PyMuPDF 1.23.15__tar.gz → 1.23.16__tar.gz

{PyMuPDF-1.23.15 → PyMuPDF-1.23.16}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: PyMuPDF
-Version: 1.23.15
+Version: 1.23.16
 Summary: A high performance Python library for data extraction, analysis, conversion & manipulation of PDF (and other) documents.
 Description-Content-Type: text/markdown
 Author: Artifex

{PyMuPDF-1.23.15 → PyMuPDF-1.23.16}/changes.txt

@@ -2,11 +2,24 @@ Change Log
 ==========
 
 
+**Changes in version 1.23.16 (2024-01-18)**
+
+* Bug fixes:
+
+  * **Fixed** `3058 <https://github.com/pymupdf/PyMuPDF/issues/3058>`_: Pixmap created from CMYK JPEG delivers RGB format
+
+* Other:
+
+  * In table detection strategy "lines_strict", exclude fill-only vector graphics.
+  * Fixed sysinstall test failure.
+  * In documentation, update feature matrix with item about text writing.
+
+
 **Changes in version 1.23.15 (2024-01-16)**
 
 * Bug fixes:
 
-  * **Fixed** `3051 <https://github.com/pymupdf/PyMuPDF/issues/3051>`_: Fix for Pixmap.set_pixel().
+  * **Fixed** `3050 <https://github.com/pymupdf/PyMuPDF/issues/3050>`_: python3.9 pix.set_pixel has something wrong in c.append( ord(i))
 
 * Other:
 

{PyMuPDF-1.23.15 → PyMuPDF-1.23.16}/scripts/sysinstall.py

@@ -291,7 +291,7 @@ def main():
     # PyMuPDF.
     command = f'. {test_venv}/bin/activate'
     command += f' && LD_LIBRARY_PATH={root_prefix}/lib PYTHONPATH={pythonpath}'
-    command += f' pytest -k "not test_color_count" {pymupdf_dir}'
+    command += f' pytest -k "not test_color_count and not test_3050" {pymupdf_dir}'
     run(command)
 
 

{PyMuPDF-1.23.15 → PyMuPDF-1.23.16}/setup.py

@@ -1072,7 +1072,7 @@ with open( f'{g_root}/READMErb.md', encoding='utf-8') as f:
 # We generate different wheels depending on g_flavour.
 #
 
-version = '1.23.15'
+version = '1.23.16'
 version_b = '1.23.9'
 
 tag_python = None

{PyMuPDF-1.23.15 → PyMuPDF-1.23.16}/src/__init__.py

@@ -2396,7 +2396,9 @@ class Colorspace:
 
     def __init__(self, type_):
         """Supported are GRAY, RGB and CMYK."""
-        if type_ == CS_GRAY:
+        if isinstance( type_, mupdf.FzColorspace):
+            self.this = type_
+        elif type_ == CS_GRAY:
             self.this = mupdf.FzColorspace(mupdf.FzColorspace.Fixed_GRAY)
         elif type_ == CS_CMYK:
             self.this = mupdf.FzColorspace(mupdf.FzColorspace.Fixed_CMYK)
@@ -7759,8 +7761,8 @@ class Page:
                 else:
                     pm = mupdf.fz_convert_pixmap(
                             arg_pix,
-                            mupdf.FzColorspace(0),
-                            mupdf.FzColorspace(0),
+                            mupdf.FzColorspace(),
+                            mupdf.FzColorspace(),
                             mupdf.FzDefaultColorspaces(None),
                             mupdf.FzColorParams(),
                             1,
@@ -9429,7 +9431,7 @@ class Pixmap:
                 self.this = mupdf.fz_convert_pixmap(
                         spix,
                         cs,
-                        mupdf.FzColorspace(0),
+                        mupdf.FzColorspace(),
                         mupdf.FzDefaultColorspaces(None),
                         mupdf.FzColorParams(),
                         1
@@ -21754,8 +21756,8 @@ def int_rc(text):
     return int(text)
 
 VersionFitz = "1.23.9" # MuPDF version.
-VersionBind = "1.23.15" # PyMuPDF version.
-VersionDate = "2024-01-16 00:00:01"
+VersionBind = "1.23.16" # PyMuPDF version.
+VersionDate = "2024-01-18 00:00:01"
 VersionDate2 = VersionDate.replace('-', '').replace(' ', '').replace(':', '')
 version = (VersionBind, VersionFitz, VersionDate2)
 pymupdf_version_tuple = tuple( [int_rc(i) for i in VersionBind.split('.')])

{PyMuPDF-1.23.15 → PyMuPDF-1.23.16}/src/table.py

@@ -1886,8 +1886,24 @@ def make_edges(page, clip=None, tset=None):
 
         parea = abs(page.rect) * 0.8  # area of the full page (80%)
 
-        # exclude graphics that are too large
-        paths = [p for p in page.get_drawings() if abs(p["rect"]) < parea]
+        # exclude irrelevant graphics
+        paths = []
+        for p in page.get_drawings():
+            if abs(p["rect"]) >= parea:
+                continue
+            if "s" in p["type"]:
+                paths.append(p)
+                continue
+            if (
+                p["rect"].width > 3
+                and p["rect"].height > 3
+                and (
+                    tset.vertical_strategy == "lines_strict"
+                    or tset.horizontal_strategy == "lines_strict"
+                )
+            ):
+                continue
+            paths.append(p)
 
         # make a list of vector graphics rectangles (IRects are sufficient)
         prects = sorted([p["rect"] for p in paths], key=lambda r: (r.y1, r.x0))
@@ -1998,14 +2014,6 @@ def make_edges(page, clip=None, tset=None):
         return line_dict
 
     for p in paths:
-        if p["type"] == "f" and p["fill"] == (1, 1, 1):
-            continue
-        if p["type"] == "f" and p["rect"].width > 3 and p["rect"].height > 3:
-            if (
-                tset.vertical_strategy == "lines_strict"
-                or tset.horizontal_strategy == "lines_strict"
-            ):
-                continue
         items = p["items"]  # items in this path
 
         # if 'closePath', add a line from last to first point

PyMuPDF-1.23.16/src_classic/version.i

@@ -0,0 +1,7 @@
+%pythoncode %{
+VersionFitz = "1.23.9" # MuPDF version.
+VersionBind = "1.23.16" # PyMuPDF version.
+VersionDate = "2024-01-18 00:00:01"
+version = (VersionBind, VersionFitz, "20240118000001")
+pymupdf_version_tuple = tuple( [int(i) for i in VersionFitz.split('.')])
+%}

{PyMuPDF-1.23.15 → PyMuPDF-1.23.16}/tests/test_pixmap.py

@@ -94,6 +94,10 @@ def test_setalpha():
     assert t == alphas
 
 def test_color_count():
+    '''
+    This is known to fail if MuPDF is built without PyMuPDF's custom config.h,
+    e.g. in Linux system installs.
+    '''
     pm = fitz.Pixmap(imgfile)
     assert pm.color_count() == 40624
 
@@ -154,6 +158,10 @@ def test_3020():
     pm4 = fitz.Pixmap(pm, pm3)
 
 def test_3050():
+    '''
+    This is known to fail if MuPDF is built without it's default third-party
+    libraries, e.g. in Linux system installs.
+    '''
     pdf_file = fitz.open(pdf)
     for page_no, page in enumerate(pdf_file):
         zoom_x = 4.0
@@ -179,3 +187,23 @@ def test_3050():
         pix.save(os.path.abspath(f'{__file__}/../../tests/test_3050_out.png'))
         assert digest1 != digest0
         assert digest1 == digest_expected
+
+def test_3058():
+    doc = fitz.Document(os.path.abspath(f'{__file__}/../../tests/resources/test_3058.pdf'))
+    images = doc[0].get_images(full=True)
+    pix = fitz.Pixmap(doc, 17)
+    
+    # First bug was that `pix.colorspace` was DeviceRGB.
+    assert str(pix.colorspace) == 'Colorspace(CS_CMYK) - DeviceCMYK'
+    
+    pix = fitz.Pixmap(fitz.csRGB, pix)
+    assert str(pix.colorspace) == 'Colorspace(CS_RGB) - DeviceRGB'
+    
+    # Second bug was that the image was converted to RGB via greyscale proofing
+    # color space, so image contained only shades of grey. This compressed
+    # easily to a .png file, so we crudely check the bug is fixed by looking at
+    # size of .png file.
+    path = os.path.abspath(f'{__file__}/../../tests/test_3058_out.png')
+    pix.save(path)
+    s = os.path.getsize(path)
+    assert 1800000 < s < 2600000, f'Unexpected size of {path}: {s}'

PyMuPDF-1.23.15/src_classic/version.i

@@ -1,7 +0,0 @@
-%pythoncode %{
-VersionFitz = "1.23.9" # MuPDF version.
-VersionBind = "1.23.15" # PyMuPDF version.
-VersionDate = "2024-01-16 00:00:01"
-version = (VersionBind, VersionFitz, "20240116000001")
-pymupdf_version_tuple = tuple( [int(i) for i in VersionFitz.split('.')])
-%}