doxyindexer.cpp File Reference

#include <cstdio>
#include <cstdlib>
#include <iostream>
#include <string>
#include <algorithm>
#include <sstream>
#include <qregexp.h>
#include <qxml.h>
#include <qfile.h>
#include <qfileinfo.h>
#include <xapian.h>

Go to the source code of this file.

Classes
class	XMLContentHandler
class	XMLErrorHandler

Macros
#define	MAX_TERM_LENGTH   245

Functions
static void	safeAddTerm (const std::string &term, Xapian::Document &doc, int wfd)
static std::string	trim (const std::string &str, const std::string &whitespace=" \t")
static std::string	reduce (const std::string &str, const std::string &fill=" ", const std::string &whitespace=" \t")
static void	addWords (const std::string &s, Xapian::Document &doc, int wfd)
static void	addIdentifiers (const std::string &s, Xapian::Document &doc, int wfd)
static void	replace_all (std::string &str, const std::string &old, const std::string &repl)
static std::string	unescapeXmlEntities (const std::string &s)
static void	usage (const char *name)
int	main (int argc, const char **argv)

Variables
static char	pathSep = '/'

Macro Definition Documentation

#define MAX_TERM_LENGTH   245

Definition at line 33 of file doxyindexer.cpp.

Function Documentation

static void addIdentifiers ( const std::string &  s, Xapian::Document &  doc, int  wfd  )

static

Adds all identifiers in s to document doc with weight wfd

Definition at line 105 of file doxyindexer.cpp.

{
  QRegExp re("[A-Z_a-z][A-Z_a-z0-9]*");
  int i,l,p=0;
  QCString qs = s.c_str();
  while ((i=re.match(qs,p,&l))!=-1)
  {
    safeAddTerm(qs.mid(p,i-p).data(),doc,wfd);
    p=i+l;
  }
}

static void addWords ( const std::string &  s, Xapian::Document &  doc, int  wfd  )

static

Adds all words in s to document doc with weight wfd

Definition at line 87 of file doxyindexer.cpp.

{
  std::istringstream iss(s);
  std::istream_iterator<std::string> begin(iss),end,it;
  for (it=begin;it!=end;++it)
  {
    std::string word = *it;
    std::string lword = word;
    std::transform(lword.begin(), lword.end(), lword.begin(), ::tolower);
    safeAddTerm(word,doc,wfd);
    if (lword!=word)
    {
      safeAddTerm(lword,doc,wfd);
    }
  }
}

int main	(	int	argc,
		const char **	argv
	)

main function to index data

Definition at line 313 of file doxyindexer.cpp.

{
  if (argc<2)
  {
    usage(argv[0]);
  }
  QString outputDir;
  for (int i=1;i<argc;i++)
  {
    if (std::string(argv[i])=="-o")
    {
      if (i>=argc-1)
      {
        std::cerr << "Error: missing parameter for -o option" << std::endl;
        usage(argv[0]);
      }
      else
      {
        i++;
        outputDir=argv[i];
        QFileInfo fi(outputDir);
        if (!fi.exists() || !fi.isDir())
        {
          std::cerr << "Error: specified output directory does not exist!" << std::endl;
          usage(argv[0]);
        }
      }
    }
    else if (std::string(argv[i])=="-h" || std::string(argv[i])=="--help")
    {
      usage(argv[0]);
    }
  }

  try
  {
    if (!outputDir.isEmpty() && outputDir.at(outputDir.length()-1)!=pathSep)
    {
      outputDir+=pathSep;
    }
    XMLContentHandler handler(outputDir);
    XMLErrorHandler errorHandler;
    for (int i=1;i<argc;i++)
    {
      if (std::string(argv[i])=="-o")
      {
        i++;
      }
      else
      {
        QString xmlFileName = argv[i];
        std::cout << "Processing " << xmlFileName.utf8() << "..." << std::endl;
        QFile xmlFile(xmlFileName);
        QXmlInputSource source(xmlFile);
        QXmlSimpleReader reader;
        reader.setContentHandler(&handler);
        reader.setErrorHandler(&errorHandler);
        reader.parse(source);
      }
    }
  }
  catch(const Xapian::Error &e) 
  {
    std::cerr << "Caught exception: " << e.get_description() << std::endl;
  }
  catch(...)
  {
    std::cerr << "Caught an unknown exception" << std::endl;
  }

  return 0;
}

static std::string reduce ( const std::string &  str, const std::string &  fill = " ", const std::string &  whitespace = " \t"  )

static

trims whitespace from start and end and replace occurrences of whitespace with fill.

Definition at line 63 of file doxyindexer.cpp.

{
  // trim first
  std::string result = trim(str, whitespace);

  // replace sub ranges
  size_t beginSpace = result.find_first_of(whitespace);
  while (beginSpace != std::string::npos)
  {
    size_t endSpace = result.find_first_not_of(whitespace, beginSpace);
    int range = endSpace - beginSpace;

    result.replace(beginSpace, range, fill);

    size_t newStart = beginSpace + fill.length();
    beginSpace = result.find_first_of(whitespace, newStart);
  }

  return result;
}

static void replace_all ( std::string &  str, const std::string &  old, const std::string &  repl  )

static

Replaces all occurrences of old with repl in string str

Definition at line 118 of file doxyindexer.cpp.

{
  size_t pos = 0;
  while ((pos = str.find(old, pos)) != std::string::npos) 
  {
    str.replace(pos, old.length(), repl);
    pos += repl.length();
  }
}

static void safeAddTerm ( const std::string &  term, Xapian::Document &  doc, int  wfd  )

static

Definition at line 41 of file doxyindexer.cpp.

{
  if (term.length()<=MAX_TERM_LENGTH) doc.add_term(term,wfd);
}

static std::string trim ( const std::string &  str, const std::string &  whitespace = " \t"  )

static

trims whitespace characters from the start and end of string str.

Definition at line 47 of file doxyindexer.cpp.

{
  size_t strBegin = str.find_first_not_of(whitespace);
  if (strBegin == std::string::npos)
    return ""; // no content

  size_t strEnd = str.find_last_not_of(whitespace);
  int strRange = strEnd - strBegin + 1;

  return str.substr(strBegin, strRange);
}

static std::string unescapeXmlEntities ( const std::string &  s )

static

Replaces all XML entities in s with their unescaped representation

Definition at line 129 of file doxyindexer.cpp.

{
  std::string result=s;
  replace_all(result,">",">");
  replace_all(result,"&lt;","<");
  replace_all(result,"&apos;","'");
  replace_all(result,"&quot;","\"");
  replace_all(result,"&amp;","&");
  return result;
}

static void usage ( const char *  name )

static

Definition at line 306 of file doxyindexer.cpp.

{
  std::cerr << "Usage: " << name << " [-o output_dir] searchdata.xml [searchdata2.xml ...]" << std::endl;
  exit(1);
}

Variable Documentation

char pathSep = '/'

static

Definition at line 38 of file doxyindexer.cpp.