Add code to create a PDF outline using PoDoFo

This commit is contained in:
Kovid Goyal 2012-08-28 16:42:08 +05:30
parent 6c5d581218
commit 7d0fcd9484
6 changed files with 210 additions and 4 deletions

View File

@ -140,6 +140,7 @@ extensions = [
[
'calibre/utils/podofo/utils.cpp',
'calibre/utils/podofo/doc.cpp',
'calibre/utils/podofo/outline.cpp',
'calibre/utils/podofo/podofo.cpp',
],
headers=[

View File

@ -98,7 +98,23 @@ def delete_all_but(path, pages):
with open(path, 'wb') as f:
f.write(raw)
if __name__ == '__main__':
f = u'/tmp/t.pdf'
delete_all_but(f, [0, 1, -2, -1])
def test_outline(src):
podofo = get_podofo()
p = podofo.PDFDoc()
with open(src, 'rb') as f:
raw = f.read()
p.load(raw)
total = p.page_count()
root = p.create_outline(u'Table of Contents')
for i in xrange(0, total):
root.create(u'Page %d'%i, i, True)
raw = p.write()
out = '/tmp/outlined.pdf'
with open(out, 'wb') as f:
f.write(raw)
print 'Outlined PDF:', out
if __name__ == '__main__':
import sys
test_outline(sys.argv[-1])

View File

@ -194,6 +194,41 @@ PDFDoc_set_box(PDFDoc *self, PyObject *args) {
Py_RETURN_NONE;
} // }}}
// create_outline() {{{
static PyObject *
PDFDoc_create_outline(PDFDoc *self, PyObject *args) {
PyObject *p;
PDFOutlineItem *ans;
PdfString *title;
if (!PyArg_ParseTuple(args, "U", &p)) return NULL;
title = podofo_convert_pystring(p);
if (title == NULL) return NULL;
ans = PyObject_New(PDFOutlineItem, &PDFOutlineItemType);
if (ans == NULL) goto error;
try {
PdfOutlines *outlines = self->doc->GetOutlines();
if (outlines == NULL) {PyErr_NoMemory(); goto error;}
ans->item = outlines->CreateRoot(*title);
if (ans->item == NULL) {PyErr_NoMemory(); goto error;}
ans->doc = self->doc;
} catch(const PdfError & err) {
podofo_set_exception(err); goto error;
} catch (...) {
PyErr_SetString(PyExc_ValueError, "An unknown error occurred while trying to create the outline");
goto error;
}
delete title;
return (PyObject*)ans;
error:
Py_XDECREF(ans); delete title;
return NULL;
} // }}}
// Properties {{{
static PyObject *
@ -430,7 +465,9 @@ static PyMethodDef PDFDoc_methods[] = {
{"set_box", (PyCFunction)PDFDoc_set_box, METH_VARARGS,
"set_box(page_num, box, left, bottom, width, height) -> Set the PDF bounding box for the page numbered nu, box must be one of: MediaBox, CropBox, TrimBox, BleedBox, ArtBox. The numbers are interpreted as pts."
},
{"create_outline", (PyCFunction)PDFDoc_create_outline, METH_VARARGS,
"create_outline(title) -> Create an outline, return the root outline item."
},
{NULL} /* Sentinel */
};

View File

@ -26,7 +26,14 @@ typedef struct {
} PDFDoc;
typedef struct {
PyObject_HEAD
PdfMemDocument *doc;
PdfOutlineItem *item;
} PDFOutlineItem;
extern PyTypeObject PDFDocType;
extern PyTypeObject PDFOutlineItemType;
extern PyObject *Error;
// Utilities

View File

@ -0,0 +1,142 @@
/*
* outline.cpp
* Copyright (C) 2012 Kovid Goyal <kovid at kovidgoyal.net>
*
* Distributed under terms of the GPL3 license.
*/
#include "global.h"
using namespace pdf;
// Constructor/destructor {{{
static void
dealloc(PDFOutlineItem* self)
{
self->ob_type->tp_free((PyObject*)self);
}
static PyObject *
new_item(PyTypeObject *type, PyObject *args, PyObject *kwds)
{
PDFOutlineItem *self;
self = (PDFOutlineItem *)type->tp_alloc(type, 0);
if (self != NULL) {
self->item = NULL;
}
return (PyObject *)self;
}
// }}}
// erase() {{{
static PyObject *
erase(PDFOutlineItem *self, PyObject *args) {
try {
self->item->Erase();
} catch(const PdfError & err) {
podofo_set_exception(err);
return NULL;
}
Py_RETURN_NONE;
} // }}}
static PyObject *
create(PDFOutlineItem *self, PyObject *args) {
PyObject *ptitle, *as_child = NULL;
PDFOutlineItem *ans;
int num;
PdfString *title;
PdfPage *page;
if (!PyArg_ParseTuple(args, "Ui|O", &ptitle, &num, &as_child)) return NULL;
title = podofo_convert_pystring(ptitle);
if (title == NULL) return NULL;
ans = PyObject_New(PDFOutlineItem, &PDFOutlineItemType);
if (ans == NULL) goto error;
ans->doc = self->doc;
try {
page = self->doc->GetPage(num);
if (page == NULL) { PyErr_Format(PyExc_ValueError, "Invalid page number: %d", num); goto error; }
PdfDestination dest(page);
if (as_child != NULL && PyObject_IsTrue(as_child)) {
ans->item = self->item->CreateChild(*title, dest);
} else
ans->item = self->item->CreateNext(*title, dest);
} catch (const PdfError &err) {
podofo_set_exception(err); goto error;
} catch (...) {
PyErr_SetString(PyExc_Exception, "An unknown error occurred while trying to create the outline item");
goto error;
}
delete title;
return (PyObject*) ans;
error:
Py_XDECREF(ans); delete title;
return NULL;
}
static PyMethodDef methods[] = {
{"create", (PyCFunction)create, METH_VARARGS,
"create(title, pagenum, as_child=False) -> Create a new outline item with title 'title', pointing to page number pagenum. If as_child is True the new item will be a child of this item otherwise it will be a sibling. Returns the newly created item."
},
{"erase", (PyCFunction)erase, METH_VARARGS,
"erase() -> Delete this item and all its children, removing it from the outline tree completely."
},
{NULL} /* Sentinel */
};
// Type definition {{{
PyTypeObject pdf::PDFOutlineItemType = {
PyObject_HEAD_INIT(NULL)
0, /*ob_size*/
"podofo.PDFOutlineItem", /*tp_name*/
sizeof(PDFOutlineItem), /*tp_basicsize*/
0, /*tp_itemsize*/
(destructor)dealloc, /*tp_dealloc*/
0, /*tp_print*/
0, /*tp_getattr*/
0, /*tp_setattr*/
0, /*tp_compare*/
0, /*tp_repr*/
0, /*tp_as_number*/
0, /*tp_as_sequence*/
0, /*tp_as_mapping*/
0, /*tp_hash */
0, /*tp_call*/
0, /*tp_str*/
0, /*tp_getattro*/
0, /*tp_setattro*/
0, /*tp_as_buffer*/
Py_TPFLAGS_DEFAULT, /*tp_flags*/
"PDF Outline items", /* tp_doc */
0, /* tp_traverse */
0, /* tp_clear */
0, /* tp_richcompare */
0, /* tp_weaklistoffset */
0, /* tp_iter */
0, /* tp_iternext */
methods, /* tp_methods */
0, /* tp_members */
0, /* tp_getset */
0, /* tp_base */
0, /* tp_dict */
0, /* tp_descr_get */
0, /* tp_descr_set */
0, /* tp_dictoffset */
0, /* tp_init */
0, /* tp_alloc */
new_item, /* tp_new */
};
// }}}

View File

@ -46,6 +46,9 @@ initpodofo(void)
if (PyType_Ready(&pdf::PDFDocType) < 0)
return;
if (PyType_Ready(&pdf::PDFOutlineItemType) < 0)
return;
pdf::Error = PyErr_NewException((char*)"podofo.Error", NULL, NULL);
if (pdf::Error == NULL) return;