[DOCS] Port sitemap extension from nightly (#19835)

* Port sitemap extension from nightly

* Bugfix requirements

* Bugfix requirements 2

.github/workflows/build_doc.yml

@@ -62,6 +62,8 @@ jobs:
           python3 -m pip install -r docs/requirements.txt --user
           cd docs/openvino_sphinx_theme
           python3 setup.py install --user
+          cd ../openvino_custom_sphinx_sitemap
+          python3 setup.py install --user
           cd ../..
           # install doxyrest
           wget https://github.com/vovkos/doxyrest/releases/download/doxyrest-2.1.3/doxyrest-2.1.3-linux-amd64.tar.xz

docs/conf.py

@@ -44,7 +44,7 @@
     'cpplexer',
     'sphinx.ext.autodoc',
     'sphinx.ext.autosummary',
-    'sphinx_sitemap'
+    'openvino_custom_sphinx_sitemap'
 ]
 
 html_baseurl = 'https://docs.openvino.ai/canonical/'
@@ -54,6 +54,19 @@
 sitemap_url_scheme = "{link}"
 site_url = f'https://docs.openvino.ai/{version_name}/'
 
+ov_sitemap_urlset = [
+    ("xmlns", "http://www.sitemaps.org/schemas/sitemap/0.9"),
+    ("xmlns:xsi", "http://www.w3.org/2001/XMLSchema-instance"),
+    ("xmlns:coveo", "https://www.coveo.com/en/company/about-us"),
+    ("xsi:schemaLocation", "http://www.sitemaps.org/schemas/sitemap/0.9 http://www.sitemaps.org/schemas/sitemap/0.9/sitemap.xsd")
+]
+
+ov_sitemap_meta = [
+    ('coveo:metadata', {
+        'ovversion': version_name,
+    })
+]
+
 # ----------------------------------------------------
 
 

docs/openvino_custom_sphinx_sitemap/openvino_custom_sphinx_sitemap/__init__.py

@@ -0,0 +1,99 @@
+import xml.etree.ElementTree as ET
+from sphinx_sitemap import setup as base_setup, get_locales, hreflang_formatter
+
+
+def setup(app):
+    app.add_config_value(
+        'ov_sitemap_urlset',
+        default=None,
+        rebuild=''
+    )
+
+    app.add_config_value(
+        'ov_sitemap_meta',
+        default=None,
+        rebuild=''
+    )
+
+    setup = base_setup(app)
+    for listener in app.events.listeners['build-finished']:
+        if listener.handler.__name__ == 'create_sitemap':
+            app.disconnect(listener.id)
+
+    app.connect('build-finished', create_sitemap)
+    return setup
+
+
+def create_sitemap(app, exception):
+    """Generates the sitemap.xml from the collected HTML page links"""
+
+    urlset = app.builder.config.ov_sitemap_urlset
+    meta = app.builder.config.ov_sitemap_meta
+
+    site_url = app.builder.config.site_url or app.builder.config.html_baseurl
+    site_url = site_url.rstrip('/') + '/'
+    if not site_url:
+        print("sphinx-sitemap error: neither html_baseurl nor site_url "
+              "are set in conf.py. Sitemap not built.")
+        return
+    if (not app.sitemap_links):
+        print("sphinx-sitemap warning: No pages generated for %s" %
+              app.config.sitemap_filename)
+        return
+
+    ET.register_namespace('xhtml', "http://www.w3.org/1999/xhtml")
+
+    root = ET.Element("urlset")
+
+    if not urlset:
+        root.set("xmlns", "http://www.sitemaps.org/schemas/sitemap/0.9")
+    else:
+        for item in urlset:
+            root.set(*item)
+
+    get_locales(app, exception)
+
+    if app.builder.config.version:
+        version = app.builder.config.version + '/'
+    else:
+        version = ""
+
+    for link in app.sitemap_links:
+        url = ET.SubElement(root, "url")
+        scheme = app.config.sitemap_url_scheme
+        if app.builder.config.language:
+            lang = app.builder.config.language + '/'
+        else:
+            lang = ""
+
+        ET.SubElement(url, "loc").text = site_url + scheme.format(
+            lang=lang, version=version, link=link
+        )
+
+        if meta:
+            for entry in meta:
+                namespace, values = entry
+                namespace_element = ET.SubElement(url, namespace)
+                for tag_name, tag_value in values.items():
+                    ET.SubElement(namespace_element, tag_name).text = tag_value
+
+        if len(app.locales) > 0:
+            for lang in app.locales:
+                lang = lang + '/'
+                linktag = ET.SubElement(
+                    url,
+                    "{http://www.w3.org/1999/xhtml}link"
+                )
+                linktag.set("rel", "alternate")
+                linktag.set("hreflang",  hreflang_formatter(lang.rstrip('/')))
+                linktag.set("href", site_url + scheme.format(
+                    lang=lang, version=version, link=link
+                ))
+
+    filename = app.outdir + "/" + app.config.sitemap_filename
+    ET.ElementTree(root).write(filename,
+                               xml_declaration=True,
+                               encoding='utf-8',
+                               method="xml")
+    print("%s was generated for URL %s in %s" % (app.config.sitemap_filename,
+          site_url, filename))

docs/openvino_custom_sphinx_sitemap/pyproject.toml

@@ -0,0 +1,13 @@
+[build-system]
+requires        = ["setuptools"]
+build-backend   = "setuptools.build_meta"
+
+[project]
+name            = "openvino_custom_sphinx_sitemap"
+version         = "0.0.1"
+description     = "Extends sphinx-sitemap plugin with additional sitemap metadata"
+dependencies    = [
+    "sphinx >= 4.5.0",
+    "sphinx-sitemap >= 2.2.0"
+]
+requires-python = ">=3.7"

docs/openvino_custom_sphinx_sitemap/setup.py

@@ -0,0 +1,9 @@
+from setuptools import setup
+
+
+setup(
+    name="openvino_custom_sphinx_sitemap",
+    version="0.0.1",
+    install_requires=['sphinx>=4.5.0', 'sphinx-sitemap>=2.2.0'],
+    packages=['openvino_custom_sphinx_sitemap'],
+)