# webster-word setup

;;; Fixes HTML from m-w.com by removing all instances of the
;;; "\240" character, which shows up as garbage on START,
;;; replacing all sense_break <span> tags with <br> tags, and
;;; adding <br>s in before every verb_class <span> tag.
(define (fix-definition-body html-str)
  (let* ((char-fixed (substitute html-str "\x09" ""))
	 (char-and-span-sense-fixed
	  (substitute char-fixed "<span class=\"sense_break\"/>" "<br>"))
	 (char-and-all-span-fixed
	  (substitute char-and-span-sense-fixed
		      "<span class=\"verb_class\">"
		      "<br><span class=\"verb_class\">"))
	 (junk-removed (substitute char-and-all-span-fixed
				    "<br /><div id=\"results_ad\"></div>"
				    ""))
	 (raw-trim (or (match "<div class=\"britannica-entry\">" junk-removed ':prefix)
		       (match "<div class=\"browse\">" junk-removed ':prefix)))
	 (trimmed (and raw-trim
		       ;; When we trim, we remove lots of </div>s.
		       ;; Put them all back on, plus some for good
		       ;; measure, and wrap an extra div element and
		       ;; use extract-html-element to trim the extras
		       ;; off again.  (The remaining wrapped div will
		       ;; do no harm.)  The extra </div>s give us a
		       ;; little wiggle room for the site format to
		       ;; change, or for trimming from the
		       ;; britannica-entry div to be different from
		       ;; trimming from the "Rhymes with" section.
		       (extract-html-element
			(string-append
			 "<div>" raw-trim
			 "</div></div></div></div></div></div></div></div></div></div></div></div></div></div></div></div>")
			"div"))))
    (remove-defs-variant (or trimmed junk-removed))
    ))

;;; Removes all <div class="defs variant"> tags without removing their
;;; content.
(define (remove-defs-variant html-str)
  (define (remove-defs-variant-helper new-str rest-str)
    (let* ((div-match (match "<div class=\"defs variant\">" rest-str)))
      (if (not div-match)
	  (string-append new-str rest-str)
	  (let* ((rest-text (match:suffix div-match))
		 (end-match (match "</div>" rest-text)))
	    (remove-defs-variant-helper
	     (string-append new-str (match:prefix div-match)
			    (make-shared-substring rest-text 0 (match:start end-match)))
	     (match:suffix end-match))))))
  (remove-defs-variant-helper "" html-str))


# webster-word DEFINITION-PAGE

(lambda (symbol)
  (define (found-definition str)
    (not (string-match "isn't in the dictionary" str)))
  (let* ((page (get-page (get "webster-word" symbol "URL")))
	 (fragment (and (found-definition page)
			(all-till-next-title symbol page 0))))
    (cond (fragment
	   (set! fragment (fix-definition-body (or (match "</h1>" fragment ':suffix)
						   fragment)))
	   (let* ((stylesheet-m (match "<link rel=\"stylesheet\"[^>]*>" page))
		  (stylesheets-start (and stylesheet-m (match:start stylesheet-m)))
		  (stylesheets-end #f))
	     (while stylesheet-m
		    (set! stylesheets-end (match:end stylesheet-m))
		    (set! stylesheet-m (match "<link rel=\"stylesheet\"[^>]*>" page
					      stylesheets-end)))
	     (qualify-relative-urls
	      (get "webster-word" symbol "SITE-URL")
	      (if stylesheets-start
		  (string-append (substring page stylesheets-start stylesheets-end)
				 fragment)
		  fragment))))
	  (else #f))))

# webster-word DEFINITION-PAGE-TERSE

(lambda (symbol)
  (let ((full-text (get "webster-word" symbol "DEFINITION-PAGE")))
    (match '("</h[0-9]>" i)
	   ;; Word appears in upper case in Webster in this context.
	   (all-till-next-title (string-append "Definition of <em>"
					       (string-upcase symbol)
					       "</em>")
				full-text 0)
	   ':suffix)))


# webster-word SITE-URL

(lambda (symbol)
  "http://www.merriam-webster.com")


# webster-word URL

; updated 25 Oct 2008, jesstess
(lambda (symbol)
  (let* ((symbol (substitute symbol " " "+"))
	 (base "http://www.merriam-webster.com/dictionary/")
	 (fullpath (string-append base symbol)))
    fullpath))