;;;; -*- Mode: Lisp; Syntax: ANSI-Common-Lisp; Base: 10 -*-
;;;; *************************************************************************
;;;; FILE IDENTIFICATION
;;;;
;;;; Name:          xml-utils.lisp
;;;; Purpose:       XML utilities
;;;; Programmer:    Kevin M. Rosenberg
;;;; Date Started:  Apr 2000
;;;;
;;;; $Id$
;;;;
;;;; This file, part of KMRCL, is Copyright (c) 2002 by Kevin M. Rosenberg
;;;;
;;;; KMRCL users are granted the rights to distribute and use this software
;;;; as governed by the terms of the Lisp Lesser GNU Public License
;;;; (http://opensource.franz.com/preamble.html), also known as the LLGPL.
;;;; *************************************************************************

(in-package #:kmrcl)


;;; XML Extraction Functions

(defun find-start-tag (tag taglen xmlstr start end)
  "Searches for the start of a tag in an xmlstring. Returns STARTPOS ATTRIBUTE-LIST)"
  (declare (simple-string tag xmlstr)
	   (fixnum taglen start end)
	   (optimize (speed 3) (safety 0) (space 0)))
  (do* ((search-str (concatenate 'string "<" tag))
	(search-len (1+ taglen))
	(bracketpos (fast-string-search search-str xmlstr search-len start end)
		    (fast-string-search search-str xmlstr search-len start end)))
       ((null bracketpos) nil)
    (let* ((endtag (+ bracketpos 1 taglen))
	   (char-after-tag (schar xmlstr endtag)))
      (when (or (char= #\> char-after-tag)
		(char= #\space char-after-tag))
	(if (char= #\> char-after-tag) 
	    (return-from find-start-tag (values (1+ endtag) nil))
	    (let ((endbrack (position-char #\> xmlstr (1+ endtag) end)))
	      (if endbrack
		  (return-from find-start-tag
		    (values (1+ endbrack)
			    (string-to-list-skip-delimiter
			     (subseq xmlstr endtag endbrack))))
		  (values nil nil)))))
      (setq start endtag))))


(defun find-end-tag (tag taglen xmlstr start end)
  (fast-string-search
   (concatenate 'string "</" tag ">") xmlstr
   (+ taglen 3) start end))
  
(defun positions-xml-tag-contents (tag xmlstr &optional (start-xmlstr 0)
				       (end-xmlstr (length xmlstr)))
  "Returns three values: the start and end positions of contents between
 the xml tags and the position following the close of the end tag."
  (let* ((taglen (length tag)))
    (multiple-value-bind (start attributes)
	(find-start-tag tag taglen xmlstr start-xmlstr end-xmlstr)
      (unless start
	(return-from positions-xml-tag-contents (values nil nil nil nil)))
      (let ((end (find-end-tag tag taglen xmlstr start end-xmlstr)))
	(unless end
	  (return-from positions-xml-tag-contents (values nil nil nil nil)))
	(values start end (+ end taglen 3) attributes)))))


(defun xml-tag-contents (tag xmlstr &optional (start-xmlstr 0)
			 (end-xmlstr (length xmlstr)))
  "Returns two values: the string between XML start and end tag 
and position of character following end tag."
  (multiple-value-bind 
      (startpos endpos nextpos attributes) 
      (positions-xml-tag-contents tag xmlstr start-xmlstr end-xmlstr)
    (if (and startpos endpos)
	(values (subseq xmlstr startpos endpos) nextpos attributes)
      (values nil nil nil))))

(defun cdata-string (str)
  (concatenate 'string "<![CDATA[" str "]]>"))

(defun write-cdata (str s)
  (declare (simple-string str) (optimize (speed 3) (safety 0) (space 0)))
  (do ((len (length str))
       (i 0 (1+ i)))
      ((= i len) str)
    (declare (fixnum i len))
    (let ((c (schar str i)))
      (case c
	(#\< (write-string "&lt;" s))
	(#\& (write-string "&amp;" s))
	(t   (write-char c s))))))

(defun xml-declaration-stream (stream &key (version "1.0") standalone encoding)
  (format stream "<?xml version=\"~A\"~A~A ?>~%"
	  version
	  (if encoding
	      (format nil " encoding=\"~A\"" encoding)
	      ""
	      )
	  (if standalone
	      (format nil " standalone=\"~A\"" standalone)
	      "")))

(defun dtd-stream (stream format &key name public-id system-id entities)
  (case format
    ((:xhtml11 :xhtml :xhtml10-strict :xhtml10-transitional :xhtml-frameset)
     (setq name "html"))
    (:html
     (setq name "HTML"))
    ((:docbook :docbook42)
     (unless name
       (setq name "book"))))

  (case format
    (:html
     (unless public-id
       (setq public-id "-//IETF//DTD HTML//EN"))
     (setq system-id nil))
    ((:xhtml11 :xhtml)
     (unless public-id
       (setq public-id  "-//W3C//DTD XHTML 1.1//EN"))
     (unless system-id
       (setq system-id "http://www.w3.org/TR/xhtml11/DTD/xhtml11.dtd")))
    (:xhtml10-strict
     (unless public-id
       (setq public-id  "-//W3C//DTD XHTML 1.0 Strict//EN"))
     (unless system-id
       (setq system-id "http://www.w3.org/TR/xhtml10/DTD/xhtml10-strict.dtd")))
    (:xhtml10-transitional
     (unless public-id
       (setq public-id  "-//W3C//DTD XHTML 1.0 Transitional//EN"))
     (unless system-id
       (setq system-id "http://www.w3.org/TR/xhtml10/DTD/xhtml10-transitional.dtd")))
    (:xhtml10-frameset
     (unless public-id
       (setq public-id  "-//W3C//DTD XHTML 1.0 Frameset//EN"))
     (unless system-id
       (setq system-id "http://www.w3.org/TR/xhtml10/DTD/xhtml10-frameset.dtd")))
    ((:docbook :docbook42)
     (unless public-id
       (setq public-id  "-//OASIS//DTD DocBook XML V4.2//EN")
     (unless system-id
       (setq system-id "http://www.oasis-open.org/docbook/xml/4.2/docbookx.dtd")))))
    
  (format stream "<!DOCTYPE ~A PUBLIC \"~A\"~A"
	  name public-id
	  (if system-id
	    (format nil " \"~A\"" system-id)
	    ""))
  
  (when entities
    (format stream " [~%~A~%]" entities))
  (write-char #\> stream)
  (write-char #\newline stream)
  stream)

(defun sgml-header-stream (stream format &key entities (encoding "iso-8859-1") standalone (version "1.0")
			  name public-id system-id)
  (when (in format :xhtml :xhtml11 :xhtml10-strict :xhtml10-transitional :xhtml10-frameset :xml :docbook)
    (xml-declaration-stream stream :version version :encoding encoding :standalone standalone))
  (unless (eq :xml format)
    (dtd-stream stream format :name name :public-id public-id :system-id system-id
		:entities entities))
  stream)