refactor: Updated version of pypdfium2 and its calls (#845)

* Switch to new pypdfium2 API

* conf.py: fix intersphinx link

* apply review suggestions

* Fix the intersphinx function reference

I need to do something to make these paths shorter.

docs/source/conf.py

@@ -39,6 +39,7 @@
 extensions = [
 	'sphinx.ext.autodoc',
 	'sphinx.ext.napoleon',
+	'sphinx.ext.intersphinx',
 	'sphinx.ext.viewcode',
     'sphinx.ext.coverage',
     'sphinx.ext.mathjax',
@@ -49,6 +50,11 @@
     'sphinx_markdown_tables',
 ]
 
+intersphinx_mapping = {
+    'python': ('https://docs.python.org/3', None),
+    'pypdfium2': ('https://pypdfium2.readthedocs.io/en/stable/', None),
+}
+
 napoleon_use_ivar = True
 
 # Add any paths that contain templates here, relative to this directory.

doctr/io/pdf.py

@@ -3,6 +3,7 @@
 # This program is licensed under the Apache License version 2.
 # See LICENSE or go to <https://www.apache.org/licenses/LICENSE-2.0.txt> for full license details.
 
+import os.path
 from pathlib import Path
 from typing import Any, List
 
@@ -24,15 +25,18 @@ def read_pdf(file: AbstractFile, scale: float = 2, **kwargs: Any) -> List[np.nda
     Args:
         file: the path to the PDF file
         scale: rendering scale (1 corresponds to 72dpi)
+        kwargs: additional parameters to :func:`pypdfium2._helpers.pdf_renderer.render_pdf_topil`
     Returns:
-        the list of pages decoded as numpy ndarray of shape H x W x 3
+        the list of pages decoded as numpy ndarray of shape H x W x C
     """
 
-    if not isinstance(file, (str, Path, bytes)):
+    if isinstance(file, Path):
+        file = str(file)
+    if not isinstance(file, (str, bytes)):
         raise TypeError("unsupported object type for argument 'file'")
 
-    if isinstance(file, (str, Path)) and not Path(file).is_file():
+    if isinstance(file, str) and not os.path.isfile(file):
         raise FileNotFoundError(f"unable to access {file}")
 
-    # Read pages with fitz and convert them to numpy ndarrays
-    return [np.asarray(img) for img, _ in pdfium.render_pdf(file, scale=scale)]
+    # Rasterise pages to PIL images with pypdfium2 and convert to numpy ndarrays
+    return [np.asarray(img) for img, _ in pdfium.render_pdf_topil(file, scale=scale, **kwargs)]

requirements-pt.txt

@@ -2,7 +2,7 @@ numpy>=1.16.0
 scipy>=1.4.0
 h5py>=3.1.0
 opencv-python>=3.4.5.20
-pypdfium2>=0.14.0
+pypdfium2>=1.0.0
 pyclipper>=1.2.0
 shapely>=1.6.0
 matplotlib>=3.1.0,<3.4.3

requirements.txt

@@ -2,7 +2,7 @@ numpy>=1.16.0
 scipy>=1.4.0
 h5py>=3.1.0
 opencv-python>=3.4.5.20
-pypdfium2>=0.14.0
+pypdfium2>=1.0.0
 pyclipper>=1.2.0
 shapely>=1.6.0
 matplotlib>=3.1.0,<3.4.3

setup.py

@@ -45,7 +45,7 @@
     "h5py>=3.1.0",
     "opencv-python>=3.4.5.20",
     "tensorflow>=2.4.0",
-    "pypdfium2>=0.14.0",
+    "pypdfium2>=1.0.0",
     "pyclipper>=1.2.0",
     "shapely>=1.6.0",
     "matplotlib>=3.1.0,<3.4.3",