source: src/Parser/TremoloParser.cpp@ 4d4d33

/*
 * Project: MoleCuilder
 * Description: creates and alters molecular systems
 * Copyright (C)  2010 University of Bonn. All rights reserved.
 * Please see the LICENSE file or "Copyright notice" in builder.cpp for details.
 */

/*
 * TremoloParser.cpp
 *
 *  Created on: Mar 2, 2010
 *      Author: metzler
 */

// include config.h
#ifdef HAVE_CONFIG_H
#include <config.h>
#endif

#include "CodePatterns/MemDebug.hpp"

#include "CodePatterns/Assert.hpp"
#include "CodePatterns/Log.hpp"
#include "CodePatterns/toString.hpp"
#include "CodePatterns/Verbose.hpp"
#include "TremoloParser.hpp"
#include "World.hpp"
#include "atom.hpp"
#include "bond.hpp"
#include "element.hpp"
#include "molecule.hpp"
#include "periodentafel.hpp"
#include "Descriptors/AtomIdDescriptor.hpp"
#include <map>
#include <vector>

#include <boost/tokenizer.hpp>
#include <iostream>
#include <iomanip>

using namespace std;

/**
 * Constructor.
 */
TremoloParser::TremoloParser() {
  knownKeys["x"] = TremoloKey::x;
  knownKeys["u"] = TremoloKey::u;
  knownKeys["F"] = TremoloKey::F;
  knownKeys["stress"] = TremoloKey::stress;
  knownKeys["Id"] = TremoloKey::Id;
  knownKeys["neighbors"] = TremoloKey::neighbors;
  knownKeys["imprData"] = TremoloKey::imprData;
  knownKeys["GroupMeasureTypeNo"] = TremoloKey::GroupMeasureTypeNo;
  knownKeys["Type"] = TremoloKey::Type;
  knownKeys["extType"] = TremoloKey::extType;
  knownKeys["name"] = TremoloKey::name;
  knownKeys["resName"] = TremoloKey::resName;
  knownKeys["chainID"] = TremoloKey::chainID;
  knownKeys["resSeq"] = TremoloKey::resSeq;
  knownKeys["occupancy"] = TremoloKey::occupancy;
  knownKeys["tempFactor"] = TremoloKey::tempFactor;
  knownKeys["segID"] = TremoloKey::segID;
  knownKeys["Charge"] = TremoloKey::Charge;
  knownKeys["charge"] = TremoloKey::charge;
  knownKeys["GrpTypeNo"] = TremoloKey::GrpTypeNo;
  knownKeys["torsion"] = TremoloKey::torsion;

  createKnownTypesByIdentity();

  // default behavior: use all possible keys on output
  for (std::map<std::string, TremoloKey::atomDataKey>::iterator iter = knownKeys.begin(); iter != knownKeys.end(); ++iter)
    usedFields.push_back(iter->first);

  // and noKey afterwards(!) such that it is not used in usedFields
  knownKeys[" "] = TremoloKey::noKey; // with this we can detect invalid keys

  // invert knownKeys for debug output
  for (std::map<std::string, TremoloKey::atomDataKey>::iterator iter = knownKeys.begin(); iter != knownKeys.end(); ++iter)
    knownKeyNames.insert( make_pair( iter->second, iter->first) );

  additionalAtomData.clear();
}

/**
 * Destructor.
 */
TremoloParser::~TremoloParser() {
  usedFields.clear();
  additionalAtomData.clear();
  atomIdMap.clear();
  knownKeys.clear();
}

/**
 * Loads atoms from a tremolo-formatted file.
 *
 * \param tremolo file
 */
void TremoloParser::load(istream* file) {
  string line;
  string::size_type location;

  // reset atomIdMap, for we now get new serials
  atomIdMap.clear();
  usedFields.clear();

  molecule *newmol = World::getInstance().createMolecule();
  newmol->ActiveFlag = true;
  // TODO: Remove the insertion into molecule when saving does not depend on them anymore. Also, remove molecule.hpp include
  World::getInstance().getMolecules()->insert(newmol);
  while (file->good()) {
    std::getline(*file, line, '\n');
    if (usedFields.empty()) {
      location = line.find("ATOMDATA", 0);
      if (location != string::npos) {
       parseAtomDataKeysLine(line, location + 8);
      }
    }
    if (line.length() > 0 && line.at(0) != '#') {
      readAtomDataLine(line, newmol);
    }
  }

  processNeighborInformation();
  adaptImprData();
  adaptTorsion();
}

/**
 * Saves the \a atoms into as a tremolo file.
 *
 * \param file where to save the state
 * \param atoms atoms to store
 */
void TremoloParser::save(ostream* file, const std::vector<atom *> &AtomList) {
  DoLog(0) && (Log() << Verbose(0) << "Saving changes to tremolo." << std::endl);

  vector<atom*>::const_iterator atomIt;
  vector<string>::iterator it;

  *file << "# ATOMDATA";
  for (it=usedFields.begin(); it < usedFields.end(); it++) {
    *file << "\t" << *it;
  }
  *file << endl;
  for (atomIt = AtomList.begin(); atomIt != AtomList.end(); atomIt++) {
    saveLine(file, *atomIt);
  }
}

/**
 * Sets the keys for which data should be written to the stream when save is
 * called.
 *
 * \param string of field names with the same syntax as for an ATOMDATA line
 *        but without the prexix "ATOMDATA"
 */
void TremoloParser::setFieldsForSave(std::string atomDataLine) {
  parseAtomDataKeysLine(atomDataLine, 0);
}


/**
 * Writes one line of tremolo-formatted data to the provided stream.
 *
 * \param stream where to write the line to
 * \param reference to the atom of which information should be written
 */
void TremoloParser::saveLine(ostream* file, atom* currentAtom) {
  vector<string>::iterator it;
  TremoloKey::atomDataKey currentField;

  DoLog(3) && (Log() << Verbose(3) << "additionalAtomData contains: " << additionalAtomData << std::endl);

  for (it = usedFields.begin(); it != usedFields.end(); it++) {
    currentField = knownKeys[it->substr(0, it->find("="))];
    switch (currentField) {
      case TremoloKey::x :
        // for the moment, assume there are always three dimensions
        DoLog(3) && (Log() << Verbose(3) << "Writing for type " << knownKeyNames[currentField] << ": " << currentAtom->getPosition() << std::endl);
        *file << currentAtom->at(0) << "\t";
        *file << currentAtom->at(1) << "\t";
        *file << currentAtom->at(2) << "\t";
        break;
      case TremoloKey::u :
        // for the moment, assume there are always three dimensions
        DoLog(3) && (Log() << Verbose(3) << "Writing for type " << knownKeyNames[currentField] << ": " << currentAtom->AtomicVelocity << std::endl);
        *file << currentAtom->AtomicVelocity[0] << "\t";
        *file << currentAtom->AtomicVelocity[1] << "\t";
        *file << currentAtom->AtomicVelocity[2] << "\t";
        break;
      case TremoloKey::Type :
        if ((!additionalAtomData.count(currentAtom->getId())) || (additionalAtomData[currentAtom->getId()].get(currentField) == "-")) {
          DoLog(3) && (Log() << Verbose(3) << "Writing for type " << knownKeyNames[currentField] << " default value: " << currentAtom->getType()->getSymbol() << std::endl);
          *file << currentAtom->getType()->getSymbol() << "\t";
        } else {
          DoLog(3) && (Log() << Verbose(3) << "Writing for type " << knownKeyNames[currentField] << ": " << additionalAtomData[currentAtom->getId()].get(currentField) << std::endl);
          *file << additionalAtomData[currentAtom->getId()].get(currentField) << "\t";
        }
        break;
      case TremoloKey::Id :
        DoLog(3) && (Log() << Verbose(3) << "Writing for type " << knownKeyNames[currentField] << ": " << currentAtom->getId()+1 << std::endl);
        *file << currentAtom->getId()+1 << "\t";
        break;
      case TremoloKey::neighbors :
        DoLog(3) && (Log() << Verbose(3) << "Writing type " << knownKeyNames[currentField] << std::endl);
        writeNeighbors(file, atoi(it->substr(it->find("=") + 1, 1).c_str()), currentAtom);
        break;
      case TremoloKey::resSeq :
        if (additionalAtomData.count(currentAtom->getId())) {
          DoLog(3) && (Log() << Verbose(3) << "Writing for type " << knownKeyNames[currentField] << ": " << additionalAtomData[currentAtom->getId()].get(currentField) << std::endl);
          *file << additionalAtomData[currentAtom->getId()].get(currentField);
        } else if (currentAtom->getMolecule() != NULL) {
          DoLog(3) && (Log() << Verbose(3) << "Writing for type " << knownKeyNames[currentField] << " its own id: " << currentAtom->getMolecule()->getId()+1 << std::endl);
          *file << setw(4) << currentAtom->getMolecule()->getId()+1;
        } else {
          DoLog(3) && (Log() << Verbose(3) << "Writing for type " << knownKeyNames[currentField] << " default value: " << defaultAdditionalData.get(currentField) << std::endl);
          *file << defaultAdditionalData.get(currentField);
        }
        *file << "\t";
        break;
      default :
        if (additionalAtomData.count(currentAtom->getId())) {
          DoLog(3) && (Log() << Verbose(3) << "Writing for type " << knownKeyNames[currentField] << ": " << additionalAtomData[currentAtom->getId()].get(currentField) << std::endl);
          *file << additionalAtomData[currentAtom->getId()].get(currentField);
        } else if (additionalAtomData.count(currentAtom->GetTrueFather()->getId())) {
          DoLog(3) && (Log() << Verbose(3) << "Writing for type " << knownKeyNames[currentField] << " stuff from father: " << additionalAtomData[currentAtom->GetTrueFather()->getId()].get(currentField) << std::endl);
          *file << additionalAtomData[currentAtom->GetTrueFather()->getId()].get(currentField);
        } else {
          DoLog(3) && (Log() << Verbose(3) << "Writing for type " << knownKeyNames[currentField] << " the default: " << defaultAdditionalData.get(currentField) << std::endl);
          *file << defaultAdditionalData.get(currentField);
        }
        *file << "\t";
        break;
    }
  }

  *file << endl;
}

/**
 * Writes the neighbor information of one atom to the provided stream.
 *
 * \param stream where to write neighbor information to
 * \param number of neighbors
 * \param reference to the atom of which to take the neighbor information
 */
void TremoloParser::writeNeighbors(ostream* file, int numberOfNeighbors, atom* currentAtom) {
  BondList::iterator currentBond = currentAtom->ListOfBonds.begin();
  for (int i = 0; i < numberOfNeighbors; i++) {
    *file << (currentBond != currentAtom->ListOfBonds.end()
        ? (*currentBond)->GetOtherAtom(currentAtom)->getId()+1 : 0) << "\t";
    if (currentBond != currentAtom->ListOfBonds.end())
      ++currentBond;
  }
}

/**
 * Stores keys from the ATOMDATA line.
 *
 * \param line to parse the keys from
 * \param with which offset the keys begin within the line
 */
void TremoloParser::parseAtomDataKeysLine(string line, int offset) {
  string keyword;
  stringstream lineStream;

  lineStream << line.substr(offset);
  usedFields.clear();
  while (lineStream.good()) {
    lineStream >> keyword;
    if (knownKeys[keyword.substr(0, keyword.find("="))] == TremoloKey::noKey) {
      // TODO: throw exception about unknown key
      cout << "Unknown key: " << keyword << " is not part of the tremolo format specification." << endl;
      break;
    }
    usedFields.push_back(keyword);
  }
  //DoLog(1) && (Log() << Verbose(1) << "INFO: " << usedFields << std::endl);
}

/**
 * Reads one data line of a tremolo file and interprets it according to the keys
 * obtained from the ATOMDATA line.
 *
 * \param line to parse as an atom
 * \param *newmol molecule to add atom to
 */
void TremoloParser::readAtomDataLine(string line, molecule *newmol = NULL) {
  vector<string>::iterator it;
  stringstream lineStream;
  atom* newAtom = World::getInstance().createAtom();
  TremoloAtomInfoContainer *atomInfo = NULL;
  additionalAtomData[newAtom->getId()] = TremoloAtomInfoContainer(); // fill with default values
  atomInfo = &additionalAtomData[newAtom->getId()];
  TremoloKey::atomDataKey currentField;
  ConvertTo<double> toDouble;
  ConvertTo<int> toInt;

  // setup tokenizer, splitting up white-spaced entries
  typedef boost::tokenizer<boost::char_separator<char> >
      tokenizer;
  boost::char_separator<char> whitespacesep(" \t");
  tokenizer tokens(line, whitespacesep);
  ASSERT(tokens.begin() != tokens.end(),
      "TremoloParser::readAtomDataLine - empty string, need at least ' '!");
  tokenizer::iterator tok_iter = tokens.begin();
  // then associate each token to each file
  for (it = usedFields.begin(); it < usedFields.end(); it++) {
    const std::string keyName = it->substr(0, it->find("="));
    currentField = knownKeys[keyName];
    const string word = *tok_iter;
    //DoLog(1) && (Log() << Verbose(1) << "INFO: Parsing key " << keyName << " with remaining data " << word << std::endl);
    switch (currentField) {
      case TremoloKey::x :
        // for the moment, assume there are always three dimensions
        for (int i=0;i<NDIM;i++) {
          ASSERT(tok_iter != tokens.end(), "TremoloParser::readAtomDataLine() - no value for x["+toString(i)+"]!");
          //DoLog(1) && (Log() << Verbose(1) << "INFO: Parsing key " << keyName << " with next token " << *tok_iter << std::endl);
          newAtom->set(i, toDouble(*tok_iter));
          tok_iter++;
        }
        break;
      case TremoloKey::u :
        // for the moment, assume there are always three dimensions
        for (int i=0;i<NDIM;i++) {
          ASSERT(tok_iter != tokens.end(), "TremoloParser::readAtomDataLine() - no value for u["+toString(i)+"]!");
          //DoLog(1) && (Log() << Verbose(1) << "INFO: Parsing key " << keyName << " with next token " << *tok_iter << std::endl);
          newAtom->AtomicVelocity[i] = toDouble(*tok_iter);
          tok_iter++;
        }
        break;
      case TremoloKey::Type :
      {
        ASSERT(tok_iter != tokens.end(), "TremoloParser::readAtomDataLine() - no value for "+keyName+"!");
        DoLog(1) && (Log() << Verbose(1) << "INFO: Parsing key " << keyName << " with next token " << *tok_iter << std::endl);
        std::string element(knownTypes[(*tok_iter)]);
        // put type name into container for later use
        atomInfo->set(currentField, *tok_iter);
        DoLog(1) && (Log() << Verbose(1) << "INFO: Parsing element " << (*tok_iter) << " as " << element << " according to KnownTypes." << std::endl);
        tok_iter++;
        newAtom->setType(World::getInstance().getPeriode()->FindElement(element));
        ASSERT(newAtom->getType(), "Type was not set for this atom");
        break;
      }
      case TremoloKey::Id :
        ASSERT(tok_iter != tokens.end(), "TremoloParser::readAtomDataLine() - no value for "+keyName+"!");
        //DoLog(1) && (Log() << Verbose(1) << "INFO: Parsing key " << keyName << " with next token " << *tok_iter << std::endl);
        atomIdMap[toInt(*tok_iter)] = newAtom->getId();
        tok_iter++;
        break;
      case TremoloKey::neighbors :
        for (int i=0;i<atoi(it->substr(it->find("=") + 1, 1).c_str());i++) {
          ASSERT(tok_iter != tokens.end(), "TremoloParser::readAtomDataLine() - no value for "+keyName+"!");
          //DoLog(1) && (Log() << Verbose(1) << "INFO: Parsing key " << keyName << " with next token " << *tok_iter << std::endl);
          lineStream << *tok_iter << "\t";
          tok_iter++;
        }
        readNeighbors(&lineStream,
            atoi(it->substr(it->find("=") + 1, 1).c_str()), newAtom->getId());
        break;
      default :
        ASSERT(tok_iter != tokens.end(), "TremoloParser::readAtomDataLine() - no value for "+keyName+"!");
        //DoLog(1) && (Log() << Verbose(1) << "INFO: Parsing key " << keyName << " with next token " << *tok_iter << std::endl);
        atomInfo->set(currentField, *tok_iter);
        tok_iter++;
        break;
    }
  }
  if (newmol != NULL) {
    //DoLog(0) && (Log() << Verbose(0) << "New Atom: " << *newAtom << " with type " << newAtom->getType()->getName() << std::endl);
    newmol->AddAtom(newAtom);
  }
}

/**
 * Reads neighbor information for one atom from the input.
 *
 * \param line stream where to read the information from
 * \param numberOfNeighbors number of neighbors to read
 * \param atomid world id of the atom the information belongs to
 */
void TremoloParser::readNeighbors(stringstream* line, int numberOfNeighbors, int atomId) {
  int neighborId = 0;
  for (int i = 0; i < numberOfNeighbors; i++) {
    *line >> neighborId;
    // 0 is used to fill empty neighbor positions in the tremolo file.
    if (neighborId > 0) {
//      DoLog(1) && (Log() << Verbose(1)
//          << "Atom with global id " << atomId
//          << " has neighbour with serial " << neighborId
//          << std::endl);
      additionalAtomData[atomId].neighbors.push_back(neighborId);
    }
  }
}

/**
 * Checks whether the provided name is within the list of used fields.
 *
 * \param field name to check
 *
 * \return true if the field name is used
 */
bool TremoloParser::isUsedField(string fieldName) {
  bool fieldNameExists = false;
  for (vector<string>::iterator usedField = usedFields.begin(); usedField != usedFields.end(); usedField++) {
    if (usedField->substr(0, usedField->find("=")) == fieldName)
      fieldNameExists = true;
  }

  return fieldNameExists;
}


/**
 * Adds the collected neighbor information to the atoms in the world. The atoms
 * are found by their current ID and mapped to the corresponding atoms with the
 * Id found in the parsed file.
 */
void TremoloParser::processNeighborInformation() {
  if (!isUsedField("neighbors")) {
    return;
  }

  for(map<int, TremoloAtomInfoContainer>::iterator currentInfo = additionalAtomData.begin();
    currentInfo != additionalAtomData.end(); currentInfo++
  ) {
    if (!currentInfo->second.neighbors_processed) {
      for(vector<int>::iterator neighbor = currentInfo->second.neighbors.begin();
        neighbor != currentInfo->second.neighbors.end(); neighbor++
      ) {
//        DoLog(1) && (Log() << Verbose(1) << "Creating bond between ("
//            << currentInfo->first
//            << ") and ("
//            << atomIdMap[*neighbor] << "|" << *neighbor << ")" << std::endl);
        World::getInstance().getAtom(AtomById(currentInfo->first))
            ->addBond(World::getInstance().getAtom(AtomById(atomIdMap[*neighbor])));
      }
      currentInfo->second.neighbors_processed = true;
    }
  }
}

/**
 * Replaces atom IDs read from the file by the corresponding world IDs. All IDs
 * IDs of the input string will be replaced; expected separating characters are
 * "-" and ",".
 *
 * \param string in which atom IDs should be adapted
 *
 * \return input string with modified atom IDs
 */
string TremoloParser::adaptIdDependentDataString(string data) {
  // there might be no IDs
  if (data == "-") {
    return "-";
  }

  char separator;
  int id;
  stringstream line, result;
  line << data;

  line >> id;
  result << atomIdMap[id];
  while (line.good()) {
    line >> separator >> id;
    result << separator << atomIdMap[id];
  }

  return result.str();
}

/**
 * Corrects the atom IDs in each imprData entry to the corresponding world IDs
 * as they might differ from the originally read IDs.
 */
void TremoloParser::adaptImprData() {
  if (!isUsedField("imprData")) {
    return;
  }

  for(map<int, TremoloAtomInfoContainer>::iterator currentInfo = additionalAtomData.begin();
    currentInfo != additionalAtomData.end(); currentInfo++
  ) {
    currentInfo->second.imprData = adaptIdDependentDataString(currentInfo->second.imprData);
  }
}

/**
 * Corrects the atom IDs in each torsion entry to the corresponding world IDs
 * as they might differ from the originally read IDs.
 */
void TremoloParser::adaptTorsion() {
  if (!isUsedField("torsion")) {
    return;
  }

  for(map<int, TremoloAtomInfoContainer>::iterator currentInfo = additionalAtomData.begin();
    currentInfo != additionalAtomData.end(); currentInfo++
  ) {
    currentInfo->second.torsion = adaptIdDependentDataString(currentInfo->second.torsion);
  }
}

/** Creates knownTypes as the identity, e.g. H -> H, He -> He, ... .
 *
 */
void TremoloParser::createKnownTypesByIdentity()
{
  // remove old mapping
  knownTypes.clear();
  // make knownTypes the identity mapping
  const periodentafel *periode = World::getInstance().getPeriode();
  for (periodentafel::const_iterator iter = periode->begin();
      iter != periode->end();
      ++iter) {
    knownTypes.insert( make_pair(iter->second->getSymbol(), iter->second->getSymbol()) );
  }
}

/** Parses a .potentials file and creates from it the knownTypes file.
 *
 * @param file input stream of .potentials file
 */
void TremoloParser::parseKnownTypes(std::istream &file)
{
  const periodentafel *periode = World::getInstance().getPeriode();
  // remove old mapping
  knownTypes.clear();

  DoLog(3) && (Log() << Verbose(3) << "additionalAtomData contains: " << additionalAtomData << std::endl);

  // parse in file
  typedef boost::tokenizer<boost::char_separator<char> >
      tokenizer;
  boost::char_separator<char> tokensep(":\t ,;");
  boost::char_separator<char> equalitysep("\t =");
  std::string line;
  while (file.good()) {
    std::getline( file, line );
    DoLog(4) && (Log() << Verbose(4) << "INFO: full line of parameters is '" << line << "'" << std::endl);
    if (line.find("particle:") != string::npos) {
      DoLog(3) && (Log() << Verbose(3) << "INFO: found line '" << line << "' containing keyword 'particle:'." << std::endl);
      tokenizer tokens(line, tokensep);
      ASSERT(tokens.begin() != tokens.end(),
          "TremoloParser::parseKnownTypes() - line with 'particle:' but no particles separated by comma.");
      // look for particle_type
      std::string particle_type("NULL");
      std::string element_type("NULL");
      for (tokenizer::iterator tok_iter = tokens.begin();
          tok_iter != tokens.end();
          ++tok_iter) {
        if ((*tok_iter).find("particle_type") != string::npos) {
          DoLog(3) && (Log() << Verbose(3) << "INFO: found line '" << line << "' containing keyword 'particle_type'." << std::endl);
          tokenizer token((*tok_iter), equalitysep);
          ASSERT(token.begin() != token.end(),
                    "TremoloParser::parseKnownTypes() - could not split particle_type by equality sign");
          tokenizer::iterator particle_iter = token.begin();
          particle_iter++;
          particle_type = *particle_iter;
        }
        if ((*tok_iter).find("element_name") != string::npos) {
          DoLog(3) && (Log() << Verbose(3) << "INFO: found line '" << line << "' containing keyword 'element_name'." << std::endl);
          tokenizer token((*tok_iter), equalitysep);
          ASSERT(token.begin() != token.end(),
                    "TremoloParser::parseKnownTypes() - could not split particle_type by equality sign");
          tokenizer::iterator element_iter = token.begin();
          element_iter++;
          element_type = *element_iter;
        }
      }
      if ((particle_type != "NULL") && (element_type != "NULL")) {
        if (periode->FindElement(element_type) != NULL) {
          DoLog(1) && (Log() << Verbose(1) << "INFO: Added Type " << particle_type << " as reference to element " << element_type << "." << std::endl);
          knownTypes.insert( make_pair (particle_type, element_type) );
        } else {
          DoeLog(1) && (Log() << Verbose(1) << "INFO: Either Type " << particle_type << " or " << element_type << " could not be recognized." << std::endl);
        }
      } else {
        DoeLog(1) && (Log() << Verbose(1) << "INFO: Desired element " << element_type << " is not known." << std::endl);
      }
    }
  }

}