Fix: apply PDF/A conversion once for mail archives (#14698)

src/paperless/parsers/mail.py

@@ -554,7 +554,11 @@ class MailDocumentParser:
         """
         archive_path = Path(self._tempdir) / "merged.pdf"
 
-        mail_pdf_file = self.generate_pdf_from_mail(mail_message)
+        # Apply PDF/A conversion once, when producing the final archive.
+        mail_pdf_file = self.generate_pdf_from_mail(
+            mail_message,
+            convert_to_pdfa=not mail_message.html,
+        )
 
         if pdf_layout is None:
             pdf_layout = MailRule.PdfLayout(settings.EMAIL_PARSE_DEFAULT_LAYOUT)
@@ -567,6 +571,7 @@ class MailDocumentParser:
             pdf_of_html_content = self.generate_pdf_from_html(
                 mail_message.html,
                 mail_message.attachments,
+                convert_to_pdfa=False,
             )
 
             logger.debug("Merging email text and HTML content into single PDF")
@@ -672,13 +677,20 @@ class MailDocumentParser:
 
         return html_file
 
-    def generate_pdf_from_mail(self, mail: MailMessage) -> Path:
+    def generate_pdf_from_mail(
+        self,
+        mail: MailMessage,
+        *,
+        convert_to_pdfa: bool = True,
+    ) -> Path:
         """Create a PDF from the email body using an HTML template and Gotenberg.
 
         Parameters
         ----------
         mail:
             Parsed mail message.
+        convert_to_pdfa:
+            Apply the configured PDF/A format to this output.
 
         Returns
         -------
@@ -707,10 +719,10 @@ class MailDocumentParser:
             ) as client,
             client.chromium.html_to_pdf() as route,
         ):
-            # Configure requested PDF/A formatting, if any
-            pdf_a_format = self._settings_to_gotenberg_pdfa()
-            if pdf_a_format is not None:
-                route.pdf_format(pdf_a_format)
+            if convert_to_pdfa:
+                pdf_a_format = self._settings_to_gotenberg_pdfa()
+                if pdf_a_format is not None:
+                    route.pdf_format(pdf_a_format)
 
             try:
                 response = (
@@ -742,6 +754,8 @@ class MailDocumentParser:
         self,
         orig_html: str,
         attachments: list[MailAttachment],
+        *,
+        convert_to_pdfa: bool = True,
     ) -> Path:
         """Generate a PDF from the HTML content of the email.
 
@@ -751,6 +765,8 @@ class MailDocumentParser:
             Raw HTML string from the email body.
         attachments:
             List of email attachments (used as inline resources).
+        convert_to_pdfa:
+            Apply the configured PDF/A format to this output.
 
         Returns
         -------
@@ -786,10 +802,10 @@ class MailDocumentParser:
             ) as client,
             client.chromium.html_to_pdf() as route,
         ):
-            # Configure requested PDF/A formatting, if any
-            pdf_a_format = self._settings_to_gotenberg_pdfa()
-            if pdf_a_format is not None:
-                route.pdf_format(pdf_a_format)
+            if convert_to_pdfa:
+                pdf_a_format = self._settings_to_gotenberg_pdfa()
+                if pdf_a_format is not None:
+                    route.pdf_format(pdf_a_format)
 
             # Add attachments as resources, cleaning the filename and replacing
             # it in the index file for inclusion

src/paperless/tests/parsers/test_mail_parser.py

@@ -594,6 +594,23 @@ class TestParser:
 
         assert mail_parser.get_archive_path() is not None
 
+    def test_generate_pdf_from_html_with_pdfa(
+        self,
+        httpx_mock: HTTPXMock,
+        mail_parser: MailDocumentParser,
+        html_email_pdf_file: Path,
+    ) -> None:
+        url = "http://localhost:3000/forms/chromium/convert/html"
+        pdf_content = html_email_pdf_file.read_bytes()
+        httpx_mock.add_response(url=url, method="POST", content=pdf_content)
+
+        archive = mail_parser.generate_pdf_from_html("<p>Email body</p>", [])
+
+        assert archive.read_bytes() == pdf_content
+        request = httpx_mock.get_request(url=url)
+        assert request is not None
+        assert b'name="pdfa"\r\n\r\nPDF/A-2b\r\n' in request.content
+
     @pytest.mark.httpx_mock(can_send_already_matched_responses=True)
     def test_generate_pdf_html_email(
         self,
@@ -635,6 +652,20 @@ class TestParser:
 
         assert mail_parser.get_archive_path() is not None
 
+        chromium_requests = httpx_mock.get_requests(
+            url="http://localhost:3000/forms/chromium/convert/html",
+        )
+        assert len(chromium_requests) == 2
+        assert all(
+            b'name="pdfa"' not in request.content for request in chromium_requests
+        )
+
+        merge_request = httpx_mock.get_request(
+            url="http://localhost:3000/forms/pdfengines/merge",
+        )
+        assert merge_request is not None
+        assert b'name="pdfa"\r\n\r\nPDF/A-2b\r\n' in merge_request.content
+
     def test_generate_pdf_html_email_html_to_pdf_failure(
         self,
         httpx_mock: HTTPXMock,