[ADD] website_url_sitemap_exclusion

setup/website_url_sitemap_exclusion/odoo/addons/website_url_sitemap_exclusion

@@ -0,0 +1 @@
+../../../../website_url_sitemap_exclusion

setup/website_url_sitemap_exclusion/setup.py

@@ -0,0 +1,6 @@
+import setuptools
+
+setuptools.setup(
+    setup_requires=['setuptools-odoo'],
+    odoo_addon=True,
+)

website_url_sitemap_exclusion/__init__.py

@@ -0,0 +1,3 @@
+#  License AGPL-3.0 or later (https://www.gnu.org/licenses/agpl).
+
+from . import models

website_url_sitemap_exclusion/__manifest__.py

@@ -0,0 +1,21 @@
+#  Copyright 2023 Simone Rubino - TAKOBI
+#  License AGPL-3.0 or later (https://www.gnu.org/licenses/agpl).
+
+{
+    "name": "Website URL Exclusion",
+    "summary": "Exclude URLs from the website sitemap",
+    "version": "12.0.1.0.0",
+    "category": "Website",
+    "license": "AGPL-3",
+    "website": "https://github.com/OCA/website"
+               "/tree/12.0/website_url_sitemap_exclusion",
+    "author": "TAKOBI, "
+              "Odoo Community Association (OCA)",
+    "depends": [
+        "website",
+    ],
+    "data": [
+        "security/ir.model.access.csv",
+        "views/website_views.xml",
+    ],
+}

website_url_sitemap_exclusion/i18n/it.po

@@ -0,0 +1,80 @@
+# Translation of Odoo Server.
+# This file contains the translation of the following modules:
+#	* website_url_sitemap_exclusion
+#
+msgid ""
+msgstr ""
+"Project-Id-Version: Odoo Server 12.0\n"
+"Report-Msgid-Bugs-To: \n"
+"POT-Creation-Date: 2023-02-27 13:13+0000\n"
+"PO-Revision-Date: 2023-02-27 13:13+0000\n"
+"Last-Translator: Simone Rubino <[email protected]>\n"
+"Language-Team: \n"
+"MIME-Version: 1.0\n"
+"Content-Type: text/plain; charset=UTF-8\n"
+"Content-Transfer-Encoding: \n"
+"Plural-Forms: \n"
+
+#. module: website_url_sitemap_exclusion
+#: model:ir.model.fields,field_description:website_url_sitemap_exclusion.field_website_forbidden_url__create_uid
+msgid "Created by"
+msgstr "Creato da"
+
+#. module: website_url_sitemap_exclusion
+#: model:ir.model.fields,field_description:website_url_sitemap_exclusion.field_website_forbidden_url__create_date
+msgid "Created on"
+msgstr "Creato il"
+
+#. module: website_url_sitemap_exclusion
+#: model:ir.model.fields,field_description:website_url_sitemap_exclusion.field_website_forbidden_url__display_name
+msgid "Display Name"
+msgstr "Nome visualizzato"
+
+#. module: website_url_sitemap_exclusion
+#: model:ir.model.fields,field_description:website_url_sitemap_exclusion.field_website__forbidden_url_ids
+msgid "Forbidden URLs regex"
+msgstr "Espressioni regolari per URL vietati"
+
+#. module: website_url_sitemap_exclusion
+#: model:ir.model.fields,field_description:website_url_sitemap_exclusion.field_website_forbidden_url__id
+msgid "ID"
+msgstr ""
+
+#. module: website_url_sitemap_exclusion
+#: model:ir.model.fields,field_description:website_url_sitemap_exclusion.field_website_forbidden_url____last_update
+msgid "Last Modified on"
+msgstr "Ultima modifica il"
+
+#. module: website_url_sitemap_exclusion
+#: model:ir.model.fields,field_description:website_url_sitemap_exclusion.field_website_forbidden_url__write_uid
+msgid "Last Updated by"
+msgstr "Ultimo aggiornamento di"
+
+#. module: website_url_sitemap_exclusion
+#: model:ir.model.fields,field_description:website_url_sitemap_exclusion.field_website_forbidden_url__write_date
+msgid "Last Updated on"
+msgstr "Ultimo aggiornamento il"
+
+#. module: website_url_sitemap_exclusion
+#: model:ir.model.fields,field_description:website_url_sitemap_exclusion.field_website_forbidden_url__regex
+msgid "Regex"
+msgstr "Espressione regolare"
+
+#. module: website_url_sitemap_exclusion
+#: code:addons/website_url_sitemap_exclusion/models/website_forbidden_url.py:35
+#, python-format
+msgid "Regular Expression `{regex}` is not valid:\n"
+"{error}"
+msgstr "L'espressione regolare `{regex}` non è valida:\n"
+"{error}"
+
+#. module: website_url_sitemap_exclusion
+#: model:ir.model,name:website_url_sitemap_exclusion.model_website_forbidden_url
+msgid "URL not allowed in the sitemap"
+msgstr "URL non consentito nella sitemap"
+
+#. module: website_url_sitemap_exclusion
+#: model:ir.model,name:website_url_sitemap_exclusion.model_website
+#: model:ir.model.fields,field_description:website_url_sitemap_exclusion.field_website_forbidden_url__website_id
+msgid "Website"
+msgstr "Sito web"

website_url_sitemap_exclusion/models/__init__.py

@@ -0,0 +1,4 @@
+#  License AGPL-3.0 or later (https://www.gnu.org/licenses/agpl).
+
+from . import website
+from . import website_forbidden_url

website_url_sitemap_exclusion/models/website.py

@@ -0,0 +1,39 @@
+#  Copyright 2023 Simone Rubino - TAKOBI
+#  License AGPL-3.0 or later (https://www.gnu.org/licenses/agpl).
+
+import re
+from odoo import api, fields, models
+
+
+class Website(models.Model):
+    _inherit = 'website'
+
+    forbidden_url_ids = fields.One2many(
+        comodel_name="website.forbidden.url",
+        inverse_name="website_id",
+        string="Forbidden URLs regex",
+    )
+
+    @api.multi
+    def enumerate_pages(self, query_string=None, force=False):
+        pages = super().enumerate_pages(query_string=query_string, force=force)
+        forbidden_url_regexes = self.mapped("forbidden_url_ids.regex")
+        forbidden_url_compiled_regexes = list(map(re.compile, forbidden_url_regexes))
+        for page in pages:
+            is_forbidden = \
+                self._is_enumerated_page_forbidden_by_regexes(
+                    page, forbidden_url_compiled_regexes,
+                )
+            if not is_forbidden:
+                yield page
+
+    def _is_enumerated_page_forbidden_by_regexes(self, page, compiled_regexes):
+        page_url = page["loc"]
+        for forbidden_url_compiled_regex in compiled_regexes:
+            is_url_forbidden = forbidden_url_compiled_regex.match(page_url)
+            if is_url_forbidden:
+                is_forbidden = True
+                break
+        else:
+            is_forbidden = False
+        return is_forbidden

website_url_sitemap_exclusion/models/website_forbidden_url.py

@@ -0,0 +1,41 @@
+#  Copyright 2023 Simone Rubino - TAKOBI
+#  License AGPL-3.0 or later (https://www.gnu.org/licenses/agpl).
+
+import re
+
+from odoo import _, api, fields, models
+from odoo.exceptions import ValidationError
+
+
+class WebsiteForbiddenUrls(models.Model):
+    _name = "website.forbidden.url"
+    _description = "URL not allowed in the sitemap"
+    _rec_name = 'regex'
+
+    regex = fields.Char(
+        string="Regular Expression",
+        required=True,
+    )
+    website_id = fields.Many2one(
+        comodel_name="website",
+        ondelete="cascade",
+    )
+
+    @api.constrains(
+        'regex',
+    )
+    def constrain_regex_valid(self):
+        """RegEx must be valid."""
+        for forbidden_url in self:
+            regex = forbidden_url.regex
+            try:
+                re.compile(regex)
+            except re.error as error:
+                raise ValidationError(
+                    _("Regular Expression `{regex}` is not valid:\n"
+                      "{error}")
+                    .format(
+                        regex=regex,
+                        error=error.msg,
+                    )
+                ) from error

website_url_sitemap_exclusion/readme/CONFIGURE.rst

@@ -0,0 +1 @@
+In the Website form, add Regular Expressions to the list `Forbidden URLs regex`.

website_url_sitemap_exclusion/readme/DESCRIPTION.rst

@@ -0,0 +1,6 @@
+Allow to exclude URLs from the sitemap.
+
+The URLs can be selected using Regular Expressions.
+
+This module can be used to exclude URLs that cannot be excluded using page properties;
+for instance, this can be used to exclude the URL `/shop` if there are no published products.

website_url_sitemap_exclusion/readme/ROADMAP.rst

@@ -0,0 +1 @@
+Model `website.forbidden.url` can be renamed to `website.blacklisted.url`.

website_url_sitemap_exclusion/security/ir.model.access.csv

@@ -0,0 +1,3 @@
+id,name,model_id:id,group_id:id,perm_read,perm_write,perm_create,perm_unlink
+access_website_forbidden_url_all,Everyone can read forbidden URLs,model_website_forbidden_url,,1,0,0,0
+manage_website_forbidden_url_designer,Designers can manage forbidden URLs,model_website_forbidden_url,website.group_website_designer,1,1,1,1

website_url_sitemap_exclusion/tests/__init__.py

@@ -0,0 +1,3 @@
+#  License AGPL-3.0 or later (https://www.gnu.org/licenses/agpl).
+
+from . import test_sitemap_exclusion

website_url_sitemap_exclusion/tests/test_sitemap_exclusion.py

@@ -0,0 +1,80 @@
+#  Copyright 2023 Simone Rubino - TAKOBI
+#  License AGPL-3.0 or later (https://www.gnu.org/licenses/agpl).
+
+import datetime
+from unittest.mock import patch
+
+from odoo import tests
+from odoo.addons.website.controllers.main import Website
+from odoo.addons.website.tools import MockRequest
+from odoo.exceptions import ValidationError
+from odoo.tests import Form
+
+
+class TestSiteMapExclusion (tests.SavepointCase):
+
+    @classmethod
+    def setUpClass(cls):
+        super().setUpClass()
+        cls.controller = Website()
+        cls.website = cls.env['website'].browse(1)
+        cls.sitemap = cls._get_sitemap(cls.env)
+
+    @classmethod
+    def _get_sitemap(cls, env):
+        """Get the updated sitemap.
+
+        Usually the sitemap is cached for 12 hours,
+        here we reset the cache so that the sitemap returned is always updated.
+        """
+        base_url = cls.env['ir.config_parameter'].sudo().get_param('web.base.url')
+
+        sitemap_cache_duration_path = 'odoo.addons.website' \
+                                      '.controllers.main.SITEMAP_CACHE_TIME'
+        no_cache = datetime.timedelta(seconds=-1)
+        with patch(sitemap_cache_duration_path, new=no_cache), \
+                MockRequest(env, website=cls.website) as request:
+            request.httprequest.url_root = base_url
+            cls.controller.sitemap_xml_index()
+            # Sitemap content is the first parameter
+            # of the first call to `make_response`
+            sitemap = request.make_response.call_args[0][0]
+        return sitemap
+
+    def test_url_exclusion(self):
+        """URLs defined in the Website are excluded from the sitemap."""
+        # Arrange: We want to exclude a URL
+        url_to_exclude = '/aboutus'
+        url_regex = '/aboutus'
+        sitemap = self.sitemap
+        # pre-condition: The URL is in the sitemap,
+        # and the regex matches the URL
+        self.assertRegex(url_to_exclude, url_regex)
+        self.assertIn(url_to_exclude, sitemap.decode())
+
+        # Act: Exclude the URL in the website
+        website_form = Form(self.website)
+        with website_form.forbidden_url_ids.new() as forbid:
+            forbid.regex = url_regex
+        website_form.save()
+
+        # Assert: The sitemap does not contain the URL
+        new_sitemap = TestSiteMapExclusion._get_sitemap(self.env)
+        self.assertNotIn(url_to_exclude, new_sitemap.decode())
+
+    def test_not_valid_regex(self):
+        """Only valid Regular Expressions can be used."""
+        # Arrange: A Regular Expression is not valid
+        not_valid_regex = '/abo[utus'
+
+        # Act: Add the Regular Expression to the website forbidden URLs
+        website_form = Form(self.website)
+        with website_form.forbidden_url_ids.new() as forbid:
+            forbid.regex = not_valid_regex
+        with self.assertRaises(ValidationError) as ve:
+            website_form.save()
+        exc_message = ve.exception.args[0]
+
+        # Assert: The validation error mentions the wrong Regular Expression
+        self.assertIn(not_valid_regex, exc_message)
+        self.assertIn('not valid', exc_message)

website_url_sitemap_exclusion/views/website_views.xml

@@ -0,0 +1,24 @@
+<?xml version="1.0" encoding="utf-8"?>
+<!--
+  ~ Copyright 2023 Simone Rubino - TAKOBI
+  ~ License AGPL-3.0 or later (https://www.gnu.org/licenses/agpl).
+  -->
+
+<odoo>
+    <record id="view_website_form" model="ir.ui.view">
+        <field name="name">Add Forbidden URLs to Website Form View</field>
+        <field name="model">website</field>
+        <field name="inherit_id" ref="website.view_website_form"/>
+        <field name="arch" type="xml">
+            <div name="other" position="after">
+                <group name="forbidden_urls">
+                    <field name="forbidden_url_ids">
+                        <tree editable="top">
+                            <field name="regex"/>
+                        </tree>
+                    </field>
+                </group>
+            </div>
+        </field>
+    </record>
+</odoo>