glib: Add poppler_page_find_text_with_options

To be able to search text with options like case sensitive, search
backwards and whole words only.

https://bugs.freedesktop.org/show_bug.cgi?id=2951

4 files changed, 62 insertions, 13 deletions

diff --git a/glib/poppler-page.cc b/glib/poppler-page.cc
index 90f9a621..8113e9ca 100644
--- a/glib/poppler-page.cc
+++ b/glib/poppler-page.cc
@@ -850,18 +850,23 @@ poppler_page_get_text (PopplerPage *page)
 }
 
 /**
- * poppler_page_find_text:
+ * poppler_page_find_text_with_options:
  * @page: a #PopplerPage
  * @text: the text to search for (UTF-8 encoded)
- * 
- * A #GList of rectangles for each occurance of the text on the page.
+ * @options: find options
+ *
+ * Finds @text in @page with the given #PopplerFindFlags options and
+ * returns a #GList of rectangles for each occurance of the text on the page.
  * The coordinates are in PDF points.
- * 
+ *
  * Return value: (element-type PopplerRectangle) (transfer full): a #GList of #PopplerRectangle,
+ *
+ * Since: 0.22
  **/
 GList *
-poppler_page_find_text (PopplerPage *page,
-			const char  *text)
+poppler_page_find_text_with_options (PopplerPage     *page,
+                                     const char      *text,
+                                     PopplerFindFlags options)
 {
   PopplerRectangle *match;
   GList *matches;
@@ -870,6 +875,7 @@ poppler_page_find_text (PopplerPage *page,
   glong ucs4_len;
   double height;
   TextPage *text_dev;
+  gboolean backwards;
 
   g_return_val_if_fail (POPPLER_IS_PAGE (page), NULL);
   g_return_val_if_fail (text != NULL, NULL);
@@ -878,17 +884,19 @@ poppler_page_find_text (PopplerPage *page,
 
   ucs4 = g_utf8_to_ucs4_fast (text, -1, &ucs4_len);
   poppler_page_get_size (page, NULL, &height);
-  
+
+  backwards = options & POPPLER_FIND_BACKWARDS;
   matches = NULL;
   xMin = 0;
-  yMin = 0;
+  yMin = backwards ? height : 0;
 
   while (text_dev->findText (ucs4, ucs4_len,
-			     gFalse, gTrue, // startAtTop, stopAtBottom
-			     gFalse, gFalse, // startAtLast, stopAtLast
-			     gFalse, gFalse, // caseSensitive, backwards
-			     gFalse, // wholeWord
-			     &xMin, &yMin, &xMax, &yMax))
+                             gFalse, gTrue, // startAtTop, stopAtBottom
+                             gTrue, gFalse, // startAtLast, stopAtLast
+                             options & POPPLER_FIND_CASE_SENSITIVE,
+                             backwards,
+                             options & POPPLER_FIND_WHOLE_WORDS_ONLY,
+                             &xMin, &yMin, &xMax, &yMax))
     {
       match = poppler_rectangle_new ();
       match->x1 = xMin;
@@ -903,6 +911,24 @@ poppler_page_find_text (PopplerPage *page,
   return g_list_reverse (matches);
 }
 
+/**
+ * poppler_page_find_text:
+ * @page: a #PopplerPage
+ * @text: the text to search for (UTF-8 encoded)
+ *
+ * Finds @text in @page with the default options (%POPPLER_FIND_DEFAULT) and
+ * returns a #GList of rectangles for each occurance of the text on the page.
+ * The coordinates are in PDF points.
+ *
+ * Return value: (element-type PopplerRectangle) (transfer full): a #GList of #PopplerRectangle,
+ **/
+GList *
+poppler_page_find_text (PopplerPage *page,
+			const char  *text)
+{
+  return poppler_page_find_text_with_options (page, text, POPPLER_FIND_DEFAULT);
+}
+
 static CairoImageOutputDev *
 poppler_page_get_image_output_dev (PopplerPage *page,
 				   GBool (*imgDrawDeviceCbk)(int img_id, void *data),
diff --git a/glib/poppler-page.h b/glib/poppler-page.h
index 6064ac5c..c081b8cc 100644
--- a/glib/poppler-page.h
+++ b/glib/poppler-page.h
@@ -60,6 +60,9 @@ PopplerPageTransition *poppler_page_get_transition       (PopplerPage        *pa
 gboolean               poppler_page_get_thumbnail_size   (PopplerPage        *page,
 							  int                *width,
 							  int                *height);
+GList             *poppler_page_find_text_with_options   (PopplerPage        *page,
+							  const  char        *text,
+							  PopplerFindFlags    options);
 GList     	      *poppler_page_find_text            (PopplerPage        *page,
 							  const  char        *text);
 void                   poppler_page_render_to_ps         (PopplerPage        *page,
diff --git a/glib/poppler.h b/glib/poppler.h
index 6c2eefd8..2d190f35 100644
--- a/glib/poppler.h
+++ b/glib/poppler.h
@@ -153,6 +153,24 @@ typedef enum /*< flags >*/
   POPPLER_PRINT_ALL               = POPPLER_PRINT_MARKUP_ANNOTS
 } PopplerPrintFlags;
 
+/**
+ * PopplerFindFlags:
+ * @POPPLER_FIND_CASE_SENSITIVE: do case sensitive search
+ * @POPPLER_FIND_BACKWARDS: search backwards
+ * @POPPLER_FIND_WHOLE_WORDS_ONLY: search only whole words
+ *
+ * Flags using while searching text in a page
+ *
+ * Since: 0.22
+ */
+typedef enum /*< flags >*/
+{
+  POPPLER_FIND_DEFAULT          = 0,
+  POPPLER_FIND_CASE_SENSITIVE   = 1 << 0,
+  POPPLER_FIND_BACKWARDS        = 1 << 1,
+  POPPLER_FIND_WHOLE_WORDS_ONLY = 1 << 2
+} PopplerFindFlags;
+
 typedef struct _PopplerDocument            PopplerDocument;
 typedef struct _PopplerIndexIter           PopplerIndexIter;
 typedef struct _PopplerFontsIter           PopplerFontsIter;
diff --git a/glib/reference/poppler-sections.txt b/glib/reference/poppler-sections.txt
index 76abf4c7..6fb14bca 100644
--- a/glib/reference/poppler-sections.txt
+++ b/glib/reference/poppler-sections.txt
@@ -33,6 +33,7 @@ poppler_page_get_selection_region
 poppler_page_selection_region_free
 poppler_page_get_selected_text
 poppler_page_find_text
+poppler_page_find_text_with_options
 poppler_page_get_text
 poppler_page_get_text_layout
 poppler_page_get_text_attributes
@@ -338,6 +339,7 @@ PopplerOrientation
 PopplerBackend
 PopplerColor
 PopplerPrintFlags
+PopplerFindFlags
 poppler_get_backend
 poppler_get_version
 poppler_date_parse