)))) + +(defun atom-generate-id (link creation-date) + "Generate a string suitable for use as an atom:id element. This +implements Mark Pilgrom's tag: URI method, using the +CREATION-DATE of the entry, and the domain part of LINK." + (format "tag:%s,%s:/%s" + (url-host (url-generic-parse-url link)) + (format-time-string "%Y-%m-%d" creation-date) + (format-time-string "%Y%m%d%H%M%S" creation-date))) + + +;;; Functions that should probably not be there + +(defun url-canonalize (address base) + "Make ADRESS an absolute URL, taking it in the BASE context." + ;; I feel such a function should exist in `url-parse'. Did I miss it? + (let ((url-base (url-generic-parse-url base)) + (url-address (url-generic-parse-url address))) + (if (url-host url-address) + address + (setf (url-filename url-base) + (expand-file-name address + (file-name-directory (url-filename url-base)))) + (url-recreate-url url-base)))) + +(defun xml-node-as-text (node) + "Return a string representing NODE, an XML structure." + (with-temp-buffer + (xml-print (xml-node-children node)) + (buffer-string))) + +(defun xml-node-create (name attrlist childlist) + "Create a new XML node." + (list name attrlist . childlist)) (provide 'atom) ;;; atom.el ends here