// -*- Mode: C++;    tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*-
// Copyright 1999 Jose M. Vidal
// Jose M. Vidal, vidal@multiagent.com, http://jmvidal.ece.sc.edu
//
// This program is free software.  You can redistribute it and/or modify
// it under the terms of the GNU General Public License
//
// $Id: XMLParser.C,v 1.13 2002/04/02 16:51:51 jmvidal Exp $
#ifdef HAVE_CONFIG_H
#include <config.h>
#endif

#include "XMLParser.H"

using namespace std;

//defined in reference.C
extern string trim(const string & s);
extern int equalNoCase(const string& s, const string &s2);

void XMLParser::startTag(string &tagName, map<string,string> & attributes){
	cout << "startTag = " << tagName << endl;
	map<string, string>::const_iterator i;
	for (i = attributes.begin(); i != attributes.end(); ++i){
		cout << "key:" << i->first << " value:" << i->second << endl;
	}
}

void XMLParser::endTag(string &tagName){
	cout << "endTag = " << tagName << endl;
}

void XMLParser::cdata(string &content){
	cout << "content = " << content << endl;
}

string XMLParser::toUpper(const string &s){
	string result = "";
	for (string::const_iterator p = s.begin(); p!= s.end(); ++p){
		result += toupper(*p);
	}
	return result;
}

void XMLParser::parse(iwebstream &in) {
	char c;
	string data = "";
	string dataUpperCase = "";
	while (!in.eof()){
		c = in.get();
		if (isspace(c))
			continue;
		if (c == '<'){ //start of a tag
			while (isspace(c = in.get())) continue; //swallow whitespace
			if (c == '/') { //it is an end tag
				while ((c = in.get()) != '>'){
					if (in.eof()) return;
					data += c;
				}
				dataUpperCase = toUpper(data);
				endTag(dataUpperCase);
				data = "";
				continue;
			}
			else if (c == '!') {
				if ((c = in.get()) == '-'){ //its a comment, swallow it
					while ((c = in.get()) != '>')
						continue;
				}
				else if (c == 'D'){ //assume its <!DOCTYPE
					while (!isspace(c = in.get()))
						if (in.eof()) return;
					doctype += c;
					while (((c = in.get()) != '>') && (!isspace(c)))
						doctype += c;
					if (c != '>')
						while ((c = in.get()) != '>')
							continue;
				}
			}
			else if (c == '?'){ //<?xml version.....>, ignore it
				while ((c = in.get()) != '>')
					continue;
			}
			else { //it is a start tag, get the argument/value pairs
				data += c; //data will hold the tagname
				c = in.get();
				while (!isspace(c) && (c != '>')){
					if (in.eof()) return;
					data += c;
					c = in.get();
				}
				map<string,string> attributeList;
        data = toUpper(data);
				if (c == '>'){
          if (data[data.length() - 2] == '/') { //its a <tag/> so call it as an end tag
            data = data.substr(0, data.length() -1);
            endTag(data);
            data = "";
            continue;
          }
          startTag(data, attributeList);
					data = "";
					continue;
				}
				string attributeName = "";
				string attributeValue = "";
        string oldAttributeName = "";
				while ((c = in.get()) != '>') {
					if (in.eof()) return;
					if (c == '='){ //start getting the attribute-value
            if (attributeName == ""){
              attributeName = oldAttributeName;
              oldAttributeName = "";
            }
						attributeValue = "";
            while (isspace(c = in.get())) continue;
            if (c == '\"') { //value surrounded by quotes
              while ((c = in.get()) != '\"'){
                if (in.eof()) return;
                attributeValue += c;
              }
            }
						else { //value not surround by quotes, stop on a space.
              while (( !isspace(c = in.get())) && (c != '>')) {
                if (in.eof()) return;
                attributeValue += c;
              }
              in.putback(c);
            }
						continue;
					}
					else if (isspace(c) && (attributeName != "")) {
						dataUpperCase = toUpper(attributeName);
						attributeList[dataUpperCase] = attributeValue;
            if (attributeValue == "")
              oldAttributeName = attributeName; //save it in case this is a "href ="
						attributeName = "";
						attributeValue = "";
						continue;
					}
					if (!isspace(c))
						attributeName += c;
				}
				if (attributeName != ""){ //add the last one. It might already be there (e.g. <a h="r" >), if so this does nothing.
          dataUpperCase = toUpper(attributeName);
					attributeList[dataUpperCase] = attributeValue;
        }
				startTag(data, attributeList);
				data = "";
			}
		}
		else { //it is content
			data += c;
			while ((c = in.get()) != '<'){
        data += c;
				if (in.eof()) {
          if (data != "" )
            cdata(data);
          return;
        }
			}
      //      data = trim(data);
			if (trim(data) != "")
				cdata(data);
			data = "";
			in.putback(c);
		}
	}
}

string XMLParser::toString(string &tagName, map<string,string> & attributes){
  string result;
  result += "<" + tagName;
  map<string, string>::const_iterator i;
  for (i = attributes.begin(); i != attributes.end(); ++i){
    result += " " + i->first + "=\"" + i->second + "\"";
  }
  result += ">";
  return result;
}

//
//
// BookmarksParser
//
//
void BookmarksParser::startTag(string &tagName, map<string,string> & attributes){
	//	XMLParser::startTag(tagName, attributes);
	
	if (tagName == "DL"){ //start a new reference tree

		if (inReference){
			currentReference = 0;
			inReference = false;
			inTitle = false;
			inComment = false;
		}

		if (currentReferenceTree == 0)
			currentReferenceTree = root;
		else
			currentReferenceTree = new referenceTree;
		
		if (root == 0)
			root = currentReferenceTree;
		rtStack.push_back(currentReferenceTree);
		
	}
	else if ((tagName == "H3") || (tagName == "A")){
		map<string, string>::const_iterator i;
		string url;
		string addDate;
		string modDate;
		string visitDate;
		bool isAliasOf = false;
		string aliasOf;
		bool isAliasId = false;
		string aliasId;
		for (i = attributes.begin(); i != attributes.end(); ++i){
			if (i->first == HREF)
				url = i->second;
			else if (i->first == ADDDATE)
				addDate = i->second;
			else if (i->first == LASTVISIT)
				visitDate = i->second;
			else if (i->first == LASTMODIFIED)
				modDate = i->second;
			else if (i->first == ALIASOF ) {
				isAliasOf = true;
				aliasOf = i->second;
			}
			else if (i->first == ALIASID ) {
				isAliasId = true;
				aliasId = i->second;
			}
		}
		bool isFolder;
		if (tagName == "H3")
			isFolder = true;
		else
			isFolder = false;
		currentReference = new reference(url, addDate, modDate, visitDate, isFolder);

		if (isAliasId) {
			currentReference->aliasID = aliasId;
		}
		else if (isAliasOf) {
			currentReference->aliasof = true;
			currentReference->aliasOf = aliasOf;
		}

		if (isFolder){
			refStack.push_back(currentReference);
		}
		
		//	currentReference->writeAsBookHelper(cout, string(""));
		inReference = true;
		inTitle = true;
	}
	else if (tagName == "DT"){
		if (inReference) {
			endReference();
		}
	}

  else if (tagName == "BR"){
    if (inComment) //replace "<BR>" with a space.
      currentReference->comment += " ";
  }
  
}

/** Add the currentReference to the currentReferenceTree */
void BookmarksParser::endReference(){
// 	cout << "Stack size = " << rtStack.size() << endl;
// 	cout << "Adding reference" << endl;
	//	currentReference->writeAsBookHelper(cout, "");
	currentReferenceTree->addReference(*currentReference);
	currentReference = 0;
	inReference = false;
	inTitle = false;
	inComment = false;
}

void BookmarksParser::endTag(string &tagName){
	//	XMLParser::endTag(tagName);
	if (tagName == "DL") {
		if (inReference) {
			endReference();
		}
		if (!refStack.empty()){
			currentReference = refStack.back();
			currentReference->children = currentReferenceTree;
			currentReferenceTree->comment = currentReference->comment;
			//			currentReferenceTree->writeAsBookmarkFile(cout, "asdf");
			//			currentReference->writeAsBookHelper(cout, "****");
			rtStack.pop_back();
			currentReferenceTree = rtStack.back();
			endReference(); //add currentReference
			refStack.pop_back();
		}
		else {
			rtStack.pop_back();
			if ( rtStack.size() > 0 ) {
				currentReferenceTree = rtStack.back();
			}
			else {
				currentReferenceTree = 0;
			}
		}
		//		cout << "Poped size=" << rtStack.size() << endl;
	}
	else if (((tagName == "A") || (tagName == "H3")) && inTitle){
		inTitle = false;
		inComment = true;
	}
}


void BookmarksParser::cdata(string &content){
	//	XMLParser::cdata(content);
	if (inTitle) {
		currentReference->title = content;
		if (currentReference->isFolder())
			currentReference->dirname = content;
	}
	else if (inComment){
		currentReference->comment += content;
	}
}


//
//
// RSSParser
//
//

void RSSParser::startTag(string &tagName, map<string,string> & attributes){
  //  XMLParser::startTag(tagName, attributes);
	if (tagName == "ITEM"){
		currentReference = new reference();
		inItem = true;
	}
	else if (tagName == "TITLE"){
    inTitle = true;
  }
  else if (tagName == "LINK"){
    inLink = true;
  }
  else if (tagName == "DESCRIPTION"){
    inDescription = true;
  }
}

void RSSParser::endTag(string &tagName){
	//	XMLParser::endTag(tagName);

  if (tagName == "ITEM"){
    root->addReference(*currentReference);
    currentReference = 0;
		inItem = false;
	}
	else if (tagName == "TITLE"){
    inTitle = false;
  }
  else if (tagName == "LINK"){
    inLink = false;
  }
  else if (tagName == "DESCRIPTION"){
    inDescription = false;
  }
}


void RSSParser::cdata(string &content){
	//	XMLParser::cdata(content);
  if (inItem){
    if (inTitle) {
      currentReference->title = content;
    }
    else if (inLink) {
      currentReference->url = content;
    }
    else if (inDescription) {
      currentReference->comment = content;
    }
  }

}

//
//
// XBEL Parser
//
//

void XBELParser::startTag(string &tagName, map<string,string> & attributes){
  //  XMLParser::startTag(tagName, attributes);
	
	if (equalNoCase(tagName,"FOLDER")){ //start a new reference tree

    currentReferenceTree = new referenceTree;
		
		if (root == 0)
			root = currentReferenceTree;

    rtStack.push_back(currentReferenceTree);
    
    currentFolder = new reference; //also, start a new reference

    currentFolder->children = currentReferenceTree;
    currentFolder->folder = true;

    if (attributes["ADDED"] != "")
      currentFolder->creationTime = atoi(attributes["ADDED"].c_str());

    refStack.push_back(currentFolder);
    
    inFolder = true;
	}
  else if (equalNoCase(tagName,"TITLE")){
    inTitle = true;
  }
  else if (equalNoCase(tagName,"DESC")){
    inDesc = true;
  }
  else if (equalNoCase(tagName,"TIME_ADDED")){
    inAdded = true;
  }
  else if (equalNoCase(tagName,"TIME_VISITED")){
    inVisited = true;
  }
  else if (equalNoCase(tagName,"TIME_MODIFIED")){
    inModified = true;
  }
	else if (equalNoCase(tagName,"BOOKMARK")){
		map<string, string>::const_iterator i;
		string url = "";
		string addDate = "";
		string modDate = "";
		string visitDate = "";
		for (i = attributes.begin(); i != attributes.end(); ++i){
			if (equalNoCase(i->first,"HREF"))
				url = i->second;
			else if (equalNoCase(i->first,"ADDED"))
				addDate = i->second;
			else if (equalNoCase(i->first,"VISITED"))
				visitDate = i->second;
			else if (equalNoCase(i->first,"MODIFIED"))
				modDate = i->second;
		}
		currentBookmark = new reference(url, addDate, modDate, visitDate, false);
		inBookmark = true;
	}
}

void XBELParser::endTag(string &tagName){
  //  XMLParser::endTag(tagName);
	if (equalNoCase(tagName,"TITLE")) {
    inTitle = false;
  }
  else if (equalNoCase(tagName,"DESC")){
    inDesc = false;
  }
  else if (equalNoCase(tagName,"TIME_ADDED")){
    inAdded = false;
  }
  else if (equalNoCase(tagName,"TIME_VISITED")){
    inVisited = false;
  }
  else if (equalNoCase(tagName,"TIME_MODIFIED")){
    inModified = false;
  }
  else if (equalNoCase(tagName,"FOLDER")){
    rtStack.pop_back();
    currentReferenceTree = rtStack.back();
    currentReferenceTree->addReference(*currentFolder);
    if (!refStack.empty()){
      refStack.pop_back();
      currentFolder = refStack.back();
    }
    inFolder = false;
  }
  else if (equalNoCase(tagName,"BOOKMARK")){
    currentReferenceTree->addReference(*currentBookmark);
    //delete currentReference;
    currentBookmark = 0;
    inBookmark = false;
  }
}


void XBELParser::cdata(string &content){
  //  XMLParser::cdata(content);
  if (inFolder && !inBookmark){
    if (inTitle) {
      currentFolder->title = content;
      currentFolder->dirname = content;
    }
    else if (inDesc){
      currentFolder->comment = content;
    }
    else if (inAdded) {
      currentFolder->creationTime = atoi(content.c_str());
    }
    else if (inVisited) {
      currentFolder->visitTime = atoi(content.c_str());
    }
    else if (inModified) {
      currentFolder->modifiedTime = atoi(content.c_str());
    }
  }
  else if (inBookmark) {
    if (inTitle) {
      currentBookmark->title = content;
      currentBookmark->dirname = content;
    }
    else if (inDesc){
      currentBookmark->comment = content;
    }
    else if (inAdded) {
      currentBookmark->creationTime = atoi(content.c_str());
    }
    else if (inVisited) {
      currentBookmark->visitTime = atoi(content.c_str());
    }
    else if (inModified) {
      currentBookmark->modifiedTime = atoi(content.c_str());
    }
  }
}


//
//
// FolderCommentParser
//
//

void FolderCommentParser::startTag(string &tagName, map<string,string> & attributes){
  //  XMLParser::startTag(tagName, attributes);
  if (tagName == "PRIVATE"){
    r.priv = true;
  }
	if (tagName == "ALIAS"){
		inAlias = true;
	}
	else if (tagName == "DIRNAME"){
    inDirname = true;
  }
  else if (tagName == "INCLUDE"){
    inInclude = true;
  }
  else if (tagName == "FOLDERHTML"){
    inFolderHTML = true;
  }
  else if (tagName == "SORT"){
    inSort = true;
    sortOrder = toUpper(attributes["ORDER"]);
    if (sortOrder == "")
      sortOrder = "NORMAL";
  }
  else if (inFolderHTML) {
    r.folderHTML += toString(tagName, attributes);
  }
  else {
    finalComment += toString(tagName, attributes);
  }
}

void FolderCommentParser::endTag(string &tagName){
	//	XMLParser::endTag(tagName);

  if (tagName == "ALIAS"){
    inAlias = false;
	}
	else if (tagName == "DIRNAME"){
    inDirname = false;
  }
  else if (tagName == "INCLUDE"){
    inInclude = false;
  }
  else if (tagName == "FOLDERHTML"){
    inFolderHTML = false;
  }
  else if (tagName == "SORT"){
    inSort = false;
  }
  else if (inFolderHTML){
    r.folderHTML += "</" + tagName + ">";
  }
  else {
    finalComment += "</" + tagName + ">";
  }
}


void FolderCommentParser::cdata(string &content){
  //  cout << "\"" << content << "\"" << endl;
  if (inAlias) {
    r.aliasof = true;
    r.aliasDir = trim(content);
  }
  else if (inDirname){
    r.dirname = trim(content);
  }
  else if (inInclude){
    string fileName = trim(content);
    iwebstream istream(fileName.c_str());
    cout << "Found INCLUDE " << fileName << " in folder " << r.title << endl;
    if (istream == 0){
      cout << "Could not open " << fileName << endl;
    }
    else {
      cout << "Reading " << fileName << endl;
      referenceTree rtc2(istream);
      r.children->addTree(rtc2);
      r.comment = "";
      r.children->comment = "";
    }
  }
  else if (inFolderHTML){
    r.folderHTML += content;
  }
  else if (inSort){
    if (sortOrder == "NORMAL"){
//       cout << "sortOrder=Normal content=" << content << endl;
//       cout << "Sorting folder " << r.title << endl;
      if (equalNoCase(content, "CREATION"))
        sort(r.children->getContents().begin(), r.children->getContents().end(), referenceCmpCreation());
      else if (equalNoCase(content, "TITLE"))
        sort(r.children->getContents().begin(), r.children->getContents().end(), referenceCmpTitle());
      else if (equalNoCase(content, "MODIFIED"))
        sort(r.children->getContents().begin(), r.children->getContents().end(), referenceCmpModified());
      else if (equalNoCase(content, "VISIT"))
        sort(r.children->getContents().begin(), r.children->getContents().end(), referenceCmpVisit());
      else if (equalNoCase(content, "HITS"))
        sort(r.children->getContents().begin(), r.children->getContents().end(), referenceCmpHits());
      else
        cerr << "ERROR: Unknown sort field: " << content << endl;
    }
    else if (sortOrder == "INVERSE"){
      if (equalNoCase(content, "CREATION"))
        sort(r.children->getContents().begin(), r.children->getContents().end(), referenceCmpCreationI());
      else if (equalNoCase(content, "TITLE"))
        sort(r.children->getContents().begin(), r.children->getContents().end(), referenceCmpTitleI());
      else if (equalNoCase(content, "MODIFIED"))
        sort(r.children->getContents().begin(), r.children->getContents().end(), referenceCmpModifiedI());
      else if (equalNoCase(content, "VISIT"))
        sort(r.children->getContents().begin(), r.children->getContents().end(), referenceCmpVisitI());
      else if (equalNoCase(content, "HITS"))
        sort(r.children->getContents().begin(), r.children->getContents().end(), referenceCmpHitsI());
      else
        cerr << "ERROR: Unknown sort field: " << content << endl;
    }
    else
      cerr << "ERROR: Unknown sort order " << sortOrder << endl;
  }
  else {
    finalComment += content;
  }
}


void FolderCommentParser::parse(string & data){
  iwebstream in("data:" + data);
  XMLParser::parse(in);
  r.comment = trim(finalComment);

  //An ugly hack to get around the bug in Netscape which swallows the last
  // characters in a comment that ends in ">", so we end in "." instead.
  if (r.comment == "."){
    r.comment = "";
  }
};