✨ document templates: manage export in PDF, HTML, xlsx, docx using LO unoconv

6 files changed, 70 insertions, 14 deletions

diff --git a/example_project/settings.py b/example_project/settings.py
index 4fa80206c..cc2e44b3d 100644
--- a/example_project/settings.py
+++ b/example_project/settings.py
@@ -328,6 +328,8 @@ DATA_UPLOAD_MAX_NUMBER_FIELDS = 10240
 DOT_BINARY = "/usr/bin/dot"
 PDFTOPPM_BINARY = ""
 
+UNOCONV_BINARY = "/usr/bin/unoconv"
+
 TEST_RUNNER = "ishtar_common.tests.ManagedModelTestRunner"
 SELENIUM_TEST = False
 CELERY_BROKER_URL = ""
diff --git a/ishtar_common/libreoffice.py b/ishtar_common/libreoffice.py
index 3869ca81a..573d99361 100644
--- a/ishtar_common/libreoffice.py
+++ b/ishtar_common/libreoffice.py
@@ -1,4 +1,4 @@
-#!/usr/bin/env python
+#!/usr/bin/env python3
 # -*- coding: utf-8 -*-
 import time
 
@@ -8,15 +8,37 @@ from com.sun.star.beans import PropertyValue
 from com.sun.star.connection import NoConnectException
 from com.sun.star.sheet.ValidationType import LIST
 
+# nosec: filename used is generated and sanitized
+import subprocess  # nosec
+
 #from com.sun.star.table import CellRangeAddress, CellAddress
 
-from ishtar_common.utils import num2col
+from ishtar_common.utils import num2col, sanitize_filepath
 
 from django.conf import settings
 
 
 RETRY = 5
 
+def get_connection():
+    return "socket,host={},port={};urp".format(
+        settings.LIBREOFFICE_HOST, settings.LIBREOFFICE_PORT
+    )
+
+
+def convert_document(document, destination_format):
+    if not document.lower().endswith(".ods") and not document.lower().endswith(".odt"):
+        return document
+    document = sanitize_filepath(document)
+    connection = get_connection() + ";StarOffice.ComponentContext"
+    args = [settings.UNOCONV_BINARY, "--connection", connection, "-f", destination_format,
+            document]
+    # nosec: filename is generated and sanitized
+    popen = subprocess.Popen(args)  # nosec
+    popen.wait(timeout=20)
+    dest_document = document[:-4] + "." + destination_format
+    return dest_document
+
 
 class UnoClient:
     def __init__(self):
@@ -30,9 +52,7 @@ class UnoClient:
 
         resolver = local_context.ServiceManager.createInstanceWithContext(
             "com.sun.star.bridge.UnoUrlResolver", local_context)
-        connection = "socket,host={},port={};urp".format(
-            settings.LIBREOFFICE_HOST, settings.LIBREOFFICE_PORT
-        )
+        connection = get_connection()
         try:
             self.service_manager = resolver.resolve(
                 "uno:{};StarOffice.ServiceManager".format(connection))
diff --git a/ishtar_common/models.py b/ishtar_common/models.py
index 73f140625..c7fe7be8a 100644
--- a/ishtar_common/models.py
+++ b/ishtar_common/models.py
@@ -2057,11 +2057,23 @@ class Dashboard(object):
                 return v
 
 
-EXPORT_FORMATS = []
+EXPORT_FORMATS = [("", "---")]
+EXPORT_FORMATS_CONTENT_TYPE = {
+    "xlsx": "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet",
+    "docx": "application/vnd.openxmlformats-officedocument.wordprocessingml.document",
+    "pdf": "application/pdf",
+    "html": "text/html",
+    "odt": "application/vnd.oasis.opendocument.text",
+    "ods": "application/vnd.oasis.opendocument.spreadsheet",
+}
 
 if settings.USE_LIBREOFFICE:
-    EXPORT_FORMATS.append(("xlsx", _("XLSX")))
-    EXPORT_FORMATS.append(("pdf", _("PDF")))
+    EXPORT_FORMATS = [
+        ("docx", _("DOCX")),
+        ("html", _("HTML")),
+        ("pdf", _("PDF")),
+        ("xlsx", _("XLSX")),
+    ]
 
 EXPORT_FORMATS_DICT = dict(EXPORT_FORMATS)
 
@@ -2124,7 +2136,7 @@ class DocumentTemplate(models.Model):
     def clean(self):
         if self.for_labels and not self.label_per_page:
             raise ValidationError(
-                _("For label template, you must provide " "number of label per page.")
+                _("For label template, you must provide number of label per page.")
             )
 
     def generate_label_template(self):
diff --git a/ishtar_common/utils.py b/ishtar_common/utils.py
index bca7ce181..d4453ff5a 100644
--- a/ishtar_common/utils.py
+++ b/ishtar_common/utils.py
@@ -1446,6 +1446,14 @@ def clean_session_cache(session):
     cache.set(cache_key_list, [], settings.CACHE_TIMEOUT)
 
 
+def sanitize_filepath(filepath):
+    #TODO: python3-pathvalidate
+    # from pathvalidate import sanitize_filepath
+    # filepath -> sanitize_filepath(filepath)
+    keep_characters = (" ", ".", "_", "-", "/")
+    return "".join(c for c in filepath if c.isalnum() or c in keep_characters).rstrip()
+
+
 def get_field_labels_from_path(model, path):
     """
     :param model: base model
diff --git a/ishtar_common/views.py b/ishtar_common/views.py
index 5b675eb52..a6c3b373d 100644
--- a/ishtar_common/views.py
+++ b/ishtar_common/views.py
@@ -91,6 +91,10 @@ from ishtar_common.utils import (
 from ishtar_common.widgets import JQueryAutoComplete
 from ishtar_common import tasks
 
+convert_document = None
+if settings.USE_LIBREOFFICE:
+    from ishtar_common.libreoffice import convert_document
+
 from .views_item import (
     CURRENT_ITEM_KEYS,
     CURRENT_ITEM_KEYS_DICT,
@@ -1535,10 +1539,19 @@ class GenerateView(IshtarMixin, LoginRequiredMixin, View):
         document = self.publish(tpl, objects)
         if not document:
             return HttpResponse(content_type="text/plain")
+        base_extension = tpl.template.name.split(".")[-1].lower()
+        extension = tpl.export_format
+        if not extension:
+            extension = base_extension
+        if not settings.USE_LIBREOFFICE and extension not in ("ods", "odt"):
+            return HttpResponseBadRequest("Invalid format type - need LibreOffice daemon")  # bad configuration
+        content_type = models.EXPORT_FORMATS_CONTENT_TYPE.get(
+            extension, "application/vnd.oasis.opendocument.text"
+        )
+        if tpl.export_format and convert_document:
+            document = convert_document(document, tpl.export_format)
         with open(document, "rb") as f:
-            response = HttpResponse(
-                f.read(), content_type="application/vnd.oasis.opendocument.text"
-            )
+            response = HttpResponse(f.read(), content_type=content_type)
             response["Content-Disposition"] = "attachment; filename={}".format(
                 document.split(os.sep)[-1]
             )
diff --git a/requirements.txt b/requirements.txt
index 2e666c9de..d5344e5b0 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -68,7 +68,8 @@ django-extensions==3.0.3
 
 pandas==1.1.5
 django-ipware==3.0.0
-
 django-axes==5.4.3
+num2words==0.5.9
+
 
-num2words==0.5.9
\ No newline at end of file
+# TODO v5: python3-pathvalidate -> libreoffice.py
\ No newline at end of file