From bf152707b305359f69638f2c572cbba981077cff Mon Sep 17 00:00:00 2001
From: Kovid Goyal <kovid@kovidgoyal.net>
Date: Tue, 17 Mar 2015 09:36:46 +0530
Subject: [PATCH] Add API to count images in a PDF file

---
 src/calibre/utils/podofo/__init__.py |  8 ++++++++
 src/calibre/utils/podofo/doc.cpp     | 28 ++++++++++++++++++++++++++++
 2 files changed, 36 insertions(+)

diff --git a/src/calibre/utils/podofo/__init__.py b/src/calibre/utils/podofo/__init__.py
index ee7a02a4be..4aa708ae66 100644
--- a/src/calibre/utils/podofo/__init__.py
+++ b/src/calibre/utils/podofo/__init__.py
@@ -119,6 +119,14 @@ def get_xmp_metadata(path):
     p.load(raw)
     return p.get_xmp_metadata()
 
+def get_image_count(path):
+    podofo = get_podofo()
+    p = podofo.PDFDoc()
+    with open(path, 'rb') as f:
+        raw = f.read()
+    p.load(raw)
+    return p.image_count()
+
 def test_outline(src):
     podofo = get_podofo()
     p = podofo.PDFDoc()
diff --git a/src/calibre/utils/podofo/doc.cpp b/src/calibre/utils/podofo/doc.cpp
index edecf44ba8..b33647336c 100644
--- a/src/calibre/utils/podofo/doc.cpp
+++ b/src/calibre/utils/podofo/doc.cpp
@@ -141,6 +141,31 @@ PDFDoc_page_count(PDFDoc *self, PyObject *args) {
     return Py_BuildValue("i", count);
 } // }}}
 
+// image_count() {{{
+static PyObject *
+PDFDoc_image_count(PDFDoc *self, PyObject *args) {
+    int count = 0;
+    const PdfObject* obj_type = NULL;
+    const PdfObject* obj_sub_type = NULL;
+    try {
+        TCIVecObjects it = self->doc->GetObjects().begin();
+         while( it != self->doc->GetObjects().end() ) {
+             if( (*it)->IsDictionary() ) {
+                 obj_type = (*it)->GetDictionary().GetKey( PdfName::KeyType );
+                 obj_sub_type = (*it)->GetDictionary().GetKey( PdfName::KeySubtype );
+                 if( ( obj_type && obj_type->IsName() && ( obj_type->GetName().GetName() == "XObject" ) ) ||
+                        ( obj_sub_type && obj_sub_type->IsName() && ( obj_sub_type->GetName().GetName() == "Image" ) ) ) count++;
+                 self->doc->FreeObjectMemory( *it );
+             }
+             it++;
+         }
+    } catch(const PdfError & err) {
+        podofo_set_exception(err);
+        return NULL;
+    }
+    return Py_BuildValue("i", count);
+} // }}}
+
 // delete_page {{{
 static PyObject *
 PDFDoc_delete_page(PDFDoc *self, PyObject *args) {
@@ -539,6 +564,9 @@ static PyMethodDef PDFDoc_methods[] = {
     {"page_count", (PyCFunction)PDFDoc_page_count, METH_VARARGS,
      "page_count() -> Number of pages in the PDF."
     },
+    {"image_count", (PyCFunction)PDFDoc_image_count, METH_VARARGS,
+     "image_count() -> Number of images in the PDF."
+    },
     {"delete_page", (PyCFunction)PDFDoc_delete_page, METH_VARARGS,
      "delete_page(page_num) -> Delete the specified page from the pdf (0 is the first page)."
     },