govoplan-connectors/tests/test_feeds.py

from __future__ import annotations

import unittest
from unittest.mock import patch

from defusedxml import ElementTree as SafeET
from fastapi import HTTPException

from govoplan_connectors.backend.feeds import (
    FEED_PRIVATE_PUBLISH_SCOPE,
    FEED_PUBLISH_SCOPE,
    ConnectorFeedProvider,
    feed_rows,
)
from govoplan_connectors.backend.router import api_render_feed
from govoplan_connectors.backend.schemas import FeedRenderPayload
from govoplan_core.auth import ApiPrincipal
from govoplan_core.core.access import PrincipalRef
from govoplan_core.core.feeds import (
    FeedCapabilityError,
    FeedEntry,
    FeedRenderRequest,
)
from govoplan_core.security.http_fetch import HttpFetchResponse


RSS = b"""<?xml version="1.0"?>
<rss version="2.0"><channel>
  <title>Decisions</title><link>https://example.test/</link>
  <description>Published decisions</description>
  <item><guid>decision-1</guid><title>Decision one</title>
    <link>https://example.test/1</link>
    <pubDate>Fri, 31 Jul 2026 10:00:00 GMT</pubDate>
    <category>planning</category>
  </item>
</channel></rss>"""

ATOM = b"""<?xml version="1.0"?>
<feed xmlns="http://www.w3.org/2005/Atom">
  <id>https://example.test/feed</id><title>Updates</title>
  <updated>2026-07-31T10:00:00Z</updated>
  <link href="https://example.test/" />
  <entry><id>update-1</id><title>Update one</title>
    <updated>2026-07-31T10:00:00Z</updated>
    <link href="https://example.test/update-1" />
  </entry>
</feed>"""


class ConnectorFeedProviderTests(unittest.TestCase):
    def setUp(self) -> None:
        self.provider = ConnectorFeedProvider()

    def test_rss_and_atom_are_normalized_to_tabular_entries(self) -> None:
        rss = self.provider.parse(RSS, source_url="https://example.test/rss")
        atom = self.provider.parse(ATOM, source_url="https://example.test/atom")

        self.assertEqual("rss", rss.format)
        self.assertEqual("decision-1", rss.entries[0].id)
        self.assertEqual("atom", atom.format)
        self.assertEqual("https://example.test/update-1", atom.entries[0].url)
        self.assertEqual("decision-1", feed_rows(rss)[0]["id"])
        self.assertEqual(64, len(rss.sha256))

    def test_fetch_records_transport_freshness_and_provenance(self) -> None:
        with patch(
            "govoplan_connectors.backend.feeds.fetch_http",
            return_value=HttpFetchResponse(
                status=200,
                headers={
                    "Content-Type": "application/rss+xml",
                    "Cache-Control": "public, max-age=600",
                    "ETag": '"feed-1"',
                },
                body=RSS,
            ),
        ):
            document = self.provider.fetch("https://example.test/rss")

        self.assertEqual('"feed-1"', document.etag)
        self.assertIsNotNone(document.acquired_at)
        self.assertEqual(600, int((document.fresh_until - document.acquired_at).total_seconds()))
        self.assertEqual(len(RSS), document.metadata["byte_count"])

    def test_render_filters_entries_by_explicit_visibility(self) -> None:
        request = FeedRenderRequest(
            format="atom",
            title="Public updates",
            feed_url="https://example.test/feed.atom",
            home_url="https://example.test/",
            entries=(
                FeedEntry(id="public-1", title="Public", visibility="public"),
                FeedEntry(id="tenant-1", title="Tenant", visibility="tenant"),
            ),
            allowed_visibilities=frozenset({"public"}),
        )
        result = self.provider.render(request)
        root = SafeET.fromstring(result.body)

        self.assertEqual(1, result.included_entries)
        self.assertEqual(1, result.excluded_entries)
        self.assertIn(b"Public", result.body)
        self.assertNotIn(b"Tenant", result.body)
        self.assertTrue(root.tag.endswith("feed"))

    def test_unsafe_xml_is_rejected(self) -> None:
        payload = b'<!DOCTYPE x [<!ENTITY y SYSTEM "file:///etc/passwd">]><rss>&y;</rss>'
        with self.assertRaisesRegex(FeedCapabilityError, "not safe or valid"):
            self.provider.parse(payload, source_url="https://example.test/rss")

    def test_render_api_derives_visibility_from_audience_and_permissions(self) -> None:
        entries = [
            {
                "id": visibility,
                "title": visibility.title(),
                "visibility": visibility,
                "source_kind": source_kind,
                "source_module": source_module,
                "source_ref": f"{source_kind}:{visibility}",
                "source_revision": "7",
            }
            for visibility, source_kind, source_module in (
                ("public", "publication", "docs"),
                ("tenant", "case", "cases"),
                ("private", "report", "reporting"),
            )
        ]
        public_payload = FeedRenderPayload(
            format="rss",
            title="Selected GovOPlaN updates",
            feed_url="https://example.test/feed.xml",
            home_url="https://example.test/",
            audience="public",
            entries=entries,
        )
        public = api_render_feed(
            public_payload,
            principal=_principal(FEED_PUBLISH_SCOPE),
        )

        self.assertEqual("public", public.headers["x-govoplan-feed-audience"])
        self.assertEqual("1", public.headers["x-govoplan-feed-included"])
        self.assertEqual("2", public.headers["x-govoplan-feed-excluded"])
        self.assertIn(b"Public", public.body)
        self.assertNotIn(b"Tenant", public.body)

        restricted_payload = public_payload.model_copy(update={"audience": "private"})
        with self.assertRaises(HTTPException) as denied:
            api_render_feed(
                restricted_payload,
                principal=_principal(FEED_PUBLISH_SCOPE),
            )
        self.assertEqual(403, denied.exception.status_code)

        restricted = api_render_feed(
            restricted_payload,
            principal=_principal(
                FEED_PUBLISH_SCOPE,
                FEED_PRIVATE_PUBLISH_SCOPE,
            ),
        )
        self.assertEqual("3", restricted.headers["x-govoplan-feed-included"])
        self.assertIn(b"Private", restricted.body)


def _principal(*scopes: str) -> ApiPrincipal:
    return ApiPrincipal(
        principal=PrincipalRef(
            account_id="account-1",
            membership_id="membership-1",
            tenant_id="tenant-1",
            scopes=frozenset(scopes),
        ),
        account=object(),
        user=object(),
    )


if __name__ == "__main__":
    unittest.main()