refactor(server): split oversized backend services

server/src/app/services/knowledge_ingest_status.py

@@ -0,0 +1,69 @@
+from __future__ import annotations
+
+from datetime import UTC, datetime
+from typing import Any
+
+from sqlalchemy import select
+from sqlalchemy.orm import Session
+
+from app.core.agent_enums import AgentRunStatus
+from app.models.agent_run import AgentRun
+from app.services.knowledge_constants import (
+    KNOWLEDGE_INGEST_STATUS_META,
+    KNOWLEDGE_INGEST_STATUS_PUBLISHED,
+    KNOWLEDGE_INGEST_SYNC_STALE_SECONDS,
+)
+
+
+def normalize_ingest_status_code(value: Any) -> int:
+    try:
+        status_code = int(value)
+    except (TypeError, ValueError):
+        return KNOWLEDGE_INGEST_STATUS_PUBLISHED
+    if status_code not in KNOWLEDGE_INGEST_STATUS_META:
+        return KNOWLEDGE_INGEST_STATUS_PUBLISHED
+    return status_code
+
+
+def is_syncing_status_stale(entry: dict[str, Any]) -> bool:
+    raw_value = str(entry.get("ingest_status_updated_at") or "").strip()
+    if not raw_value:
+        return True
+    try:
+        updated_at = datetime.fromisoformat(raw_value)
+    except ValueError:
+        return True
+    if updated_at.tzinfo is None:
+        updated_at = updated_at.replace(tzinfo=UTC)
+    age_seconds = (datetime.now(UTC) - updated_at.astimezone(UTC)).total_seconds()
+    return age_seconds >= KNOWLEDGE_INGEST_SYNC_STALE_SECONDS
+
+
+def should_preserve_syncing_status(entry: dict[str, Any], *, db: Session | None) -> bool:
+    agent_run_id = str(entry.get("ingest_agent_run_id") or "").strip()
+    if not agent_run_id or db is None:
+        return not is_syncing_status_stale(entry)
+
+    run = db.scalar(select(AgentRun).where(AgentRun.run_id == agent_run_id))
+    if run is None:
+        return not is_syncing_status_stale(entry)
+    if run.status != AgentRunStatus.RUNNING.value:
+        return False
+
+    heartbeat_at = str((run.route_json or {}).get("heartbeat_at") or "").strip()
+    if heartbeat_at:
+        probe_entry = {"ingest_status_updated_at": heartbeat_at}
+        return not is_syncing_status_stale(probe_entry)
+
+    return not is_syncing_status_stale(entry)
+
+
+def resolve_linked_ingest_run_status(entry: dict[str, Any], *, db: Session | None) -> str:
+    agent_run_id = str(entry.get("ingest_agent_run_id") or "").strip()
+    if not agent_run_id or db is None:
+        return ""
+
+    run = db.scalar(select(AgentRun).where(AgentRun.run_id == agent_run_id))
+    if run is None:
+        return ""
+    return str(run.status or "").strip()