2011-10-12 03:11:59 +02:00
|
|
|
#include "csv_datasource.hpp"
|
|
|
|
|
|
|
|
// boost
|
|
|
|
#include <boost/make_shared.hpp>
|
|
|
|
#include <boost/tokenizer.hpp>
|
|
|
|
#include <boost/lexical_cast.hpp>
|
|
|
|
#include <boost/algorithm/string.hpp>
|
2011-10-15 05:28:23 +02:00
|
|
|
#include <boost/spirit/include/qi.hpp>
|
2011-10-12 03:11:59 +02:00
|
|
|
|
|
|
|
// mapnik
|
|
|
|
#include <mapnik/feature_layer_desc.hpp>
|
|
|
|
#include <mapnik/feature_factory.hpp>
|
|
|
|
#include <mapnik/geometry.hpp>
|
|
|
|
#include <mapnik/memory_featureset.hpp>
|
|
|
|
#include <mapnik/wkt/wkt_factory.hpp>
|
2012-01-15 07:35:40 +01:00
|
|
|
#include <mapnik/util/geometry_to_ds_type.hpp>
|
2011-11-14 04:33:57 +01:00
|
|
|
#include <mapnik/ptree_helpers.hpp> // mapnik::boolean
|
2011-10-12 03:11:59 +02:00
|
|
|
|
|
|
|
// stl
|
|
|
|
#include <sstream>
|
2011-11-10 01:45:18 +01:00
|
|
|
#include <fstream>
|
|
|
|
#include <iostream>
|
2011-11-14 04:33:57 +01:00
|
|
|
#include <vector>
|
2011-10-12 03:11:59 +02:00
|
|
|
#include <string>
|
|
|
|
|
|
|
|
using mapnik::datasource;
|
|
|
|
using mapnik::parameters;
|
2011-10-15 05:28:23 +02:00
|
|
|
using namespace boost::spirit;
|
2011-10-12 03:11:59 +02:00
|
|
|
|
|
|
|
DATASOURCE_PLUGIN(csv_datasource)
|
|
|
|
|
|
|
|
csv_datasource::csv_datasource(parameters const& params, bool bind)
|
2011-11-14 04:33:57 +01:00
|
|
|
: datasource(params),
|
|
|
|
desc_(*params_.get<std::string>("type"), *params_.get<std::string>("encoding", "utf-8")),
|
|
|
|
extent_(),
|
|
|
|
filename_(),
|
|
|
|
inline_string_(),
|
|
|
|
file_length_(0),
|
|
|
|
row_limit_(*params_.get<int>("row_limit", 0)),
|
|
|
|
features_(),
|
|
|
|
escape_(*params_.get<std::string>("escape", "")),
|
|
|
|
separator_(*params_.get<std::string>("separator", "")),
|
|
|
|
quote_(*params_.get<std::string>("quote", "")),
|
|
|
|
headers_(),
|
|
|
|
manual_headers_(boost::trim_copy(*params_.get<std::string>("headers", ""))),
|
|
|
|
strict_(*params_.get<mapnik::boolean>("strict", false)),
|
|
|
|
quiet_(*params_.get<mapnik::boolean>("quiet", false)),
|
|
|
|
filesize_max_(*params_.get<float>("filesize_max", 20.0)) // MB
|
2011-10-12 03:11:59 +02:00
|
|
|
{
|
|
|
|
/* TODO:
|
2011-11-14 04:33:57 +01:00
|
|
|
general:
|
|
|
|
- refactor parser into generic class
|
|
|
|
- tests of grid_renderer output
|
|
|
|
- ensure that the attribute desc_ matches the first feature added
|
|
|
|
alternate large file pipeline:
|
|
|
|
- stat file, detect > 15 MB
|
|
|
|
- build up csv line-by-line iterator
|
|
|
|
- creates opportunity to filter attributes by map query
|
|
|
|
speed:
|
|
|
|
- add properties for wkt/lon/lat at parse time
|
|
|
|
- remove boost::lexical_cast
|
|
|
|
- add ability to pass 'filter' keyword to drop attributes at layer init
|
|
|
|
- create quad tree on the fly for small/med size files
|
|
|
|
- memory map large files for reading
|
|
|
|
- smaller features (less memory overhead)
|
|
|
|
usability:
|
|
|
|
- enforce column names without leading digit
|
|
|
|
- better error messages (add filepath) if not reading from string
|
|
|
|
- move to spirit to tokenize and add character level error feedback:
|
|
|
|
http://boost-spirit.com/home/articles/qi-example/tracking-the-input-position-while-parsing/
|
2011-10-12 03:11:59 +02:00
|
|
|
*/
|
2011-11-14 04:33:57 +01:00
|
|
|
|
2011-10-12 03:11:59 +02:00
|
|
|
boost::optional<std::string> inline_string = params_.get<std::string>("inline");
|
|
|
|
if (inline_string)
|
|
|
|
{
|
|
|
|
inline_string_ = *inline_string;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
boost::optional<std::string> file = params_.get<std::string>("file");
|
|
|
|
if (!file) throw mapnik::datasource_exception("CSV Plugin: missing <file> parameter");
|
2011-11-14 04:33:57 +01:00
|
|
|
|
2011-10-12 03:11:59 +02:00
|
|
|
boost::optional<std::string> base = params_.get<std::string>("base");
|
|
|
|
if (base)
|
|
|
|
filename_ = *base + "/" + *file;
|
|
|
|
else
|
|
|
|
filename_ = *file;
|
|
|
|
}
|
2011-11-14 04:33:57 +01:00
|
|
|
|
2011-10-12 03:11:59 +02:00
|
|
|
if (bind)
|
|
|
|
{
|
|
|
|
this->bind();
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
csv_datasource::~csv_datasource() { }
|
|
|
|
|
|
|
|
void csv_datasource::bind() const
|
|
|
|
{
|
|
|
|
if (is_bound_) return;
|
2011-11-14 04:33:57 +01:00
|
|
|
|
2011-10-12 03:11:59 +02:00
|
|
|
if (!inline_string_.empty())
|
|
|
|
{
|
|
|
|
std::istringstream in(inline_string_);
|
2011-10-15 05:28:23 +02:00
|
|
|
parse_csv(in,escape_, separator_, quote_);
|
2011-10-12 03:11:59 +02:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
2011-10-15 05:28:23 +02:00
|
|
|
std::ifstream in(filename_.c_str(),std::ios_base::in | std::ios_base::binary);
|
2011-10-12 03:11:59 +02:00
|
|
|
if (!in.is_open())
|
|
|
|
throw mapnik::datasource_exception("CSV Plugin: could not open: '" + filename_ + "'");
|
2011-10-15 05:28:23 +02:00
|
|
|
parse_csv(in,escape_, separator_, quote_);
|
2011-10-12 03:11:59 +02:00
|
|
|
in.close();
|
|
|
|
}
|
|
|
|
is_bound_ = true;
|
|
|
|
}
|
|
|
|
|
|
|
|
template <typename T>
|
2011-10-15 05:28:23 +02:00
|
|
|
void csv_datasource::parse_csv(T& stream,
|
|
|
|
std::string const& escape,
|
|
|
|
std::string const& separator,
|
|
|
|
std::string const& quote) const
|
2011-10-12 03:11:59 +02:00
|
|
|
{
|
2011-11-14 04:33:57 +01:00
|
|
|
stream.seekg(0, std::ios::end);
|
2011-11-01 16:55:23 +01:00
|
|
|
file_length_ = stream.tellg();
|
2011-11-14 04:33:57 +01:00
|
|
|
|
2011-10-17 20:03:50 +02:00
|
|
|
if (filesize_max_ > 0)
|
|
|
|
{
|
|
|
|
double file_mb = static_cast<double>(file_length_)/1048576;
|
2011-11-14 04:33:57 +01:00
|
|
|
|
2011-10-17 20:03:50 +02:00
|
|
|
// throw if this is an unreasonably large file to read into memory
|
|
|
|
if (file_mb > filesize_max_)
|
|
|
|
{
|
|
|
|
std::ostringstream s;
|
|
|
|
s << "CSV Plugin: csv file is greater than " << filesize_max_ << "MB "
|
|
|
|
<< " - you should use a more efficient data format like sqlite, postgis or a shapefile "
|
|
|
|
<< " to render this data (set 'filesize_max=0' to disable this restriction if you have lots of memory)";
|
|
|
|
throw mapnik::datasource_exception(s.str());
|
|
|
|
}
|
|
|
|
}
|
2011-10-15 05:28:23 +02:00
|
|
|
|
2011-10-19 03:21:19 +02:00
|
|
|
// set back to start
|
2011-11-14 04:33:57 +01:00
|
|
|
stream.seekg(0, std::ios::beg);
|
2011-10-12 03:11:59 +02:00
|
|
|
|
2011-10-15 05:28:23 +02:00
|
|
|
// autodetect newlines
|
2011-10-19 03:21:19 +02:00
|
|
|
char newline = '\n';
|
|
|
|
int newline_count = 0;
|
|
|
|
int carriage_count = 0;
|
2011-11-14 04:33:57 +01:00
|
|
|
for (unsigned idx = 0; idx < file_length_; idx++)
|
2011-10-12 03:11:59 +02:00
|
|
|
{
|
2011-10-19 03:21:19 +02:00
|
|
|
char c = static_cast<char>(stream.get());
|
|
|
|
if (c == '\n')
|
|
|
|
{
|
|
|
|
++newline_count;
|
|
|
|
}
|
|
|
|
else if (c == '\r')
|
|
|
|
{
|
|
|
|
++carriage_count;
|
|
|
|
}
|
|
|
|
// read at least 2000 bytes before testing
|
|
|
|
if (idx == file_length_-1 || idx > 4000)
|
|
|
|
{
|
|
|
|
if (newline_count > carriage_count)
|
|
|
|
{
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
else if (carriage_count > newline_count)
|
|
|
|
{
|
|
|
|
newline = '\r';
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
2011-10-15 05:28:23 +02:00
|
|
|
}
|
2011-10-19 03:21:19 +02:00
|
|
|
|
2011-10-15 05:28:23 +02:00
|
|
|
// set back to start
|
2011-11-14 04:33:57 +01:00
|
|
|
stream.seekg(0, std::ios::beg);
|
|
|
|
|
2011-10-19 03:21:19 +02:00
|
|
|
// get first line
|
|
|
|
std::string csv_line;
|
|
|
|
std::getline(stream,csv_line,newline);
|
|
|
|
|
|
|
|
// if user has not passed a separator manually
|
2011-10-15 05:28:23 +02:00
|
|
|
// then attempt to detect by reading first line
|
|
|
|
std::string sep = boost::trim_copy(separator);
|
|
|
|
if (sep.empty())
|
|
|
|
{
|
|
|
|
// default to ','
|
|
|
|
sep = ",";
|
2011-11-02 02:11:10 +01:00
|
|
|
int num_commas = std::count(csv_line.begin(), csv_line.end(), ',');
|
2011-10-15 05:28:23 +02:00
|
|
|
// detect tabs
|
|
|
|
int num_tabs = std::count(csv_line.begin(), csv_line.end(), '\t');
|
|
|
|
if (num_tabs > 0)
|
|
|
|
{
|
|
|
|
if (num_tabs > num_commas)
|
|
|
|
{
|
|
|
|
sep = "\t";
|
|
|
|
#ifdef MAPNIK_DEBUG
|
|
|
|
std::clog << "CSV Plugin: auto detected tab separator\n";
|
|
|
|
#endif
|
|
|
|
}
|
|
|
|
}
|
2011-11-02 02:11:10 +01:00
|
|
|
else // pipes
|
|
|
|
{
|
|
|
|
int num_pipes = std::count(csv_line.begin(), csv_line.end(), '|');
|
|
|
|
if (num_pipes > num_commas)
|
|
|
|
{
|
|
|
|
sep = "|";
|
|
|
|
#ifdef MAPNIK_DEBUG
|
|
|
|
std::clog << "CSV Plugin: auto detected '|' separator\n";
|
|
|
|
#endif
|
|
|
|
}
|
|
|
|
else // semicolons
|
|
|
|
{
|
|
|
|
int num_semicolons = std::count(csv_line.begin(), csv_line.end(), ';');
|
|
|
|
if (num_semicolons > num_commas)
|
|
|
|
{
|
|
|
|
sep = ";";
|
2011-11-14 04:33:57 +01:00
|
|
|
#ifdef MAPNIK_DEBUG
|
2011-11-02 02:11:10 +01:00
|
|
|
std::clog << "CSV Plugin: auto detected ';' separator\n";
|
2011-11-14 04:33:57 +01:00
|
|
|
#endif
|
2011-11-02 02:11:10 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
2011-10-15 05:28:23 +02:00
|
|
|
}
|
2011-10-19 03:21:19 +02:00
|
|
|
|
|
|
|
// set back to start
|
2011-11-14 04:33:57 +01:00
|
|
|
stream.seekg(0, std::ios::beg);
|
|
|
|
|
2011-10-15 05:28:23 +02:00
|
|
|
typedef boost::escaped_list_separator<char> escape_type;
|
2011-10-12 03:11:59 +02:00
|
|
|
|
2011-10-15 05:28:23 +02:00
|
|
|
std::string esc = boost::trim_copy(escape);
|
|
|
|
if (esc.empty()) esc = "\\";
|
2011-11-14 04:33:57 +01:00
|
|
|
|
2011-10-15 05:28:23 +02:00
|
|
|
std::string quo = boost::trim_copy(quote);
|
|
|
|
if (quo.empty()) quo = "\"";
|
2011-10-12 03:11:59 +02:00
|
|
|
|
2011-10-15 05:28:23 +02:00
|
|
|
#ifdef MAPNIK_DEBUG
|
|
|
|
std::clog << "CSV Plugin: csv grammer: sep: '" << sep << "' quo: '" << quo << "' esc: '" << esc << "'\n";
|
|
|
|
#endif
|
|
|
|
|
|
|
|
boost::escaped_list_separator<char> grammer;
|
|
|
|
try
|
|
|
|
{
|
2011-11-14 04:33:57 +01:00
|
|
|
// grammer = boost::escaped_list_separator<char>('\\', ',', '\"');
|
2011-10-15 05:28:23 +02:00
|
|
|
grammer = boost::escaped_list_separator<char>(esc, sep, quo);
|
|
|
|
}
|
2011-11-14 04:33:57 +01:00
|
|
|
catch(const std::exception & ex)
|
2011-10-15 05:28:23 +02:00
|
|
|
{
|
|
|
|
std::ostringstream s;
|
|
|
|
s << "CSV Plugin: " << ex.what();
|
|
|
|
throw mapnik::datasource_exception(s.str());
|
|
|
|
}
|
2011-11-14 04:33:57 +01:00
|
|
|
|
2011-10-17 20:18:44 +02:00
|
|
|
typedef boost::tokenizer< escape_type > Tokenizer;
|
2011-10-15 05:28:23 +02:00
|
|
|
|
|
|
|
int line_number(1);
|
|
|
|
bool has_wkt_field = false;
|
|
|
|
bool has_lat_field = false;
|
|
|
|
bool has_lon_field = false;
|
|
|
|
unsigned wkt_idx;
|
|
|
|
unsigned lat_idx;
|
|
|
|
unsigned lon_idx;
|
|
|
|
|
|
|
|
if (!manual_headers_.empty())
|
|
|
|
{
|
2011-10-17 20:18:44 +02:00
|
|
|
Tokenizer tok(manual_headers_, grammer);
|
2011-10-15 05:28:23 +02:00
|
|
|
Tokenizer::iterator beg = tok.begin();
|
|
|
|
unsigned idx(0);
|
|
|
|
for (; beg != tok.end(); ++beg)
|
2011-10-12 03:11:59 +02:00
|
|
|
{
|
2011-10-15 05:28:23 +02:00
|
|
|
std::string val = boost::trim_copy(*beg);
|
|
|
|
std::string lower_val = boost::algorithm::to_lower_copy(val);
|
2011-11-04 12:18:40 +01:00
|
|
|
if (lower_val == "wkt"
|
|
|
|
|| (lower_val.find("geom") != std::string::npos))
|
2011-10-12 03:11:59 +02:00
|
|
|
{
|
2011-10-15 05:28:23 +02:00
|
|
|
wkt_idx = idx;
|
|
|
|
has_wkt_field = true;
|
|
|
|
}
|
2011-10-19 04:27:23 +02:00
|
|
|
if (lower_val == "x"
|
|
|
|
|| lower_val == "lon"
|
|
|
|
|| lower_val == "long"
|
|
|
|
|| (lower_val.find("longitude") != std::string::npos))
|
2011-10-15 05:28:23 +02:00
|
|
|
{
|
|
|
|
lon_idx = idx;
|
|
|
|
has_lon_field = true;
|
|
|
|
}
|
2011-10-19 04:27:23 +02:00
|
|
|
if (lower_val == "y"
|
|
|
|
|| lower_val == "lat"
|
|
|
|
|| (lower_val.find("latitude") != std::string::npos))
|
2011-10-15 05:28:23 +02:00
|
|
|
{
|
|
|
|
lat_idx = idx;
|
|
|
|
has_lat_field = true;
|
|
|
|
}
|
|
|
|
++idx;
|
|
|
|
headers_.push_back(val);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
else // parse first line as headers
|
|
|
|
{
|
|
|
|
while (std::getline(stream,csv_line,newline))
|
|
|
|
{
|
|
|
|
try
|
|
|
|
{
|
2011-10-17 20:18:44 +02:00
|
|
|
Tokenizer tok(csv_line, grammer);
|
2011-10-15 05:28:23 +02:00
|
|
|
Tokenizer::iterator beg = tok.begin();
|
2011-11-14 09:34:26 +01:00
|
|
|
std::string val;
|
|
|
|
if (beg != tok.end())
|
|
|
|
val = boost::trim_copy(*beg);
|
2011-11-14 04:33:57 +01:00
|
|
|
|
2011-10-15 05:28:23 +02:00
|
|
|
// skip blank lines
|
|
|
|
if (val.empty())
|
2011-10-12 03:11:59 +02:00
|
|
|
{
|
2011-10-15 05:28:23 +02:00
|
|
|
// do nothing
|
|
|
|
++line_number;
|
2011-10-12 03:11:59 +02:00
|
|
|
}
|
2011-10-15 05:28:23 +02:00
|
|
|
else
|
2011-10-12 03:11:59 +02:00
|
|
|
{
|
2011-10-15 05:28:23 +02:00
|
|
|
int idx = -1;
|
|
|
|
for (; beg != tok.end(); ++beg)
|
|
|
|
{
|
|
|
|
++idx;
|
|
|
|
val = boost::trim_copy(*beg);
|
|
|
|
if (val.empty())
|
|
|
|
{
|
2011-11-02 01:33:05 +01:00
|
|
|
if (strict_)
|
|
|
|
{
|
|
|
|
std::ostringstream s;
|
|
|
|
s << "CSV Plugin: expected a column header at line "
|
|
|
|
<< line_number << ", column " << idx
|
|
|
|
<< " - ensure this row contains valid header fields: '"
|
|
|
|
<< csv_line << "'\n";
|
|
|
|
throw mapnik::datasource_exception(s.str());
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
// create a placeholder for the empty header
|
|
|
|
std::ostringstream s;
|
|
|
|
s << "_" << idx;
|
|
|
|
headers_.push_back(s.str());
|
|
|
|
}
|
2011-10-15 05:28:23 +02:00
|
|
|
}
|
|
|
|
else
|
2011-11-14 04:33:57 +01:00
|
|
|
{
|
2011-10-15 05:28:23 +02:00
|
|
|
std::string lower_val = boost::algorithm::to_lower_copy(val);
|
2011-11-04 12:18:40 +01:00
|
|
|
if (lower_val == "wkt"
|
|
|
|
|| (lower_val.find("geom") != std::string::npos))
|
2011-10-15 05:28:23 +02:00
|
|
|
{
|
|
|
|
wkt_idx = idx;
|
|
|
|
has_wkt_field = true;
|
|
|
|
}
|
2011-10-19 04:27:23 +02:00
|
|
|
if (lower_val == "x"
|
|
|
|
|| lower_val == "lon"
|
|
|
|
|| lower_val == "long"
|
|
|
|
|| (lower_val.find("longitude") != std::string::npos))
|
2011-10-15 05:28:23 +02:00
|
|
|
{
|
|
|
|
lon_idx = idx;
|
|
|
|
has_lon_field = true;
|
|
|
|
}
|
2011-10-19 04:27:23 +02:00
|
|
|
if (lower_val == "y"
|
|
|
|
|| lower_val == "lat"
|
|
|
|
|| (lower_val.find("latitude") != std::string::npos))
|
2011-10-15 05:28:23 +02:00
|
|
|
{
|
|
|
|
lat_idx = idx;
|
|
|
|
has_lat_field = true;
|
|
|
|
}
|
|
|
|
headers_.push_back(val);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
++line_number;
|
|
|
|
break;
|
2011-10-12 03:11:59 +02:00
|
|
|
}
|
|
|
|
}
|
2011-11-14 04:33:57 +01:00
|
|
|
catch(const std::exception & ex)
|
2011-10-15 05:28:23 +02:00
|
|
|
{
|
|
|
|
std::ostringstream s;
|
|
|
|
s << "CSV Plugin: error parsing headers: " << ex.what();
|
|
|
|
throw mapnik::datasource_exception(s.str());
|
|
|
|
}
|
2011-10-12 03:11:59 +02:00
|
|
|
}
|
2011-10-15 05:28:23 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
if (!has_wkt_field && (!has_lon_field || !has_lat_field) )
|
|
|
|
{
|
|
|
|
std::ostringstream s;
|
2011-10-19 03:21:19 +02:00
|
|
|
s << "CSV Plugin: could not detect column headers with the name of wkt ,x/y, or latitude/longitude - this is required for reading geometry data";
|
2011-10-15 05:28:23 +02:00
|
|
|
throw mapnik::datasource_exception(s.str());
|
|
|
|
}
|
|
|
|
|
2011-10-29 06:50:31 +02:00
|
|
|
int feature_count(1);
|
2011-10-15 05:28:23 +02:00
|
|
|
bool extent_initialized = false;
|
2012-01-17 19:34:08 +01:00
|
|
|
std::size_t num_headers = headers_.size();
|
|
|
|
|
2012-01-19 23:36:27 +01:00
|
|
|
ctx_ = boost::make_shared<mapnik::context_type>();
|
2012-01-17 19:34:08 +01:00
|
|
|
for (std::size_t i = 0; i < headers_.size(); ++i)
|
|
|
|
{
|
|
|
|
ctx_->push(headers_[i]);
|
|
|
|
}
|
|
|
|
|
2011-10-15 05:28:23 +02:00
|
|
|
mapnik::transcoder tr(desc_.get_encoding());
|
|
|
|
|
|
|
|
while (std::getline(stream,csv_line,newline))
|
|
|
|
{
|
|
|
|
if ((row_limit_ > 0) && (line_number > row_limit_))
|
2011-10-12 03:11:59 +02:00
|
|
|
{
|
2011-10-15 05:28:23 +02:00
|
|
|
#ifdef MAPNIK_DEBUG
|
|
|
|
std::clog << "CSV Plugin: row limit hit, exiting at feature: " << feature_count << "\n";
|
|
|
|
#endif
|
|
|
|
break;
|
|
|
|
}
|
2011-11-14 04:33:57 +01:00
|
|
|
|
2011-11-04 12:18:40 +01:00
|
|
|
unsigned line_length = csv_line.length();
|
2011-10-15 05:28:23 +02:00
|
|
|
|
2011-10-17 20:03:50 +02:00
|
|
|
// skip blank lines
|
2011-11-04 12:18:40 +01:00
|
|
|
if (line_length < 5)
|
|
|
|
{
|
|
|
|
std::string trimmed = csv_line;
|
|
|
|
boost::trim_if(trimmed,boost::algorithm::is_any_of("\",'\r\n"));
|
|
|
|
if (trimmed.empty()){
|
|
|
|
++line_number;
|
|
|
|
continue;
|
2011-10-17 20:03:50 +02:00
|
|
|
#ifdef MAPNIK_DEBUG
|
2011-11-04 12:18:40 +01:00
|
|
|
std::clog << "CSV Plugin: empty row encountered at line: " << line_number << "\n";
|
2011-10-17 20:03:50 +02:00
|
|
|
#endif
|
2011-11-04 12:18:40 +01:00
|
|
|
}
|
2011-10-17 20:03:50 +02:00
|
|
|
}
|
|
|
|
|
2011-10-15 05:28:23 +02:00
|
|
|
try
|
|
|
|
{
|
2011-10-17 20:18:44 +02:00
|
|
|
Tokenizer tok(csv_line, grammer);
|
|
|
|
Tokenizer::iterator beg = tok.begin();
|
2011-11-14 04:33:57 +01:00
|
|
|
|
2011-10-15 05:28:23 +02:00
|
|
|
// early return for strict mode
|
|
|
|
if (strict_)
|
2011-10-12 03:11:59 +02:00
|
|
|
{
|
2011-11-01 16:55:23 +01:00
|
|
|
unsigned num_fields = std::distance(beg,tok.end());
|
2011-10-15 05:28:23 +02:00
|
|
|
if (num_fields != num_headers)
|
|
|
|
{
|
|
|
|
std::ostringstream s;
|
|
|
|
s << "CSV Plugin: # of headers != # of values parsed for row " << line_number << "\n";
|
|
|
|
throw mapnik::datasource_exception(s.str());
|
|
|
|
}
|
2011-10-12 03:11:59 +02:00
|
|
|
}
|
2011-11-14 04:33:57 +01:00
|
|
|
|
2012-01-17 19:34:08 +01:00
|
|
|
mapnik::feature_ptr feature(mapnik::feature_factory::create(ctx_,feature_count));
|
2011-10-15 05:28:23 +02:00
|
|
|
double x(0);
|
|
|
|
double y(0);
|
|
|
|
bool parsed_x = false;
|
|
|
|
bool parsed_y = false;
|
|
|
|
bool parsed_wkt = false;
|
|
|
|
bool null_geom = false;
|
|
|
|
std::vector<std::string> collected;
|
2011-11-14 04:33:57 +01:00
|
|
|
|
2011-11-01 16:55:23 +01:00
|
|
|
for (unsigned i = 0; i < num_headers; ++i)
|
2011-10-12 03:11:59 +02:00
|
|
|
{
|
2011-11-01 00:09:29 +01:00
|
|
|
std::string fld_name(headers_.at(i));
|
|
|
|
collected.push_back(fld_name);
|
|
|
|
std::string value;
|
|
|
|
if (beg == tok.end())
|
2011-10-15 05:28:23 +02:00
|
|
|
{
|
2012-01-17 19:34:08 +01:00
|
|
|
feature->put(fld_name,tr.transcode(value.c_str()));
|
2011-11-04 12:18:40 +01:00
|
|
|
null_geom = true;
|
|
|
|
if (feature_count == 1)
|
|
|
|
{
|
|
|
|
desc_.add_descriptor(mapnik::attribute_descriptor(fld_name,mapnik::String));
|
|
|
|
}
|
2011-11-01 00:09:29 +01:00
|
|
|
continue;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
value = boost::trim_copy(*beg);
|
|
|
|
++beg;
|
2011-10-15 05:28:23 +02:00
|
|
|
}
|
2011-11-01 00:09:29 +01:00
|
|
|
|
2011-10-15 05:28:23 +02:00
|
|
|
int value_length = value.length();
|
2011-11-14 04:33:57 +01:00
|
|
|
|
2011-10-12 03:11:59 +02:00
|
|
|
// parse wkt
|
2011-10-15 05:28:23 +02:00
|
|
|
if (has_wkt_field)
|
2011-10-12 03:11:59 +02:00
|
|
|
{
|
2011-11-14 04:33:57 +01:00
|
|
|
if (i == wkt_idx)
|
|
|
|
{
|
|
|
|
// skip empty geoms
|
|
|
|
if (value.empty())
|
|
|
|
{
|
|
|
|
null_geom = true;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
// optimize simple "POINT (x y)"
|
|
|
|
// using this shaved 2 seconds off csv that took 8 seconds total to parse
|
|
|
|
if (value.find("POINT") == 0)
|
|
|
|
{
|
|
|
|
using boost::phoenix::ref;
|
|
|
|
using boost::spirit::qi::_1;
|
|
|
|
std::string::const_iterator str_beg = value.begin();
|
|
|
|
std::string::const_iterator str_end = value.end();
|
|
|
|
bool r = qi::phrase_parse(str_beg,str_end,
|
2012-02-02 02:37:35 +01:00
|
|
|
(
|
|
|
|
qi::lit("POINT") >> '('
|
|
|
|
>> double_[ref(x) = _1]
|
|
|
|
>> double_[ref(y) = _1] >> ')'
|
|
|
|
),
|
|
|
|
ascii::space);
|
2011-11-14 04:33:57 +01:00
|
|
|
|
|
|
|
if (r && (str_beg == str_end))
|
|
|
|
{
|
|
|
|
mapnik::geometry_type * pt = new mapnik::geometry_type(mapnik::Point);
|
|
|
|
pt->move_to(x,y);
|
|
|
|
feature->add_geometry(pt);
|
|
|
|
parsed_wkt = true;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
std::ostringstream s;
|
|
|
|
s << "CSV Plugin: expected well known text geometry: could not parse row "
|
|
|
|
<< line_number
|
|
|
|
<< ",column "
|
|
|
|
<< i << " - found: '"
|
|
|
|
<< value << "'";
|
|
|
|
if (strict_)
|
|
|
|
{
|
|
|
|
throw mapnik::datasource_exception(s.str());
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
if (!quiet_) std::clog << s.str() << "\n";
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
if (mapnik::from_wkt(value, feature->paths()))
|
|
|
|
{
|
|
|
|
parsed_wkt = true;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
std::ostringstream s;
|
|
|
|
s << "CSV Plugin: expected well known text geometry: could not parse row "
|
|
|
|
<< line_number
|
|
|
|
<< ",column "
|
|
|
|
<< i << " - found: '"
|
|
|
|
<< value << "'";
|
|
|
|
if (strict_)
|
|
|
|
{
|
|
|
|
throw mapnik::datasource_exception(s.str());
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
if (!quiet_) std::clog << s.str() << "\n";
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
2011-10-12 03:11:59 +02:00
|
|
|
}
|
2011-10-15 05:28:23 +02:00
|
|
|
else
|
2011-10-12 03:11:59 +02:00
|
|
|
{
|
2011-10-15 05:28:23 +02:00
|
|
|
// longitude
|
|
|
|
if (i == lon_idx)
|
2011-10-12 03:11:59 +02:00
|
|
|
{
|
2011-10-15 05:28:23 +02:00
|
|
|
// skip empty geoms
|
|
|
|
if (value.empty())
|
|
|
|
{
|
|
|
|
null_geom = true;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
2011-11-14 04:33:57 +01:00
|
|
|
try
|
2011-10-15 05:28:23 +02:00
|
|
|
{
|
|
|
|
x = boost::lexical_cast<double>(value);
|
|
|
|
parsed_x = true;
|
|
|
|
}
|
2011-11-14 04:33:57 +01:00
|
|
|
catch(boost::bad_lexical_cast & ex)
|
2011-10-15 05:28:23 +02:00
|
|
|
{
|
|
|
|
std::ostringstream s;
|
|
|
|
s << "CSV Plugin: expected a float value for longitude: could not parse row "
|
|
|
|
<< line_number
|
|
|
|
<< ", column "
|
|
|
|
<< i << " - found: '"
|
|
|
|
<< value << "'";
|
|
|
|
if (strict_)
|
|
|
|
{
|
2011-11-14 04:33:57 +01:00
|
|
|
throw mapnik::datasource_exception(s.str());
|
2011-10-15 05:28:23 +02:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
if (!quiet_) std::clog << s.str() << "\n";
|
|
|
|
}
|
|
|
|
}
|
2011-10-12 03:11:59 +02:00
|
|
|
}
|
2011-10-15 05:28:23 +02:00
|
|
|
// latitude
|
|
|
|
else if (i == lat_idx)
|
2011-10-12 03:11:59 +02:00
|
|
|
{
|
2011-10-15 05:28:23 +02:00
|
|
|
// skip empty geoms
|
|
|
|
if (value.empty())
|
|
|
|
{
|
|
|
|
null_geom = true;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
2011-11-14 04:33:57 +01:00
|
|
|
try
|
2011-10-15 05:28:23 +02:00
|
|
|
{
|
|
|
|
y = boost::lexical_cast<double>(value);
|
|
|
|
parsed_y = true;
|
|
|
|
}
|
2011-11-14 04:33:57 +01:00
|
|
|
catch(boost::bad_lexical_cast & ex)
|
2011-10-15 05:28:23 +02:00
|
|
|
{
|
|
|
|
std::ostringstream s;
|
|
|
|
s << "CSV Plugin: expected a float value for latitude: could not parse row "
|
|
|
|
<< line_number
|
|
|
|
<< ", column "
|
|
|
|
<< i << " - found: '"
|
|
|
|
<< value << "'";
|
|
|
|
if (strict_)
|
|
|
|
{
|
2011-11-14 04:33:57 +01:00
|
|
|
throw mapnik::datasource_exception(s.str());
|
2011-10-15 05:28:23 +02:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
if (!quiet_) std::clog << s.str() << "\n";
|
|
|
|
}
|
|
|
|
}
|
2011-10-12 03:11:59 +02:00
|
|
|
}
|
|
|
|
}
|
2011-11-14 04:33:57 +01:00
|
|
|
|
2011-12-13 00:55:33 +01:00
|
|
|
// now, add all values as attributes
|
|
|
|
/* First we detect likely strings, then try parsing likely numbers,
|
|
|
|
finally falling back to string type
|
|
|
|
* We intentionally do not try to detect boolean or null types
|
2012-02-02 02:37:35 +01:00
|
|
|
since they are not common in csv
|
2011-12-13 00:55:33 +01:00
|
|
|
* Likely strings are either empty values, very long values
|
2012-02-02 02:37:35 +01:00
|
|
|
or value with leading zeros like 001 (which are not safe
|
|
|
|
to assume are numbers)
|
2011-12-13 00:55:33 +01:00
|
|
|
*/
|
|
|
|
|
|
|
|
bool has_dot = value.find(".") != std::string::npos;
|
|
|
|
if (value.empty() ||
|
2012-02-02 02:37:35 +01:00
|
|
|
(value_length > 20) ||
|
|
|
|
(value_length > 1 && !has_dot && value[0] == '0'))
|
2011-10-12 03:11:59 +02:00
|
|
|
{
|
2012-01-17 19:34:08 +01:00
|
|
|
feature->put(fld_name,tr.transcode(value.c_str()));
|
2011-11-01 00:09:29 +01:00
|
|
|
if (feature_count == 1)
|
2011-11-02 01:48:30 +01:00
|
|
|
{
|
2011-11-01 00:09:29 +01:00
|
|
|
desc_.add_descriptor(mapnik::attribute_descriptor(fld_name,mapnik::String));
|
2011-11-02 01:48:30 +01:00
|
|
|
}
|
2011-10-12 03:11:59 +02:00
|
|
|
}
|
2011-10-15 05:28:23 +02:00
|
|
|
else if ((value[0] >= '0' && value[0] <= '9') || value[0] == '-')
|
2011-10-12 03:11:59 +02:00
|
|
|
{
|
2011-10-15 05:28:23 +02:00
|
|
|
double float_val = 0.0;
|
|
|
|
std::string::const_iterator str_beg = value.begin();
|
|
|
|
std::string::const_iterator str_end = value.end();
|
|
|
|
bool r = qi::phrase_parse(str_beg,str_end,qi::double_,ascii::space,float_val);
|
2011-11-01 00:09:29 +01:00
|
|
|
if (r && (str_beg == str_end))
|
2011-10-12 03:11:59 +02:00
|
|
|
{
|
2011-12-13 00:55:33 +01:00
|
|
|
if (has_dot)
|
2011-10-12 03:11:59 +02:00
|
|
|
{
|
2012-01-17 19:34:08 +01:00
|
|
|
feature->put(fld_name,float_val);
|
2011-10-29 06:50:31 +02:00
|
|
|
if (feature_count == 1)
|
2011-11-02 01:48:30 +01:00
|
|
|
{
|
2011-12-13 00:55:33 +01:00
|
|
|
desc_.add_descriptor(
|
|
|
|
mapnik::attribute_descriptor(
|
|
|
|
fld_name,mapnik::Double));
|
2011-11-02 01:48:30 +01:00
|
|
|
}
|
2011-10-12 03:11:59 +02:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
2012-01-17 19:34:08 +01:00
|
|
|
feature->put(fld_name,static_cast<int>(float_val));
|
2011-10-29 06:50:31 +02:00
|
|
|
if (feature_count == 1)
|
2011-11-02 01:48:30 +01:00
|
|
|
{
|
2011-12-13 00:55:33 +01:00
|
|
|
desc_.add_descriptor(
|
|
|
|
mapnik::attribute_descriptor(
|
|
|
|
fld_name,mapnik::Integer));
|
2011-11-02 01:48:30 +01:00
|
|
|
}
|
2011-10-12 03:11:59 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
2011-10-15 05:28:23 +02:00
|
|
|
// fallback to normal string
|
2012-01-17 19:34:08 +01:00
|
|
|
feature->put(fld_name,tr.transcode(value.c_str()));
|
2011-10-29 06:50:31 +02:00
|
|
|
if (feature_count == 1)
|
2011-11-02 01:48:30 +01:00
|
|
|
{
|
2011-12-13 00:55:33 +01:00
|
|
|
desc_.add_descriptor(
|
|
|
|
mapnik::attribute_descriptor(
|
|
|
|
fld_name,mapnik::String));
|
2011-11-02 01:48:30 +01:00
|
|
|
}
|
2011-10-15 05:28:23 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
2011-11-02 16:07:59 +01:00
|
|
|
// fallback to normal string
|
2012-01-17 19:34:08 +01:00
|
|
|
feature->put(fld_name,tr.transcode(value.c_str()));
|
2011-11-02 16:07:59 +01:00
|
|
|
if (feature_count == 1)
|
2011-10-15 05:28:23 +02:00
|
|
|
{
|
2011-12-13 00:55:33 +01:00
|
|
|
desc_.add_descriptor(
|
|
|
|
mapnik::attribute_descriptor(
|
|
|
|
fld_name,mapnik::String));
|
2011-10-12 03:11:59 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
2011-11-14 04:33:57 +01:00
|
|
|
|
2011-11-01 00:09:29 +01:00
|
|
|
if (null_geom)
|
2011-10-15 05:28:23 +02:00
|
|
|
{
|
|
|
|
++line_number;
|
|
|
|
std::ostringstream s;
|
|
|
|
s << "CSV Plugin: null geometry encountered for line "
|
|
|
|
<< line_number;
|
|
|
|
if (strict_)
|
|
|
|
{
|
|
|
|
throw mapnik::datasource_exception(s.str());
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
if (!quiet_) std::clog << s.str() << "\n";
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
}
|
2011-11-14 04:33:57 +01:00
|
|
|
|
2011-10-12 03:11:59 +02:00
|
|
|
if (has_wkt_field)
|
|
|
|
{
|
|
|
|
if (parsed_wkt)
|
|
|
|
{
|
2011-10-15 05:28:23 +02:00
|
|
|
if (!extent_initialized)
|
2011-10-12 03:11:59 +02:00
|
|
|
{
|
|
|
|
extent_initialized = true;
|
|
|
|
extent_ = feature->envelope();
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
extent_.expand_to_include(feature->envelope());
|
|
|
|
}
|
|
|
|
features_.push_back(feature);
|
2011-10-29 06:50:31 +02:00
|
|
|
++feature_count;
|
2011-10-12 03:11:59 +02:00
|
|
|
}
|
2011-10-15 05:28:23 +02:00
|
|
|
else
|
2011-10-12 03:11:59 +02:00
|
|
|
{
|
|
|
|
std::ostringstream s;
|
2011-10-15 05:28:23 +02:00
|
|
|
s << "CSV Plugin: could not read WKT geometry "
|
2011-11-14 04:33:57 +01:00
|
|
|
<< "for line " << line_number << " - found " << headers_.size()
|
2011-10-15 05:28:23 +02:00
|
|
|
<< " with values like: " << csv_line << "\n";
|
|
|
|
if (strict_)
|
|
|
|
{
|
|
|
|
throw mapnik::datasource_exception(s.str());
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
if (!quiet_) std::clog << s.str() << "\n";
|
|
|
|
continue;
|
|
|
|
}
|
2011-10-12 03:11:59 +02:00
|
|
|
}
|
2011-10-15 05:28:23 +02:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
if (parsed_x && parsed_y)
|
2011-10-12 03:11:59 +02:00
|
|
|
{
|
2011-10-15 05:28:23 +02:00
|
|
|
mapnik::geometry_type * pt = new mapnik::geometry_type(mapnik::Point);
|
|
|
|
pt->move_to(x,y);
|
|
|
|
feature->add_geometry(pt);
|
|
|
|
features_.push_back(feature);
|
|
|
|
++feature_count;
|
2011-11-14 04:33:57 +01:00
|
|
|
|
2011-10-15 05:28:23 +02:00
|
|
|
if (!extent_initialized)
|
|
|
|
{
|
|
|
|
extent_initialized = true;
|
|
|
|
extent_ = feature->envelope();
|
2011-11-14 04:33:57 +01:00
|
|
|
|
2011-10-15 05:28:23 +02:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
extent_.expand_to_include(feature->envelope());
|
|
|
|
}
|
2011-10-12 03:11:59 +02:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
2011-10-15 05:28:23 +02:00
|
|
|
std::ostringstream s;
|
|
|
|
if (!parsed_x)
|
2011-10-12 03:11:59 +02:00
|
|
|
{
|
2011-10-15 05:28:23 +02:00
|
|
|
s << "CSV Plugin: does your csv have valid headers?\n"
|
|
|
|
<< "Could not detect or parse any rows named 'x' or 'longitude' "
|
2011-11-14 04:33:57 +01:00
|
|
|
<< "for line " << line_number << " but found " << headers_.size()
|
2011-10-15 05:28:23 +02:00
|
|
|
<< " with values like: " << csv_line << "\n"
|
|
|
|
<< "for: " << boost::algorithm::join(collected, ",") << "\n";
|
|
|
|
}
|
|
|
|
if (!parsed_y)
|
|
|
|
{
|
|
|
|
s << "CSV Plugin: does your csv have valid headers?\n"
|
|
|
|
<< "Could not detect or parse any rows named 'y' or 'latitude' "
|
2011-11-14 04:33:57 +01:00
|
|
|
<< "for line " << line_number << " but found " << headers_.size()
|
2011-10-15 05:28:23 +02:00
|
|
|
<< " with values like: " << csv_line << "\n"
|
|
|
|
<< "for: " << boost::algorithm::join(collected, ",") << "\n";
|
|
|
|
}
|
|
|
|
if (strict_)
|
|
|
|
{
|
|
|
|
throw mapnik::datasource_exception(s.str());
|
2011-10-12 03:11:59 +02:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
2011-10-15 05:28:23 +02:00
|
|
|
if (!quiet_) std::clog << s.str() << "\n";
|
|
|
|
continue;
|
2011-10-12 03:11:59 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
2011-10-15 05:28:23 +02:00
|
|
|
++line_number;
|
|
|
|
}
|
2011-11-14 04:33:57 +01:00
|
|
|
catch(const mapnik::datasource_exception & ex )
|
2011-11-01 00:09:29 +01:00
|
|
|
{
|
|
|
|
if (strict_)
|
|
|
|
{
|
|
|
|
throw mapnik::datasource_exception(ex.what());
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
if (!quiet_) std::clog << ex.what() << "\n";
|
|
|
|
}
|
|
|
|
}
|
2011-11-14 04:33:57 +01:00
|
|
|
catch(const std::exception & ex)
|
2011-10-15 05:28:23 +02:00
|
|
|
{
|
|
|
|
std::ostringstream s;
|
|
|
|
s << "CSV Plugin: unexpected error parsing line: " << line_number
|
|
|
|
<< " - found " << headers_.size() << " with values like: " << csv_line << "\n"
|
|
|
|
<< " and got error like: " << ex.what();
|
|
|
|
if (strict_)
|
|
|
|
{
|
|
|
|
throw mapnik::datasource_exception(s.str());
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
if (!quiet_) std::clog << s.str() << "\n";
|
|
|
|
}
|
2011-10-12 03:11:59 +02:00
|
|
|
}
|
|
|
|
}
|
2011-10-29 06:50:31 +02:00
|
|
|
if (!feature_count > 0)
|
|
|
|
{
|
|
|
|
if (!quiet_) std::clog << "CSV Plugin: could not parse any lines of data\n";
|
|
|
|
}
|
2011-10-12 03:11:59 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
std::string csv_datasource::name()
|
|
|
|
{
|
|
|
|
return "csv";
|
|
|
|
}
|
|
|
|
|
2012-01-17 07:09:46 +01:00
|
|
|
datasource::datasource_t csv_datasource::type() const
|
2011-10-12 03:11:59 +02:00
|
|
|
{
|
|
|
|
return datasource::Vector;
|
|
|
|
}
|
|
|
|
|
|
|
|
mapnik::box2d<double> csv_datasource::envelope() const
|
|
|
|
{
|
|
|
|
if (!is_bound_) bind();
|
|
|
|
|
|
|
|
return extent_;
|
|
|
|
}
|
|
|
|
|
2012-01-17 07:09:46 +01:00
|
|
|
boost::optional<mapnik::datasource::geometry_t> csv_datasource::get_geometry_type() const
|
2012-01-15 07:35:40 +01:00
|
|
|
{
|
2012-01-17 07:09:46 +01:00
|
|
|
if (! is_bound_) bind();
|
|
|
|
boost::optional<mapnik::datasource::geometry_t> result;
|
2012-01-15 07:35:40 +01:00
|
|
|
int multi_type = 0;
|
|
|
|
unsigned num_features = features_.size();
|
2012-02-13 20:01:58 +01:00
|
|
|
for (unsigned i = 0; i < num_features && i < 5; ++i)
|
2012-01-15 07:35:40 +01:00
|
|
|
{
|
2012-01-17 07:09:46 +01:00
|
|
|
mapnik::util::to_ds_type(features_[i]->paths(),result);
|
|
|
|
if (result)
|
2012-01-15 07:35:40 +01:00
|
|
|
{
|
2012-01-17 07:09:46 +01:00
|
|
|
int type = static_cast<int>(*result);
|
|
|
|
if (multi_type > 0 && multi_type != type)
|
|
|
|
{
|
|
|
|
result.reset(mapnik::datasource::Collection);
|
|
|
|
return result;
|
|
|
|
}
|
|
|
|
multi_type = type;
|
2012-01-15 07:35:40 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
return result;
|
|
|
|
}
|
|
|
|
|
2011-10-12 03:11:59 +02:00
|
|
|
mapnik::layer_descriptor csv_datasource::get_descriptor() const
|
|
|
|
{
|
|
|
|
if (!is_bound_) bind();
|
2011-11-14 04:33:57 +01:00
|
|
|
|
2011-10-12 03:11:59 +02:00
|
|
|
return desc_;
|
|
|
|
}
|
|
|
|
|
|
|
|
mapnik::featureset_ptr csv_datasource::features(mapnik::query const& q) const
|
|
|
|
{
|
|
|
|
if (!is_bound_) bind();
|
2011-11-14 04:33:57 +01:00
|
|
|
|
2011-12-05 21:03:38 +01:00
|
|
|
const std::set<std::string>& attribute_names = q.property_names();
|
|
|
|
std::set<std::string>::const_iterator pos = attribute_names.begin();
|
|
|
|
while (pos != attribute_names.end())
|
|
|
|
{
|
|
|
|
bool found_name = false;
|
2012-01-17 19:34:08 +01:00
|
|
|
for (std::size_t i = 0; i < headers_.size(); ++i)
|
2011-12-05 21:03:38 +01:00
|
|
|
{
|
|
|
|
if (headers_[i] == *pos)
|
|
|
|
{
|
|
|
|
found_name = true;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (! found_name)
|
|
|
|
{
|
|
|
|
std::ostringstream s;
|
|
|
|
|
|
|
|
s << "CSV Plugin: no attribute '" << *pos << "'. Valid attributes are: "
|
|
|
|
<< boost::algorithm::join(headers_, ",") << ".";
|
|
|
|
|
|
|
|
throw mapnik::datasource_exception(s.str());
|
|
|
|
}
|
|
|
|
++pos;
|
|
|
|
}
|
2011-11-14 04:33:57 +01:00
|
|
|
|
2011-10-12 03:11:59 +02:00
|
|
|
return boost::make_shared<mapnik::memory_featureset>(q.get_bbox(),features_);
|
|
|
|
}
|
|
|
|
|
|
|
|
mapnik::featureset_ptr csv_datasource::features_at_point(mapnik::coord2d const& pt) const
|
|
|
|
{
|
|
|
|
if (!is_bound_) bind();
|
2011-11-14 04:33:57 +01:00
|
|
|
|
2011-10-12 03:11:59 +02:00
|
|
|
throw mapnik::datasource_exception("CSV Plugin: features_at_point is not supported yet");
|
|
|
|
}
|