From e30c0f17b565b0488f8ad67c13a3974f00aa43d9 Mon Sep 17 00:00:00 2001
From: "Jorj X. McKie" <jorj.x.mckie@outlook.de>
Date: Wed, 6 Mar 2024 08:40:59 -0400
Subject: [PATCH] Add support for redacting vector graphics

Add support for redacting vector graphics

Also add a quick fix for property Annot.irt_xref

Update test_2548.py
---
 src/__init__.py      |  6 ++++--
 src/utils.py         | 19 +++++++++++++------
 tests/test_2548.py   |  2 +-
 tests/test_annots.py | 17 ++++++++++++++++-
 4 files changed, 34 insertions(+), 10 deletions(-)

diff --git a/src/__init__.py b/src/__init__.py
index 434f3d7e7..ee819f628 100644
--- a/src/__init__.py
+++ b/src/__init__.py
@@ -778,7 +778,8 @@ def info(self):
 
         return res
 
-    def irt_xref():
+    @property
+    def irt_xref(self):
         '''
         annotation IRT xref
         '''
@@ -7605,11 +7606,12 @@ def _addWidget(self, field_type, field_name):
         JM_add_annot_id(annot, "W")
         return Annot(annot)
 
-    def _apply_redactions(self, images):
+    def _apply_redactions(self, images, graphics):
         page = self._pdf_page()
         opts = mupdf.PdfRedactOptions()
         opts.black_boxes = 0  # no black boxes
         opts.image_method = images  # how to treat images
+        opts.line_art = graphics  # how to treat vector graphics
         ASSERT_PDF(page)
         success = mupdf.pdf_redact_page(page.doc(), page, opts)
         return success
diff --git a/src/utils.py b/src/utils.py
index 696afc780..0a00d8677 100644
--- a/src/utils.py
+++ b/src/utils.py
@@ -1213,7 +1213,7 @@ def set_metadata(doc: fitz.Document, m: dict) -> None:
         info_xref = 0
     else:
         info_xref = int(temp.replace("0 R", ""))
-    
+
     if m == {} and info_xref == 0:  # nothing to do
         return
 
@@ -4187,13 +4187,20 @@ def commit(self, overlay: bool = True) -> None:
         return
 
 
-def apply_redactions(page: fitz.Page, images: int = 2) -> bool:
+def apply_redactions(page: fitz.Page, images: int = 2, graphics: int = 1) -> bool:
     """Apply the redaction annotations of the page.
 
     Args:
         page: the PDF page.
-        images: 0 - ignore images, 1 - remove complete overlapping image,
-                2 - blank out overlapping image parts.
+        images:
+              0 - ignore images
+              1 - remove all overlapping images
+              2 - blank out overlapping image parts
+              3 - remove image unless invisible
+        graphics:
+              0 - ignore graphics
+              1 - remove graphics if contained in rectangle
+              2 - remove all overlapping graphics
     """
 
     def center_rect(annot_rect, text, font, fsize):
@@ -4246,7 +4253,7 @@ def center_rect(annot_rect, text, font, fsize):
     if redact_annots == []:  # any redactions on this page?
         return False  # no redactions
 
-    rc = page._apply_redactions(images)  # call MuPDF redaction process step
+    rc = page._apply_redactions(images, graphics)  # call MuPDF
     if not rc:  # should not happen really
         raise ValueError("Error applying redactions.")
 
@@ -4504,7 +4511,7 @@ def append_this(pos, text):
                 pos, text, font=font, fontsize=fontsize, small_caps=small_caps
                 )
         return ret
-        
+
     tolerance = fontsize * 0.2  # extra distance to left border
     space_len = textlen(" ")
     std_width = rect.width - tolerance
diff --git a/tests/test_2548.py b/tests/test_2548.py
index 5af77e054..9aa4212a1 100644
--- a/tests/test_2548.py
+++ b/tests/test_2548.py
@@ -36,7 +36,7 @@ def test_2548():
     # versions with updated MuPDF also fix the bug.
     
     if fitz.mupdf_version_tuple >= (1, 24):
-        expected = 'format error: cycle in structure tree\nstructure tree broken, assume tree is missing'
+        expected = 'Loop found in structure tree. Ignoring structure.'
         assert wt == expected, f'expected:\n    {expected!r}\nwt:\n    {wt!r}\n'
         assert not e
     elif (0
diff --git a/tests/test_annots.py b/tests/test_annots.py
index bcc923cc3..fbf78f79b 100644
--- a/tests/test_annots.py
+++ b/tests/test_annots.py
@@ -149,7 +149,7 @@ def test_stamp():
     page = doc.reload_page(page)
 
 
-def test_redact():
+def test_redact1():
     doc = fitz.open()
     page = doc.new_page()
     annot = page.add_redact_annot(r, text="Hello")
@@ -168,6 +168,21 @@ def test_redact():
     assert s == r
     page.apply_redactions()
 
+
+def test_redact2():
+    """Test removal of graphics (line art)."""
+    if not hasattr(fitz, "mupdf"):
+        print("Not executing 'test_redact2' in classic")
+        return
+    doc = fitz.open()
+    page = doc.new_page()
+    rect = fitz.Rect(100, 100, 200, 200)
+    page.draw_rect(rect)
+    page.add_redact_annot(rect)
+    page.apply_redactions(graphics=2)
+    assert page.get_drawings() == []
+
+
 def test_1645():
     '''
     Test fix for #1645.