🗃️(backend) export to docx

We can now export our document to a docx file.
This is done by converting the html to a docx
file using the pypandoc and pandoc library.
We added the "format" param to the
generate-document endpoint, "format" accept
"pdf" or "docx" as value.

.github/workflows/impress.yml

@@ -198,7 +198,7 @@ jobs:
       - name: Install gettext (required to compile messages)
         run: |
           sudo apt-get update
-          sudo apt-get install -y gettext
+          sudo apt-get install -y gettext pandoc
 
       - name: Generate a MO file from strings extracted from the project
         run: python manage.py compilemessages

Dockerfile

@@ -75,6 +75,7 @@ RUN apt-get update && \
       libgdk-pixbuf2.0-0 \
       libpango-1.0-0 \
       libpangocairo-1.0-0 \
+      pandoc \
       shared-mime-info && \
   rm -rf /var/lib/apt/lists/*
 

src/backend/core/api/serializers.py

@@ -180,6 +180,12 @@ class DocumentGenerationSerializer(serializers.Serializer):
         required=False,
         default="html",
     )
+    format = serializers.ChoiceField(
+        choices=["pdf", "docx"],
+        label=_("Format"),
+        required=False,
+        default="pdf",
+    )
 
 
 class InvitationSerializer(serializers.ModelSerializer):

src/backend/core/api/viewsets.py

@@ -1,13 +1,11 @@
 """API endpoints"""
-from io import BytesIO
-
 from django.contrib.postgres.aggregates import ArrayAgg
 from django.db.models import (
     OuterRef,
     Q,
     Subquery,
 )
-from django.http import FileResponse, Http404
+from django.http import Http404
 
 from botocore.exceptions import ClientError
 from rest_framework import (
@@ -460,7 +458,16 @@ class TemplateViewSet(
     # pylint: disable=unused-argument
     def generate_document(self, request, pk=None):
         """
-        Generate and return pdf for this template with the content passed.
+        Generate and return a document for this template around the
+        body passed as argument.
+
+        2 types of body are accepted:
+        - HTML: body_type = "html"
+        - Markdown: body_type = "markdown"
+
+        2 types of documents can be generated:
+        - PDF: format = "pdf"
+        - Docx: format = "docx"
         """
         serializer = serializers.DocumentGenerationSerializer(data=request.data)
 
@@ -471,13 +478,10 @@ class TemplateViewSet(
 
         body = serializer.validated_data["body"]
         body_type = serializer.validated_data["body_type"]
+        export_format = serializer.validated_data["format"]
 
         template = self.get_object()
-        pdf_content = template.generate_document(body, body_type)
-
-        response = FileResponse(BytesIO(pdf_content), content_type="application/pdf")
-        response["Content-Disposition"] = f"attachment; filename={template.title}.pdf"
-        return response
+        return template.generate_document(body, body_type, export_format)
 
 
 class TemplateAccessViewSet(

src/backend/core/models.py

@@ -2,10 +2,13 @@
 Declare and configure the models for the impress core application
 """
 import hashlib
+import os
 import smtplib
+import tempfile
 import textwrap
 import uuid
 from datetime import timedelta
+from io import BytesIO
 from logging import getLogger
 
 from django.conf import settings
@@ -16,6 +19,7 @@ from django.core import exceptions, mail, validators
 from django.core.files.base import ContentFile
 from django.core.files.storage import default_storage
 from django.db import models
+from django.http import FileResponse
 from django.template.base import Template as DjangoTemplate
 from django.template.context import Context
 from django.template.loader import render_to_string
@@ -26,10 +30,10 @@ from django.utils.translation import override
 
 import frontmatter
 import markdown
+import pypandoc
+import weasyprint
 from botocore.exceptions import ClientError
 from timezone_field import TimeZoneField
-from weasyprint import CSS, HTML
-from weasyprint.text.fonts import FontConfiguration
 
 logger = getLogger(__name__)
 
@@ -564,10 +568,90 @@ class Template(BaseModel):
             "retrieve": can_get,
         }
 
-    def generate_document(self, body, body_type):
+    def generate_pdf(self, body_html, metadata):
         """
-        Generate and return a PDF document for this template around the
+        Generate and return a pdf document wrapped around the current template
+        """
+        document_html = weasyprint.HTML(
+            string=DjangoTemplate(self.code).render(
+                Context({"body": html.format_html(body_html), **metadata})
+            )
+        )
+        css = weasyprint.CSS(
+            string=self.css,
+            font_config=weasyprint.text.fonts.FontConfiguration(),
+        )
+
+        pdf_content = document_html.write_pdf(stylesheets=[css], zoom=1)
+        response = FileResponse(BytesIO(pdf_content), content_type="application/pdf")
+        response["Content-Disposition"] = f"attachment; filename={self.title}.pdf"
+
+        return response
+
+    def generate_word(self, body_html, metadata):
+        """
+        Generate and return a docx document wrapped around the current template
+        """
+        template_string = DjangoTemplate(self.code).render(
+            Context({"body": html.format_html(body_html), **metadata})
+        )
+
+        html_string = f"""
+        <!DOCTYPE html>
+        <html>
+        <head>
+            <style>
+                {self.css}
+            </style>
+        </head>
+        <body>
+            {template_string}
+        </body>
+        </html>
+        """
+
+        reference_docx = "core/static/reference.docx"
+
+        # Convert the HTML to a temporary docx file
+        with tempfile.NamedTemporaryFile(delete=False, suffix=".docx") as tmp_file:
+            output_path = tmp_file.name
+
+        pypandoc.convert_text(
+            html_string,
+            "docx",
+            format="html",
+            outputfile=output_path,
+            extra_args=["--reference-doc", reference_docx],
+        )
+
+        # Create a BytesIO object to store the output of the temporary docx file
+        with open(output_path, "rb") as f:
+            output = BytesIO(f.read())
+
+        # Remove the temporary docx file
+        os.remove(output_path)
+
+        output.seek(0)
+
+        response = FileResponse(
+            output,
+            content_type="application/vnd.openxmlformats-officedocument.wordprocessingml.document",
+        )
+        response["Content-Disposition"] = f"attachment; filename={self.title}.docx"
+        return response
+
+    def generate_document(self, body, body_type, export_format):
+        """
+        Generate and return a document for this template around the
         body passed as argument.
+
+        2 types of body are accepted:
+        - HTML: body_type = "html"
+        - Markdown: body_type = "markdown"
+
+        2 types of documents can be generated:
+        - PDF: export_format = "pdf"
+        - Docx: export_format = "docx"
         """
         document = frontmatter.loads(body)
         metadata = document.metadata
@@ -580,16 +664,10 @@ class Template(BaseModel):
                 markdown.markdown(textwrap.dedent(strip_body)) if strip_body else ""
             )
 
-        document_html = HTML(
-            string=DjangoTemplate(self.code).render(
-                Context({"body": html.format_html(body_html), **metadata})
-            )
-        )
-        css = CSS(
-            string=self.css,
-            font_config=FontConfiguration(),
-        )
-        return document_html.write_pdf(stylesheets=[css], zoom=1)
+        if export_format == "pdf":
+            return self.generate_pdf(body_html, metadata)
+
+        return self.generate_word(body_html, metadata)
 
 
 class TemplateAccess(BaseAccess):

src/backend/core/tests/templates/test_api_templates_generate_document.py

@@ -178,3 +178,26 @@ def test_api_templates_generate_document_type_unknown():
             '"unknown" is not a valid choice.',
         ]
     }
+
+
+def test_api_templates_generate_document_export_docx():
+    """Generate pdf document with the body type html."""
+    user = factories.UserFactory()
+
+    client = APIClient()
+    client.force_login(user)
+
+    template = factories.TemplateFactory(is_public=True)
+    data = {"body": "<p>Test body</p>", "body_type": "html", "format": "docx"}
+
+    response = client.post(
+        f"/api/v1.0/templates/{template.id!s}/generate-document/",
+        data,
+        format="json",
+    )
+
+    assert response.status_code == 200
+    assert (
+        response.headers["content-type"]
+        == "application/vnd.openxmlformats-officedocument.wordprocessingml.document"
+    )

src/backend/demo/data/template/code.txt

@@ -1,7 +1,7 @@
 <page size="A4">
   <div class="header">
-    <img
-      src="https://upload.wikimedia.org/wikipedia/fr/7/72/Logo_du_Gouvernement_de_la_R%C3%A9publique_fran%C3%A7aise_%282020%29.svg"
+    <img width="200"
+      src="https://impress-staging.beta.numerique.gouv.fr/assets/logo-gouv.png"
     />
   </div>
   <div class="content">

src/backend/demo/data/template/css.txt

@@ -1,18 +1,20 @@
 body {
   background: white; 
-  font-family: arial
-}
-.header {
-  display: flex;
-  justify-content: space-between;
+  font-family: arial;
 }
 .header img {
   width: 5cm;
   margin-left: -0.4cm;
 }
 .body{
-  margin-top: 1.5rem
+  margin-top: 1.5rem;
 }
 img {
   max-width: 100%;
 }
+[custom-style="center"] {
+  text-align: center;
+}
+[custom-style="right"] {
+  text-align: right;
+}

src/backend/pyproject.toml

@@ -49,6 +49,7 @@ dependencies = [
     "nested-multipart-parser==1.5.0",
     "psycopg[binary]==3.1.14",
     "PyJWT==2.8.0",
+    "pypandoc==1.13",
     "python-frontmatter==1.0.1",
     "requests==2.32.2",
     "sentry-sdk==2.8.0",