From 4372ce9bf443ceca12449bcaf1a89460e85c05db Mon Sep 17 00:00:00 2001
From: rich <rich>
Date: Thu, 3 Aug 2006 13:33:15 +0000
Subject: [PATCH] Make the 404 page more informative - it now actually performs
 the search for you.

---
 html/_graphics/searching.gif   | Bin 0 -> 1547 bytes
 html/_graphics/searching.xcf   | Bin 0 -> 5133 bytes
 scripts/page.ml                | 134 ++++++++++++++++++++++++++++++++++++++---
 templates/page_404.html        |  61 ++++++++++++++-----
 templates/page_404_header.html |  26 ++++++++
 5 files changed, 198 insertions(+), 23 deletions(-)
 create mode 100644 html/_graphics/searching.gif
 create mode 100644 html/_graphics/searching.xcf
 create mode 100644 templates/page_404_header.html

diff --git a/html/_graphics/searching.gif b/html/_graphics/searching.gif
new file mode 100644
index 0000000000000000000000000000000000000000..40a85ff0be0f5adff2c3120d181c09f570ff09e4
GIT binary patch
literal 1547
zcmZ?wbhEHb3}+BvIK}`1jEsy-Oiawo%q%P{tgNiu+}u1oJOTm&!otELA|j%qqT=G>
z5)u-Ul9EzVQqt1Wva+&ra&q$W@(KzHii(O#N=hm!DypigYHDig>gpOA8k(A#T3TA#
z+S<Chx`u{^CMG6kW@c7aR<^db4h{~Ej*hObuAZKrK0ZGF{{Df1fk8n*!NI{{VPR2G
zQE_o`2?+^FNlD4c$*HNSxw*M{d3mL!rDbJh<>lp-m6g@i)wQ*?O-)U$t*vcsZ5<sQ
z-QC?iJv|dAPMkDp(zI#Qrca+fYu2pUvuDqlGiUDHx%21GU$kh^iWMtXu3Wix?b?kS
zH}2ZCYtNoNr%s(Zefsp-vuDqrKY#J!#Y>kiUA}zz>eZ{)u3fu+{rZg?H*Vg%dF$4#
z+qZAuxpU|4-Mjbi-+%Dn!J|iyo<4p0?Af#D&!4||@#591S8v|DdHeS5yLa#2zkmPX
z!-tO_KYsrF`RmuO-@bkO{{8!pA3uKm`t|$w?>~S3{Qdj)-@kwV|NjSvI+RfS&+X?L
z671|4;A*62z|05~R{SUGT$GwvlA5AWo>`Ki5R#Fq;O^-gz@Ye(g)@bro<Rqs7ZiUC
z?Ef07o0?l%+uA!iySjUN`}!wLoHTjL)M?Xa%$(KiCndt;J0p_0HCJo))GU^;>iq19
zRc(tW=a#K#%V3OZ$<`MU(nzYd6BUsQTx>1oUY(|W)FzV2KwLPeT6#&Otcf_kf0HY(
zl-Psqt>$dTAx&x?)m09P)hXND#3C8It1GLOtg1fDIVqHIN=`}Q+Zm-3e4eegd0YvJ
zMGx*M8d&iCF<^FRKF%%ltspb;p%Y`%wr^7y1=269v5nx+xZ;*|ekxP-A%+dkM;lyS
z0)!?oerTI8ht>IAPu0$Jo(4<3Y`rC#8PnthY*UW>@YLv5P?Y)BvZC<#vbhU0zo|$b
zDB*9Hi(uTnXO7iot(7(oVJs_yIr_LmO57Y)6eOLNGtrSrQ23bY&&MiXl9Taz>5*nG
zvu8n9_q3mF;S_YxP+ani`TFI~Rd-G<W?_7Fz3ax3jtPn_mN#}3zrL5NAut-&7-1cY
z9@gFa>NPJ)GA=ySsat(vLqS91k?FRUA~hE_E?hfZh+pf*4DYF*<;;ZDTo@L%Oqj*r
z>Q@}AyW-Sb8!v$k^VW9wc5z5L%~*VP_xAVq+n5+f0~{m36R-q0=g(4?be0w_^=)q)
z4<BM=6N>-#B;bSN@wRPoUN$QdlJ!E*SKsiNA;8!o<dkJ6;<Y!*G(05q-IKtk2U~O9
Wa)Mk=2pqce_Wprp?syRn25SIv?1IAp

literal 0
HcmV?d00001

diff --git a/html/_graphics/searching.xcf b/html/_graphics/searching.xcf
new file mode 100644
index 0000000000000000000000000000000000000000..19a2822d3d5c10ccadb91bf0c6acec1e197bf207
GIT binary patch
literal 5133
zcmeHKUr1A76#wovb-Jl5|MU;_ZfTp@s*ywlK^YN3(O(LBi01Wf-g294vm6nEPz3c*
zJrraT^i&TLy+-dv7NM8QvPdfXWB-ikkJIgR&h=iKIfL?{h<mu_e!rdXJLfyxvvbb(
z`K3S;-y-;UpVTNaq@Bn%0!TBEISyY1$mW_8j1eOhBAbvik&BR*Am{qglv@Y|0%9<V
zzZoeOuMLads94V*kfIIz_6Ct(U%PQL%@!k*30j3FxNSK8rU{#}qZIJ^MYlgJ)uRP6
z$C+J`Xsl7>rC?MHM?@hig@TdF9GA~8gc?I(-lYugu)ogBFI~J;p)|mwPF136T^dzw
zo}U_!YD=WaD@Z{<@6q-`VJw|cmARXPQWTvR8btECCFJu(#OSz#q&-<QCYh*8-$NyC
z7z4P7u{@6^!BGn}g<xilSJ>wdhnj=+ybx-N%~uv0)YTMc3Kk%nRZLphD#n53q+m-5
zwx;0R6r7KkY-yB*fc79t{eO)~nBlStG_P^3;X0b&C&~?djAh**Su2+b@k>sA)<r0W
zglq&OB;Id=Sui|kx&uRRAs>=SFey_ra4HVjF#3|a3kK+ZyNfTaeGG@~eAQLic+hbc
z2DaGwL$YCadD%5^89`nsme1`3I6d3x=!6`&)oc;E0M@ia@@O^m$Ki}aX6GRR%VWvW
z7czGfPE`Q<Aah;NZif#UJr}HCh2g&a)*}PV$vHoim%{yRj`9qdivv7mpMZncr~o}C
zMGsMT20UW%E(_#A{7OMT8$wUNel9vWA{%ib2cu8e7fp+{QXU$Tp0>I@V1YL>!0}3`
z4vau|yq~>-bEIg1=ap^A(O!hAj*1dTw;cOMU+v=#d=GwIInTcC9=zN-r()GZI8f-0
zxfJbZ4_42$2PbL2D=RDUbfERbpc|G!{Pswq!v-s0_}zw*1=XF<$_|4bBKqQbi1aLi
z>12_KVK%){^OUmHsw5p|E~cLF=W|>W75pk>tx#*lmx{mbr<Q>t<v{Q^8bt<eK1SwV
z6==SqKjtB9eH9bs*(lHbNHV<&=~YPY<$5pw|BqWKjiNwl4tvL&!<4^&-2^N@ngEs3
zzjaa2E-L6kUYh!!hxNaZdKQt|7xJ<G7wGjL(|+vHPoerLbXuoS{kr|n-{q+OV?9r0
JEQsR={O>tv=ji|d

literal 0
HcmV?d00001

diff --git a/scripts/page.ml b/scripts/page.ml
index 4a3b020..47e6ae9 100644
--- a/scripts/page.ml
+++ b/scripts/page.ml
@@ -1,7 +1,7 @@
 (* COCANWIKI - a wiki written in Objective CAML.
  * Written by Richard W.M. Jones <rich@merjis.com>.
  * Copyright (C) 2004 Merjis Ltd.
- * $Id: page.ml,v 1.49 2006/08/01 14:50:47 rich Exp $
+ * $Id: page.ml,v 1.50 2006/08/03 13:33:15 rich Exp $
  *
  * This program is free software; you can redistribute it and/or modify
  * it under the terms of the GNU General Public License as published by
@@ -34,6 +34,7 @@ open Cocanwiki_date
 open Cocanwiki_server_settings
 open Cocanwiki_links
 open Cocanwiki_extensions
+open Cocanwiki_strings
 
 type fp_status = FPOK of int32 * string * string * Calendar.t * bool
 	       | FPInternalRedirect of string
@@ -74,6 +75,7 @@ let run r (q : cgi) dbh hostid
   let template_page = get_template ~page dbh hostid "page.html" in
 
   (* This is the simpler template for 404 pages. *)
+  let template_404_header  = get_template dbh hostid "page_404_header.html" in
   let template_404  = get_template dbh hostid "page_404.html" in
 
   (* Host-specific fields. *)
@@ -397,8 +399,8 @@ let run r (q : cgi) dbh hostid
   let make_404 () =
     Request.set_status r 404;		(* Return a 404 error code. *)
 
-    let t = template_404 in
-    t#set "page" page;
+    let th = template_404_header in
+    th#set "page" page;
 
     let search_terms =
       String.map
@@ -406,13 +408,129 @@ let run r (q : cgi) dbh hostid
              ('a'..'z' | 'A'..'Z' | '0'..'9') as c -> c
            | _ -> ' ') page in
 
-    t#set "search_terms" search_terms;
+    th#set "search_terms" search_terms;
 
-    t#conditional "can_edit" can_edit;
-    t#conditional "can_manage_users" can_manage_users;
-    t#conditional "has_stats" has_stats;
+    (* Flush out the header while we start the search. *)
+    q#header ();
+    ignore (print_string r th#to_string);
+    ignore (Request.rflush r);
+
+    let t = template_404 in
+    t#set "query" search_terms;
+    t#set "canonical_hostname" host.canonical_hostname;
+
+    (* This is a simplified version of the code in search.ml. *)
+    let have_results =
+      (* Get the keywords from the query string. *)
+      let keywords = Pcre.split ~rex:split_words search_terms in
+      let keywords =
+	List.filter (fun s -> not (string_is_whitespace s)) keywords in
+      let keywords = List.map String.lowercase keywords in
+
+      (* Turn the keywords into a tsearch2 ts_query string. *)
+      let tsquery = String.concat "&" keywords in
+
+      (* Search the titles first. *)
+      let rows =
+	PGSQL(dbh)
+	    "select url, title, last_modified_date,
+                    (lower (title) = lower ($search_terms)) as exact
+               from pages
+              where hostid = $hostid
+	        and url is not null
+	        and redirect is null
+                and title_description_fti @@ to_tsquery ('default', $tsquery)
+              order by exact desc, last_modified_date desc, title" in
+
+      let titles =
+	List.map (function
+		  | (Some url, title, last_modified, _) ->
+		      url, title, last_modified
+		  | _ -> assert false) rows in
+
+      let have_titles = titles <> [] in
+      t#conditional "have_titles" have_titles;
+
+      (* Search the contents. *)
+      let rows =
+	PGSQL(dbh)
+	  "select c.id, p.url, p.title, p.last_modified_date
+             from contents c, pages p
+            where c.pageid = p.id
+              and p.hostid = $hostid
+              and url is not null
+              and p.redirect is null
+              and c.content_fti @@ to_tsquery ('default', $tsquery)
+            order by p.last_modified_date desc, p.title
+            limit 50" in
+
+      let contents =
+	List.map (function
+		  | (contentid, Some url, title, last_modified) ->
+		      contentid, url, title, last_modified
+		  | _ -> assert false) rows in
+
+      let have_contents = contents <> [] in
+      t#conditional "have_contents" have_contents;
+
+      (* Pull out the actual text which matched so we can generate a summary.
+       * XXX tsearch2 can actually do better than this by emboldening
+       * the text which maps.
+       *)
+      let content_map =
+	if contents = [] then []
+	else (
+	  let rows =
+	    let contentids =
+	      List.map (fun (contentid, _,_,_) -> contentid) contents in
+	    PGSQL(dbh)
+	      "select id, sectionname, content from contents
+                where id in $@contentids" in
+	  List.map (fun (id, sectionname, content) ->
+		      id, (sectionname, content)) rows
+	) in
 
-    q#template t
+      (* Generate the final tables. *)
+      let table =
+	List.map (fun (url, title, last_modified) ->
+		    let last_modified = printable_date last_modified in
+		    [ "url", Template.VarString url;
+		      "title", Template.VarString title;
+		      "last_modified", Template.VarString last_modified ]
+		 ) titles in
+      t#table "titles" table;
+
+      let table =
+	List.map
+	  (fun (contentid, url, title, last_modified) ->
+	     let sectionname, content = List.assoc contentid content_map in
+	     let have_sectionname, sectionname =
+	       match sectionname with
+		 None -> false, ""
+	       | Some sectionname -> true, sectionname in
+	     let content =
+	       truncate 160
+		 (Wikilib.text_of_xhtml
+		    (Wikilib.xhtml_of_content r dbh hostid content)) in
+	     let linkname = linkname_of_sectionname sectionname in
+	     let last_modified = printable_date last_modified in
+	     [ "url", Template.VarString url;
+	       "title", Template.VarString title;
+	       "have_sectionname", Template.VarConditional have_sectionname;
+	       "sectionname", Template.VarString sectionname;
+	       "linkname", Template.VarString linkname;
+	       "content", Template.VarString content;
+	       "last_modified", Template.VarString last_modified ]
+	  ) contents in
+      t#table "contents" table;
+
+      (* Do we have any results? *)
+      let have_results = have_titles || have_contents in
+      have_results in
+    t#conditional "have_results" have_results;
+
+    (* Deliver the rest of the page. *)
+    ignore (print_string r t#to_string)
   in
 
   (* Fetch a page by name.  This function can give three answers:
diff --git a/templates/page_404.html b/templates/page_404.html
index d6dda8d..f799265 100644
--- a/templates/page_404.html
+++ b/templates/page_404.html
@@ -1,24 +1,55 @@
-<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
-<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
-<head>
-<title>Page not found</title>
-<meta name="robots" content="noindex,nofollow"/>
-<meta name="author" content="http://www.merjis.com/" />
-<link rel="stylesheet" href="::theme_css_html_tag::" type="text/css" title="Standard"/>
-<link rel="alternate stylesheet" href="/_css/easytoread.css" type="text/css" title="High contrast, big fonts"/>
-</head><body>
-
-<h1><span>Page not found</span></h1>
+::if(have_results)::
+
+::if(have_titles)::
+<div id="titles">
+<ul id="titles">
+::table(titles)::
+<li>
+  <a href="/::url_html_tag::">::title_html::</a>
+  <span class="last_modified_date">- Last change: ::last_modified_html::</span>
+</li>
+::end::
+</ul>
+</div>
+::end::
+
+::if(have_contents)::
+<ul id="contents">
+::table(contents)::
+<li>
+  <a href="/::url_html_tag::">::title_html::</a>
+  ::if(have_sectionname)::
+    (<a href="/::url_html_tag::#::linkname_html_tag::">::sectionname_html::</a>)
+  ::end:: <br/>
+  <span class="content">::content::</span> <br/>
+  <span class="last_modified_date">Last change: ::last_modified_html::</span>
+</li>
+::end::
+</ul>
+::end::
+
+::else::
 
 <p>
-Search our site for this page:
+<strong>There are no similar pages found.</strong>
 </p>
 
-<form method="post" action="/_search">
 <p>
-<input name="q" value="::search_terms_html_tag::" size="60"/><input type="submit" value="Search"/>
+Some tips for finding what you want:
 </p>
-</form>
+
+<ul>
+<li> Use fewer words. </li>
+<li> <a href="/_sitemap">Try browsing the sitemap.</a> </li>
+<li> <a href="http://www.google.com/search?q=site&#58;::canonical_hostname_url::+::query_url::">Try using Google to find similar pages.</a>
+</ul>
+
+::end::
+
+<script type="text/javascript"><!--
+var img = document.getElementById ("searching_img");
+img.style.display = "none";
+//--></script>
 
 ::include(footer.html)::
 </body>
diff --git a/templates/page_404_header.html b/templates/page_404_header.html
new file mode 100644
index 0000000..f666742
--- /dev/null
+++ b/templates/page_404_header.html
@@ -0,0 +1,26 @@
+<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
+<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
+<head>
+<title>Page not found</title>
+<meta name="robots" content="noindex,nofollow"/>
+<meta name="author" content="http://www.merjis.com/" />
+<link rel="stylesheet" href="::theme_css_html_tag::" type="text/css" title="Standard"/>
+<link rel="stylesheet" href="/_css/search.css" type="text/css" title="Standard"/>
+<link rel="alternate stylesheet" href="/_css/easytoread.css" type="text/css" title="High contrast, big fonts"/>
+</head><body onload="document.f.q.focus ()">
+
+<h1><span>Page not found</span></h1>
+
+<p>
+Search our site for this page:
+</p>
+
+<form method="post" action="/_search" name="f">
+<p>
+<input name="q" value="::search_terms_html_tag::" size="60"/><input type="submit" value="Search"/>
+</p>
+</form>
+
+<p>
+<img width="87" height="16" src="/_graphics/searching.gif" alt="Searching ..." id="searching_img"/>
+</p>
\ No newline at end of file
-- 
1.8.3.1