-;;; atom.el --- Create an Atom feed
+;;; atom.el --- Create an Atom feed -*- lexical-binding: t -*-
;; Copyright (C) 2011 Frédéric Perrin
;;; 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")))
-;; (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"
-;; "<p>One can also use <acronym>XHTML</acronym> in the
-;; entries.</p>")
-;; (atom-print my-atom-feed))
+;; ; 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"
+;; "<p>One can also use <acronym>XHTML</acronym> in the entries.</p>")
+;;
+;; (atom-print my-atom-feed)
+;; ;; If you prefer RSS feeds:
+;; (atom-print-as-rss my-atom-feed))
+
+;; Full documentation is available at <http://tar-jx.bz/code/atom.html>.
+;; See atom-tests.el for usage examples.
;;; Code:
-(defun atom-create (title link &optional author updated id)
+(require 'xml)
+(require 'url-parse)
+
+(defun atom-create (title link &optional props)
"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.
+PROPS is an optional plist with the following properties:
+
+- :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. If missing, the resulting
+ feed is non-conforming.
+
+- :id is a unique identifier for this feed. If not given, it
+ 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
+variable `user-full-name' and `user-mail-address'.
-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))))
+- :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 (plist-get props :author)))
+ (if (plist-member props :subtitle)
+ (atom-modify-entry atom-feed 'subtitle (plist-get props :subtitle)))
+ (if (plist-member props :self)
+ (atom-modify-entry atom-feed 'link
+ `(((href . ,(plist-get props :self)) (rel . "self")
+ (type . "application/atom+xml")))))
+ (atom-modify-entry atom-feed 'updated (atom-format-time (plist-get props :updated)))
+ (atom-modify-entry atom-feed 'id (or (plist-get props :id) (plist-get props :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)))))
-
-(defun atom-add-entry (atom title link content
- &optional summary updated id)
- "Add an entry to the atom flux ATOM. Return the newly added
-entry.
+ (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 props)
+ "Add an entry to the atom flux ATOM.
+
+Return the newly adde dentry.
TITLE is a short, text-only, human readable string.
CONTENT is the content of the entry; use `atom-add-html-entry'
or `atom-add-xhtml-entry' when CONTENT is not text-only.
-If SUMMARY is not given, the entry will not contain any summary.
+PROPS is an optional plist with the following properties:
+
+- :summary, if is not given, the entry will not contain any summary.
+
+- :updated defaults to `(current-time)'.
-UPDATED defaults to `(current-time)' if omitted, which is
-probably not a very good default.
+- :published, if given, is the earliest availability of the
+ entry. It is optional, and shouldn't change even if the entry
+ content (etc.) updated after the initial publication.
-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 is a unique ID for the entry; defaulting to LINK. RFC4287
+ has specific requirements about valid IRI that may be used,
+ which this library does not try to enforce."
(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))
- (atom-modify-entry entry 'updated (atom-format-time updated))
- (if summary (atom-modify-entry entry 'summary summary))
+ (atom-modify-entry entry 'id (or (plist-get props :id) link))
+ (atom-modify-entry entry 'updated (atom-format-time (plist-get props :updated)))
+ (if (plist-member props :published)
+ (atom-modify-entry entry 'published (atom-format-time (plist-get props :published))))
+ (if (plist-member props :summary)
+ (atom-modify-entry entry 'summary (plist-get props :summary)))
(atom-modify-entry entry 'content content)
(atom-push-entry atom entry)
entry))
"Add an entry to ATOM, with a textual content. See
`atom-add-entry' for details.")
-(defun atom-add-html-entry (atom title link content
- &optional summary updated id)
- "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))
-
-(defun atom-add-xhtml-entry (atom title link content
- &optional summary updated id)
- "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))
+(defun atom-add-html-entry (atom title link content &optional props)
+ "Add an entry to ATOM, with some HTML content.
+
+TITLE, LINK, PROPS as in `atom-add-entry'. CONTENT should be a string
+enconding a valid HTML fragment. See `atom-add-entry' for
+additional details."
+ (if (plist-member props :summary)
+ (plist-put props :summary (atom-massage-html (plist-get props :summary))))
+ (atom-add-entry atom title link (atom-massage-html content) props))
+
+(defun atom-add-xhtml-entry (atom title link content &optional props)
+ "Add an entry to ATOM, with some XHTML content.
+
+TITLE, LINK, PROPS as in `atom-add-entry'. 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."
+ (if (plist-member props :summary)
+ (plist-put props :summary (atom-massage-xhtml (plist-get props :summary))))
+ (atom-add-entry atom title link (atom-massage-xhtml content) props))
+
+(defvar atom-xml-declaration "<?xml version=\"1.0\" encoding=\"utf-8\"?>\n")
(defun atom-print (atom)
"Print the Atom feed ATOM in the current buffer."
- (insert "<?xml version=\"1.0\" encoding=\"utf-8\"?>\n")
+ (insert atom-xml-declaration)
(insert "<feed xmlns=\"http://www.w3.org/2005/Atom\">\n")
(xml-print atom)
(insert "\n</feed>"))
+(defun atom-write-file (atom filename)
+ "Writes the feed ATOM to FILENAME."
+ (with-temp-buffer
+ (atom-print atom)
+ (write-file filename)))
+
\f
+(defun atom-to-rss (atom &optional rss-self)
+ "Translate Atom feed ATOM into an RSS one, returning the translation.
+
+If RSS-SELF is given, it is used as self link of the RSS feed.
+
+Some information may be lost or approximated."
+ (let ((rss (list (assoc 'title atom))))
+ (if rss-self
+ (atom-modify-entry rss 'atom:link
+ `(((href . ,rss-self) (rel . "self")
+ (type . "application/atom+xml")))))
+ (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 the Atom entry 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)))
+ (setcar (cdr descr) nil))
+ `(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 &optional rss-self)
+ "Convert Atom feed ATOM to RSS in the current buffer.
+
+If RSS-SELF is given, it is used as self link of the RSS feed."
+ (let ((rss (atom-to-rss atom rss-self)))
+ (insert atom-xml-declaration)
+ ;; xmlns:atom included in order to allow the atom:link rel=self element
+ (insert "<rss version=\"2.0\" xmlns:atom=\"http://www.w3.org/2005/Atom\">\n")
+ (insert " <channel>\n")
+ (xml-print rss " ")
+ (insert "\n </channel>\n")
+ (insert "</rss>")))
+
+(defun atom-to-rss-time (time)
+ "Translate TIME from the format used by Atom into the format used by RSS.
+
+TIME is a string."
+ (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 &optional rss-self)
+ "Save ATOM as a RSS feed into FILENAME.
+
+If RSS-SELF is given, it is used as self link of the RSS feed."
+ (with-temp-buffer
+ (atom-print-as-rss atom rss-self)
+ (write-file filename)))
+
+\f
+(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")
+
(defun atom-format-time (&optional time)
- "Format a time according to RFC3339."
+ "Format time value 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 string TIME as specified in RFC3339 into Emacs's native format."
+ ;; Same remark as in `atom-format-time': RFC3339 wants a colon between hour
+ ;; and minute parts of the timezome, so remove it before `date-to-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
-Atom feed. CONTENT must be a string."
+ "Massage CONTENT so it can be used as an HTML fragment in an Atom feed.
+
+CONTENT must be a string."
(list '((type . "html")) content))
(defun atom-string-to-xml (string)
"Convert STRING into a Lisp structure as used by `xml.el'."
- (with-temp-buffer
- (insert string)
- (xml-parse-region (point-min) (point-max))))
+ (require 'xml-xhtml-entities)
+ (let ((xml-entity-alist xml-xhtml-entities)
+ (xml-validating-parser t))
+ (with-temp-buffer
+ (insert "<div xmlns=\"" atom-xhtml-namespace "\">")
+ (insert string)
+ (insert "</div>")
+ ;; `xml-parse-region' returns a list of elements, even though it
+ ;; requires an only root node. 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."
+ "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;
+- nil, in which case variables `user-full-name' and `user-mail-address'
+ are used;
+- 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))))
+
+\f
+;;; Functions that should probably not be there
+
+(defun url-canonalize (address base)
+ "Make ADDRESS 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)))
(provide 'atom)
;;; atom.el ends here
projects / atom.git / blobdiff