Add image conversion backends

camelot/backends/__init__.py

@@ -0,0 +1,3 @@
+# -*- coding: utf-8 -*-
+
+from .image_conversion import ImageConversionBackend

camelot/backends/ghostscript_backend.py

@@ -0,0 +1,17 @@
+# -*- coding: utf-8 -*-
+
+import ghostscript
+
+
+class GhostscriptBackend(object):
+    def convert(self, pdf_path, png_path, resolution=300):
+        gs_args = [
+            "gs",
+            "-q",
+            "-sDEVICE=png16m",
+            "-o",
+            png_path,
+            f"-r{resolution}",
+            pdf_path,
+        ]
+        ghostscript.Ghostscript(*gs_args)

camelot/backends/image_conversion.py

@@ -0,0 +1,15 @@
+# -*- coding: utf-8 -*-
+
+from .poppler_backend import PopplerBackend
+from .ghostscript_backend import GhostscriptBackend
+
+backends = {"poppler": PopplerBackend, "ghostscript": GhostscriptBackend}
+
+
+class ImageConversionBackend(object):
+    def __init__(self, backend="poppler"):
+        self.backend = backend
+
+    def convert(self, pdf_path, png_path):
+        converter = backends[self.backend]()
+        converter.convert(pdf_path, png_path)

camelot/backends/poppler_backend.py

@@ -0,0 +1,8 @@
+# -*- coding: utf-8 -*-
+
+from pdftopng import pdftopng
+
+
+class PopplerBackend(object):
+    def convert(self, pdf_path, png_path):
+        pdftopng.convert(pdf_path, png_path)

camelot/parsers/lattice.py

@@ -29,6 +29,7 @@ from ..image_processing import (
     find_contours,
     find_joints,
 )
+from ..backends import ImageConversionBackend
 
 
 logger = logging.getLogger("camelot")
@@ -111,7 +112,7 @@ class Lattice(BaseParser):
         threshold_constant=-2,
         iterations=0,
         resolution=300,
-        **kwargs
+        **kwargs,
     ):
         self.table_regions = table_regions
         self.table_areas = table_areas
@@ -208,9 +209,8 @@ class Lattice(BaseParser):
         return t
 
     def _generate_image(self):
-        from pdftopng import pdftopng
+        converter = ImageConversionBackend()
-
+        converter.convert(self.filename, self.imagename)
-        pdftopng.convert(pdf_path=self.filename, png_path=self.imagename)
 
     def _generate_table_bbox(self):
         def scale_areas(areas):

setup.py

@@ -24,10 +24,7 @@ requires = [
     "tabulate>=0.8.9",
 ]
 
-base_requires = [
+base_requires = ["ghostscript>=0.7", "opencv-python>=3.4.2.17", "pdftopng>=0.1.1"]
-    'opencv-python>=3.4.2.17',
-    'pdftopng>=0.1.1'
-]
 
 plot_requires = [
     "matplotlib>=2.2.3",
@@ -48,36 +45,38 @@ dev_requires = dev_requires + all_requires
 
 
 def setup_package():
-    metadata = dict(name=about['__title__'],
+    metadata = dict(
-                    version=about['__version__'],
+        name=about["__title__"],
-                    description=about['__description__'],
+        version=about["__version__"],
-                    long_description=readme,
+        description=about["__description__"],
-                    long_description_content_type="text/markdown",
+        long_description=readme,
-                    url=about['__url__'],
+        long_description_content_type="text/markdown",
-                    author=about['__author__'],
+        url=about["__url__"],
-                    author_email=about['__author_email__'],
+        author=about["__author__"],
-                    license=about['__license__'],
+        author_email=about["__author_email__"],
-                    packages=find_packages(exclude=('tests',)),
+        license=about["__license__"],
-                    install_requires=requires,
+        packages=find_packages(exclude=("tests",)),
-                    extras_require={
+        install_requires=requires,
-                        'all': all_requires,
+        extras_require={
-                        'base': base_requires,
+            "all": all_requires,
-                        'dev': dev_requires,
+            "base": base_requires,
-                        'plot': plot_requires
+            "dev": dev_requires,
-                    },
+            "plot": plot_requires,
-                    entry_points={
+        },
-                        'console_scripts': [
+        entry_points={
-                            'camelot = camelot.cli:cli',
+            "console_scripts": [
-                        ],
+                "camelot = camelot.cli:cli",
-                    },
+            ],
-                    classifiers=[
+        },
-                        # Trove classifiers
+        classifiers=[
-                        # Full list: https://pypi.python.org/pypi?%3Aaction=list_classifiers
+            # Trove classifiers
-                        'License :: OSI Approved :: MIT License',
+            # Full list: https://pypi.python.org/pypi?%3Aaction=list_classifiers
-                        'Programming Language :: Python :: 3.6',
+            "License :: OSI Approved :: MIT License",
-                        'Programming Language :: Python :: 3.7',
+            "Programming Language :: Python :: 3.6",
-                        'Programming Language :: Python :: 3.8'
+            "Programming Language :: Python :: 3.7",
-                    ])
+            "Programming Language :: Python :: 3.8",
+        ],
+    )
 
     try:
         from setuptools import setup