feat: check source links during scheduled fetches

This commit is contained in:
ik
2026-09-13 16:32:10 +07:00
parent bc3ceb5dfe
commit 1305cccfa5
6 changed files with 110 additions and 7 deletions
+47 -2
View File
@@ -2,7 +2,9 @@ from __future__ import annotations
from datetime import datetime, timezone
from typing import Literal
from urllib.error import HTTPError
from sqlalchemy import select
from sqlalchemy.orm import Session
from .models import ExternalObservation, ModerationStatus
@@ -11,6 +13,16 @@ from .models import ExternalObservation, ModerationStatus
SourceCheckStatus = Literal["available", "missing", "temporary_error", "blocked"]
def classify_source_failure(exc: Exception) -> SourceCheckStatus:
"""Classify the result of the scheduled request without retrying it."""
if isinstance(exc, HTTPError):
if exc.code in {404, 410}:
return "missing"
if exc.code in {401, 403, 429}:
return "blocked"
return "temporary_error"
def record_source_check(
session: Session,
observation: ExternalObservation,
@@ -24,7 +36,18 @@ def record_source_check(
data. Transient errors and access blocks remain diagnostic and never remove
an observation from activity.
"""
current = checked_at or datetime.now(timezone.utc)
_apply_source_check(observation, status, checked_at or datetime.now(timezone.utc))
session.commit()
return observation
def _apply_source_check(
observation: ExternalObservation,
status: SourceCheckStatus,
checked_at: datetime,
) -> None:
"""Mutate one observation; the caller owns the transaction boundary."""
current = checked_at
observation.source_check_status = status
observation.source_checked_at = current
if status == "missing" and observation.status != "withdrawn":
@@ -34,5 +57,27 @@ def record_source_check(
observation.review_note = "Source record missing; withdrawn pending moderator review"
observation.reviewed_at = current
observation.moderation_version += 1
def record_scheduled_source_check(
session: Session,
*,
source_system: str,
source_url: str,
status: SourceCheckStatus,
checked_at: datetime | None = None,
) -> int:
"""Apply one scheduled request result only to observations with that exact URL.
Aggregate pages cannot prove that an omitted record was deleted, so absence
from a parsed listing is deliberately ignored.
"""
observations = list(session.scalars(select(ExternalObservation).where(
ExternalObservation.source_system == source_system,
ExternalObservation.source_url == source_url,
)))
current = checked_at or datetime.now(timezone.utc)
for observation in observations:
_apply_source_check(observation, status, current)
session.commit()
return observation
return len(observations)