ughost 1.0.0__py3-none-any.whl

ughost/__init__.py

@@ -0,0 +1,27 @@
+#==============================================================================
+# C O P Y R I G H T
+#------------------------------------------------------------------------------
+# Copyright (c) 2021-2023 by Helmut Konrad Schewe. All rights reserved.
+# This file is property of Helmut Konrad Schewe. Any unauthorized copy,
+# use or distribution is an offensive act against international law and may
+# be prosecuted under federal law. Its content is company confidential.
+#==============================================================================
+
+import os
+
+import utilo
+
+from ughost.extract import images
+from ughost.optimize import small
+from ughost.parts import Part
+from ughost.parts import bounding_convert
+from ughost.parts import run
+from ughost.utils import pdfwrite
+
+__version__ = '0.9.1'
+
+ROOT = os.path.abspath(os.path.join(os.path.dirname(__file__), '..'))
+PROCESS = 'ughost'
+
+INSTALLED = utilo.hasprog('gs') or utilo.hasprog('gswin64c')
+HAS_GHOST = INSTALLED

ughost/cli.py

@@ -0,0 +1,89 @@
+# =============================================================================
+# C O P Y R I G H T
+# -----------------------------------------------------------------------------
+# Copyright (c) 2022-2023 by Helmut Konrad Schewe. All rights reserved.
+# This file is property of Helmut Konrad Schewe. Any unauthorized copy,
+# use or distribution is an offensive act against international law and may
+# be prosecuted under federal law. Its content is company confidential.
+# =============================================================================
+
+import pdflog
+import utilo
+import utilo.cli
+
+import ughost
+
+DESCRIPTION = ''
+CONFIG = utilo.ParserConfiguration(
+    inputparameter=True,
+    outputparameter=True,
+    prefix=False,
+    multiprocessed=False,
+    cacheflag=False,
+    waitingflag=False,
+)
+DPI = 216.0
+
+
+@utilo.saveme
+def main():
+    inpath, outpath, dpi, pages = eval_cli()
+    write_images(
+        inpath,
+        outpath=outpath,
+        dpi=dpi,
+        pages=pages,
+    )
+    return utilo.SUCCESS
+
+
+def eval_cli():
+    parser = utilo.cli.create_parser(
+        config=CONFIG,
+        description=DESCRIPTION,
+        todo=[
+            utilo.Parameter(
+                longcut='dpi',
+                message='use 216 as default',
+                args={'default': DPI},
+            )
+        ],
+        prog=ughost.PROCESS,
+        version=ughost.__version__,
+    )
+    args = utilo.parse(parser)
+    inpath, outpath = utilo.sources(args, singleinput=True)  # pylint:disable=W0632
+    # It is only single path supported. Run program multiple times if more
+    # than one analysis is required.
+    inpath = inpath[0]
+    dpi = args.get('dpi', DPI)
+    pages = parse_pages(
+        args.get('pages', None),
+        inpath=inpath,
+    )
+    return inpath, outpath, dpi, pages
+
+
+def write_images(inpath, outpath, dpi: float, pages: tuple = None):
+    root = ughost.pdfwrite(
+        source=inpath,
+        dpi=dpi,
+        pages=pages,
+    )
+    written = utilo.file_list(root, include='png', absolute=True)
+    for path, filename in zip(written, pages):
+        filename = f'{str(filename).zfill(3)}.png'
+        dst = utilo.join(outpath, filename)
+        utilo.debug(f'write {dst}')
+        utilo.file_copy(path, dst=dst)
+
+
+def parse_pages(pages: tuple, inpath: str) -> tuple:
+    pagecount = pdflog.pagecount(inpath)
+    if not pages:
+        return utilo.rtuple(pagecount)
+    pages = utilo.parse_pages(
+        pages[0],
+        pagecount=pagecount,
+    )
+    return pages

ughost/extract.py

@@ -0,0 +1,52 @@
+# =============================================================================
+# C O P Y R I G H T
+# -----------------------------------------------------------------------------
+# Copyright (c) 2021-2023 by Helmut Konrad Schewe. All rights reserved.
+# This file is property of Helmut Konrad Schewe. Any unauthorized copy,
+# use or distribution is an offensive act against international law and may
+# be prosecuted under federal law. Its content is company confidential.
+# =============================================================================
+
+import io
+import os
+
+import iamraw
+import PIL.Image
+import utilo
+
+import ughost
+
+DPI = 72
+RENDERER = 300
+
+
+def images(source: str, boundings: iamraw.ImageInformations, dpi=DPI) -> list:
+    # ensure that bounding box matches with correct page
+    pages = sorted(set(item.page for item in boundings))
+    root = ughost.pdfwrite(source, dpi=RENDERER, pages=pages)
+    pagenr = {page: index for index, page in enumerate(pages, start=1)}
+    loaded = [
+        load_image(
+            bounding,
+            path=os.path.join(root, f'{pagenr[bounding.page]}.png'),
+            dpi=dpi,
+        ) for bounding in boundings
+    ]
+    return loaded
+
+
+def load_image(bounding: iamraw.ImageInformation, path: str, dpi=DPI) -> bytes:
+    raw = io.BytesIO()
+    with PIL.Image.open(path, formats=('png',)) as loaded:
+        # left, upper, right, lower
+        bounding = utilo.tuple_mult(
+            bounding.bounding,
+            value=RENDERER / dpi,
+        )
+        croped = loaded.crop(bounding)
+        croped.save(raw, format='png')
+    # rewind the buffer
+    raw.seek(0)
+    # convert to bytes
+    result = raw.getvalue()
+    return result

ughost/optimize.py

@@ -0,0 +1,25 @@
+# =============================================================================
+# C O P Y R I G H T
+# -----------------------------------------------------------------------------
+# Copyright (c) 2021-2023 by Helmut Konrad Schewe. All rights reserved.
+# This file is property of Helmut Konrad Schewe. Any unauthorized copy,
+# use or distribution is an offensive act against international law and may
+# be prosecuted under federal law. Its content is company confidential.
+# =============================================================================
+"""Optimize PDF
+============
+
+Use ughostScript to optimize pdf file.
+"""
+
+import utilo
+
+import ughost.utils
+
+
+def small(source: str, destination: str, pages: tuple = None):
+    pages = ughost.utils.gpages_fromtuple(pages)
+    config = '-sDEVICE=pdfwrite -dBATCH -dNOPAUSE -SAFE'
+    source = f'"{source}"'
+    cmd = f'{ughost.utils.GHOST} {config} {pages} -sOutputFile={destination} {source}'
+    utilo.run(cmd)

ughost/parts.py

@@ -0,0 +1,161 @@
+# =============================================================================
+# C O P Y R I G H T
+# -----------------------------------------------------------------------------
+# Copyright (c) 2022-2023 by Helmut Konrad Schewe. All rights reserved.
+# This file is property of Helmut Konrad Schewe. Any unauthorized copy,
+# use or distribution is an offensive act against international law and may
+# be prosecuted under federal law. Its content is company confidential.
+# =============================================================================
+
+import dataclasses
+import io
+
+import PIL.Image
+import PIL.ImageChops
+import PIL.ImageDraw
+import utilo
+
+import ughost
+import ughost.cli
+
+
+@dataclasses.dataclass
+class Part:
+    page: int = None
+    bounding: tuple = None
+    color: tuple = None
+    name: str = None
+
+    def __getitem__(self, index):
+        if not index:
+            return self.page
+        if index == 1:
+            return self.bounding
+        if index == 2:
+            return self.color
+        if index == 3:
+            return self.name
+        raise IndexError
+
+
+def run(src: str, dst: str, boundings: list) -> list:
+    result = []
+    extracted = parts(source=src, boundings=boundings)
+    for image, bounding in zip(extracted, boundings):
+        name = bounding.name
+        if name is None:
+            name = utilo.tmpname()
+        outpath = utilo.join(dst, f'{name}.png')
+        utilo.debug(outpath)
+        # TODO: SECURITY: BEFORE RELASE: USE PRIVATE LATER
+        utilo.file_replace_binary(outpath, content=image)
+        result.append(outpath)
+    return result
+
+
+# TODO: HOLY VALUE
+DPI_PDF = 72.0
+
+
+def bounding_convert(pdf: tuple, dpi: int = ughost.cli.DPI) -> tuple:
+    """\
+    >>> bounding_convert((120, 120, 520, 520))
+    (360.0, 360.0, 1560.0, 1560.0)
+    """
+    assert dpi and DPI_PDF, f'invalid dpi: {dpi}, {DPI_PDF}'
+    result = utilo.tuple_mult(
+        pdf,
+        value=dpi / DPI_PDF,
+    )
+    return result
+
+
+def parts(source: str, boundings: list) -> list:
+    extracted = extract(
+        source=source,
+        boundings=boundings,
+    )
+    result = []
+    for item in boundings:
+        page, bounding = item.page, item.bounding
+        color = item.color
+        path = extracted[page]
+        part = extract_part(path, bounding, color)
+        result.append(part)
+    return result
+
+
+def extract(source, boundings: list) -> dict:
+    workdir = utilo.tmpdir(root=ughost.ROOT)
+    pages = set(item[0] for item in boundings)
+    pages = sorted(pages)
+    pagesraw: str = ','.join(str(item) for item in pages)
+    cmd = f'{ughost.PROCESS} -i {source} -o {workdir} --pages={pagesraw}'
+    utilo.run(cmd)
+    files = utilo.file_list(
+        workdir,
+        absolute=True,
+    )
+    result = dict(zip(
+        pages,
+        files,
+    ))
+    return result
+
+
+def extract_part(path: str, bounding: tuple, color=None) -> bytes:
+    raw = io.BytesIO()
+    with PIL.Image.open(path, formats=('png',)) as loaded:
+        size = loaded._size  # pylint:disable=W0212
+        mask = create_mask(
+            bounding,
+            size=size,
+        )
+        image = PIL.ImageChops.multiply(loaded, mask)
+        image = colorize(image, color=color)
+        image.save(raw, format='png')
+    # rewind the buffer
+    raw.seek(0)
+    # convert to bytes
+    result = raw.getvalue()
+    return result
+
+
+def create_mask(bounding: tuple, size: tuple) -> PIL.Image:
+    """Color list of rectangles defined in `bounding`."""
+    single = isinstance(bounding[0], (int, float))
+    if single:
+        bounding = [bounding]
+    image = PIL.Image.new(
+        mode='RGBA',
+        size=size,
+    )
+    draw = PIL.ImageDraw.Draw(image)
+    color = (0, 0, 0, 255)
+    for bbox in bounding:
+        # TODO: VERYIFY OVERLAPPING RECTANLGES
+        draw.rectangle(
+            bbox,
+            fill=color,
+        )
+    return image
+
+
+def colorize(image, color):
+    if color is None:
+        color = (255, 0, 0)
+    # split the image into individual bands
+    source = image.split()
+    alpha = 3
+    # select regions where alpha is selected
+    mask = source[alpha].point(lambda i: i)
+    # process the green band
+    for band, col in enumerate(color):
+        source[band].paste(
+            source[band].point(lambda i: col),  # pylint:disable=cell-var-from-loop
+            None,
+            mask,
+        )
+    # build a new multiband image
+    image = PIL.Image.merge(image.mode, source)
+    return image

ughost/utils.py

@@ -0,0 +1,56 @@
+# =============================================================================
+# C O P Y R I G H T
+# -----------------------------------------------------------------------------
+# Copyright (c) 2021-2023 by Helmut Konrad Schewe. All rights reserved.
+# This file is property of Helmut Konrad Schewe. Any unauthorized copy,
+# use or distribution is an offensive act against international law and may
+# be prosecuted under federal law. Its content is company confidential.
+# =============================================================================
+
+import os
+
+import utilo
+
+import ughost
+
+GHOST = 'gswin64c' if utilo.iswin() else 'gs'
+
+
+def pdfwrite(
+    source: str,
+    dpi: int = 300,
+    formats: str = 'pngalpha',
+    root: str = None,
+    pages: tuple = None,
+):
+    root = utilo.tmpdir(root=ughost.ROOT) if root is None else root
+    if isinstance(pages, int):  # pylint:disable=W0160
+        destination = os.path.join(root, f'{pages}.png')
+    else:
+        destination = os.path.join(root, '%d.png')
+    pages = gpages_fromtuple(pages)
+    config = f'-sDEVICE={formats} -r{dpi} -dBATCH -dNOPAUSE -SAFE'
+    source = f'"{source}"'
+    cmd = f'{GHOST} {config} {pages} -sOutputFile={destination} {source}'
+    utilo.run(cmd)
+    return root
+
+
+def gpages_fromtuple(pages: tuple = None) -> str:
+    """\
+    >>> gpages_fromtuple((1, 2, 3))
+    '-sPageList=2,3,4'
+    >>> gpages_fromtuple()
+    ''
+    """
+    if pages is None:
+        return ''
+    if isinstance(pages, int):
+        pages = (pages,)
+    # ughost requires sorted page numbers
+    pages = sorted(pages)
+    # -sPageList=1,3,5
+    pages = utilo.tuple_plus(pages, value=1)
+    pages: str = utilo.from_tuple(pages, separator=',')
+    pages: str = f'-sPageList={pages}'
+    return pages

ughost-1.0.0.dist-info/METADATA

@@ -0,0 +1,27 @@
+Metadata-Version: 2.4
+Name: ughost
+Version: 1.0.0
+Author-email: Helmut Konrad Schewe <helmutus@outlook.com>
+License-Expression: MIT
+Project-URL: Homepage, https://github.com/anaticulae/ughost
+Project-URL: Repository, https://github.com/anaticulae/ughost
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Programming Language :: Python :: 3.14
+Requires-Python: >=3.12
+Description-Content-Type: text/markdown
+Requires-Dist: utilo<3.0.0,>=2.107.5
+Requires-Dist: iamraw<5.0.0,>=4.91.4
+Requires-Dist: Pillow<13.0.0,>=12.2.0
+Requires-Dist: pdflog<2.0.0,>=1.0.2
+Provides-Extra: dev
+Requires-Dist: utilotest<2.0.0,>=1.0.3; extra == "dev"
+Requires-Dist: hoverpower<2.0.0,>=1.1.0; extra == "dev"
+Requires-Dist: upainter<2.0.0,>=1.0.0; extra == "dev"
+
+# ughost
+
+```
+    minidocks/ughostscript
+    minidocks/python
+```

ughost-1.0.0.dist-info/RECORD

@@ -0,0 +1,11 @@
+ughost/__init__.py,sha256=KrSDVN4ReaJcZhGNQlZ9kg2s-OOcpE3K62VFQIP0ssI,978
+ughost/cli.py,sha256=RaJzpo083YwWzmDftlVWmaSrhrRyIqmJMfqvtbtIv8Q,2497
+ughost/extract.py,sha256=mMSBRgCc7m0_sU6BW_Tse9y8qXWgz2vjXNTq3z5S5GY,1693
+ughost/optimize.py,sha256=_wv0iBGcDvZwW44T-PhvwZEXoDrQfomxYra7xZgAIxU,964
+ughost/parts.py,sha256=xURqlPCOhhwe2NwSi51J0TnuWucKI0kzTaMc16Zm9GE,4452
+ughost/utils.py,sha256=WhgV7_jpBUq8liy5eU2TK-IvYy_ddjzjfCOLzmqKroo,1759
+ughost-1.0.0.dist-info/METADATA,sha256=O-sF1vEec-eQ2w1qFn6AhWLYCKSh4H01If6O7HOUVEs,871
+ughost-1.0.0.dist-info/WHEEL,sha256=aeYiig01lYGDzBgS8HxWXOg3uV61G9ijOsup-k9o1sk,91
+ughost-1.0.0.dist-info/entry_points.txt,sha256=vkkyU1gm3piJUkQmliNkz0R38mNuvuQcYa6z3jMO4rk,42
+ughost-1.0.0.dist-info/top_level.txt,sha256=FodITgpiFEkTILdICCFDfFIf2DZBK8gPhh8cWnITS00,7
+ughost-1.0.0.dist-info/RECORD,,

ughost-1.0.0.dist-info/WHEEL

@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: setuptools (82.0.1)
+Root-Is-Purelib: true
+Tag: py3-none-any
+

ughost-1.0.0.dist-info/entry_points.txt

@@ -0,0 +1,2 @@
+[console_scripts]
+ghost = ughost.cli:main

ughost-1.0.0.dist-info/top_level.txt

@@ -0,0 +1 @@
+ughost