✨(marion) add pdf_options argument to AbstractDocument.create method

We want Marion consumers be able to use new pdf options offered by weasyprint 59
.0. In order to do that, we add a new `pdf_options` argument to the
`AbstractDocument.create` method.
Furthermore, weasyprint 59.0 with pdf compression enable is no more compatible
with pdfminer extract text. So until this issue is fixed, we disable by default
the pdf compression. But consumers are able to enable this option at their own
risks through the `pdf_options` argument.

CHANGELOG.md

@@ -11,6 +11,7 @@ Changelog](https://keepachangelog.com/en/1.0.0/), and this project adheres to
 ### Added
 
 - Add Django 4.2 compatibility
+- Add `pdf_options` argument to `AbstractDocument.create` method
 
 ## [0.4.0] - 2022-08-05
 

src/marion/marion/issuers/base.py

@@ -14,7 +14,7 @@
 
 from pydantic import BaseModel
 from pydantic.error_wrappers import ValidationError
-from weasyprint import CSS, HTML
+from weasyprint import CSS, DEFAULT_OPTIONS, HTML
 from weasyprint.document import DocumentMetadata
 from weasyprint.text.fonts import FontConfiguration
 
@@ -323,7 +323,16 @@ def get_django_context(self) -> Context:
         """Get the Django Context instance from the context model instance."""
         return Context(self.context.dict())
 
-    def create(self, persist=True):
+    @staticmethod
+    def _clean_pdf_options(options: dict) -> dict:
+        """Clean pdf options.
+
+        Remove any pdf options that is not in the DEFAULT_OPTIONS list.
+
+        """
+        return {key: value for key, value in options.items() if key in DEFAULT_OPTIONS}
+
+    def create(self, persist=True, pdf_options: DEFAULT_OPTIONS = None):
         """Create document.
 
         Given an HTML template, a CSS template and the required context to
@@ -342,6 +351,12 @@ def create(self, persist=True):
             When persist is False, document is created without persisting. In
             this case create returns the PDF document as bytes.
 
+        - pdf_options<dict>
+
+            Additional options to pass to Weasyprint's write_pdf method.
+            Check to see all available options:
+            https://doc.courtbouillon.org/weasyprint/stable/api_reference.html#weasyprint.DEFAULT_OPTIONS
+
         """
 
         if self.context is None:
@@ -358,10 +373,22 @@ def create(self, persist=True):
         document = html.render(stylesheets=[css], font_config=font_config)
         document.metadata = self.metadata
 
+        common_options = {
+            "zoom": 1,
+            # MARK: Disable PDF compression by default until this issue is fixed:
+            # https://github.com/Kozea/WeasyPrint/issues/1885
+            "uncompressed_pdf": True,
+        }
+        cleaned_pdf_options = (
+            self._clean_pdf_options(pdf_options) if pdf_options else {}
+        )
+
         if persist is False:
-            return document.write_pdf(zoom=1)
+            return document.write_pdf(**common_options, **cleaned_pdf_options)
 
         document_path = self.get_document_path()
-        document.write_pdf(target=document_path, zoom=1)
+        document.write_pdf(
+            target=document_path, **common_options, **cleaned_pdf_options
+        )
 
         return document_path

src/marion/marion/tests/issuers/test_base.py

@@ -12,7 +12,7 @@
 import pytest
 from pdfminer.high_level import extract_text as pdf_extract_text
 from pydantic import BaseModel
-from weasyprint.document import DocumentMetadata
+from weasyprint.document import Document, DocumentMetadata
 
 from marion.defaults import DOCUMENTS_ROOT
 from marion.exceptions import (
@@ -469,6 +469,103 @@ def fetch_context(self):
     )
 
 
+def test_abstract_document_create_with_pdf_options():
+    """Test AbstractDocument create method with pdf_options"""
+
+    # pylint: disable=missing-class-docstring
+    class ContextModel(BaseModel):
+        pass
+
+    # pylint: disable=missing-class-docstring
+    class ContextQueryModel(BaseModel):
+        pass
+
+    class TestDocument(AbstractDocument):
+        context_model = ContextModel
+        context_query_model = ContextQueryModel
+
+        def get_html(self):
+            return Template("<body>An empty document</body>")
+
+        def get_css(self):
+            return Template("")
+
+        def fetch_context(self):
+            return {}
+
+    test_document = TestDocument()
+
+    with patch.object(Document, "write_pdf") as mocked_write_pdf:
+        test_document.create(
+            pdf_options={
+                "presentational_hints": True,
+                "pdf_version": "1.5",
+                "jpeg_quality": 50,
+                "optimize_images": True,
+            }
+        )
+
+    mocked_write_pdf.assert_called_once()
+    kwargs = mocked_write_pdf.call_args[1]
+    assert kwargs["presentational_hints"] is True
+    assert kwargs["optimize_images"] is True
+    assert kwargs["pdf_version"] == "1.5"
+    assert kwargs["jpeg_quality"] == 50
+
+
+def test_abstract_document_clean_pdf_options():
+    """
+    When pdf_options are passed to the create method, they should be cleaned.
+    """
+
+    # pylint: disable=missing-class-docstring
+    class ContextModel(BaseModel):
+        pass
+
+    # pylint: disable=missing-class-docstring
+    class ContextQueryModel(BaseModel):
+        pass
+
+    class TestDocument(AbstractDocument):
+        context_model = ContextModel
+        context_query_model = ContextQueryModel
+
+        def get_html(self):
+            return Template("<body>An empty document</body>")
+
+        def get_css(self):
+            return Template("")
+
+        def fetch_context(self):
+            return {}
+
+    test_document = TestDocument()
+
+    with patch.object(Document, "write_pdf") as mocked_write_pdf:
+        test_document_file_path = test_document.create(
+            pdf_options={
+                "target": "unknown.pdf",
+                "zoom": 2,
+                "unknown_option": "unknown",
+                "jpeg_quality": 50,
+                "presentational_hints": True,
+            }
+        )
+
+    mocked_write_pdf.assert_called_once()
+    kwargs = mocked_write_pdf.call_args[1]
+
+    # Valid options should have been kept
+    assert kwargs["presentational_hints"] is True
+    assert kwargs["jpeg_quality"] == 50
+
+    # Invalid options should have been removed
+    assert kwargs.get("unknown_option") is None
+    assert kwargs["zoom"] == 1
+    assert kwargs["target"] != "unknown.pdf"
+    assert kwargs["target"] == test_document_file_path
+
+
 def test_abstract_document_jinja_template_engine(settings):
     """Test document rendering with jinja templates"""