devplacepy/tests/api/sitemapxml.py

# retoor <retoor@molodetz.nl>

import time
import pytest
import requests
from tests.conftest import BASE_URL
from devplacepy.database import get_table
from devplacepy.utils import clear_user_cache
from devplacepy import project_files
from devplacepy.project_files import ProjectFileError
from devplacepy.services.devii.actions.dispatcher import (
    confirmation_error,
    _is_confirmed,
)
_counter_project_visibility = [0]
def _signup_project_visibility():
    _counter_project_visibility[0] += 1
    name = f"pv{int(time.time() * 1000)}{_counter_project_visibility[0]}"
    session = requests.Session()
    session.post(
        f"{BASE_URL}/auth/signup",
        data={
            "username": name,
            "email": f"{name}@t.dev",
            "password": "secret123",
            "confirm_password": "secret123",
            "birth_date": "1990-01-01",
            "accept_terms": "1",
        },
        allow_redirects=True,
    )
    row = get_table("users").find_one(username=name)
    return name, row["uid"], row["api_key"]
def _make_admin_project_visibility():
    name, uid, key = _signup_project_visibility()
    get_table("users").update({"uid": uid, "role": "Admin"}, ["uid"])
    clear_user_cache(uid)
    return name, uid, key
def _h_project_visibility(key=None):
    headers = {"Accept": "application/json"}
    if key:
        headers["X-API-KEY"] = key
    return headers
def _create_project_project_visibility(key, title, is_private=False):
    data = {
        "title": title,
        "description": "visibility test",
        "project_type": "software",
        "status": "In Development",
    }
    if is_private:
        data["is_private"] = "on"
    r = requests.post(f"{BASE_URL}/projects/create", headers=_h_project_visibility(key), data=data)
    assert r.status_code == 200, r.text
    return r.json()["data"]
def _project_uid(slug):
    return get_table("projects").find_one(slug=slug)["uid"]
def _write_project_visibility(key, slug, path, content):
    return requests.post(
        f"{BASE_URL}/projects/{slug}/files/write",
        headers=_h_project_visibility(key),
        data={"path": path, "content": content},
        allow_redirects=False,
    )
def _set_private(key, slug, value):
    return requests.post(
        f"{BASE_URL}/projects/{slug}/private",
        headers=_h_project_visibility(key),
        data={"value": 1 if value else 0},
        allow_redirects=False,
    )
def _set_readonly(key, slug, value):
    return requests.post(
        f"{BASE_URL}/projects/{slug}/readonly",
        headers=_h_project_visibility(key),
        data={"value": 1 if value else 0},
        allow_redirects=False,
    )
def _list_slugs(key=None, user_uid=None):
    params = {"user_uid": user_uid} if user_uid else None
    r = requests.get(f"{BASE_URL}/projects", headers=_h_project_visibility(key), params=params)
    return [p["slug"] for p in r.json()["projects"]]
def _seed_news_seo():
    from datetime import datetime, timezone
    from devplacepy.database import get_table
    from devplacepy.utils import generate_uid, make_combined_slug

    uid = generate_uid()
    title = f"SEO Test News Article {uid.split('-')[-1]}"
    slug = make_combined_slug(title, uid)
    get_table("news").insert(
        {
            "deleted_at": None,
            "deleted_by": None,
            "uid": uid,
            "slug": slug,
            "title": title,
            "description": "A seeded news article for SEO tests.",
            "content": "Body content for the seeded article.",
            "url": "https://example.com/article",
            "source_name": "ExampleSource",
            "status": "published",
            "synced_at": datetime.now(timezone.utc).isoformat(),
            "show_on_landing": 0,
            "grade": 8,
        }
    )
    return slug, uid
def _seed_owner():
    from datetime import datetime, timezone
    from uuid import uuid4
    from devplacepy.database import get_table

    uid = str(uuid4())
    get_table("users").insert(
        {
            "uid": uid,
            "username": f"seo_{uid[:8]}",
            "terms_version": "1",
            "email": f"{uid[:8]}@seo.test",
            "password_hash": "x",
            "role": "Member",
            "is_active": True,
            "created_at": datetime.now(timezone.utc).isoformat(),
        }
    )
    return uid
def _seed_post_seo(image=None):
    from datetime import datetime, timezone
    from uuid import uuid4
    from devplacepy.database import get_table
    from devplacepy.utils import make_combined_slug

    owner = _seed_owner()
    uid = str(uuid4())
    slug = make_combined_slug("SEO Detail Post", uid)
    get_table("posts").insert(
        {
            "deleted_at": None,
            "deleted_by": None,
            "uid": uid,
            "user_uid": owner,
            "slug": slug,
            "title": "SEO Detail Post",
            "content": "Body text for the SEO detail post.",
            "topic": "general",
            "project_uid": None,
            "image": image,
            "stars": 0,
            "created_at": datetime.now(timezone.utc).isoformat(),
        }
    )
    return slug, uid
def _seed_gist():
    from datetime import datetime, timezone
    from uuid import uuid4
    from devplacepy.database import get_table
    from devplacepy.utils import make_combined_slug

    owner = _seed_owner()
    uid = str(uuid4())
    slug = make_combined_slug("SEO Detail Gist", uid)
    get_table("gists").insert(
        {
            "deleted_at": None,
            "deleted_by": None,
            "uid": uid,
            "user_uid": owner,
            "slug": slug,
            "title": "SEO Detail Gist",
            "description": "Gist description for SEO tests.",
            "source_code": "print('seo')",
            "language": "python",
            "stars": 0,
            "created_at": datetime.now(timezone.utc).isoformat(),
        }
    )
    return slug, uid
def _seed_project():
    from datetime import datetime, timezone
    from uuid import uuid4
    from devplacepy.database import get_table
    from devplacepy.utils import make_combined_slug

    owner = _seed_owner()
    uid = str(uuid4())
    slug = make_combined_slug("SEO Detail Project", uid)
    get_table("projects").insert(
        {
            "deleted_at": None,
            "deleted_by": None,
            "uid": uid,
            "user_uid": owner,
            "slug": slug,
            "title": "SEO Detail Project",
            "description": "Project description for SEO tests.",
            "project_type": "software",
            "platforms": "Linux",
            "status": "Released",
            "stars": 0,
            "created_at": datetime.now(timezone.utc).isoformat(),
        }
    )
    return slug, uid
def _seed_news_image(news_uid):
    from devplacepy.database import get_table

    get_table("news_images").insert(
        {
            "deleted_at": None,
            "deleted_by": None,
            "news_uid": news_uid,
            "url": "https://example.com/seo-news-image.jpg",
        }
    )
def _seed_feed_posts(count):
    from datetime import datetime, timezone, timedelta
    from uuid import uuid4
    from devplacepy.database import get_table
    from devplacepy.utils import make_combined_slug

    owner = _seed_owner()
    topic = f"seopag{owner[:8]}"
    base = datetime(2026, 1, 1, tzinfo=timezone.utc)
    posts = get_table("posts")
    for i in range(count):
        uid = str(uuid4())
        posts.insert(
            {
                "deleted_at": None,
                "deleted_by": None,
                "uid": uid,
                "user_uid": owner,
                "slug": make_combined_slug(f"seo pag {i}", uid),
                "title": None,
                "content": f"seo pag post {i}",
                "topic": topic,
                "project_uid": None,
                "image": None,
                "stars": 0,
                "created_at": (base - timedelta(seconds=i)).isoformat(),
            }
        )
    return topic


def test_private_project_excluded_from_sitemap(app_server):
    _, _, key = _signup_project_visibility()
    public_slug = _create_project_project_visibility(key, "Sitemap Public")["slug"]
    private_slug = _create_project_project_visibility(key, "Sitemap Private", is_private=True)["slug"]
    xml = requests.get(f"{BASE_URL}/sitemap.xml").text
    assert f"/projects/{public_slug}" in xml
    assert f"/projects/{private_slug}" not in xml


def test_sitemap_xml_exists(app_server):
    r = requests.get(f"{BASE_URL}/sitemap.xml")
    assert r.status_code == 200
    assert "<urlset" in r.text
    assert "<loc>" in r.text


def test_sitemap_headers_and_static_entries(app_server):
    r = requests.get(f"{BASE_URL}/sitemap.xml")
    assert r.headers["content-type"].startswith("application/xml")
    assert "cache-control" in {k.lower() for k in r.headers}
    assert f"{BASE_URL}/gists" in r.text


def test_sitemap_lists_the_public_moderation_surfaces(app_server):
    xml = requests.get(f"{BASE_URL}/sitemap.xml").text
    assert f"<loc>{BASE_URL}/reports/reasons</loc>" in xml
    assert f"<loc>{BASE_URL}/workspaces/index</loc>" in xml
    assert f"{BASE_URL}/reports/mine" not in xml


def test_sitemap_lists_each_legal_document_exactly_once(app_server):
    from devplacepy.seo import LEGAL_DOC_SLUGS

    xml = requests.get(f"{BASE_URL}/sitemap.xml").text
    for slug in LEGAL_DOC_SLUGS:
        loc = f"<loc>{BASE_URL}/docs/{slug}.html</loc>"
        assert xml.count(loc) == 1, f"{slug} listed {xml.count(loc)} times"


def test_sitemap_never_repeats_a_documentation_url(app_server):
    import re

    xml = requests.get(f"{BASE_URL}/sitemap.xml").text
    locs = [loc for loc in re.findall(r"<loc>(.*?)</loc>", xml) if "/docs/" in loc]
    assert locs, "the sitemap lists no documentation pages"
    duplicates = sorted({loc for loc in locs if locs.count(loc) > 1})
    assert duplicates == [], duplicates


def test_sitemap_includes_published_news(app_server):
    slug, _ = _seed_news_seo()
    r = requests.get(f"{BASE_URL}/sitemap.xml")
    assert f"/news/{slug}" in r.text