Private
Public Access
Compare commits
5
Commits
| Author | SHA1 | Date | |
|---|---|---|---|
|
|
7afda7d0ba | ||
|
|
37e658c49a | ||
|
|
e5f1277014 | ||
|
|
8d2900424a | ||
|
|
70b0d4f277 |
No files matched your search
@@ -45,4 +45,5 @@ def _setup_exempt_paths():
|
||||
reverse("pwa:service_worker"),
|
||||
reverse("pwa:offline"),
|
||||
reverse("core:health"),
|
||||
reverse("core:health_ready"),
|
||||
}
|
||||
@@ -72,3 +72,21 @@ def test_too_many_failed_attempts_against_one_username_blocks_that_username(clie
|
||||
response = client.post(reverse("accounts:login"), {"username": "bob", "password": "pw-bob-123"})
|
||||
assert response.status_code == 302
|
||||
assert response.wsgi_request.user == bob
|
||||
|
||||
|
||||
@pytest.mark.django_db
|
||||
def test_login_still_works_when_the_throttle_cache_is_down(client, monkeypatch):
|
||||
from unittest.mock import Mock
|
||||
|
||||
from django.contrib.auth import get_user_model
|
||||
|
||||
get_user_model().objects.create_user("alice", password="pw-alice-123")
|
||||
broken = Mock(side_effect=ConnectionError("redis down"))
|
||||
monkeypatch.setattr("apps.accounts.views.cache.get", broken)
|
||||
monkeypatch.setattr("apps.accounts.views.cache.incr", broken)
|
||||
monkeypatch.setattr("apps.accounts.views.cache.set", broken)
|
||||
|
||||
bad = client.post(reverse("accounts:login"), {"username": "alice", "password": "wrong"})
|
||||
assert bad.status_code == 200
|
||||
good = client.post(reverse("accounts:login"), {"username": "alice", "password": "pw-alice-123"})
|
||||
assert good.status_code == 302
|
||||
@@ -1,3 +1,5 @@
|
||||
import logging
|
||||
|
||||
from django.contrib import messages
|
||||
from django.contrib.auth import login
|
||||
from django.contrib.auth.decorators import user_passes_test
|
||||
@@ -10,6 +12,8 @@ from .forms import CreateUserForm, SetupForm
|
||||
from .middleware import SetupWizardMiddleware
|
||||
from .models import SetupLock, User
|
||||
|
||||
logger = logging.getLogger(__name__)
|
||||
|
||||
is_admin = user_passes_test(lambda u: u.is_authenticated and u.is_staff)
|
||||
|
||||
# Two independent throttles, either of which can block an attempt: by
|
||||
@@ -41,14 +45,24 @@ def _attempts_key(kind: str, value: str) -> str:
|
||||
|
||||
|
||||
def _too_many_attempts(key: str, limit: int) -> bool:
|
||||
# Fails open if the cache (Redis) is unreachable: logging in must keep
|
||||
# working through a cache outage, at the cost of no throttling until
|
||||
# it's back — an outage shouldn't lock every user out of the app.
|
||||
try:
|
||||
return cache.get(key, 0) >= limit
|
||||
except Exception:
|
||||
logger.warning("Login throttle cache unavailable; not rate limiting", exc_info=True)
|
||||
return False
|
||||
|
||||
|
||||
def _record_failed_attempt(key: str) -> None:
|
||||
try:
|
||||
try:
|
||||
cache.incr(key)
|
||||
except ValueError:
|
||||
cache.set(key, 1, LOGIN_RATE_LIMIT_WINDOW_SECONDS)
|
||||
except Exception:
|
||||
logger.warning("Login throttle cache unavailable; attempt not recorded", exc_info=True)
|
||||
|
||||
|
||||
class LoginView(DjangoLoginView):
|
||||
|
||||
@@ -1,3 +1,5 @@
|
||||
import logging
|
||||
|
||||
from django.core.cache import cache
|
||||
from django.db.models import Count
|
||||
|
||||
@@ -5,6 +7,8 @@ from apps.core.cache_utils import SIDEBAR_UNREAD_CACHE_TTL_SECONDS, sidebar_unre
|
||||
|
||||
from .models import SAVED_TAG_NAME, STARRED_TAG_NAME, Article, ArticleTag, UserArticleState
|
||||
|
||||
logger = logging.getLogger(__name__)
|
||||
|
||||
|
||||
def with_user_state(articles, user):
|
||||
"""Annotates an iterable of Article instances with is_read/is_starred/
|
||||
@@ -56,7 +60,14 @@ def unread_counts_by_feed(user):
|
||||
the current user's own read state, so this is rarely visibly stale.
|
||||
"""
|
||||
cache_key = sidebar_unread_cache_key(user.pk)
|
||||
# The cache is an optimization here, never a dependency: every
|
||||
# authenticated page render goes through this, so a Redis outage must
|
||||
# degrade to recomputing the count, not to a 500 on every page.
|
||||
try:
|
||||
counts = cache.get(cache_key)
|
||||
except Exception:
|
||||
logger.warning("Sidebar cache unavailable; computing unread counts directly")
|
||||
counts = None
|
||||
if counts is not None:
|
||||
return counts
|
||||
|
||||
@@ -68,5 +79,8 @@ def unread_counts_by_feed(user):
|
||||
.annotate(unread=Count("id"))
|
||||
)
|
||||
counts = {row["feed_id"]: row["unread"] for row in rows}
|
||||
try:
|
||||
cache.set(cache_key, counts, SIDEBAR_UNREAD_CACHE_TTL_SECONDS)
|
||||
except Exception:
|
||||
logger.warning("Sidebar cache unavailable; skipping cache write")
|
||||
return counts
|
||||
+38
-2
@@ -1,10 +1,14 @@
|
||||
from celery import shared_task
|
||||
from datetime import timedelta
|
||||
|
||||
from celery import shared_task
|
||||
from django.utils import timezone
|
||||
|
||||
from apps.core.models import SiteSettings
|
||||
from apps.feeds.models import Subscription
|
||||
from apps.rules.engine import evaluate_rules_for_article
|
||||
from apps.rules.models import FilterRule
|
||||
|
||||
from .models import Article, SubscriptionTag
|
||||
from .models import Article, ArticleTag, SubscriptionTag
|
||||
|
||||
|
||||
@shared_task
|
||||
@@ -57,3 +61,35 @@ def apply_filter_rules_for_feed(feed_id: int, article_ids: list[int]):
|
||||
evaluated += 1
|
||||
|
||||
return evaluated
|
||||
|
||||
|
||||
PURGE_CHUNK_SIZE = 500
|
||||
|
||||
|
||||
@shared_task
|
||||
def purge_old_articles():
|
||||
"""Deletes articles older than SiteSettings.retention_days (C7), in
|
||||
chunks so one run never holds a huge delete in a single transaction.
|
||||
Anything a user starred or saved for later (a system ArticleTag) is
|
||||
kept — those are explicit "keep this" signals. UserArticleState and
|
||||
ArticleTag rows go with their article via ON DELETE CASCADE, and the
|
||||
SQLite FTS triggers drop the search entries. 0 days means keep
|
||||
everything forever (the default).
|
||||
"""
|
||||
retention_days = SiteSettings.load().retention_days
|
||||
if retention_days <= 0:
|
||||
return 0
|
||||
|
||||
cutoff = timezone.now() - timedelta(days=retention_days)
|
||||
keep = ArticleTag.objects.filter(tag__is_system=True).values("article_id")
|
||||
deleted_total = 0
|
||||
while True:
|
||||
chunk = list(
|
||||
Article.objects.filter(fetched_at__lt=cutoff)
|
||||
.exclude(pk__in=keep)
|
||||
.values_list("pk", flat=True)[:PURGE_CHUNK_SIZE]
|
||||
)
|
||||
if not chunk:
|
||||
return deleted_total
|
||||
Article.objects.filter(pk__in=chunk).delete()
|
||||
deleted_total += len(chunk)
|
||||
@@ -58,3 +58,19 @@ def test_unread_counts_by_feed_cache_is_per_user(alice, feed):
|
||||
|
||||
assert unread_counts_by_feed(alice) == {feed.pk: 1}
|
||||
assert unread_counts_by_feed(bob) == {feed.pk: 1}
|
||||
|
||||
|
||||
@pytest.mark.django_db
|
||||
def test_unread_counts_still_work_when_the_cache_is_down(alice, feed, monkeypatch):
|
||||
"""A Redis outage must degrade the sidebar to recomputing counts, not
|
||||
500 every authenticated page."""
|
||||
from unittest.mock import Mock
|
||||
|
||||
Subscription.objects.create(user=alice, feed=feed)
|
||||
Article.objects.create(feed=feed, guid="a-1")
|
||||
broken = Mock(side_effect=ConnectionError("redis down"))
|
||||
monkeypatch.setattr("apps.articles.queries.cache.get", broken)
|
||||
monkeypatch.setattr("apps.articles.queries.cache.set", broken)
|
||||
|
||||
assert unread_counts_by_feed(alice) == {feed.pk: 1}
|
||||
invalidate_sidebar_unread_cache(alice.pk)
|
||||
@@ -0,0 +1,74 @@
|
||||
from datetime import timedelta
|
||||
|
||||
import pytest
|
||||
from django.contrib.auth import get_user_model
|
||||
from django.utils import timezone
|
||||
|
||||
from apps.core.models import SiteSettings
|
||||
from apps.feeds.models import Feed
|
||||
|
||||
from .models import STARRED_TAG_NAME, Article, ArticleTag, UserArticleState, get_system_tag
|
||||
from .tasks import purge_old_articles
|
||||
|
||||
User = get_user_model()
|
||||
|
||||
|
||||
def _article(feed, guid, age_days):
|
||||
article = Article.objects.create(feed=feed, guid=guid, title=guid)
|
||||
Article.objects.filter(pk=article.pk).update(
|
||||
fetched_at=timezone.now() - timedelta(days=age_days)
|
||||
)
|
||||
return article
|
||||
|
||||
|
||||
@pytest.fixture
|
||||
def feed(db):
|
||||
return Feed.objects.create(feed_url="https://example.com/feed.xml")
|
||||
|
||||
|
||||
@pytest.mark.django_db
|
||||
def test_purge_is_a_noop_with_the_default_keep_forever_setting(feed):
|
||||
_article(feed, "ancient", age_days=5000)
|
||||
|
||||
assert purge_old_articles() == 0
|
||||
assert Article.objects.count() == 1
|
||||
|
||||
|
||||
@pytest.mark.django_db
|
||||
def test_purge_deletes_only_articles_older_than_the_retention_window(feed):
|
||||
SiteSettings.objects.create(retention_days=30)
|
||||
old = _article(feed, "old", age_days=45)
|
||||
recent = _article(feed, "recent", age_days=5)
|
||||
|
||||
assert purge_old_articles() == 1
|
||||
assert not Article.objects.filter(pk=old.pk).exists()
|
||||
assert Article.objects.filter(pk=recent.pk).exists()
|
||||
|
||||
|
||||
@pytest.mark.django_db
|
||||
def test_purge_keeps_starred_articles_and_cascades_user_state(feed):
|
||||
SiteSettings.objects.create(retention_days=30)
|
||||
user = User.objects.create_user("alice", password="pw-alice-123")
|
||||
starred = _article(feed, "starred-old", age_days=90)
|
||||
ArticleTag.objects.create(
|
||||
user=user, article=starred, tag=get_system_tag(user, STARRED_TAG_NAME)
|
||||
)
|
||||
doomed = _article(feed, "plain-old", age_days=90)
|
||||
state = UserArticleState.objects.create(user=user, article=doomed, is_read=True)
|
||||
|
||||
purge_old_articles()
|
||||
|
||||
assert Article.objects.filter(pk=starred.pk).exists()
|
||||
assert not Article.objects.filter(pk=doomed.pk).exists()
|
||||
assert not UserArticleState.objects.filter(pk=state.pk).exists()
|
||||
|
||||
|
||||
@pytest.mark.django_db
|
||||
def test_purge_works_through_more_than_one_chunk(feed, monkeypatch):
|
||||
monkeypatch.setattr("apps.articles.tasks.PURGE_CHUNK_SIZE", 2)
|
||||
SiteSettings.objects.create(retention_days=30)
|
||||
for i in range(5):
|
||||
_article(feed, f"old-{i}", age_days=60)
|
||||
|
||||
assert purge_old_articles() == 5
|
||||
assert Article.objects.count() == 0
|
||||
+1
-1
@@ -14,7 +14,7 @@ class SiteSettingsAdmin(admin.ModelAdmin):
|
||||
install. No changelist/add/delete views exist at all here.
|
||||
"""
|
||||
|
||||
fields = ["fetch_interval_minutes"]
|
||||
fields = ["fetch_interval_minutes", "retention_days"]
|
||||
|
||||
def has_add_permission(self, request):
|
||||
return False
|
||||
|
||||
@@ -1,5 +1,9 @@
|
||||
import logging
|
||||
|
||||
from django.core.cache import cache
|
||||
|
||||
logger = logging.getLogger(__name__)
|
||||
|
||||
# Short enough that a missed invalidation point self-heals within one
|
||||
# page load's worth of staleness; long enough to actually absorb repeat
|
||||
# authenticated-page renders (apps.core.context_processors.sidebar runs
|
||||
@@ -19,4 +23,7 @@ def invalidate_sidebar_unread_cache(user_id: int) -> None:
|
||||
out the TTL above. Rule-driven state changes during background feed
|
||||
ingestion aren't invalidated this way — those self-heal within the
|
||||
TTL, same as a missed call site here would."""
|
||||
try:
|
||||
cache.delete(sidebar_unread_cache_key(user_id))
|
||||
except Exception:
|
||||
logger.warning("Sidebar cache unavailable; nothing to invalidate")
|
||||
@@ -0,0 +1,18 @@
|
||||
# Generated by Django 5.2.18 on 2026-11-01 15:48
|
||||
|
||||
from django.db import migrations, models
|
||||
|
||||
|
||||
class Migration(migrations.Migration):
|
||||
|
||||
dependencies = [
|
||||
('core', '0001_initial'),
|
||||
]
|
||||
|
||||
operations = [
|
||||
migrations.AddField(
|
||||
model_name='sitesettings',
|
||||
name='retention_days',
|
||||
field=models.PositiveIntegerField(default=0, help_text="Articles older than this many days (by when they were fetched) are deleted daily, along with everyone's read state and tags for them — except articles anyone has starred or saved for later. 0 keeps everything forever."),
|
||||
),
|
||||
]
|
||||
@@ -18,6 +18,16 @@ class SiteSettings(models.Model):
|
||||
),
|
||||
)
|
||||
|
||||
retention_days = models.PositiveIntegerField(
|
||||
default=0,
|
||||
help_text=(
|
||||
"Articles older than this many days (by when they were fetched) are "
|
||||
"deleted daily, along with everyone's read state and tags for them — "
|
||||
"except articles anyone has starred or saved for later. 0 keeps "
|
||||
"everything forever."
|
||||
),
|
||||
)
|
||||
|
||||
class Meta:
|
||||
verbose_name = "Site settings"
|
||||
verbose_name_plural = "Site settings"
|
||||
|
||||
@@ -22,6 +22,48 @@ def test_health_check_works_before_first_run_setup(client):
|
||||
assert response.status_code == 200
|
||||
|
||||
|
||||
@pytest.mark.django_db
|
||||
def test_health_ready_is_degraded_until_a_poll_tick_has_happened(client):
|
||||
response = client.get(reverse("core:health_ready"))
|
||||
assert response.status_code == 503
|
||||
body = response.json()
|
||||
assert body["status"] == "degraded"
|
||||
assert body["checks"]["database"] is True
|
||||
assert body["checks"]["redis"] is True
|
||||
assert body["checks"]["poll_heartbeat"] is False
|
||||
|
||||
|
||||
@pytest.mark.django_db
|
||||
def test_health_ready_is_ok_after_poll_due_feeds_runs(client):
|
||||
from apps.feeds.tasks import poll_due_feeds
|
||||
|
||||
poll_due_feeds()
|
||||
|
||||
response = client.get(reverse("core:health_ready"))
|
||||
assert response.status_code == 200
|
||||
assert response.json()["checks"] == {
|
||||
"database": True,
|
||||
"redis": True,
|
||||
"poll_heartbeat": True,
|
||||
}
|
||||
|
||||
|
||||
@pytest.mark.django_db
|
||||
def test_health_ready_flags_a_stale_poll_heartbeat(client):
|
||||
from datetime import timedelta
|
||||
|
||||
from django.core.cache import cache
|
||||
from django.utils import timezone
|
||||
|
||||
from apps.feeds.tasks import POLL_HEARTBEAT_CACHE_KEY
|
||||
|
||||
cache.set(POLL_HEARTBEAT_CACHE_KEY, timezone.now() - timedelta(minutes=30))
|
||||
|
||||
response = client.get(reverse("core:health_ready"))
|
||||
assert response.status_code == 503
|
||||
assert response.json()["checks"]["poll_heartbeat"] is False
|
||||
|
||||
|
||||
@pytest.mark.django_db
|
||||
def test_dashboard_requires_staff(client, django_user_model):
|
||||
django_user_model.objects.create_user("regular", password="pw-regular-123")
|
||||
|
||||
@@ -6,6 +6,7 @@ app_name = "core"
|
||||
|
||||
urlpatterns = [
|
||||
path("health/", views.health, name="health"),
|
||||
path("health/ready/", views.health_ready, name="health_ready"),
|
||||
path("dashboard/", views.dashboard, name="dashboard"),
|
||||
path("dashboard/feeds/<int:feed_id>/retry/", views.retry_feed, name="retry_feed"),
|
||||
]
|
||||
+50
-3
@@ -1,20 +1,27 @@
|
||||
from django.contrib import messages
|
||||
from django.contrib.auth.decorators import user_passes_test
|
||||
from django.core.cache import cache
|
||||
from django.db import connection
|
||||
from django.db.models import Count, Q
|
||||
from django.db.models import Count, Min, Q
|
||||
from django.http import JsonResponse
|
||||
from django.shortcuts import get_object_or_404, redirect, render
|
||||
from django.utils import timezone
|
||||
from django.views.decorators.http import require_POST
|
||||
|
||||
from apps.articles.models import Article
|
||||
from apps.feeds.models import Feed, Subscription
|
||||
from apps.feeds.tasks import fetch_feed, reset_feed_backoff
|
||||
from apps.feeds.tasks import POLL_HEARTBEAT_CACHE_KEY, fetch_feed, reset_feed_backoff
|
||||
|
||||
is_admin = user_passes_test(lambda u: u.is_authenticated and u.is_staff)
|
||||
|
||||
|
||||
# Beat ticks poll_due_feeds every minute; this many minutes without a tick
|
||||
# means beat or the worker (or the broker between them) isn't working.
|
||||
POLL_HEARTBEAT_MAX_AGE_SECONDS = 5 * 60
|
||||
|
||||
|
||||
def health(request):
|
||||
"""Unauthenticated liveness/readiness check — reachable even before
|
||||
"""Unauthenticated liveness check — reachable even before
|
||||
first-run setup (see SetupWizardMiddleware). Used by scripts/upgrade.sh
|
||||
to confirm a freshly restarted deploy actually came up, and is a
|
||||
reasonable target for any external uptime monitor too."""
|
||||
@@ -26,6 +33,45 @@ def health(request):
|
||||
return JsonResponse({"status": "ok"})
|
||||
|
||||
|
||||
def health_ready(request):
|
||||
"""Readiness: liveness plus everything background fetching depends on
|
||||
(C8) — the database, the Redis cache/broker, and a recent
|
||||
poll_due_feeds tick, which proves beat is scheduling AND a worker is
|
||||
consuming. /health/ above stays DB-only on purpose: scripts/upgrade.sh
|
||||
gates a deploy on it right after restart, before beat has had a
|
||||
chance to tick. Point an uptime monitor at this one instead."""
|
||||
checks = {"database": False, "redis": False, "poll_heartbeat": False}
|
||||
detail = {}
|
||||
|
||||
try:
|
||||
with connection.cursor() as cursor:
|
||||
cursor.execute("SELECT 1")
|
||||
checks["database"] = True
|
||||
except Exception:
|
||||
pass
|
||||
|
||||
try:
|
||||
cache.set("health-ready-probe", "1", timeout=10)
|
||||
checks["redis"] = cache.get("health-ready-probe") == "1"
|
||||
except Exception:
|
||||
pass
|
||||
|
||||
try:
|
||||
last_tick = cache.get(POLL_HEARTBEAT_CACHE_KEY)
|
||||
except Exception:
|
||||
last_tick = None
|
||||
if last_tick is not None:
|
||||
age = (timezone.now() - last_tick).total_seconds()
|
||||
detail["poll_heartbeat_age_seconds"] = int(age)
|
||||
checks["poll_heartbeat"] = age <= POLL_HEARTBEAT_MAX_AGE_SECONDS
|
||||
|
||||
ok = all(checks.values())
|
||||
return JsonResponse(
|
||||
{"status": "ok" if ok else "degraded", "checks": checks, **detail},
|
||||
status=200 if ok else 503,
|
||||
)
|
||||
|
||||
|
||||
@is_admin
|
||||
def dashboard(request):
|
||||
feeds = Feed.objects.annotate(subscriber_count=Count("subscriptions"))
|
||||
@@ -39,6 +85,7 @@ def dashboard(request):
|
||||
"error_feed_count": erroring_feeds.count(),
|
||||
"subscription_count": Subscription.objects.count(),
|
||||
"article_count": Article.objects.count(),
|
||||
"oldest_article_at": Article.objects.aggregate(oldest=Min("fetched_at"))["oldest"],
|
||||
}
|
||||
|
||||
return render(
|
||||
|
||||
@@ -18,6 +18,7 @@ import feedparser
|
||||
import nh3
|
||||
import requests
|
||||
from django.conf import settings
|
||||
from django.db import IntegrityError, transaction
|
||||
from django.utils import timezone
|
||||
|
||||
from apps.articles.models import Article
|
||||
@@ -131,6 +132,26 @@ def _entry_content_html(entry) -> str:
|
||||
return getattr(entry, "summary", "") or ""
|
||||
|
||||
|
||||
UPDATABLE_ARTICLE_FIELDS = (
|
||||
"title",
|
||||
"link",
|
||||
"author",
|
||||
"summary",
|
||||
"content_html",
|
||||
"published_at",
|
||||
"content_hash",
|
||||
)
|
||||
|
||||
|
||||
def _ensure_pks(feed: Feed, articles: list[Article]) -> list[int]:
|
||||
"""bulk_create only populates primary keys on backends that can
|
||||
RETURN them (Postgres, SQLite 3.35+); look the rest up by guid."""
|
||||
if all(a.pk for a in articles):
|
||||
return [a.pk for a in articles]
|
||||
guids = [a.guid for a in articles]
|
||||
return list(Article.objects.filter(feed=feed, guid__in=guids).values_list("pk", flat=True))
|
||||
|
||||
|
||||
def effective_interval_minutes(feed: Feed, configured_interval_minutes: int) -> int:
|
||||
"""The actual polling interval to use for `feed`, applying exponential
|
||||
backoff on top of the site-wide `configured_interval_minutes` (see
|
||||
@@ -217,52 +238,69 @@ def fetch_and_ingest(feed: Feed) -> FetchResult:
|
||||
)
|
||||
entries = entries[:MAX_ENTRIES_PER_POLL]
|
||||
|
||||
# One prefetch + bulk writes for the whole poll (C2), instead of a
|
||||
# get_or_create (SELECT, then INSERT) per entry. Entries are keyed by
|
||||
# guid so a feed that repeats one within a single response is only
|
||||
# processed once.
|
||||
parsed_entries: dict[str, dict] = {}
|
||||
for entry in entries:
|
||||
guid = _entry_guid(entry)
|
||||
if guid in parsed_entries:
|
||||
continue
|
||||
title = (getattr(entry, "title", "") or "")[:1000]
|
||||
link = (getattr(entry, "link", "") or "")[:2000]
|
||||
author = (getattr(entry, "author", "") or "")[:500]
|
||||
summary = (getattr(entry, "summary", "") or "")[:5000]
|
||||
content_html = sanitize_html(_entry_content_html(entry))
|
||||
published_at = _entry_published_at(entry)
|
||||
content_hash = _content_hash(title, link, author, summary, content_html, published_at)
|
||||
|
||||
article, created = Article.objects.get_or_create(
|
||||
feed=feed,
|
||||
guid=guid,
|
||||
defaults={
|
||||
"link": link,
|
||||
parsed_entries[guid] = {
|
||||
"title": title,
|
||||
"link": link,
|
||||
"author": author,
|
||||
"content_html": content_html,
|
||||
"summary": summary,
|
||||
"content_hash": content_hash,
|
||||
"content_html": content_html,
|
||||
"published_at": published_at,
|
||||
},
|
||||
)
|
||||
"content_hash": _content_hash(title, link, author, summary, content_html, published_at),
|
||||
}
|
||||
|
||||
if created:
|
||||
new_ids.append(article.pk)
|
||||
elif article.content_hash != content_hash:
|
||||
article.title = title
|
||||
article.link = link
|
||||
article.author = author
|
||||
article.summary = summary
|
||||
article.content_html = content_html
|
||||
article.published_at = published_at
|
||||
article.content_hash = content_hash
|
||||
article.save(
|
||||
update_fields=[
|
||||
"title",
|
||||
"link",
|
||||
"author",
|
||||
"summary",
|
||||
"content_html",
|
||||
"published_at",
|
||||
"content_hash",
|
||||
existing = {a.guid: a for a in Article.objects.filter(feed=feed, guid__in=parsed_entries)}
|
||||
|
||||
to_create = [
|
||||
Article(feed=feed, guid=guid, **fields)
|
||||
for guid, fields in parsed_entries.items()
|
||||
if guid not in existing
|
||||
]
|
||||
to_update = []
|
||||
for guid, article in existing.items():
|
||||
fields = parsed_entries[guid]
|
||||
if article.content_hash != fields["content_hash"]:
|
||||
for name, value in fields.items():
|
||||
setattr(article, name, value)
|
||||
to_update.append(article)
|
||||
|
||||
if to_update:
|
||||
Article.objects.bulk_update(to_update, list(UPDATABLE_ARTICLE_FIELDS), batch_size=100)
|
||||
updated_ids.extend(a.pk for a in to_update)
|
||||
|
||||
if to_create:
|
||||
try:
|
||||
with transaction.atomic():
|
||||
Article.objects.bulk_create(to_create)
|
||||
except IntegrityError:
|
||||
# Another worker inserted some of these guids between the
|
||||
# prefetch above and now (two fetches of the same feed can
|
||||
# overlap) — fall back to row-by-row so nothing is lost or
|
||||
# duplicated, at the cost of the old per-entry round trips
|
||||
# for this rare case only.
|
||||
for article in to_create:
|
||||
fields = {name: getattr(article, name) for name in UPDATABLE_ARTICLE_FIELDS}
|
||||
obj, created = Article.objects.get_or_create(
|
||||
feed=feed, guid=article.guid, defaults=fields
|
||||
)
|
||||
updated_ids.append(article.pk)
|
||||
if created:
|
||||
new_ids.append(obj.pk)
|
||||
else:
|
||||
new_ids.extend(_ensure_pks(feed, to_create))
|
||||
|
||||
return FetchResult(
|
||||
status=Feed.STATUS_OK, new_article_ids=new_ids, updated_article_ids=updated_ids
|
||||
|
||||
+15
-1
@@ -4,6 +4,7 @@ from io import BytesIO
|
||||
|
||||
from celery import shared_task
|
||||
from django.contrib.auth import get_user_model
|
||||
from django.core.cache import cache
|
||||
from django.db.models import Q
|
||||
from django.utils import timezone
|
||||
|
||||
@@ -27,6 +28,12 @@ CLAIM_STALE_MINUTES = 10
|
||||
# across several beat ticks instead of enqueuing everything — and every
|
||||
# bounded Feed query below stays index-only work, not a full scan (C1).
|
||||
POLL_BATCH_SIZE = 500
|
||||
RULE_DISPATCH_CHUNK_SIZE = 50
|
||||
|
||||
# Written on every poll_due_feeds tick; apps.core.views.health_ready
|
||||
# reads it to tell whether beat is scheduling and a worker is consuming.
|
||||
POLL_HEARTBEAT_CACHE_KEY = "poll-due-feeds-heartbeat"
|
||||
POLL_HEARTBEAT_TTL_SECONDS = 3600
|
||||
|
||||
|
||||
@shared_task
|
||||
@@ -39,6 +46,7 @@ def poll_due_feeds():
|
||||
here, not via one Celery Beat entry per feed.
|
||||
"""
|
||||
now = timezone.now()
|
||||
cache.set(POLL_HEARTBEAT_CACHE_KEY, now, timeout=POLL_HEARTBEAT_TTL_SECONDS)
|
||||
claim_stale_before = now - timedelta(minutes=CLAIM_STALE_MINUTES)
|
||||
due_feed_ids = list(
|
||||
Feed.objects.filter(is_active=True)
|
||||
@@ -125,7 +133,13 @@ def fetch_feed(feed_id: int):
|
||||
len(result.new_article_ids),
|
||||
len(result.updated_article_ids),
|
||||
)
|
||||
apply_filter_rules_for_feed.delay(feed_id, changed_article_ids)
|
||||
# Chunked so one huge poll (a first fetch of a long archive feed)
|
||||
# becomes several bounded rule-evaluation tasks, each well inside
|
||||
# the worker's time limits, instead of one unbounded one (C2).
|
||||
for start in range(0, len(changed_article_ids), RULE_DISPATCH_CHUNK_SIZE):
|
||||
apply_filter_rules_for_feed.delay(
|
||||
feed_id, changed_article_ids[start : start + RULE_DISPATCH_CHUNK_SIZE]
|
||||
)
|
||||
|
||||
return {
|
||||
"status": result.status,
|
||||
|
||||
@@ -264,3 +264,77 @@ def test_fetch_only_processes_the_first_max_entries_per_poll(feed):
|
||||
assert len(result.new_article_ids) == MAX_ENTRIES_PER_POLL
|
||||
assert Article.objects.filter(feed=feed).count() == MAX_ENTRIES_PER_POLL
|
||||
assert not Article.objects.filter(feed=feed, guid=f"guid-{MAX_ENTRIES_PER_POLL}").exists()
|
||||
|
||||
|
||||
def _feed_xml(count, title_suffix=""):
|
||||
items = "".join(
|
||||
f"<item><title>Item {i}{title_suffix}</title><guid>guid-{i}</guid></item>"
|
||||
for i in range(count)
|
||||
)
|
||||
return (
|
||||
f'<?xml version="1.0"?><rss version="2.0"><channel><title>T</title>{items}</channel></rss>'
|
||||
)
|
||||
|
||||
|
||||
@pytest.mark.django_db
|
||||
@responses.activate
|
||||
def test_ingest_uses_a_constant_number_of_article_queries(feed):
|
||||
"""C2: one prefetch + bulk writes per poll, not a get_or_create per
|
||||
entry — query count must not grow with the number of entries."""
|
||||
from django.db import connection
|
||||
from django.test.utils import CaptureQueriesContext
|
||||
|
||||
responses.add(responses.GET, FEED_URL, body=_feed_xml(40), status=200)
|
||||
with CaptureQueriesContext(connection) as ctx:
|
||||
result = fetch_and_ingest(feed)
|
||||
|
||||
assert len(result.new_article_ids) == 40
|
||||
assert len(ctx.captured_queries) < 10
|
||||
assert Article.objects.filter(feed=feed).count() == 40
|
||||
|
||||
|
||||
@pytest.mark.django_db
|
||||
@responses.activate
|
||||
def test_ingest_dedups_a_guid_repeated_within_one_response(feed):
|
||||
xml = (
|
||||
'<?xml version="1.0"?><rss version="2.0"><channel><title>T</title>'
|
||||
"<item><title>A</title><guid>same</guid></item>"
|
||||
"<item><title>B</title><guid>same</guid></item>"
|
||||
"</channel></rss>"
|
||||
)
|
||||
responses.add(responses.GET, FEED_URL, body=xml, status=200)
|
||||
|
||||
result = fetch_and_ingest(feed)
|
||||
|
||||
assert len(result.new_article_ids) == 1
|
||||
assert Article.objects.filter(feed=feed).count() == 1
|
||||
|
||||
|
||||
@pytest.mark.django_db
|
||||
@responses.activate
|
||||
def test_ingest_reports_updates_separately_from_new_articles(feed):
|
||||
responses.add(responses.GET, FEED_URL, body=_feed_xml(3), status=200)
|
||||
fetch_and_ingest(feed)
|
||||
|
||||
responses.replace(responses.GET, FEED_URL, body=_feed_xml(4, " edited"), status=200)
|
||||
result = fetch_and_ingest(feed)
|
||||
|
||||
assert len(result.new_article_ids) == 1
|
||||
assert len(result.updated_article_ids) == 3
|
||||
assert Article.objects.get(feed=feed, guid="guid-0").title == "Item 0 edited"
|
||||
|
||||
|
||||
@pytest.mark.django_db
|
||||
@responses.activate
|
||||
def test_ingest_falls_back_to_row_by_row_when_a_concurrent_insert_wins(feed):
|
||||
from unittest.mock import patch
|
||||
|
||||
from django.db import IntegrityError
|
||||
|
||||
responses.add(responses.GET, FEED_URL, body=_feed_xml(3), status=200)
|
||||
|
||||
with patch.object(Article.objects, "bulk_create", side_effect=IntegrityError("race")):
|
||||
result = fetch_and_ingest(feed)
|
||||
|
||||
assert len(result.new_article_ids) == 3
|
||||
assert Article.objects.filter(feed=feed).count() == 3
|
||||
@@ -208,3 +208,23 @@ def test_fetch_feed_schedules_the_next_fetch_using_the_site_wide_interval():
|
||||
assert feed.next_fetch_at is not None
|
||||
expected = before + timedelta(minutes=120)
|
||||
assert abs((feed.next_fetch_at - expected).total_seconds()) < 5
|
||||
|
||||
|
||||
@pytest.mark.django_db
|
||||
@responses.activate
|
||||
def test_fetch_feed_dispatches_rule_evaluation_in_bounded_chunks(monkeypatch):
|
||||
monkeypatch.setattr("apps.feeds.tasks.RULE_DISPATCH_CHUNK_SIZE", 2)
|
||||
feed = Feed.objects.create(feed_url="https://example.com/feed.xml")
|
||||
items = "".join(f"<item><title>T{i}</title><guid>g{i}</guid></item>" for i in range(5))
|
||||
responses.add(
|
||||
responses.GET,
|
||||
feed.feed_url,
|
||||
body=f'<?xml version="1.0"?><rss version="2.0"><channel>{items}</channel></rss>',
|
||||
status=200,
|
||||
)
|
||||
|
||||
with patch("apps.feeds.tasks.apply_filter_rules_for_feed.delay") as mock_delay:
|
||||
fetch_feed(feed.id)
|
||||
|
||||
chunk_sizes = [len(call.args[1]) for call in mock_delay.call_args_list]
|
||||
assert chunk_sizes == [2, 2, 1]
|
||||
@@ -343,3 +343,133 @@ def test_sidebar_unread_count_updates_immediately_after_mark_all_read(client, al
|
||||
|
||||
response = client.get(reverse("reader:stream"))
|
||||
assert "unread-count" not in response.content.decode()
|
||||
|
||||
|
||||
def _stream_article_ids(client, query):
|
||||
response = client.get(reverse("reader:stream"), query)
|
||||
content = response.content.decode()
|
||||
ids = [int(m) for m in re.findall(r'<li id="article-(\d+)"', content)]
|
||||
return ids, content
|
||||
|
||||
|
||||
def _link_query(content, label):
|
||||
match = re.search(rf'<a href="\?([^"]+)">{label}', content)
|
||||
if not match:
|
||||
return None
|
||||
from urllib.parse import parse_qsl
|
||||
|
||||
return dict(parse_qsl(match.group(1).replace("&", "&")))
|
||||
|
||||
|
||||
@pytest.mark.django_db
|
||||
def test_stream_keyset_pagination_pages_through_tied_timestamps_without_gaps(client, alice):
|
||||
"""Keyset pagination (C4) on (sort_at, id): articles sharing an exact
|
||||
timestamp must neither repeat nor go missing across page boundaries,
|
||||
which offset paging over a non-unique order can't guarantee."""
|
||||
from django.utils import timezone
|
||||
|
||||
from apps.reader.views import PAGE_SIZE
|
||||
|
||||
feed = Feed.objects.create(feed_url="https://example.com/feed.xml")
|
||||
Subscription.objects.create(user=alice, feed=feed)
|
||||
same_moment = timezone.now()
|
||||
total = PAGE_SIZE * 2 + 5
|
||||
created = [
|
||||
Article.objects.create(feed=feed, guid=f"g{i}", title=f"T{i}", published_at=same_moment)
|
||||
for i in range(total)
|
||||
]
|
||||
|
||||
client.login(username="alice", password="pw-alice-123")
|
||||
seen = []
|
||||
query = {"filter": "all"}
|
||||
pages = 0
|
||||
while query:
|
||||
ids, content = _stream_article_ids(client, query)
|
||||
assert len(ids) <= PAGE_SIZE
|
||||
seen.extend(ids)
|
||||
query = _link_query(content, "Older")
|
||||
pages += 1
|
||||
assert pages < 10
|
||||
|
||||
assert pages == 3
|
||||
assert len(seen) == total
|
||||
assert sorted(seen) == sorted(a.pk for a in created)
|
||||
|
||||
|
||||
@pytest.mark.django_db
|
||||
def test_stream_newer_link_returns_to_the_previous_page(client, alice):
|
||||
from django.utils import timezone
|
||||
|
||||
from apps.reader.views import PAGE_SIZE
|
||||
|
||||
feed = Feed.objects.create(feed_url="https://example.com/feed.xml")
|
||||
Subscription.objects.create(user=alice, feed=feed)
|
||||
now = timezone.now()
|
||||
for i in range(PAGE_SIZE + 5):
|
||||
Article.objects.create(
|
||||
feed=feed,
|
||||
guid=f"g{i}",
|
||||
title=f"T{i}",
|
||||
published_at=now - timezone.timedelta(minutes=i),
|
||||
)
|
||||
|
||||
client.login(username="alice", password="pw-alice-123")
|
||||
first_ids, first_content = _stream_article_ids(client, {"filter": "all"})
|
||||
assert _link_query(first_content, "Newer") is None
|
||||
|
||||
second_ids, second_content = _stream_article_ids(client, _link_query(first_content, "Older"))
|
||||
assert len(second_ids) == 5
|
||||
assert _link_query(second_content, "Older") is None
|
||||
|
||||
back_ids, _ = _stream_article_ids(client, _link_query(second_content, "Newer"))
|
||||
assert back_ids == first_ids
|
||||
|
||||
|
||||
@pytest.mark.django_db
|
||||
def test_stream_pagination_links_keep_the_feed_scope(client, alice):
|
||||
from apps.reader.views import PAGE_SIZE
|
||||
|
||||
feed = Feed.objects.create(feed_url="https://example.com/feed.xml")
|
||||
Subscription.objects.create(user=alice, feed=feed)
|
||||
for i in range(PAGE_SIZE + 1):
|
||||
Article.objects.create(feed=feed, guid=f"g{i}", title=f"T{i}")
|
||||
|
||||
client.login(username="alice", password="pw-alice-123")
|
||||
_, content = _stream_article_ids(client, {"filter": "all", "feed": feed.pk})
|
||||
|
||||
assert _link_query(content, "Older")["feed"] == str(feed.pk)
|
||||
|
||||
|
||||
def test_templates_use_no_inline_event_handlers():
|
||||
"""Production's CSP (script-src 'self') blocks inline on*= handlers —
|
||||
they work in dev, where no CSP is sent, and silently do nothing in
|
||||
production. Behavior belongs in static/js."""
|
||||
from pathlib import Path
|
||||
|
||||
from django.conf import settings
|
||||
|
||||
pattern = re.compile(r"""<[^>]+\s(on[a-z]+)\s*=""", re.IGNORECASE)
|
||||
offenders = []
|
||||
for template in Path(settings.BASE_DIR, "templates").rglob("*.html"):
|
||||
for match in pattern.finditer(template.read_text()):
|
||||
offenders.append(f"{template.name}: {match.group(1)}")
|
||||
assert offenders == []
|
||||
|
||||
|
||||
@pytest.mark.django_db
|
||||
def test_article_detail_back_link_keeps_the_stream_scope(client, alice):
|
||||
feed = Feed.objects.create(feed_url="https://example.com/feed.xml")
|
||||
Subscription.objects.create(user=alice, feed=feed)
|
||||
article = Article.objects.create(feed=feed, guid="g1", title="Hello")
|
||||
|
||||
client.login(username="alice", password="pw-alice-123")
|
||||
stream = client.get(
|
||||
reverse("reader:stream"), {"filter": "all", "feed": feed.pk}
|
||||
).content.decode()
|
||||
link = re.search(r'href="(/articles/\d+/\?[^"]+)"', stream)
|
||||
assert link is not None
|
||||
|
||||
detail = client.get(link.group(1).replace("&", "&")).content.decode()
|
||||
assert f"filter=all&feed={feed.pk}" in detail
|
||||
assert "Back to stream" in detail
|
||||
assert article.title in detail
|
||||
+66
-13
@@ -1,8 +1,10 @@
|
||||
from datetime import datetime
|
||||
from itertools import groupby
|
||||
from urllib.parse import urlencode
|
||||
|
||||
from django.contrib.auth.decorators import login_required
|
||||
from django.db.models import Max, Q
|
||||
from django.db.models.functions import Coalesce
|
||||
from django.http import Http404
|
||||
from django.shortcuts import get_object_or_404, redirect, render
|
||||
from django.urls import reverse
|
||||
@@ -157,26 +159,68 @@ def _filtered_queryset(request, active_filter):
|
||||
return _apply_scope(request, queryset)
|
||||
|
||||
|
||||
def _encode_cursor(article) -> str:
|
||||
return f"{article.sort_at.isoformat()},{article.pk}"
|
||||
|
||||
|
||||
def _decode_cursor(raw):
|
||||
"""(sort_at, id) from a cursor produced by _encode_cursor, or None for
|
||||
a missing/garbled one (treated as "no cursor" — the first page)."""
|
||||
if not raw:
|
||||
return None
|
||||
try:
|
||||
timestamp, pk = raw.rsplit(",", 1)
|
||||
return datetime.fromisoformat(timestamp), int(pk)
|
||||
except ValueError:
|
||||
return None
|
||||
|
||||
|
||||
@login_required
|
||||
def stream(request):
|
||||
active_filter = _resolve_active_filter(request)
|
||||
queryset, active_tag, active_feed = _filtered_queryset(request, active_filter)
|
||||
|
||||
try:
|
||||
page = max(int(request.GET.get("page", 1)), 1)
|
||||
except ValueError:
|
||||
page = 1
|
||||
start, end = (page - 1) * PAGE_SIZE, page * PAGE_SIZE
|
||||
|
||||
total_count = queryset.count()
|
||||
articles = with_user_state(queryset[start:end], request.user)
|
||||
|
||||
scope_params = {"filter": active_filter}
|
||||
if active_tag:
|
||||
scope_params["tag"] = active_tag.pk
|
||||
if active_feed:
|
||||
scope_params["feed"] = active_feed.pk
|
||||
|
||||
# Keyset pagination (C4) over (sort_at, id) — sort_at falls back to
|
||||
# fetched_at for entries a feed gave no published date, and id breaks
|
||||
# ties so pages neither shift nor repeat articles that share a
|
||||
# timestamp. No COUNT(*): fetching one row past the page size is
|
||||
# enough to know whether another page exists in that direction.
|
||||
ordered = queryset.annotate(sort_at=Coalesce("published_at", "fetched_at"))
|
||||
after = _decode_cursor(request.GET.get("after"))
|
||||
before = _decode_cursor(request.GET.get("before"))
|
||||
|
||||
if after:
|
||||
sort_at, pk = after
|
||||
rows = list(
|
||||
ordered.filter(Q(sort_at__gt=sort_at) | Q(sort_at=sort_at, id__gt=pk)).order_by(
|
||||
"sort_at", "id"
|
||||
)[: PAGE_SIZE + 1]
|
||||
)
|
||||
has_newer_page = len(rows) > PAGE_SIZE
|
||||
has_older_page = True
|
||||
rows = rows[:PAGE_SIZE][::-1]
|
||||
else:
|
||||
if before:
|
||||
sort_at, pk = before
|
||||
ordered = ordered.filter(Q(sort_at__lt=sort_at) | Q(sort_at=sort_at, id__lt=pk))
|
||||
rows = list(ordered.order_by("-sort_at", "-id")[: PAGE_SIZE + 1])
|
||||
has_older_page = len(rows) > PAGE_SIZE
|
||||
has_newer_page = bool(before)
|
||||
rows = rows[:PAGE_SIZE]
|
||||
|
||||
articles = with_user_state(rows, request.user)
|
||||
|
||||
newer_qs = older_qs = ""
|
||||
if articles:
|
||||
newer_qs = urlencode({**scope_params, "after": _encode_cursor(articles[0])})
|
||||
older_qs = urlencode({**scope_params, "before": _encode_cursor(articles[-1])})
|
||||
|
||||
context = {
|
||||
"date_groups": _group_by_date(articles),
|
||||
"active_filter": active_filter,
|
||||
@@ -184,15 +228,17 @@ def stream(request):
|
||||
"active_tag": active_tag,
|
||||
"active_feed": active_feed,
|
||||
"user_tags": Tag.objects.filter(user=request.user),
|
||||
"page": page,
|
||||
"has_next_page": end < total_count,
|
||||
"has_previous_page": page > 1,
|
||||
"has_newer_page": has_newer_page and bool(articles),
|
||||
"has_older_page": has_older_page and bool(articles),
|
||||
"newer_qs": newer_qs,
|
||||
"older_qs": older_qs,
|
||||
# Lets the "new articles" banner ask, on this exact same scope,
|
||||
# whether anything with a higher id has shown up since this page
|
||||
# was rendered — id (not published_at) because it's monotonic
|
||||
# regardless of how a feed backdates an entry's published date.
|
||||
"latest_article_id": queryset.aggregate(Max("id"))["id__max"] or 0,
|
||||
"new_check_qs": urlencode(scope_params),
|
||||
"scope_qs": urlencode(scope_params),
|
||||
}
|
||||
return render(request, "reader/stream.html", context)
|
||||
|
||||
@@ -266,7 +312,14 @@ def _load_article_and_mark_read(user, article_id):
|
||||
@login_required
|
||||
def article_detail(request, article_id):
|
||||
article = _load_article_and_mark_read(request.user, article_id)
|
||||
return render(request, "reader/article_detail.html", {"article": article})
|
||||
# The stream passes its filter/tag/feed scope through the link so
|
||||
# "Back to stream" returns to it instead of resetting to every feed.
|
||||
back_params = {k: request.GET[k] for k in ("filter", "tag", "feed") if request.GET.get(k)}
|
||||
return render(
|
||||
request,
|
||||
"reader/article_detail.html",
|
||||
{"article": article, "back_qs": urlencode(back_params)},
|
||||
)
|
||||
|
||||
|
||||
@login_required
|
||||
|
||||
@@ -1,27 +1,28 @@
|
||||
from django.contrib.postgres.search import SearchQuery, SearchRank, SearchVector
|
||||
|
||||
from apps.articles.models import Article
|
||||
|
||||
from .base import SearchBackend
|
||||
|
||||
# Matches the stored search_vector column, GIN index, and maintaining
|
||||
# trigger created by apps/search/migrations/0002_postgres_search_vector.py
|
||||
# — the query reads that column directly instead of recomputing a
|
||||
# tsvector over every article's full text on every search (C5).
|
||||
_VECTOR = "articles_article.search_vector"
|
||||
_QUERY = "plainto_tsquery('english', %s)"
|
||||
|
||||
|
||||
class PostgresSearchBackend(SearchBackend):
|
||||
"""Computes the search vector on the fly at query time rather than
|
||||
storing/indexing it on Article — keeps the schema identical to the
|
||||
SQLite backend. Revisit with a stored+GIN-indexed column only if an
|
||||
instance's archive grows large enough to need it."""
|
||||
|
||||
def search(self, query: str, user, limit: int = 50):
|
||||
if not query.strip():
|
||||
return Article.objects.none()
|
||||
|
||||
vector = SearchVector("title", weight="A") + SearchVector("content_html", weight="B")
|
||||
search_query = SearchQuery(query)
|
||||
|
||||
return (
|
||||
Article.objects.filter(feed__subscriptions__user=user)
|
||||
.select_related("feed")
|
||||
.annotate(search=vector, rank=SearchRank(vector, search_query))
|
||||
.filter(search=search_query)
|
||||
.extra(
|
||||
select={"rank": f"ts_rank({_VECTOR}, {_QUERY})"},
|
||||
select_params=[query],
|
||||
where=[f"{_VECTOR} @@ {_QUERY}"],
|
||||
params=[query],
|
||||
)
|
||||
.order_by("-rank", "-published_at")[:limit]
|
||||
)
|
||||
@@ -1,12 +1,14 @@
|
||||
import logging
|
||||
import sqlite3
|
||||
|
||||
from django.db import connection
|
||||
|
||||
from apps.articles.models import Article
|
||||
from apps.feeds.models import Subscription
|
||||
|
||||
from .base import SearchBackend
|
||||
|
||||
logger = logging.getLogger(__name__)
|
||||
|
||||
|
||||
def _fts_phrase_query(raw_query: str) -> str:
|
||||
"""Wraps the whole user input as a single FTS5 phrase so arbitrary
|
||||
@@ -21,24 +23,28 @@ class SQLiteFTSBackend(SearchBackend):
|
||||
if not query.strip():
|
||||
return []
|
||||
|
||||
feed_ids = list(Subscription.objects.filter(user=user).values_list("feed_id", flat=True))
|
||||
if not feed_ids:
|
||||
return []
|
||||
|
||||
placeholders = ",".join(["%s"] * len(feed_ids))
|
||||
sql = f"""
|
||||
# Scoped through a subquery on Subscription rather than an IN
|
||||
# list built from every subscribed feed id (C6): a user with
|
||||
# thousands of feeds would otherwise exceed SQLite's bound-
|
||||
# parameter limit and the whole search would fail.
|
||||
sql = """
|
||||
SELECT a.id
|
||||
FROM articles_article a
|
||||
JOIN articles_fts f ON f.rowid = a.id
|
||||
WHERE articles_fts MATCH %s AND a.feed_id IN ({placeholders})
|
||||
WHERE articles_fts MATCH %s
|
||||
AND a.feed_id IN (SELECT feed_id FROM feeds_subscription WHERE user_id = %s)
|
||||
ORDER BY rank
|
||||
LIMIT %s
|
||||
"""
|
||||
try:
|
||||
with connection.cursor() as cursor:
|
||||
cursor.execute(sql, [_fts_phrase_query(query), *feed_ids, limit])
|
||||
cursor.execute(sql, [_fts_phrase_query(query), user.pk, limit])
|
||||
ordered_ids = [row[0] for row in cursor.fetchall()]
|
||||
except sqlite3.OperationalError:
|
||||
# A genuinely bad FTS query is rare (the input is quoted as a
|
||||
# phrase), so this is far more likely a real operational
|
||||
# failure — log it instead of silently showing "no results".
|
||||
logger.exception("SQLite FTS search failed for user %s", user.pk)
|
||||
return []
|
||||
|
||||
if not ordered_ids:
|
||||
|
||||
@@ -0,0 +1,61 @@
|
||||
"""Stored, GIN-indexed tsvector for Postgres search (C5), kept current by a
|
||||
trigger so no ingestion/update code path can forget to maintain it.
|
||||
|
||||
No-ops entirely on SQLite — its FTS5 table (0001_sqlite_fts) already plays
|
||||
this role. The column is deliberately not a Django model field: a
|
||||
SearchVectorField would break SQLite, and nothing but the search backend
|
||||
ever reads it (apps/search/backends/postgres.py).
|
||||
"""
|
||||
|
||||
from django.db import migrations
|
||||
|
||||
VECTOR_EXPR = (
|
||||
"setweight(to_tsvector('english', coalesce({row}title, '')), 'A') || "
|
||||
"setweight(to_tsvector('english', coalesce({row}content_html, '')), 'B')"
|
||||
)
|
||||
|
||||
CREATE = f"""
|
||||
ALTER TABLE articles_article ADD COLUMN IF NOT EXISTS search_vector tsvector;
|
||||
|
||||
UPDATE articles_article SET search_vector = {VECTOR_EXPR.format(row="")};
|
||||
|
||||
CREATE INDEX IF NOT EXISTS articles_article_search_vector_gin
|
||||
ON articles_article USING GIN (search_vector);
|
||||
|
||||
CREATE OR REPLACE FUNCTION articles_article_search_vector_update() RETURNS trigger AS $$
|
||||
BEGIN
|
||||
NEW.search_vector := {VECTOR_EXPR.format(row="NEW.")};
|
||||
RETURN NEW;
|
||||
END
|
||||
$$ LANGUAGE plpgsql;
|
||||
|
||||
DROP TRIGGER IF EXISTS articles_article_search_vector_trg ON articles_article;
|
||||
CREATE TRIGGER articles_article_search_vector_trg
|
||||
BEFORE INSERT OR UPDATE OF title, content_html ON articles_article
|
||||
FOR EACH ROW EXECUTE FUNCTION articles_article_search_vector_update();
|
||||
"""
|
||||
|
||||
DROP = """
|
||||
DROP TRIGGER IF EXISTS articles_article_search_vector_trg ON articles_article;
|
||||
DROP FUNCTION IF EXISTS articles_article_search_vector_update();
|
||||
DROP INDEX IF EXISTS articles_article_search_vector_gin;
|
||||
ALTER TABLE articles_article DROP COLUMN IF EXISTS search_vector;
|
||||
"""
|
||||
|
||||
|
||||
def create_vector(apps, schema_editor):
|
||||
if schema_editor.connection.vendor != "postgresql":
|
||||
return
|
||||
schema_editor.execute(CREATE)
|
||||
|
||||
|
||||
def drop_vector(apps, schema_editor):
|
||||
if schema_editor.connection.vendor != "postgresql":
|
||||
return
|
||||
schema_editor.execute(DROP)
|
||||
|
||||
|
||||
class Migration(migrations.Migration):
|
||||
dependencies = [("search", "0001_sqlite_fts"), ("articles", "0001_initial")]
|
||||
|
||||
operations = [migrations.RunPython(create_vector, drop_vector)]
|
||||
@@ -17,4 +17,8 @@ app.conf.beat_schedule = {
|
||||
"task": "apps.feeds.tasks.poll_due_feeds",
|
||||
"schedule": crontab(minute="*/1"),
|
||||
},
|
||||
"purge-old-articles": {
|
||||
"task": "apps.articles.tasks.purge_old_articles",
|
||||
"schedule": crontab(hour=3, minute=30),
|
||||
},
|
||||
}
|
||||
@@ -77,6 +77,21 @@ ASGI_APPLICATION = "config.asgi.application"
|
||||
# postgres://user:pass@host:5432/dbname
|
||||
DATABASES = {"default": env.db_url("DATABASE_URL", default=f"sqlite:///{BASE_DIR / 'db.sqlite3'}")}
|
||||
|
||||
# gunicorn, Celery workers, and the FTS triggers all write to the same
|
||||
# SQLite file (C6): without WAL, readers and a writer block each other,
|
||||
# and without a busy timeout a contended write fails immediately with
|
||||
# "database is locked" instead of waiting its turn. IMMEDIATE takes the
|
||||
# write lock at transaction start so a read-then-write transaction can't
|
||||
# deadlock upgrading it mid-way.
|
||||
if DATABASES["default"]["ENGINE"] == "django.db.backends.sqlite3":
|
||||
DATABASES["default"].setdefault("OPTIONS", {}).update(
|
||||
{
|
||||
"timeout": 20,
|
||||
"transaction_mode": "IMMEDIATE",
|
||||
"init_command": "PRAGMA journal_mode=WAL; PRAGMA synchronous=NORMAL;",
|
||||
}
|
||||
)
|
||||
|
||||
AUTH_USER_MODEL = "accounts.User"
|
||||
|
||||
AUTH_PASSWORD_VALIDATORS = [
|
||||
|
||||
@@ -36,7 +36,7 @@ SECURE_PROXY_SSL_HEADER = ("HTTP_X_FORWARDED_PROTO", "https")
|
||||
# over loopback (see ALLOWED_HOSTS above) it has no X-Forwarded-Proto
|
||||
# header to prove it's "secure", and there's no TLS listener on gunicorn's
|
||||
# own port to redirect to anyway.
|
||||
SECURE_REDIRECT_EXEMPT = [r"^health/$"]
|
||||
SECURE_REDIRECT_EXEMPT = [r"^health/$", r"^health/ready/$"]
|
||||
|
||||
# WhiteNoise serves /static/ directly from gunicorn instead of relying on
|
||||
# the reverse proxy to alias the staticfiles directory on disk — the proxy
|
||||
|
||||
+20
-2
@@ -313,8 +313,26 @@ surfaces any feeds that are failing to fetch, which is the fastest way to catch
|
||||
misconfigured feed or an outage at the source.
|
||||
|
||||
`/health/` is an unauthenticated liveness check (DB connectivity only) —
|
||||
`scripts/upgrade.sh` uses it to confirm a deploy actually came back up, and
|
||||
it's a reasonable target for an external uptime monitor too.
|
||||
`scripts/upgrade.sh` uses it to confirm a deploy actually came back up.
|
||||
Point an external uptime monitor at `/health/ready/` instead: it also checks
|
||||
Redis and that the background poller has ticked in the last five minutes
|
||||
(proving Celery beat is scheduling *and* a worker is consuming), and returns
|
||||
503 with a per-check breakdown when any of them fails — so "the site loads
|
||||
but nothing is being fetched" actually alerts.
|
||||
|
||||
**Retention.** By default every fetched article is kept forever. In the Django
|
||||
admin, "Site settings" has a *Retention days* field: set it to delete articles
|
||||
fetched longer ago than that, nightly (03:30 UTC, via Celery beat). Articles
|
||||
anyone has starred or saved for later are always kept; read state and tags for
|
||||
deleted articles go with them. The dashboard shows the article count and the
|
||||
age of the oldest one, so you can see what retention is doing.
|
||||
|
||||
**SQLite.** SQLite installs run in WAL mode with a 20-second busy timeout (set
|
||||
automatically in `config/settings/base.py`) so gunicorn, the Celery worker, and
|
||||
the full-text-search triggers can share the file without "database is locked"
|
||||
errors. That's plenty for a personal or small-team instance; for many users or
|
||||
a very large archive, use PostgreSQL (its search uses a stored, GIN-indexed
|
||||
vector kept current by a trigger, added by migration automatically).
|
||||
|
||||
## Backups
|
||||
|
||||
|
||||
+23
-5
@@ -1,8 +1,11 @@
|
||||
function toggleArticle(event, id) {
|
||||
event.preventDefault();
|
||||
var link = event.currentTarget;
|
||||
// Delegated listeners, not inline onclick="" attributes: production's
|
||||
// Content Security Policy (config/settings/prod.py, script-src 'self')
|
||||
// blocks inline event handlers, which silently turned the title click into
|
||||
// a plain link to the detail page and the new-articles banner into a no-op.
|
||||
|
||||
function toggleArticle(link) {
|
||||
var row = link.closest("li.article-row");
|
||||
var body = document.getElementById("article-body-" + id);
|
||||
var body = row.querySelector(".article-inline-body");
|
||||
var wasOpen = row.classList.contains("is-expanded");
|
||||
|
||||
document.querySelectorAll("li.article-row.is-expanded").forEach(function (openRow) {
|
||||
@@ -25,5 +28,20 @@ function toggleArticle(event, id) {
|
||||
);
|
||||
}
|
||||
}
|
||||
return false;
|
||||
}
|
||||
|
||||
document.addEventListener("click", function (event) {
|
||||
var link = event.target.closest("a.article-title[data-inline-url]");
|
||||
if (link) {
|
||||
// Leave modified clicks alone so "open in new tab" still works.
|
||||
if (event.button !== 0 || event.ctrlKey || event.metaKey || event.shiftKey || event.altKey) {
|
||||
return;
|
||||
}
|
||||
event.preventDefault();
|
||||
toggleArticle(link);
|
||||
return;
|
||||
}
|
||||
if (event.target.closest(".new-articles-banner")) {
|
||||
location.reload();
|
||||
}
|
||||
});
|
||||
@@ -6,7 +6,7 @@
|
||||
<ul class="dashboard-stats">
|
||||
<li>{{ stats.feed_count }} feed(s) — {{ stats.active_feed_count }} active</li>
|
||||
<li>{{ stats.subscription_count }} subscription(s)</li>
|
||||
<li>{{ stats.article_count }} article(s)</li>
|
||||
<li>{{ stats.article_count }} article(s){% if stats.oldest_article_at %}, oldest fetched <time datetime="{{ stats.oldest_article_at|date:"c" }}" data-local-format="long">{{ stats.oldest_article_at|date:"M j, Y H:i" }}</time>{% endif %}</li>
|
||||
<li>{{ stats.error_feed_count }} feed(s) currently erroring</li>
|
||||
</ul>
|
||||
|
||||
|
||||
@@ -10,10 +10,10 @@
|
||||
</ul>
|
||||
|
||||
<nav class="pagination">
|
||||
{% if has_previous_page %}
|
||||
<a href="?filter={{ active_filter }}{% if active_tag %}&tag={{ active_tag.pk }}{% endif %}&page={{ page|add:"-1" }}">← Newer</a>
|
||||
{% if has_newer_page %}
|
||||
<a href="?{{ newer_qs }}">← Newer</a>
|
||||
{% endif %}
|
||||
{% if has_next_page %}
|
||||
<a href="?filter={{ active_filter }}{% if active_tag %}&tag={{ active_tag.pk }}{% endif %}&page={{ page|add:"1" }}">Older →</a>
|
||||
{% if has_older_page %}
|
||||
<a href="?{{ older_qs }}">Older →</a>
|
||||
{% endif %}
|
||||
</nav>
|
||||
@@ -6,9 +6,8 @@
|
||||
<img src="{{ article.feed.favicon_url }}" alt="" class="feed-favicon">
|
||||
{% endif %}
|
||||
<span class="article-feed-name">{{ article.feed.title|default:article.feed.feed_url }}</span>
|
||||
<a href="{% url 'reader:article_detail' article.pk %}" class="article-title"
|
||||
data-inline-url="{% url 'reader:article_inline' article.pk %}"
|
||||
onclick="return toggleArticle(event, {{ article.pk }})">{{ article.title|default:"(untitled)" }}</a>
|
||||
<a href="{% url 'reader:article_detail' article.pk %}{% if scope_qs %}?{{ scope_qs }}{% endif %}" class="article-title"
|
||||
data-inline-url="{% url 'reader:article_inline' article.pk %}">{{ article.title|default:"(untitled)" }}</a>
|
||||
{% if article.published_at %}
|
||||
<time class="article-time" datetime="{{ article.published_at|date:"c" }}" data-local-format="short">{{ article.published_at|date:"g:i A" }}</time>
|
||||
{% endif %}
|
||||
|
||||
@@ -1,5 +1,5 @@
|
||||
{% if new_count %}
|
||||
<button type="button" class="new-articles-banner" onclick="location.reload()">
|
||||
<button type="button" class="new-articles-banner">
|
||||
{{ new_count }} new article{{ new_count|pluralize }} — click to show
|
||||
</button>
|
||||
{% endif %}
|
||||
@@ -1,7 +1,7 @@
|
||||
{% extends "base.html" %}
|
||||
{% block title %}{{ article.title|default:"Article" }} — Evection{% endblock %}
|
||||
{% block content %}
|
||||
<p><a href="{% url 'reader:stream' %}">← Back to stream</a></p>
|
||||
<p><a href="{% url 'reader:stream' %}{% if back_qs %}?{{ back_qs }}{% endif %}">← Back to stream</a></p>
|
||||
<article>
|
||||
{% include "reader/_article_actions.html" %}
|
||||
{% include "reader/_article_content.html" %}
|
||||
|
||||
Reference in new issue
Block a user