// TextDataset.cxx

#include "dataset_base/TextDataset.h"
#include <iostream>
#include <sstream>
#include <memory>
#include "dataset_util/FileStatus.h"
#include "dataset_util/WorkingDirectory.h"

using std::string;
using std::cerr;
using std::ostringstream;
using std::endl;
using dset::GenericDataset;
using dset::TextDataset;

typedef Text::WordList WordList;

//**********************************************************************
// Local definitions.
//**********************************************************************

namespace {

// Register creator.
int SingleFileDataset_xml_stat = GenericDataset::
  register_promoter("TextDataset",
                    dset::promote<TextDataset>);

// Register ID.
// Returns 0 for success.
int regid(const char* name) {
  static int stat = 1;
  if ( stat != 1 ) return stat;
  stat = 2;
  for ( int i=100; i<1000; ++i ) {
    ContentId cid = ContentId::register_id(i, name);
    if ( cid.is_valid() ) {
      stat = 0;
      break;
    }
  }
  return stat;
}

// Force registration
int CRSRTAT = regid("Text");

}  // end unnamed namespace.

//**********************************************************************
// Member functions.
//**********************************************************************

// Constructor.

TextDataset::TextDataset(string clab, string snames) {
  assert( regid("Text") == 0 );
  set_fulltype("TextDataset");
  set_id();
  // No location.
  // But *not* a virtual dataset.
  set_location_implicit();
  // Create a text object for each name.
  ContentIdList cids;
  // Lock the directory.
  WorkingDirectory wd;
  // Loop over names and construct a text object for each.
  WordList names = Text::split(snames);
  for ( WordList::const_iterator inam=names.begin();
        inam!=names.end(); ++ inam ) {
    if ( ! clab.size() ) {
      set_error(10);
      break;
    }
    Text::Name name = *inam;
    ContentId cid("Text", name);
    // Make sure name is not already used.
    if ( cids.find(cid) != cids.end() ) {
      set_error(11);
      break;
    }
    // Make sure the filename is relative.
    if ( ! FileName(name).is_relative() ) {
      set_error(12);
      break;
    }
    // Create text.
    Text txt(name, true, true);
    if ( ! txt.is_valid() ) {
      set_error(13);
      break;
    }
    assert( txt.name() == name );
    // All OK--insert the ID and text object.
    cids.insert(cid);
    m_texts.push_back(txt);
  }
  Content con("TextDataset", clab, cids);
  set_content(con);
  set_evstate_none();
}

//**********************************************************************

// Promoter.

TextDataset::TextDataset(const GenericDataset& gdst)
: GenericDataset(gdst) {
  const XmlElement* pxtra = extra_xml();
  if ( pxtra != 0 ) {
    // Fetch XML for text objexts.
    const XmlElement::ElementList& xtxts =
      pxtra->children(Text::xml_name());
    for ( XmlElement::ElementList::const_iterator iele=xtxts.begin();
          iele!=xtxts.end(); ++iele ) {
      Text txt(**iele);
      m_texts.push_back(txt);
    }
  } else {
    cerr << "TextDataset::ctor (promoter): Unable to find extra data"
         << endl;
  }
  // This might be a good place to check that the content ID's and
  // text names are consistent.
}

//**********************************************************************

// Merge.

int TextDataset::merge(const Dataset& dst, std::string) {
  if ( ! is_valid() ) {
    return 1;
  }
  if ( is_locked() ) {
    return 2;
  }
  if ( ! dst.is_valid() ) {
    return 3;
  }
  if ( ! dst.is_locked() ) {
    return 4;
  }
  // Check type of input dataset.
  if ( dst.fulltype() != fulltype() ) {
    return 5;
  }
  // Content label must match.
  string oldclab = content().front().name();
  string addclab = dst.content().front().name();
  if ( addclab == oldclab ) {
    return 6;
  }
  // Merge content ID lists.
  const ContentIdList& oldcids = content().front().content_ids();
  const ContentIdList& addcids = dst.content().front().content_ids();
  ContentIdList newcids;
  newcids.insert(oldcids.begin(), oldcids.end());
  newcids.insert(addcids.begin(), addcids.end());
  ContentIdList::size_type sumcids = oldcids.size() + addcids.size();
  // Content may not overlap.
  if ( newcids.size() != sumcids ) {
    return 7;
  }
  Content con("TextDataset", oldclab, newcids);
  int stat = reset_content(con);
  if ( stat != 0 ) {
    return 100 + stat;
  }
  return 0;
}

//**********************************************************************

// Create extra XML.

const XmlElement* TextDataset::create_extra_xml() const {
  if ( ! is_valid() ) return 0;
  // Create top XML elment.
  std::auto_ptr<XmlElement> pele(new XmlElement("ExtraData"));
  // Add text files.
  for ( TextList::const_iterator itxt=texts().begin();
        itxt!=texts().end(); ++itxt ) {
    XmlElement* pele_txt = itxt->xml();
    if ( pele_txt == 0 ) return 0;
    pele->add_child(pele_txt);
  }
  // Return the XML.
  return pele.release();
}


//**********************************************************************

// Web page.

Text TextDataset::
web_page(string baseurl, string repurl, string entry) const {
  Text wp;
  wp.append("<HTML>");
  // Display dataset.
  if ( entry == "" ) {
    string header = "TextDataset " + id().to_string();
    wp.append("<TITLE>" + header + "</TITLE>");
    //wp.append("<H3>" + header + "</H3>");
    wp.append("<pre>");
    ostringstream sout;
    ostr(sout, "", repurl);
    wp.append(sout.str());
    wp.append("</pre>");
    wp.append("<b>Text objects</b>:");
    for ( TextList::const_iterator itxt=texts().begin();
          itxt!=texts().end(); ++itxt ) {
      const Text& txt = *itxt;
      wp.append("<br>&nbsp;&nbsp;<a href=" + baseurl + "?text=" + 
                txt.name() + ">" + txt.name() + "</a>");
    }
  } else if ( entry.size() > 5 && entry.substr(0,5) == "text=" ) {
    string name = entry.substr(5);
    string header = "TextDataset " + id().to_string() + " text " + name;
    wp.append("<TITLE>" + header + "</TITLE>");
    //wp.append("<H3>" + header + "</H3>");
    TextList::const_iterator itxt = texts().begin();
    for ( ; itxt!=texts().end(); ++itxt ) {
      if ( itxt->name() == name ) break;
    }
    if ( itxt == texts().end() ) {
      wp.append("Dataset does not have text objeect named " + name);
    } else {
      wp.append("<pre>");
      ostringstream sout;
      sout << *itxt;
      wp.append(sout.str());
      wp.append("</pre>");
    }
  } else {
    wp.append("Dataset::web_page: Invalid entry " + entry);
  }
  wp.append("</HTML>");
  return wp;
}

//**********************************************************************