XMLHandler.h

Go to the documentation of this file.

// Copyright (c) 2002-present, OpenMS Inc. -- EKU Tuebingen, ETH Zurich, and FU Berlin
// SPDX-License-Identifier: BSD-3-Clause
//
// --------------------------------------------------------------------------
// $Maintainer: Chris Bielow $
// $Authors: Marc Sturm, Chris Bielow $
// --------------------------------------------------------------------------
 
#pragma once
 
#include <OpenMS/CONCEPT/Types.h>
#include <OpenMS/CONCEPT/Macros.h>
 
 
#include <OpenMS/DATASTRUCTURES/DateTime.h>
#include <OpenMS/DATASTRUCTURES/DataValue.h>
#include <OpenMS/DATASTRUCTURES/ListUtils.h>
 
#include <xercesc/sax2/Attributes.hpp>
#include <xercesc/sax2/DefaultHandler.hpp>
#include <xercesc/util/XMLString.hpp>
 
#include <iosfwd>
#include <string>
#include <memory>
 
 
namespace OpenMS
{
  class ControlledVocabulary;
  class CVTerm;
  class MetaInfoInterface;
  class ProteinIdentification;
 
  namespace Internal
  {
 
    #define CONST_XMLCH(s) reinterpret_cast<const ::XMLCh*>(u ## s)
 
    static_assert(sizeof(::XMLCh) == sizeof(char16_t),
                  "XMLCh is not sized correctly for UTF-16.");
 
    //Adapted from https://www.codeproject.com/articles/99551/redux-raii-adapter-for-xerces
    //Copyright 2010 Orjan Westin
    //Under BSD license
    //========================================================================================================
    template<typename T>
    class OPENMS_DLLAPI shared_xerces_ptr
    {
      // Function to release Xerces data type with a release member function
      template<typename U>
      static void doRelease_(U* item)
      {
        // Only release this if it has no owner
        if (nullptr == item->getOwnerDocument())
          item->release();
      }
 
      static void doRelease_(char* item);
      static void doRelease_(XMLCh* item);
 
      // The actual data we're holding
      std::shared_ptr<T> item_;
    public:
      // Default constructor
      shared_xerces_ptr() = default;
      // Assignment constructor
      shared_xerces_ptr(T* item)
          : item_(item, doRelease_ )
      {}
      // Assignment of data to guard
      shared_xerces_ptr& operator=(T* item)
      {
        assign(item);
        return *this;
      }
      // Give up hold on data
      void reset()
      {
        item_.reset();
      }
      // Release currently held data, if any, to hold another
      void assign(T* item)
      {
        item_.reset(item, doRelease_ );
      }
      // Get pointer to the currently held data, if any
      T* get()
      {
        return item_.get();
      }
      const T* get() const
      {
        return item_.get();
      }
      // Return true if no data is held
      bool is_released() const
      {
        return (nullptr == item_.get());
      }
    };
 
    template <typename T>
    class OPENMS_DLLAPI unique_xerces_ptr
    {
    private:
 
      template<typename U>
      static void doRelease_(U*& item)
      {
        // Only release this if it has no parent (otherwise
        // parent will release it)
        if (nullptr == item->getOwnerDocument())
          item->release();
      }
 
      static void doRelease_(char*& item);
      static void doRelease_(XMLCh*& item);
 
      T* item_;
 
    public:
 
      // Hide copy constructor and assignment operator
      unique_xerces_ptr(const unique_xerces_ptr<T>&) = delete;
      unique_xerces_ptr& operator=(const unique_xerces_ptr<T>&) = delete;
 
      unique_xerces_ptr()
          : item_(nullptr)
      {}
 
      explicit unique_xerces_ptr(T* i)
          : item_(i)
      {}
 
      ~unique_xerces_ptr()
      {
        xerces_release();
      }
 
      unique_xerces_ptr(unique_xerces_ptr<T>&& other) noexcept
          : item_(nullptr)
      {
        this->swap(other);
      }
 
      void swap(unique_xerces_ptr<T>& other) noexcept
      {
        std::swap(item_, other.item_);
      }
 
      // Assignment of data to guard (not chainable)
      void operator=(T* i)
      {
        reassign(i);
      }
 
      // Release held data (i.e. delete/free it)
      void xerces_release()
      {
        if (!is_released())
        {
          // Use type-specific release mechanism
          doRelease_(item_);
          item_ = nullptr;
        }
      }
 
      // Give up held data (i.e. return data without releasing)
      T* yield()
      {
        T* tempItem = item_;
        item_ = nullptr;
        return tempItem;
      }
 
      // Release currently held data, if any, to hold another
      void assign(T* i)
      {
        xerces_release();
        item_ = i;
      }
 
      // Get pointer to the currently held data, if any
      T* get() const
      {
        return item_;
      }
 
      // Return true if no data is held
      bool is_released() const
      {
        return (nullptr == item_);
      }
    };
 
    //========================================================================================================
 
    /*
     * @brief Helper class for XML parsing that handles the conversions of Xerces strings
     *
     * It provides the convert() function which internally calls
     * XMLString::transcode and ensures that the memory is released properly
     * through XMLString::release internally. It returns a std::string or
     * std::basic_string<XMLCh> to the caller who takes ownership of the data.
     *
    */
    class OPENMS_DLLAPI StringManager
    {
 
      typedef std::basic_string<XMLCh> XercesString;
 
      inline static unique_xerces_ptr<XMLCh> fromNative_(const char* str)
      {
        return unique_xerces_ptr<XMLCh>(xercesc::XMLString::transcode(str));
      }
 
      inline static unique_xerces_ptr<XMLCh> fromNative_(const std::string& str)
      {
        return fromNative_(str.c_str());
      }
 
      inline static std::string toNative_(const XMLCh* str)
      { 
        std::string r;
        XMLSize_t l = strLength(str);
        if(isASCII(str, l))
        {
          appendASCII(str,l,r);
        }
        else
        {
          r = (unique_xerces_ptr<char>(xercesc::XMLString::transcode(str)).get());
        }
        return r;
      }
 
      inline static std::string toNative_(const unique_xerces_ptr<XMLCh>& str)
      {
        return toNative_(str.get());
      }
 
protected:
      static void compress64_ (const XMLCh * input_it, char* output_it);
 
public:
      StringManager();
 
      ~StringManager();
 
      // https://github.com/OpenMS/OpenMS/issues/8122
      #if defined(__GNUC__)
      __attribute__((no_sanitize("address")))
      #elif defined(_MSC_VER)
      __declspec(no_sanitize_address) 
      #endif
      static XMLSize_t strLength(const XMLCh* input_ptr);
 
      inline static XercesString convert(const char * str)
      {
        return fromNative_(str).get();
      }
 
      inline static XercesString convert(const std::string & str)
      {
        return fromNative_(str.c_str()).get();
      }
 
      inline static unique_xerces_ptr<XMLCh> convertPtr(const char * str)
      {
        return fromNative_(str);
      }
 
      inline static unique_xerces_ptr<XMLCh> convertPtr(const std::string & str)
      {
        return fromNative_(str.c_str());
      }
 
      inline static std::string convert(const XMLCh * str)
      {
        return toNative_(str);
      }
      static bool isASCII(const XMLCh * chars, const XMLSize_t length);
 
      
      
      static void appendASCII(const XMLCh * str, const XMLSize_t length, std::string & result);
 
    };
 
    class OPENMS_DLLAPI XMLHandler :
      public xercesc::DefaultHandler
    {
public:
 
      class OPENMS_DLLAPI EndParsingSoftly :
        public Exception::BaseException
      {
      public:
        EndParsingSoftly(const char * file, int line, const char * function) :
          Exception::BaseException(file, line, function)
        {
        }
 
      };
 
      enum ActionMode
      {
        LOAD,               
        STORE               
      };
 
      enum LOADDETAIL 
      {  
        LD_ALLDATA,       // default; load all data
        LD_RAWCOUNTS,     // only count the total number of spectra and chromatograms (usually very fast)
        LD_COUNTS_WITHOPTIONS // count the number of spectra, while respecting PeakFileOptions (msLevel and RTRange) and chromatograms (fast)
      };
 
 
      XMLHandler(const std::string & filename, const std::string & version);
      ~XMLHandler() override;
 
      void reset();
 
 
      void fatalError(const xercesc::SAXParseException & exception) override;
      void error(const xercesc::SAXParseException & exception) override;
      void warning(const xercesc::SAXParseException & exception) override;
 
      void fatalError(ActionMode mode, const std::string & msg, UInt line = 0, UInt column = 0) const;
      void error(ActionMode mode, const std::string & msg, UInt line = 0, UInt column = 0) const;
      void warning(ActionMode mode, const std::string & msg, UInt line = 0, UInt column = 0) const;
 
      void characters(const XMLCh * const chars, const XMLSize_t length) override;
      void startElement(const XMLCh * const uri, const XMLCh * const localname, const XMLCh * const qname, const xercesc::Attributes & attrs) override;
      void endElement(const XMLCh * const uri, const XMLCh * const localname, const XMLCh * const qname) override;
 
      virtual void writeTo(std::ostream & /*os*/);
 
      virtual LOADDETAIL getLoadDetail() const;
 
      virtual void setLoadDetail(const LOADDETAIL d);
 
      static std::string writeXMLEscape(const std::string& to_escape)
      {
        std::string _copy = to_escape;
        // has() is cheap, so check before calling substitute(), since substitute() will usually happen rarely
        if (StringUtils::has(_copy, '&')) StringUtils::substitute(_copy, "&","&amp;");
        if (StringUtils::has(_copy, '>')) StringUtils::substitute(_copy, ">","&gt;");
        if (StringUtils::has(_copy, '"')) StringUtils::substitute(_copy, "\"","&quot;");
        if (StringUtils::has(_copy, '<')) StringUtils::substitute(_copy, "<","&lt;");
        if (StringUtils::has(_copy, '\'')) StringUtils::substitute(_copy, "'","&apos;");
 
        return _copy;
      }
 
      static DataValue fromXSDString(const std::string& type, const std::string& value)
      {
        DataValue data_value;
        // float type
        if (type == "xsd:double" || type == "xsd:float" || type == "xsd:decimal")
        {
          data_value = DataValue(StringUtils::toDouble(value));
        }
        // <=32 bit integer types
        else if (type == "xsd:byte" ||          // 8bit signed
                 type == "xsd:int" ||           // 32bit signed
                 type == "xsd:unsignedShort" || // 16bit unsigned
                 type == "xsd:short" ||         // 16bit signed
                 type == "xsd:unsignedByte" || type == "xsd:unsignedInt")
        {
          data_value = DataValue(StringUtils::toInt32(value));
        }
        // 64 bit integer types
        else if (type == "xsd:long" || type == "xsd:unsignedLong" ||       // 64bit signed or unsigned respectively
                 type == "xsd:integer" || type == "xsd:negativeInteger" || // any 'integer' has arbitrary size... but we have to cope with 64bit for now.
                 type == "xsd:nonNegativeInteger" || type == "xsd:nonPositiveInteger" || type == "xsd:positiveInteger")
        {
          data_value = DataValue(StringUtils::toInt64(value)); // internally a signed 64-bit integer. So if someone uses 2^64-1 as value, toInt64() will raise an exception...
        }
        // everything else is treated as a string
        else
        {
          data_value = DataValue(value);
        }
        return data_value;
      }
 
 
      DataValue cvParamToValue(const ControlledVocabulary& cv, const std::string& parent_tag, 
                               const std::string& accession, const std::string& name, const std::string& value,
                               const std::string& unit_accession) const;
 
      DataValue cvParamToValue(const ControlledVocabulary& cv, const CVTerm& raw_term) const;
 
      void checkUniqueIdentifiers_(const std::vector<ProteinIdentification>& prot_ids) const;
 
protected:
      std::string file_;
 
      std::string version_;
 
      StringManager sm_;
 
      std::vector<std::string> open_tags_;
 
      LOADDETAIL load_detail_; 
 
 
      inline bool equal_(const XMLCh * a, const XMLCh * b) const
      {
        return xercesc::XMLString::compareString(a, b) == 0;
      }
 
 
 
      void writeUserParam_(const std::string & tag_name, std::ostream & os, const MetaInfoInterface & meta, UInt indent) const;
 
 
 
 
      std::vector<std::vector<std::string> > cv_terms_;
 
      SignedSize cvStringToEnum_(const Size section, const std::string & term, const char * message, const SignedSize result_on_error = 0);
 
 
 
 
      inline Int asInt_(const std::string & in) const
      {
        Int res = 0;
        try
        {
          res = StringUtils::toInt32(in);
        }
        catch (Exception::ConversionError&)
        {
          error(LOAD,std::string("Int conversion error of \"") + in + "\"");
        }
        return res;
      }
 
      inline Int asInt_(const XMLCh * in) const
      {
        return xercesc::XMLString::parseInt(in);
      }
 
      inline UInt asUInt_(const std::string & in) const
      {
        UInt res = 0;
        try
        {
          Int tmp = StringUtils::toInt32(in);
          if (tmp < 0)
          {
            throw Exception::ConversionError(__FILE__, __LINE__, OPENMS_PRETTY_FUNCTION, "");
          }
          res = UInt(tmp);
        }
        catch (Exception::ConversionError& )
        {
          error(LOAD,std::string("UInt conversion error of \"") + in + "\"");
        }
        return res;
      }
 
      inline double asDouble_(const std::string & in) const
      {
        double res = 0.0;
        try
        {
          res = StringUtils::toDouble(in);
        }
        catch (Exception::ConversionError& )
        {
          error(LOAD,std::string("Double conversion error of \"") + in + "\"");
        }
        return res;
      }
 
      inline float asFloat_(const std::string & in) const
      {
        float res = 0.0;
        try
        {
          res = StringUtils::toFloat(in);
        }
        catch (Exception::ConversionError& )
        {
          error(LOAD,std::string("Float conversion error of \"") + in + "\"");
        }
        return res;
      }
 
      inline bool asBool_(const std::string & in) const
      {
        if (in == "true" || in == "TRUE" || in == "True" || in == "1")
        {
          return true;
        }
        else if (in == "false" || in == "FALSE" || in == "False" || in == "0")
        {
          return false;
        }
        else
        {
          error(LOAD,std::string("Boolean conversion error of \"") + in + "\"");
        }
        return false;
      }
 
      inline DateTime asDateTime_(std::string date_string) const
      {
        DateTime date_time;
        if (!date_string.empty())
        {
          try
          {
            //strip away milliseconds
            StringUtils::trim(date_string);
            date_string = StringUtils::substr(date_string, 0, 19);
            date_time.set(date_string);
          }
          catch (Exception::ParseError& /*err*/ )
          {
            error(LOAD,std::string("DateTime conversion error of \"") + date_string + "\"");
          }
        }
        return date_time;
      }
 
 
 
 
      inline std::string attributeAsString_(const xercesc::Attributes & a, const char * name) const
      {
        const XMLCh * val = a.getValue(sm_.convertPtr(name).get());
        if (val == nullptr) fatalError(LOAD,std::string("Required attribute '") + name + "' not present!");
        return sm_.convert(val);
      }
 
      inline Int attributeAsInt_(const xercesc::Attributes & a, const char * name) const
      {
        const XMLCh * val = a.getValue(sm_.convertPtr(name).get());
        if (val == nullptr) fatalError(LOAD,std::string("Required attribute '") + name + "' not present!");
        return xercesc::XMLString::parseInt(val);
      }
 
      inline double attributeAsDouble_(const xercesc::Attributes & a, const char * name) const
      {
        const XMLCh * val = a.getValue(sm_.convertPtr(name).get());
        if (val == nullptr) fatalError(LOAD,std::string("Required attribute '") + name + "' not present!");
        return StringUtils::toDouble(sm_.convert(val));
      }
 
      inline DoubleList attributeAsDoubleList_(const xercesc::Attributes & a, const char * name) const
      {
        std::string tmp(expectList_(attributeAsString_(a, name)));
        return ListUtils::create<double>(StringUtils::substr(tmp, 1, tmp.size() - 2));
      }
 
      inline IntList attributeAsIntList_(const xercesc::Attributes & a, const char * name) const
      {
        std::string tmp(expectList_(attributeAsString_(a, name)));
        return ListUtils::create<Int>(StringUtils::substr(tmp, 1, tmp.size() - 2));
      }
 
      inline StringList attributeAsStringList_(const xercesc::Attributes & a, const char * name) const
      {
        std::string tmp(expectList_(attributeAsString_(a, name)));         
        StringList tmp_list = ListUtils::create<std::string>(StringUtils::substr(tmp, 1, tmp.size() - 2)); // between [ and ]
  
        if (StringUtils::hasSubstring(tmp, "\\|")) // check full string for escaped comma
        {
          for (std::string& s : tmp_list)
          {
            StringUtils::substitute(s, "\\|", ",");
          }          
        }
        return tmp_list;
      }
 
      inline bool optionalAttributeAsString_(std::string & value, const xercesc::Attributes & a, const char * name) const
      {
        const XMLCh * val = a.getValue(sm_.convertPtr(name).get());
        if (val != nullptr)
        {
          value = sm_.convert(val);
          return true;
        }
        return false;
      }
 
      inline bool optionalAttributeAsInt_(Int & value, const xercesc::Attributes & a, const char * name) const
      {
        const XMLCh * val = a.getValue(sm_.convertPtr(name).get());
        if (val != nullptr)
        {
          value = xercesc::XMLString::parseInt(val);
          return true;
        }
        return false;
      }
 
      inline bool optionalAttributeAsUInt_(UInt & value, const xercesc::Attributes & a, const char * name) const
      {
        const XMLCh * val = a.getValue(sm_.convertPtr(name).get());
        if (val != nullptr)
        {
          value = xercesc::XMLString::parseInt(val);
          return true;
        }
        return false;
      }
 
      inline bool optionalAttributeAsDouble_(double & value, const xercesc::Attributes & a, const char * name) const
      {
        const XMLCh * val = a.getValue(sm_.convertPtr(name).get());
        if (val != nullptr)
        {
          value =StringUtils::toDouble(sm_.convert(val));
          return true;
        }
        return false;
      }
 
      inline bool optionalAttributeAsDoubleList_(DoubleList & value, const xercesc::Attributes & a, const char * name) const
      {
        const XMLCh * val = a.getValue(sm_.convertPtr(name).get());
        if (val != nullptr)
        {
          value = attributeAsDoubleList_(a, name);
          return true;
        }
        return false;
      }
 
      inline bool optionalAttributeAsStringList_(StringList & value, const xercesc::Attributes & a, const char * name) const
      {
        const XMLCh * val = a.getValue(sm_.convertPtr(name).get());
        if (val != nullptr)
        {
          value = attributeAsStringList_(a, name);
          return true;
        }
        return false;
      }
 
      inline bool optionalAttributeAsIntList_(IntList & value, const xercesc::Attributes & a, const char * name) const
      {
        const XMLCh * val = a.getValue(sm_.convertPtr(name).get());
        if (val != nullptr)
        {
          value = attributeAsIntList_(a, name);
          return true;
        }
        return false;
      }
 
      inline std::string attributeAsString_(const xercesc::Attributes & a, const XMLCh * name) const
      {
        const XMLCh * val = a.getValue(name);
        if (val == nullptr) fatalError(LOAD,std::string("Required attribute '") + sm_.convert(name) + "' not present!");
        return sm_.convert(val);
      }
 
      inline Int attributeAsInt_(const xercesc::Attributes & a, const XMLCh * name) const
      {
        const XMLCh * val = a.getValue(name);
        if (val == nullptr) fatalError(LOAD,std::string("Required attribute '") + sm_.convert(name) + "' not present!");
        return xercesc::XMLString::parseInt(val);
      }
 
      inline double attributeAsDouble_(const xercesc::Attributes & a, const XMLCh * name) const
      {
        const XMLCh * val = a.getValue(name);
        if (val == nullptr) fatalError(LOAD,std::string("Required attribute '") + sm_.convert(name) + "' not present!");
        return StringUtils::toDouble(sm_.convert(val));
      }
 
      inline DoubleList attributeAsDoubleList_(const xercesc::Attributes & a, const XMLCh * name) const
      {
        std::string tmp(expectList_(attributeAsString_(a, name)));
        return ListUtils::create<double>(StringUtils::substr(tmp, 1, tmp.size() - 2));
      }
 
      inline IntList attributeAsIntList_(const xercesc::Attributes & a, const XMLCh * name) const
      {
        std::string tmp(expectList_(attributeAsString_(a, name)));
        return ListUtils::create<Int>(StringUtils::substr(tmp, 1, tmp.size() - 2));
      }
 
      inline StringList attributeAsStringList_(const xercesc::Attributes & a, const XMLCh * name) const
      {
        std::string tmp(expectList_(attributeAsString_(a, name)));
        StringList tmp_list = ListUtils::create<std::string>(StringUtils::substr(tmp, 1, tmp.size() - 2)); // between [ and ]
 
        if (StringUtils::hasSubstring(tmp, "\\|")) // check full string for escaped comma
        {
          for (std::string& s : tmp_list)
          {
            StringUtils::substitute(s, "\\|", ",");
          }          
        }
        return tmp_list;
      }
 
      inline bool optionalAttributeAsString_(std::string& value, const xercesc::Attributes & a, const XMLCh * name) const
      {
        const XMLCh * val = a.getValue(name);
        if (val != nullptr)
        {
          value = sm_.convert(val);
          return !value.empty();
        }
        return false;
      }
 
      inline bool optionalAttributeAsInt_(Int & value, const xercesc::Attributes & a, const XMLCh * name) const
      {
        const XMLCh * val = a.getValue(name);
        if (val != nullptr)
        {
          value = xercesc::XMLString::parseInt(val);
          return true;
        }
        return false;
      }
 
      inline bool optionalAttributeAsUInt_(UInt & value, const xercesc::Attributes & a, const XMLCh * name) const
      {
        const XMLCh * val = a.getValue(name);
        if (val != nullptr)
        {
          value = xercesc::XMLString::parseInt(val);
          return true;
        }
        return false;
      }
 
      inline bool optionalAttributeAsDouble_(double & value, const xercesc::Attributes & a, const XMLCh * name) const
      {
        const XMLCh * val = a.getValue(name);
        if (val != nullptr)
        {
          value = StringUtils::toDouble(sm_.convert(val));
          return true;
        }
        return false;
      }
 
      inline bool optionalAttributeAsDoubleList_(DoubleList & value, const xercesc::Attributes & a, const XMLCh * name) const
      {
        const XMLCh * val = a.getValue(name);
        if (val != nullptr)
        {
          value = attributeAsDoubleList_(a, name);
          return true;
        }
        return false;
      }
 
      inline bool optionalAttributeAsIntList_(IntList & value, const xercesc::Attributes & a, const XMLCh * name) const
      {
        const XMLCh * val = a.getValue(name);
        if (val != nullptr)
        {
          value = attributeAsIntList_(a, name);
          return true;
        }
        return false;
      }
 
      inline bool optionalAttributeAsStringList_(StringList & value, const xercesc::Attributes & a, const XMLCh * name) const
      {
        const XMLCh * val = a.getValue(name);
        if (val != nullptr)
        {
          value = attributeAsStringList_(a, name);
          return true;
        }
        return false;
      }
 
 
private:
      XMLHandler();
 
      inline const std::string& expectList_(const std::string& str) const
      {
        if (!(StringUtils::hasPrefix(str, '[') && StringUtils::hasSuffix(str, ']')))
        {
          fatalError(LOAD,std::string("List argument is not a string representation of a list!"));
        }
        return str;
      }
 
    };
 
  }   // namespace Internal
} // namespace OpenMS