[poppler] poppler/gilib: poppler_page_find_text() returns wrong PopplerRectangle

[Kenji Okimoto]

Hi,

I'm parsing some PDF files by following code:

Use this PDF, for example.
https://www.sedl.org/afterschool/toolkits/science/pdf/ast_sci_data_tables_sample.pdf

#include <stdio.h>
#include <stdlib.h>
#include <glib/poppler.h>

int main(void)
{
  GError *error = NULL;
  PopplerDocument *doc =
poppler_document_new_from_file("file:///path/to/ast_sci_data_tables_sample.pdf",
NULL, &error);
  if (!doc) {
    printf("%s\n", error->message);
    return 0;
  }
  PopplerPage *page = poppler_document_get_page(doc, 1);
  // char *content = poppler_page_get_text(page);
  char *content;
  double height;
  poppler_page_get_size(page, NULL, &height);
  GList *list = poppler_page_find_text(page, "Wingfoot Express");
  for (GList *node = list; node != NULL; node = node->next) {
    PopplerRectangle *rec = (PopplerRectangle *)node->data;
    content = poppler_page_get_text_for_area(page, rec);
    printf("%s\n", content); // Displays unexpected wrong text
    printf("x1=%f, y1=%f, x2=%f, y2=%f\n", rec->x1, rec->y1, rec->x2,
rec->y2);
    rec->y1 = height - rec->y1;
    rec->y2 = height - rec->y2;
    content = poppler_page_get_text_for_area(page, rec);
    printf("%s\n", content); // Displays expected text
    printf("x1=%f, y1=%f, x2=%f, y2=%f\n", rec->x1, rec->y1, rec->x2,
rec->y2);
  }
  return 0;
}

I'm confusing that I cannot use PopplerRectangle returned by
poppler_page_find_text() with poppler_page_get_text_for_area().

I think that poppler_page_get_text_for_area() should return proper text
without editing poppler_page_find_text() result.

In C++ version,
I can use poppler::rectangle returned by page->search() with page->text().

#include <iostream>
#include <cstdlib>
#include <cstring>
#include <memory>
#include <iostream>
#include <sstream>

#include <poppler-document.h>
#include <poppler-page.h>
#include <poppler-rectangle.h>

int main(int argx, char **argv)
{
  const std::string path("/path/to/ast_sci_data_tables_sample.pdf");
  poppler::document *doc = poppler::document::load_from_file(path);
  poppler::page *page = doc->create_page(1);
  poppler::rectangle <double>r;
  poppler::ustring text = poppler::ustring::from_latin1("Wingfoot Express");
  bool matched = page->search(text, r,

poppler::page::search_direction_enum::search_from_top,

poppler::case_sensitivity_enum::case_sensitive);
  std::cout << text.to_latin1() << std::endl
            << matched << std::endl
            << r.x() << "," << r.y() << std::endl
            << r.left() << "," << r.right() << std::endl
            << r.top() << "," << r.bottom() << std::endl
            << r.width() << "," << r.height() << std::endl
            << r << std::endl;
  poppler::ustring t = page->text(r);
  std::cout << t.to_latin1() << std::endl;
  return 0;
}

Thanks

-- 
Kenji Okimoto <okimoto at clear-code.com>