/* Copyright (C) 2007 J.F.Dockes
* 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
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the
* Free Software Foundation, Inc.,
* 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
*/
#ifdef HAVE_CONFIG_H
#include "autoconfig.h"
#endif
#include <stdio.h>
#include <stdlib.h>
#include <math.h>
#include <time.h>
#include <sstream>
using std::ostringstream;
using std::endl;
#include "cstr.h"
#include "reslistpager.h"
#include "debuglog.h"
#include "rclconfig.h"
#include "smallut.h"
#include "plaintorich.h"
#include "mimehandler.h"
// Default highlighter. No need for locking, this is query-only.
static const string cstr_hlfontcolor("<font color=\"blue\">");
static const string cstr_hlendfont("</font>");
class PlainToRichHtReslist : public PlainToRich {
public:
virtual string startMatch(unsigned int)
{
return cstr_hlfontcolor;
}
virtual string endMatch()
{
return cstr_hlendfont;
}
};
static PlainToRichHtReslist g_hiliter;
ResListPager::ResListPager(int pagesize)
: m_pagesize(pagesize),
m_newpagesize(pagesize),
m_resultsInCurrentPage(0),
m_winfirst(-1),
m_hasNext(true),
m_hiliter(&g_hiliter)
{
}
void ResListPager::resultPageNext()
{
if (m_docSource.isNull()) {
LOGDEB(("ResListPager::resultPageNext: null source\n"));
return;
}
int resCnt = m_docSource->getResCnt();
LOGDEB(("ResListPager::resultPageNext: rescnt %d, winfirst %d\n",
resCnt, m_winfirst));
if (m_winfirst < 0) {
m_winfirst = 0;
} else {
m_winfirst += m_respage.size();
}
// Get the next page of results. Note that we look ahead by one to
// determine if there is actually a next page
vector<ResListEntry> npage;
int pagelen = m_docSource->getSeqSlice(m_winfirst, m_pagesize + 1, npage);
// If page was truncated, there is no next
m_hasNext = (pagelen == m_pagesize + 1);
// Get rid of the possible excess result
if (pagelen == m_pagesize + 1) {
npage.resize(m_pagesize);
pagelen--;
}
if (pagelen <= 0) {
// No results ? This can only happen on the first page or if the
// actual result list size is a multiple of the page pref (else
// there would have been no Next on the last page)
if (m_winfirst > 0) {
// Have already results. Let them show, just disable the
// Next button. We'd need to remove the Next link from the page
// too.
// Restore the m_winfirst value, let the current result vector alone
m_winfirst -= m_respage.size();
} else {
// No results at all (on first page)
m_winfirst = -1;
}
return;
}
m_resultsInCurrentPage = pagelen;
m_respage = npage;
}
void ResListPager::resultPageFor(int docnum)
{
if (m_docSource.isNull()) {
LOGDEB(("ResListPager::resultPageFor: null source\n"));
return;
}
int resCnt = m_docSource->getResCnt();
LOGDEB(("ResListPager::resultPageFor(%d): rescnt %d, winfirst %d\n",
docnum, resCnt, m_winfirst));
m_winfirst = (docnum / m_pagesize) * m_pagesize;
// Get the next page of results.
vector<ResListEntry> npage;
int pagelen = m_docSource->getSeqSlice(m_winfirst, m_pagesize, npage);
// If page was truncated, there is no next
m_hasNext = (pagelen == m_pagesize);
if (pagelen <= 0) {
m_winfirst = -1;
return;
}
m_respage = npage;
}
void ResListPager::displayDoc(RclConfig *config, int i, Rcl::Doc& doc,
const HighlightData& hdata, const string& sh)
{
ostringstream chunk;
int percent;
if (doc.pc == -1) {
percent = 0;
// Document not available, maybe other further, will go on.
doc.meta[Rcl::Doc::keyabs] = string(trans("Unavailable document"));
} else {
percent = doc.pc;
}
// Determine icon to display if any
string iconurl = iconUrl(config, doc);
// Printable url: either utf-8 if transcoding succeeds, or url-encoded
string url;
printableUrl(config->getDefCharset(), doc.url, url);
// Make title out of file name if none yet
string titleOrFilename;
string utf8fn;
doc.getmeta(Rcl::Doc::keytt, &titleOrFilename);
doc.getmeta(Rcl::Doc::keyfn, &utf8fn);
if (utf8fn.empty()) {
utf8fn = path_getsimple(url);
}
if (titleOrFilename.empty()) {
titleOrFilename = utf8fn;
}
// Result number
char numbuf[20];
int docnumforlinks = m_winfirst + 1 + i;
sprintf(numbuf, "%d", docnumforlinks);
// Document date: either doc or file modification time
char datebuf[100];
datebuf[0] = 0;
if (!doc.dmtime.empty() || !doc.fmtime.empty()) {
time_t mtime = doc.dmtime.empty() ?
atol(doc.fmtime.c_str()) : atol(doc.dmtime.c_str());
struct tm *tm = localtime(&mtime);
strftime(datebuf, 99, dateFormat().c_str(), tm);
}
// Size information. We print both doc and file if they differ a lot
off_t fsize = -1, dsize = -1;
if (!doc.dbytes.empty())
dsize = atol(doc.dbytes.c_str());
if (!doc.fbytes.empty())
fsize = atol(doc.fbytes.c_str());
string sizebuf;
if (dsize > 0) {
sizebuf = displayableBytes(dsize);
if (fsize > 10 * dsize && fsize - dsize > 1000)
sizebuf += string(" / ") + displayableBytes(fsize);
} else if (fsize >= 0) {
sizebuf = displayableBytes(fsize);
}
string richabst;
bool needabstract = parFormat().find("%A") != string::npos;
if (needabstract && m_docSource.isNotNull()) {
vector<string> vabs;
m_docSource->getAbstract(doc, vabs);
m_hiliter->set_inputhtml(false);
for (vector<string>::const_iterator it = vabs.begin();
it != vabs.end(); it++) {
if (!it->empty()) {
// No need to call escapeHtml(), plaintorich handles it
list<string> lr;
// There may be data like page numbers before the snippet text.
// will be in brackets.
string::size_type bckt = it->find("]");
if (bckt == string::npos) {
m_hiliter->plaintorich(*it, lr, hdata);
} else {
m_hiliter->plaintorich(it->substr(bckt), lr, hdata);
lr.front() = it->substr(0, bckt) + lr.front();
}
richabst += lr.front();
richabst += absSep();
}
}
}
// Links;
ostringstream linksbuf;
if (canIntern(doc.mimetype, config)) {
linksbuf << "<a href=\"P" << docnumforlinks << "\">"
<< trans("Preview") << "</a> ";
}
string apptag;
doc.getmeta(Rcl::Doc::keyapptg, &apptag);
if (!config->getMimeViewerDef(doc.mimetype, apptag, false).empty()) {
linksbuf << "<a href=\"E" << docnumforlinks << "\">"
<< trans("Open") << "</a>";
}
ostringstream snipsbuf;
if (doc.haspages) {
snipsbuf << "<a href=\"A" << docnumforlinks << "\">"
<< trans("Snippets") << "</a> ";
linksbuf << " " << snipsbuf.str();
}
// Build the result list paragraph:
// Subheader: this is used by history
if (!sh.empty())
chunk << "<p style='clear: both;'><b>" << sh << "</p>\n<p>";
else
chunk << "<p style='margin: 0px;padding: 0px;clear: both;'>";
// Configurable stuff
map<string, string> subs;
subs["A"] = !richabst.empty() ? richabst : "";
subs["D"] = datebuf;
subs["E"] = snipsbuf.str();
subs["I"] = iconurl;
subs["i"] = doc.ipath;
subs["K"] = !doc.meta[Rcl::Doc::keykw].empty() ?
string("[") + escapeHtml(doc.meta[Rcl::Doc::keykw]) + "]" : "";
subs["L"] = linksbuf.str();
subs["N"] = numbuf;
subs["M"] = doc.mimetype;
subs["R"] = doc.meta[Rcl::Doc::keyrr];
subs["S"] = sizebuf;
subs["T"] = escapeHtml(titleOrFilename);
subs["t"] = escapeHtml(doc.meta[Rcl::Doc::keytt]);
subs["U"] = url;
// Let %(xx) access all metadata. HTML-neuter everything:
for (map<string,string>::iterator it = doc.meta.begin();
it != doc.meta.end(); it++) {
if (!it->first.empty())
subs[it->first] = escapeHtml(it->second);
}
string formatted;
pcSubst(parFormat(), formatted, subs);
chunk << formatted;
chunk << "</p>" << endl;
// This was to force qt 4.x to clear the margins (which it should do
// anyway because of the paragraph's style), but we finally took
// the table approach for 1.15 for now (in guiutils.cpp)
// chunk << "<br style='clear:both;height:0;line-height:0;'>" << endl;
LOGDEB2(("Chunk: [%s]\n", (const char *)chunk.rdbuf()->str().c_str()));
append(chunk.rdbuf()->str(), i, doc);
}
bool ResListPager::getDoc(int num, Rcl::Doc& doc)
{
if (m_winfirst < 0 || m_respage.size() == 0)
return false;
if (num < m_winfirst || num >= m_winfirst + int(m_respage.size()))
return false;
doc = m_respage[num-m_winfirst].doc;
return true;
}
void ResListPager::displayPage(RclConfig *config)
{
LOGDEB(("ResListPager::displayPage\n"));
if (m_docSource.isNull()) {
LOGDEB(("ResListPager::displayPage: null source\n"));
return;
}
if (m_winfirst < 0 && !pageEmpty()) {
LOGDEB(("ResListPager::displayPage: sequence error: winfirst < 0\n"));
return;
}
ostringstream chunk;
// Display list header
// We could use a <title> but the textedit doesnt display
// it prominently
// Note: have to append text in chunks that make sense
// html-wise. If we break things up too much, the editor
// gets confused. Hence the use of the 'chunk' text
// accumulator
// Also note that there can be results beyond the estimated resCnt.
chunk << "<html><head>" << endl
<< "<meta http-equiv=\"content-type\""
<< " content=\"text/html; charset=utf-8\">" << endl
<< headerContent()
<< "</head><body>" << endl
<< pageTop()
<< "<p><font size=+1><b>"
<< m_docSource->title()
<< "</b></font> ";
if (pageEmpty()) {
chunk << trans("<p><b>No results found</b><br>");
string reason = m_docSource->getReason();
if (!reason.empty()) {
chunk << "<blockquote>" << escapeHtml(reason) <<
"</blockquote></p>";
} else {
HighlightData hldata;
m_docSource->getTerms(hldata);
vector<string> uterms(hldata.uterms.begin(), hldata.uterms.end());
if (!uterms.empty()) {
map<string, vector<string> > spellings;
suggest(uterms, spellings);
if (!spellings.empty()) {
if (o_index_stripchars) {
chunk <<
trans("<p><i>Alternate spellings (accents suppressed): </i>")
<< "<br /><blockquote>";
} else {
chunk <<
trans("<p><i>Alternate spellings: </i>")
<< "<br /><blockquote>";
}
for (map<string, vector<string> >::const_iterator it0 =
spellings.begin(); it0 != spellings.end(); it0++) {
chunk << "<b>" << it0->first << "</b> : ";
for (vector<string>::const_iterator it =
it0->second.begin();
it != it0->second.end(); it++) {
chunk << *it << " ";
}
chunk << "<br />";
}
chunk << "</blockquote></p>";
}
}
}
} else {
unsigned int resCnt = m_docSource->getResCnt();
if (m_winfirst + m_respage.size() < resCnt) {
chunk << trans("Documents") << " <b>" << m_winfirst + 1
<< "-" << m_winfirst + m_respage.size() << "</b> "
<< trans("out of at least") << " "
<< resCnt << " " << trans("for") << " " ;
} else {
chunk << trans("Documents") << " <b>"
<< m_winfirst + 1 << "-" << m_winfirst + m_respage.size()
<< "</b> " << trans("for") << " ";
}
}
chunk << detailsLink();
if (hasPrev() || hasNext()) {
chunk << " ";
if (hasPrev()) {
chunk << "<a href=\"" + prevUrl() + "\"><b>"
<< trans("Previous")
<< "</b></a> ";
}
if (hasNext()) {
chunk << "<a href=\""+ nextUrl() + "\"><b>"
<< trans("Next")
<< "</b></a>";
}
}
chunk << "</p>" << endl;
append(chunk.rdbuf()->str());
chunk.rdbuf()->str("");
if (pageEmpty())
return;
HighlightData hdata;
m_docSource->getTerms(hdata);
// Emit data for result entry paragraph. Do it in chunks that make sense
// html-wise, else our client may get confused
for (int i = 0; i < (int)m_respage.size(); i++) {
Rcl::Doc& doc(m_respage[i].doc);
string& sh(m_respage[i].subHeader);
displayDoc(config, i, doc, hdata, sh);
}
// Footer
chunk << "<p align=\"center\">";
if (hasPrev() || hasNext()) {
if (hasPrev()) {
chunk << "<a href=\"" + prevUrl() + "\"><b>"
<< trans("Previous")
<< "</b></a> ";
}
if (hasNext()) {
chunk << "<a href=\""+ nextUrl() + "\"><b>"
<< trans("Next")
<< "</b></a>";
}
}
chunk << "</p>" << endl;
chunk << "</body></html>" << endl;
append(chunk.rdbuf()->str());
}
// Default implementations for things that should be implemented by
// specializations
string ResListPager::nextUrl()
{
return "n-1";
}
string ResListPager::prevUrl()
{
return "p-1";
}
string ResListPager::iconUrl(RclConfig *config, Rcl::Doc& doc)
{
string apptag;
doc.getmeta(Rcl::Doc::keyapptg, &apptag);
return cstr_fileu + config->getMimeIconPath(doc.mimetype, apptag);
}
bool ResListPager::append(const string& data)
{
fprintf(stderr, "%s", data.c_str());
return true;
}
string ResListPager::trans(const string& in)
{
return in;
}
string ResListPager::detailsLink()
{
string chunk = "<a href=\"H-1\">";
chunk += trans("(show query)") + "</a>";
return chunk;
}
const string &ResListPager::parFormat()
{
static const string cstr_format("<img src=\"%I\" align=\"left\">"
"%R %S %L <b>%T</b><br>"
"%M %D <i>%U</i><br>"
"%A %K");
return cstr_format;
}
const string &ResListPager::dateFormat()
{
static const string cstr_format(" %Y-%m-%d %H:%M:%S %z");
return cstr_format;
}