bag0.cpp

Go to the documentation of this file.

#include "eleve/config.hpp"

#include "eleve/src/utilities.hpp"

#include <iostream>
#include <string>
#include <list>
#include <vector>
#include <map>

#include <typeinfo>

#include <boost/bind.hpp>
#include <boost/algorithm/string/trim.hpp>
#include <boost/tokenizer.hpp>
#include <boost/multi_array.hpp>
#include <boost/date_time/gregorian/formatters.hpp>

#include "eleve/src/loader0.hpp"
#include "eleve/src/scal0.hpp"

#include "bag1.hpp"

namespace eepgwde { namespace detail {

  // Our test bag.
  static Bag bag0;

}; };

#include "eleve/src/typing0.hpp"
#include "eleve/src/typing3.hpp"
#include "eleve/src/typing1.hpp"

// Needed to avoid the having to specify the defaults for the template.
std::string to0(const boost::gregorian::date &d) {
  return boost::gregorian::to_simple_string(d);
}

// You need to give a filename of a good square CSV file - all column headings present.
int main(int argc, char *argv[]) {
  using namespace std;

  using namespace boost;
  using namespace eepgwde::detail;

  // A test string for embedded commas.
  string s = "Field 1,\"putting quotes around fields, allows commas\",Field 3";

  if (argc < 2) return(1);
  string n(argv[1]);
  loader l(n);          //< this will read the file.

  char * csv;
  string s1;
  int cols_header = 0; //< columns in the header.

  // This is a typical input loop.
  // We have to count the columns and store it into cols_header.
  while ( csv = l.get() ) {
    string s(csv);
    cerr << "line: " << s << endl;
    // I've never checked if tokenizer can be cleared.
    tokenizer<escaped_list_separator<char> > tok(s);
    tokenizer<escaped_list_separator<char> >::iterator beg;

    int pos;        //< the column number - follow AWK with first field as 1.

    // Tokenizer doesn't count tokens, so we them first count the number of tokens/columns.
    for(beg=tok.begin(), pos=1; beg!=tok.end(); ++beg, pos++)
      ;
    int cols_counted = pos;

    // Here I set pos to 1.
    for(beg=tok.begin(), pos=1; beg!=tok.end(); ++beg, pos++) {
      if (l.line() == 1) {
    cols_header = cols_counted; //< store the number of columns permitted.
    s1 = *beg;
    trim(s1);
    cerr << "header: " << s1 << endl;
     
    bag0.add(pos, s1);  //< store the column name against the position.
    continue;       //< skip.
      }
      if (cols_counted == cols_header) //< only if a data line has enough cols, take it in
    bag0.append(pos, *beg);
    }
  }

  // Report what the bag has.

  // List of mappings (column number, column name).
  cerr << "Bag header mappings\n";
  cerr << bag0.header.left;
  cerr << "\n";

  // For each column, by name (names are on the right).
  for(Bag::r_itr hi = bag0.header.right.begin(); hi != bag0.header.right.end(); ++hi) {

    // Get the list by name. Give name, size and counts.
    std::list<boost::any> * l0 = bag0.get(hi->first);
    cerr << hi->first << " "
     << "size: " << l0->size() << " " 
     << "count non-empty <boost::any>: " << eepgwde::detail::count<boost::any>(l0)  << " " 
     << "count int: " << eepgwde::detail::count<int>(l0)  << " " 
     << "count double: " << eepgwde::detail::count<double>(l0)  << " " 
     << "count date: " << eepgwde::detail::count<posixdate>(l0)  << " " 
     << endl;

    // There is flag in render called fatal. When it's zero then no exceptions.

    int fails = 0;
    // Can we render it as a vector of int?
    {
      vector<int> v;
      cerr << "int" << endl;
      try {
    eepgwde::daemon::watcher = 1;
    render(&v, l0, fails);
    cerr << "render: size: " << v.size() << endl;
    copy(v.begin(), v.end(), ostream_iterator<int>(cerr, "\t"));
    cerr << endl;
      } catch (const boost::bad_any_cast & e) {
    cerr << hi->first << " : fails1: " << fails << " : " << e.what() << endl;
      } catch (const boost::bad_lexical_cast & e1) {
    cerr << hi->first << " : fails2: " << fails << " : " << e1.what() << endl;
      }
    }

    // Can we render it as a vector of double?
    {
      vector<double> v;
      cerr << "double" << endl;
      try {
    render(&v, l0, fails);
    cerr << "render: size: " << v.size() << endl;
    copy(v.begin(), v.end(), ostream_iterator<double>(cerr, "\t"));
    cerr << endl;
      } catch (const boost::bad_any_cast & e) {
    cerr << hi->first << " : fails: " << fails << " : " << e.what() << endl;
      } catch (const boost::bad_lexical_cast & e1) {
    cerr << hi->first << " : fails: " << fails << " : " << e1.what() << endl;
      }
    }

    // Can we render it as a vector of date?
    {
      vector<posixdate> v;
      cerr << "date" << endl;
      try {
    render(&v, l0, fails);
    cerr << "render: size: " << v.size() << endl;
    vector<std::string> s(v.size());
    transform(v.begin(), v.end(),
          s.begin(),
          (boost::bind(&to0,_1)) );
    copy(s.begin(), s.end(), ostream_iterator<std::string>(cerr, "\t"));
    cerr << endl;
      } catch (const boost::bad_any_cast & e) {
    cerr << hi->first << " : fails: " << fails << " : " << e.what() << endl;
      } catch (const boost::bad_lexical_cast & e1) {
    cerr << hi->first << " : fails: " << fails << " : " << e1.what() << endl;
      }
    }

    // Use the singleton tester.
    using eepgwde::daemon::Q;

    for (std::list<boost::any>::iterator j = l0->begin();
     j != l0->end(); ++j) {
      cerr << hi->first << " " << boost::any_cast<string>(*j)
       << " is_double: "
       << is<double>(*j)
       << "\n";
    }
  }

  // Do the same again, but now overwrite the list.

  cerr << "over-render" << endl;

  // For each column, by name (names are on the right).
  for(Bag::r_itr hi = bag0.header.right.begin(); hi != bag0.header.right.end(); ++hi) {

    // Get the list by name. Give name, size and counts.
    std::list<boost::any> * l0 = bag0.get(hi->first);
    cerr << hi->first << " "
     << "size: " << l0->size() << " " 
     << "count non-empty: " << eepgwde::detail::count<boost::any>(l0)  << " " 
     << "count int: " << eepgwde::detail::count<int>(l0)  << " " 
     << "count double: " << eepgwde::detail::count<double>(l0)  << " " 
     << "count date: " << eepgwde::detail::count<posixdate>(l0)  << " " 
     << endl;

    // There is flag in render called fatal. When it's zero then no exceptions.

    int fails = 0;
    // Render as double where possible - specify the type
    if (eepgwde::detail::count<double>(l0) > 0) {
      try {
    render<double>(l0, fails);
    cerr << "render: fails: " << fails << endl;
      } catch (const boost::bad_any_cast & e) {
    cerr << hi->first << " : fails: " << fails << " : " << e.what() << endl;
      } catch (const boost::bad_lexical_cast & e1) {
    cerr << hi->first << " : fails: " << fails << " : " << e1.what() << endl;
      }
    } else if (eepgwde::detail::count<posixdate>(l0) > 0) {
      try {
    render<posixdate>(l0, fails);
    cerr << "render: fails: " << fails << endl;
      } catch (const boost::bad_any_cast & e) {
    cerr << hi->first << " : fails: " << fails << " : " << e.what() << endl;
      } catch (const boost::bad_lexical_cast & e1) {
    cerr << hi->first << " : fails: " << fails << " : " << e1.what() << endl;
      }
    }

    // Use the singleton tester and now use an any_cast.
    using eepgwde::daemon::Q;

    std::string s2;
    for (std::list<boost::any>::iterator j = l0->begin();
     j != l0->end(); ++j) {
      s2 = "un-castable";
      if (is<posixdate>(*j)) 
    s2 = boost::gregorian::to_simple_string(any_cast<posixdate>(*j));
      if (is<double>(*j)) 
    s2 = boost::lexical_cast<std::string>( any_cast<double>(*j) );
      if (is<std::string>(*j)) 
    s2 = boost::lexical_cast<std::string>( any_cast<std::string>(*j) );
      cerr << hi->first << " "
       << demangle( (j->type()).name() ) << " "
       << s2
       << "\n";
    }
  }

  return 0;
}

---