PyMuPDF 1.23.22__tar.gz → 1.23.24__tar.gz

{PyMuPDF-1.23.22 → PyMuPDF-1.23.24}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: PyMuPDF
-Version: 1.23.22
+Version: 1.23.24
 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-1.23.24/READMErb.md

@@ -0,0 +1,7 @@
+# PyMuPDF 1.23.22
+
+This wheel contains [MuPDF](https://mupdf.readthedocs.io/) shared libraries for
+use by [PyMuPDF](https://pymupdf.readthedocs.io/).
+
+This wheel is shared by PyMuPDF wheels that are specific to different Python
+versions, significantly reducing the total size of a release.

{PyMuPDF-1.23.22 → PyMuPDF-1.23.24}/changes.txt

@@ -2,6 +2,45 @@ Change Log
 ==========
 
 
+**Changes in version 1.23.24 (2024-02-19)**
+
+* Fixed issues:
+
+  * **Fixed** `3148 <https://github.com/pymupdf/PyMuPDF/issues/3148>`_: Table extraction - vertical text not handled correctly
+  * **Fixed** `3179 <https://github.com/pymupdf/PyMuPDF/issues/3179>`_: Table Detection: Incorrect Separation of Vector Graphics Clusters
+  * **Fixed** `3180 <https://github.com/pymupdf/PyMuPDF/issues/3180>`_: Cannot show optional content group: AttributeError: module 'fitz.mupdf' has no attribute 'pdf_array_push_drop'
+
+* Other:
+
+  * Be able to test system install using `sudo pip install` instead of a venv.
+
+
+**Changes in version 1.23.23 (2024-02-18)**
+
+* Fixed issues:
+
+  * **Fixed** `3126 <https://github.com/pymupdf/PyMuPDF/issues/3126>`_: Initialising Archive with a pathlib.Path fails.
+  * **Fixed** `3131 <https://github.com/pymupdf/PyMuPDF/issues/3131>`_: Calling the next attribute of an Annot raises a "No attribute .parent" warning
+  * **Fixed** `3134 <https://github.com/pymupdf/PyMuPDF/issues/3134>`_: Using an IRect as clip parameter in Page.get_pixmap no longer works since 1.23.9
+  * **Fixed** `3140 <https://github.com/pymupdf/PyMuPDF/issues/3140>`_: PDF document stays in use after closing
+  * **Fixed** `3150 <https://github.com/pymupdf/PyMuPDF/issues/3150>`_: doc.select() hangs on this doc.
+  * **Fixed** `3163 <https://github.com/pymupdf/PyMuPDF/issues/3163>`_: AssertionError on using fitz.IRect
+  * **Fixed** `3177 <https://github.com/pymupdf/PyMuPDF/issues/3177>`_: fitz.Pixmap(None, pix) Unrecognised args for constructing Pixmap
+
+* Other:
+
+  *
+    Improved `Document.select() by using new MuPDF function
+    `pdf_rearrange_pages()`. This is a more complete (and faster)
+    implementation of what needs to be done here in that not only pages will
+    be rearranged, but also consequential changes will be made to the table
+    of contents, links to removed pages and affected entries in the Optional
+    Content definitions.
+  * `TextWriter.appendv()`: added `small_caps` arg.
+  * Fixed some valgrind errors with MuPDF master.
+  * Fixed `Document.insert_image()` when build with MuPDF master.
+
+
 **Changes in version 1.23.22 (2024-02-12)**
 
 * Fixed issues:

{PyMuPDF-1.23.22 → PyMuPDF-1.23.24}/scripts/gh_release.py

@@ -434,6 +434,7 @@ def venv( command=None, packages=None, quick=False):
         log(f'{quick=}: Not creating venv because directory already exists: {venv_name}')
         command2 += 'true'
     else:
+        quick = False
         command2 += f'{sys.executable} -m venv{ssp} {venv_name}'
     if platform.system() == 'Windows':
         command2 += f' && {venv_name}\\Scripts\\activate'
@@ -471,7 +472,7 @@ def test( project, package, valgrind):
         run(
                 f'{sys.executable} {project}/tests/run_compound.py'
                     f' valgrind --suppressions={project}/valgrind.supp --error-exitcode=100 --errors-for-leak-kinds=none --fullpath-after='
-                    f' pytest -s -vv {project}/tests'
+                    f' pytest {project}/tests'
                     ,
                 env_extra=dict(
                     PYTHONMALLOC='malloc',
@@ -479,7 +480,7 @@ def test( project, package, valgrind):
                     ),
                 )
     else:
-        run(f'{sys.executable} {project}/tests/run_compound.py pytest -s {project}/tests')
+        run(f'{sys.executable} {project}/tests/run_compound.py pytest {project}/tests')
     
 
 def pyodide_setup(clean=False):

{PyMuPDF-1.23.22 → PyMuPDF-1.23.24}/scripts/sysinstall.py

@@ -24,28 +24,39 @@ Args:
     --mupdf-git <git_args>
         Get or update `mupdf_dir` using git. If `mupdf_dir` already
         exists we run `git pull` in it; otherwise we run `git
-        clone` with `git_args`. For example:
+        clone` with `<git_args> <mupdf_dir>`. For example:
             --mupdf-git "--branch master https://github.com/ArtifexSoftware/mupdf.git"
-    --pymupdf-dir <pymupdf_dir>
-        Path of PyMuPDF checkout; default is 'PyMuPDF'.
+    --mupdf-so-mode <mode>
+        Used with `install -m <mode> ...` when installing MuPDF. For example
+        `--mupdf-so-mode 744`.
+    --packages 0|1
+        If 1 (the default) we install required system packages such as
+        `libfreetype-dev`.
+    --pip 0|venv|sudo
+        Whether/how to install Python packages.
+        If '0' we assume required packages are already available.
+        If 'sudo' we install required Python packages using `sudo pip install
+        ...`.
+        If 'venv' (the default) we install Python packages and run installer
+        and test commands inside venv's.
     --prefix:
         Directory within `root`; default is `/usr/local`. Must start with `/`.
+    --pymupdf-dir <pymupdf_dir>
+        Path of PyMuPDF checkout; default is 'PyMuPDF'.
     --root <root>
         Root of install directory; default is `/`.
     --tesseract5 0|1
         If 1 (the default), we force installation of libtesseract-dev version
         5 (which is not available as a default package in Ubuntu-22.04) from
         package repository ppa:alex-p/tesseract-ocr-devel.
-    --test-venv <venv_path>
-        Run tests in specified venv; the default is a hard-coded venv name. The
-        venv will be created and required packages installed using `pip`. If
-        `venv-path` is empty string, we do not create or use a venv, and will
-        instead attempt to install required packages with `pip` in the current
-        Python environment.
+    --test-venv <test_venv>
+        Set the name of the venv in which we run tests (only with `--pip
+        venv`); the default is a hard-coded venv name. The venv will be
+        created, and required packages installed using `pip`.
     --use-installer 0|1
         If 1 (the default), we use `python -m installer` to install PyMuPDF
-        from a generated wheel. Otherwise we use `pip install`, which refuses
-        to do a system install with `--root /`, referencing PEP-668.
+        from a generated wheel. [Otherwise we use `pip install`, which refuses
+        to do a system install with `--root /`, referencing PEP-668.]
     -m 0|1
         If 1 (the default) we build and install MuPDF, otherwise we just show
         what command we would have run.
@@ -61,12 +72,14 @@ To only show what commands would be run, but not actually run them, specify `-m
 '''
 
 import glob
+import multiprocessing
 import os
 import platform
 import subprocess
 import sys
 import sysconfig
 
+import test as test_py
 
 # Requirements for a system build and install:
 #
@@ -107,6 +120,8 @@ def main():
     mupdf = True
     mupdf_dir = 'mupdf'
     mupdf_git = None
+    mupdf_so_mode = None
+    packages = True
     prefix = '/usr/local'
     pymupdf = True
     pymupdf_dir = os.path.abspath( f'{__file__}/../..')
@@ -114,6 +129,7 @@ def main():
     tesseract5 = True
     test = True
     test_venv = 'venv-pymupdf-sysinstall-test'
+    pip = 'venv'
     
     # Parse command-line.
     #
@@ -128,12 +144,15 @@ def main():
             return
         elif arg == '--mupdf-dir':      mupdf_dir = next(args)
         elif arg == '--mupdf-git':      mupdf_git = next(args)
+        elif arg == '--mupdf-so-mode':  mupdf_so_mode = next(args)
+        elif arg == '--packages':       packages = int(next(args))
         elif arg == '--prefix':         prefix = next(args)
         elif arg == '--pymupdf-dir':    pymupdf_dir = next(args)
         elif arg == '--root':           root = next(args)
         elif arg == '--tesseract5':     tesseract5 = int(next(args))
         elif arg == '--test-venv':      test_venv = next(args)
         elif arg == '--use-installer':  use_installer = int(next(args))
+        elif arg == '--pip':            pip = next(args)
         elif arg == '-m':               mupdf = int(next(args))
         elif arg == '-p':               pymupdf = int(next(args))
         elif arg == '-t':               test = int(next(args))
@@ -141,6 +160,8 @@ def main():
             assert 0, f'Unrecognised arg: {arg!r}'
     
     assert prefix.startswith('/')
+    pip_values = ('0', 'sudo', 'venv')
+    assert pip in pip_values, f'Unrecognised --pip value {pip!r} should be one of: {pip_values!r}'
     root = os.path.abspath(root)
     root_prefix = f'{root}{prefix}'.replace('//', '/')
     
@@ -161,44 +182,54 @@ def main():
             print(f'## Clone MuPDF into {mupdf_dir}.')
             run(f'git clone --recursive --depth 1 --shallow-submodules {mupdf_git} {mupdf_dir}')
     
-    # Install required system packages. We assume a Debian package system.
-    #
-    print('## Install system packages required by MuPDF.')
-    run(f'sudo apt update')
-    run(f'sudo apt install {" ".join(g_sys_packages)}')
-    # Ubuntu-22.04 has freeglut3-dev, not libglut-dev.
-    run(f'sudo apt install libglut-dev | sudo apt install freeglut3-dev')
-    if tesseract5:
-        print(f'## Force installation of libtesseract-dev version 5.')
-        # https://stackoverflow.com/questions/76834972/how-can-i-run-pytesseract-python-library-in-ubuntu-22-04
+    if packages:
+        # Install required system packages. We assume a Debian package system.
         #
-        run('sudo apt install -y software-properties-common')
-        run('sudo add-apt-repository ppa:alex-p/tesseract-ocr-devel')
-        run('sudo apt update')
-        run('sudo apt install -y libtesseract-dev')
-    else:
-        run('sudo apt install libtesseract-dev')
+        print('## Install system packages required by MuPDF.')
+        run(f'sudo apt update')
+        run(f'sudo apt install {" ".join(g_sys_packages)}')
+        # Ubuntu-22.04 has freeglut3-dev, not libglut-dev.
+        run(f'sudo apt install libglut-dev | sudo apt install freeglut3-dev')
+        if tesseract5:
+            print(f'## Force installation of libtesseract-dev version 5.')
+            # https://stackoverflow.com/questions/76834972/how-can-i-run-pytesseract-python-library-in-ubuntu-22-04
+            #
+            run('sudo apt install -y software-properties-common')
+            run('sudo add-apt-repository ppa:alex-p/tesseract-ocr-devel')
+            run('sudo apt update')
+            run('sudo apt install -y libtesseract-dev')
+        else:
+            run('sudo apt install libtesseract-dev')
     
     # Build+install MuPDF. We use mupd:Makefile's install-shared-python target.
     #
+    if pip == 'sudo':
+        print('## Installing Python packages required for building MuPDF and PyMuPDF.')
+        run(f'sudo pip install --upgrade pip')
+        names = (''
+                + test_py.get_pyproject_required(os.path.abspath(f'{__file__}/../../pyproject.toml'))
+                + ' '
+                + test_py.get_pyproject_required(os.path.abspath(f'{mupdf_dir}/pyproject.toml'))
+                )
+        run(f'sudo pip install {names}')
+    
     print('## Build and install MuPDF.')
-    if 1:
-        # Current MuPDF creates softlinks with `ln -s` which breaks if there
-        # was a previous build; it should do `ln -sf`. We make things work by
-        # deleting any existing softlinks here.
-        run(f'rm {root_prefix}/lib/libmupdf.so || true')
-        run(f'rm {root_prefix}/lib/libmupdfcpp.so || true')
     command = f'cd {mupdf_dir}'
     command += f' && {sudo}make'
+    command += f' -j {multiprocessing.cpu_count()}'
     #command += f' EXE_LDFLAGS=-Wl,--trace' # Makes linker generate diagnostics as it runs.
     command += f' DESTDIR={root}'
     command += f' HAVE_LEPTONICA=yes'
     command += f' HAVE_TESSERACT=yes'
     command += f' USE_SYSTEM_LIBS=yes'
+    command += f' VENV_FLAG={"--venv" if pip == "venv" else ""}'
+    if mupdf_so_mode:
+        command += f' SO_INSTALL_MODE={mupdf_so_mode}'
     command += f' build_prefix=system-libs-'
     command += f' prefix={prefix}'
     command += f' verbose=yes'
     command += f' install-shared-python'
+    command += f' INSTALL_MODE=755'
     run( command)
     
     # Build+install PyMuPDF.
@@ -217,23 +248,34 @@ def main():
     env += f'PYMUPDF_SETUP_IMPLEMENTATIONS=b'   # Only build the rebased implementation.
     if use_installer:
         print(f'## Building wheel.')
+        if pip == 'venv':
+            venv_name = 'venv-pymupdf-sysinstall'
         run(f'pwd')
         run(f'rm dist/* || true')
-        run(f'{env} pip wheel -vv -w dist {os.path.abspath(pymupdf_dir)}')
+        if pip == 'venv':
+            run(f'{sys.executable} -m venv {venv_name}')
+            run(f'. {venv_name}/bin/activate && pip install --upgrade pip')
+            run(f'. {venv_name}/bin/activate && pip install --upgrade installer')
+            run(f'{env} {venv_name}/bin/python -m pip wheel -vv -w dist {os.path.abspath(pymupdf_dir)}')
+        elif pip == 'sudo':
+            run(f'sudo pip install --upgrade pip')
+            run(f'sudo pip install installer')
+            run(f'{env} pip wheel -vv -w dist {os.path.abspath(pymupdf_dir)}')
+        else:
+            log(f'Not installing "installer" because {pip=}.')
         wheel = glob.glob(f'dist/*')
         assert len(wheel) == 1, f'{wheel=}'
         wheel = wheel[0]
         print(f'## Installing wheel using `installer`.')
-        venv = 'venv-pymupdf-sysinstall'
-        run(f'{sys.executable} -m venv {venv}')
-        run(f'. {venv}/bin/activate && pip install --upgrade pip')
-        run(f'. {venv}/bin/activate && pip install --upgrade installer')
         pv = '.'.join(platform.python_version_tuple()[:2])
         p = f'{root_prefix}/lib/python{pv}'
         # `python -m installer` fails to overwrite existing files.
         run(f'{sudo}rm -r {p}/site-packages/fitz || true')
         run(f'{sudo}rm -r {p}/site-packages/PyMuPDF-*.dist-info || true')
-        run(f'{sudo}{venv}/bin/python -m installer --destdir {root} --prefix {prefix} {wheel}')
+        if pip == 'venv':
+            run(f'{sudo}{venv_name}/bin/python -m installer --destdir {root} --prefix {prefix} {wheel}')
+        else:
+            run(f'{sudo}{sys.executable} -m installer --destdir {root} --prefix {prefix} {wheel}')
         # It seems that MuPDF Python bindings are installed into
         # `.../dist-packages` (from mupdf:Mafile's call of `$(shell python3
         # -c "import sysconfig; print(sysconfig.get_path('platlib'))")` while
@@ -267,7 +309,7 @@ def main():
         del sys.path[0]
         pythonpath = pipcl.install_dir(root)
     
-    # Show contents of installation director. This is very slow on github,
+    # Show contents of installation directory. This is very slow on github,
     # where /usr/local contains lots of things.
     #run(f'find {root_prefix}|sort')
         
@@ -277,20 +319,27 @@ def main():
     def run(command):
         return run_command(command, doit=test)
     import gh_release
-    # Create venv.
-    run(f'{sys.executable} -m venv {test_venv}')
-    # Install required packages.
-    command = f'. {test_venv}/bin/activate'
-    command += f' && pip install --upgrade pip'
-    command += f' && pip install --upgrade {gh_release.test_packages}'
-    run(command)
+    if pip == 'venv':
+        # Create venv.
+        run(f'{sys.executable} -m venv {test_venv}')
+        # Install required packages.
+        command = f'. {test_venv}/bin/activate'
+        command += f' && pip install --upgrade pip'
+        command += f' && pip install --upgrade {gh_release.test_packages}'
+        run(command)
+    elif pip == 'sudo':
+        run(f'sudo pip install --upgrade {gh_release.test_packages}')
+    else:
+        log(f'Not installing packages for testing because {pip=}.')
     # Run pytest.
     #
     # We need to set PYTHONPATH and LD_LIBRARY_PATH. In particular we
     # use pipcl.install_dir() to find where pipcl will have installed
     # PyMuPDF.
-    command = f'. {test_venv}/bin/activate'
-    command += f' && LD_LIBRARY_PATH={root_prefix}/lib PYTHONPATH={pythonpath}'
+    command = ''
+    if pip == 'venv':
+        command += f'. {test_venv}/bin/activate &&'
+    command += f' LD_LIBRARY_PATH={root_prefix}/lib PYTHONPATH={pythonpath}'
     command += f' pytest -k "not test_color_count and not test_3050" {pymupdf_dir}'
     run(command)
 

{PyMuPDF-1.23.22 → PyMuPDF-1.23.24}/scripts/test.py

@@ -58,11 +58,14 @@ Options:
         PYMUPDF_SETUP_MUPDF_BUILD, which is used by PyMuPDF/setup.py. If not
         specifed PyMuPDF will download its default mupdf .tgz.]
     -p <pytest-options>
-        Set pytest options; default is '-s'.
-    -t <name>
-        Pytest test name. Should be relative to PyMuPDF directory. For example:
+        Set pytest options; default is ''.
+    -t <names>
+        Pytest test names, comma-separated. Should be relative to PyMuPDF
+        directory. For example:
             -t tests/test_general.py
-            -t tests/test_general.py::test_subset_fonts
+            -t tests/test_general.py::test_subset_fonts.
+        To specify multiple tests, use comma-separated list and/or multiple `-t
+        <names>` args.
     -v
         Avoid delay if venv directory already exists. We assume the existing
         directory was created by us earlier and is a valid venv containing all
@@ -87,6 +90,10 @@ Commands:
     test
         Runs PyMuPDF's pytest tests in venv. Default is to test classic, rebased and
         unoptimised rebased; use `-i` to change this.
+
+Environment:
+    PYMUDF_SCRIPTS_TEST_options
+        Is prepended to command line args.
 '''
 
 import gh_release
@@ -94,6 +101,7 @@ import gh_release
 import os
 import platform
 import re
+import shlex
 import subprocess
 import sys
 import textwrap
@@ -113,12 +121,15 @@ def main(argv):
     build_type = None
     gdb = False
     implementations = None
-    test_name = None
+    test_names = list()
     venv_quick = False
     pytest_options = None
     timeout = None
     
-    args = iter( argv[1:])
+    options = os.environ.get('PYMUDF_SCRIPTS_TEST_options', '')
+    options = shlex.split(options)
+    
+    args = iter(options + argv[1:])
     i = 0
     while 1:
         try:
@@ -148,7 +159,7 @@ def main(argv):
         elif arg == '-p':
             pytest_options  = next(args)
         elif arg == '-t':
-            test_name = next(args)
+            test_names += next(args).split(',')
         elif arg == '--timeout':
             timeout = float(next(args))
         elif arg == '-v':
@@ -193,7 +204,7 @@ def main(argv):
                 implementations=implementations,
                 valgrind=valgrind,
                 venv_quick=venv_quick,
-                test_name=test_name,
+                test_names=test_names,
                 pytest_options=pytest_options,
                 timeout=timeout,
                 gdb=gdb,
@@ -213,6 +224,17 @@ def main(argv):
             assert 0
 
 
+def get_env_bool(name, default=0):
+    v = os.environ.get(name)
+    if v in ('1', 'true'):
+        return 1
+    elif v in ('0', 'false'):
+        return 0
+    elif v is None:
+        return default
+    else:
+        assert 0, f'Bad environ {name=} {v=}'
+
 def show_help():
     print(__doc__)
     print(venv_info())
@@ -307,7 +329,7 @@ def test(
         implementations,
         valgrind,
         venv_quick=False,
-        test_name=None,
+        test_names=None,
         pytest_options=None,
         timeout=None,
         gdb=False,
@@ -320,7 +342,7 @@ def test(
             See top-level option `--valgrind`.
         venv_quick:
             See top-level option `-v`.
-        test_name:
+        test_names:
             See top-level option `-t`.
         pytest_options:
             See top-level option `-p`.
@@ -332,10 +354,13 @@ def test(
         if valgrind:
             pytest_options = '-s -vv'
         else:
-            pytest_options = '-s'
-    pytest_arg = pymupdf_dir_rel
-    if test_name:
-        pytest_arg += f'/{test_name}'
+            pytest_options = ''
+    pytest_arg = ''
+    if test_names:
+        for test_name in test_names:
+            pytest_arg += f' {pymupdf_dir_rel}/{test_name}'
+    else:
+        pytest_arg += f' {pymupdf_dir_rel}'
     python = gh_release.relpath(sys.executable)
     log('Running tests with tests/run_compound.py and pytest.')
     try:
@@ -359,7 +384,7 @@ def test(
             command = (
                     f'{python} {pymupdf_dir_rel}/tests/run_compound.py{run_compound_args}'
                         f' valgrind --suppressions={pymupdf_dir_rel}/valgrind.supp --error-exitcode=100 --errors-for-leak-kinds=none --fullpath-after='
-                        f' {python} -m pytest {pytest_options} {pytest_arg}'
+                        f' {python} -m pytest {pytest_options}{pytest_arg}'
                         )
             env_extra=dict(
                     PYTHONMALLOC='malloc',
@@ -385,12 +410,13 @@ def test(
         log('\n' + venv_info(pytest_args=f'{pytest_options} {pytest_arg}'))
 
 
-def get_pyproject_required():
+def get_pyproject_required(ppt=None):
     '''
     Returns space-separated names of required packages in pyproject.toml.  We
     do not do a proper parse and rely on the packages being in a single line.
     '''
-    ppt = os.path.abspath(f'{__file__}/../../pyproject.toml')
+    if ppt is None:
+        ppt = os.path.abspath(f'{__file__}/../../pyproject.toml')
     with open(ppt) as f:
         for line in f:
             m = re.match('^requires = \\[(.*)\\]$', line)

{PyMuPDF-1.23.22 → PyMuPDF-1.23.24}/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 = '1.23.22'
+version = '1.23.24'
 version_b = '1.23.22'
 
 tag_python = None