.github/workflows/release.yml

@@ -21,6 +21,10 @@ jobs:
           python-version: "3.12"
       - uses: astral-sh/setup-uv@v4
       - run: uv sync --extra dev
+      - name: Lint
+        run: uv run ruff check
+      - name: Test
+        run: uv run pytest
       - name: Semantic Release
         id: release
         run: uv run semantic-release publish

app/api/routers/run_manual.py

@@ -1,7 +1,7 @@
 from __future__ import annotations
 
 import logging
-from typing import Any
+from typing import Any, NoReturn
 
 from fastapi import Request
 from sqlalchemy.exc import IntegrityError
@@ -215,7 +215,7 @@ async def execute_manual_run(
         raise_manual_failed(exc=exc, user_id=user_id)
 
 
-def raise_manual_blocked_safety(*, exc, user_id: int) -> None:
+def raise_manual_blocked_safety(*, exc, user_id: int) -> NoReturn:
     structured_log(
         logger,
         "info",
@@ -233,7 +233,7 @@ def raise_manual_blocked_safety(*, exc, user_id: int) -> None:
     ) from exc
 
 
-def raise_manual_failed(*, exc: Exception, user_id: int) -> None:
+def raise_manual_failed(*, exc: Exception, user_id: int) -> NoReturn:
     structured_log(
         logger,
         "exception",

app/api/schemas/admin.py

@@ -35,7 +35,7 @@ class AdminUsersListEnvelope(BaseModel):
 
 class AdminUserCreateRequest(BaseModel):
     email: str = Field(max_length=254)
-    password: str = Field(max_length=128)
+    password: str = Field(min_length=8, max_length=128)
     is_admin: bool = False
 
     model_config = ConfigDict(extra="forbid")
@@ -55,7 +55,7 @@ class AdminUserEnvelope(BaseModel):
 
 
 class AdminResetPasswordRequest(BaseModel):
-    new_password: str = Field(max_length=128)
+    new_password: str = Field(min_length=8, max_length=128)
 
     model_config = ConfigDict(extra="forbid")
 
@@ -64,7 +64,7 @@ class AdminScholarHttpSettingsData(BaseModel):
     user_agent: str
     rotate_user_agent: bool
     accept_language: str
-    cookie: str
+    cookie: str | None = None
 
     model_config = ConfigDict(extra="forbid")
 
@@ -80,7 +80,7 @@ class AdminScholarHttpSettingsUpdateRequest(BaseModel):
     user_agent: str
     rotate_user_agent: bool
     accept_language: str
-    cookie: str
+    cookie: str | None = None
 
     model_config = ConfigDict(extra="forbid")
 

app/api/schemas/auth.py

@@ -45,7 +45,7 @@ class CsrfBootstrapEnvelope(BaseModel):
 
 class LoginRequest(BaseModel):
     email: str = Field(max_length=254)
-    password: str = Field(max_length=128)
+    password: str = Field(min_length=8, max_length=128)
 
     model_config = ConfigDict(extra="forbid")
 
@@ -66,8 +66,8 @@ class LoginEnvelope(BaseModel):
 
 
 class ChangePasswordRequest(BaseModel):
-    current_password: str = Field(max_length=128)
+    current_password: str = Field(min_length=8, max_length=128)
-    new_password: str = Field(max_length=128)
+    new_password: str = Field(min_length=8, max_length=128)
-    confirm_password: str = Field(max_length=128)
+    confirm_password: str = Field(min_length=8, max_length=128)
 
     model_config = ConfigDict(extra="forbid")

app/main.py

@@ -81,6 +81,26 @@ async def lifespan(_: FastAPI):
             error=str(exc),
         )
 
+    try:
+        session_factory = get_session_factory()
+        async with session_factory() as session:
+            await session.execute(
+                text(
+                    "UPDATE publication_pdf_jobs SET status = 'queued'"
+                    " WHERE status = 'running'"
+                    " AND (last_attempt_at IS NULL OR last_attempt_at < NOW() - INTERVAL '10 minutes')"
+                )
+            )
+            await session.commit()
+            structured_log(logger, "info", "app.startup_stuck_pdf_jobs_recovered")
+    except Exception as exc:
+        structured_log(
+            logger,
+            "error",
+            "app.startup_stuck_pdf_jobs_recovery_failed",
+            error=str(exc),
+        )
+
     await scheduler_service.start()
     yield
     await scheduler_service.stop()

app/services/arxiv/rate_limit.py

@@ -68,8 +68,10 @@ async def _run_serialized_fetch(
     source_path: str,
 ) -> tuple[httpx.Response, bool]:
     session_factory = get_session_factory()
+    async with session_factory() as lock_session:
+        await _acquire_arxiv_lock(lock_session)
+        try:
             async with session_factory() as db_session, db_session.begin():
-        await _acquire_arxiv_lock(db_session)
                 runtime_state = await _load_runtime_state_for_update(db_session)
                 wait_seconds = await _wait_for_allowed_slot_or_raise(
                     runtime_state,
@@ -88,6 +90,9 @@ async def _run_serialized_fetch(
                     response_status=int(response.status_code),
                     source_path=source_path,
                 )
+                cooldown_until_value = runtime_state.cooldown_until
+        finally:
+            await _release_arxiv_lock(lock_session)
     structured_log(
         logger,
         "info",
@@ -95,7 +100,7 @@ async def _run_serialized_fetch(
         status_code=int(response.status_code),
         wait_seconds=wait_seconds,
         cooldown_remaining_seconds=_cooldown_remaining_seconds(
-            runtime_state.cooldown_until, now_utc=datetime.now(UTC)
+            cooldown_until_value, now_utc=datetime.now(UTC)
         ),
         source_path=source_path,
     )
@@ -104,7 +109,17 @@ async def _run_serialized_fetch(
 
 async def _acquire_arxiv_lock(db_session: AsyncSession) -> None:
     await db_session.execute(
-        text("SELECT pg_advisory_xact_lock(:namespace, :lock_key)"),
+        text("SELECT pg_advisory_lock(:namespace, :lock_key)"),
+        {
+            "namespace": ARXIV_RATE_LIMIT_LOCK_NAMESPACE,
+            "lock_key": ARXIV_RATE_LIMIT_LOCK_KEY,
+        },
+    )
+
+
+async def _release_arxiv_lock(db_session: AsyncSession) -> None:
+    await db_session.execute(
+        text("SELECT pg_advisory_unlock(:namespace, :lock_key)"),
         {
             "namespace": ARXIV_RATE_LIMIT_LOCK_NAMESPACE,
             "lock_key": ARXIV_RATE_LIMIT_LOCK_KEY,

app/services/crossref/application.py

@@ -20,7 +20,7 @@ if TYPE_CHECKING:
     from app.services.publications.types import PublicationListItem, UnreadPublicationItem
 
 TOKEN_RE = re.compile(r"[a-z0-9]+")
-NON_ALNUM_RE = re.compile(r"[^a-z0-9\\s]+")
+NON_ALNUM_RE = re.compile(r"[^a-z0-9\s]+")
 STOP_WORDS = {"the", "and", "for", "with", "from", "method", "study", "analysis"}
 _RATE_LOCK = threading.Lock()
 _LAST_REQUEST_AT = 0.0
@@ -37,6 +37,7 @@ def _rate_limit_wait(min_interval_seconds: float) -> None:
         remaining = interval - elapsed
     if remaining > 0:
         time.sleep(remaining)
+    with _RATE_LOCK:
         _LAST_REQUEST_AT = time.monotonic()
 
 

app/services/dbops/near_duplicate_repair.py

@@ -162,7 +162,10 @@ async def _complete_job(
 
 
 async def _fail_job(db_session: AsyncSession, *, job: DataRepairJob, error: Exception) -> None:
+    from sqlalchemy.orm import make_transient
+
     await db_session.rollback()
+    make_transient(job)
     job.status = REPAIR_STATUS_FAILED
     job.error_text = str(error)
     job.finished_at = _utcnow()

app/services/ingestion/application.py

@@ -401,6 +401,7 @@ class ScholarIngestionService:
                 if run_to_fail:
                     run_to_fail.status = RunStatus.FAILED
                     run_to_fail.end_dt = datetime.now(UTC)
+                    run_to_fail.error_log = run_to_fail.error_log or {}
                     run_to_fail.error_log["terminal_exception"] = str(exc)
                     await cleanup_session.commit()
         except Exception:

app/services/ingestion/publication_upsert.py

@@ -176,6 +176,7 @@ async def upsert_profile_publications(
     seen_publication_ids: set[int] = set()
     discovered_count = 0
 
+    try:
         for candidate in publications:
             publication = await resolve_publication(db_session, candidate)
             if publication.id in seen_publication_ids:
@@ -201,7 +202,7 @@ async def upsert_profile_publications(
             db_session.add(link)
             discovered_count += 1
 
-        await commit_discovered_publication(
+            await flush_discovered_publication(
                 db_session,
                 run=run,
                 scholar=scholar,
@@ -211,10 +212,15 @@ async def upsert_profile_publications(
         if not scholar.baseline_completed:
             scholar.baseline_completed = True
 
+        await db_session.commit()
+    except Exception:
+        await db_session.rollback()
+        raise
+
     return discovered_count
 
 
-async def commit_discovered_publication(
+async def flush_discovered_publication(
     db_session: AsyncSession,
     *,
     run: CrawlRun,
@@ -222,7 +228,7 @@ async def commit_discovered_publication(
     publication: Publication,
 ) -> None:
     run.new_pub_count = int(run.new_pub_count or 0) + 1
-    await db_session.commit()
+    await db_session.flush()
     await run_events.publish(
         run_id=run.id,
         event_type="publication_discovered",

app/services/ingestion/queue.py

@@ -196,6 +196,7 @@ async def list_due_jobs(
             IngestionQueueItem.id.asc(),
         )
         .limit(limit)
+        .with_for_update(skip_locked=True)
     )
     rows = list(result.scalars().all())
     jobs: list[ContinuationQueueJob] = []

app/services/publications/pdf_queue_common.py

@@ -47,4 +47,5 @@ def queued_job(
         status=PDF_STATUS_QUEUED,
         queued_at=now,
         last_requested_by_user_id=user_id,
+        attempt_count=0,
     )

app/services/publications/pdf_queue_resolution.py

@@ -46,7 +46,7 @@ async def _mark_attempt_started(
             db_session.add(job)
         job.status = PDF_STATUS_RUNNING
         job.last_attempt_at = utcnow()
-        job.attempt_count = int(job.attempt_count) + 1
+        job.attempt_count = int(job.attempt_count or 0) + 1
         db_session.add(
             event_row(
                 publication_id=publication_id,
@@ -240,6 +240,26 @@ async def _run_resolution_task(
                 detail="Stopping PDF resolution batch — OpenAlex daily budget exhausted",
             )
             break
+        except Exception:
+            structured_log(
+                logger,
+                "exception",
+                "pdf_queue.row_failed",
+                publication_id=row.publication_id,
+            )
+            try:
+                await _persist_outcome(
+                    publication_id=row.publication_id,
+                    user_id=user_id,
+                    outcome=_failed_outcome(row=row),
+                )
+            except Exception:
+                structured_log(
+                    logger,
+                    "exception",
+                    "pdf_queue.row_fail_persist_error",
+                    publication_id=row.publication_id,
+                )
 
 
 def _register_task(task: asyncio.Task[None]) -> None:

app/services/runs/events.py

@@ -8,6 +8,8 @@ from app.logging_utils import structured_log
 
 logger = logging.getLogger(__name__)
 
+_SUBSCRIBER_QUEUE_MAXSIZE = 256
+
 
 class RunEventPublisher:
     def __init__(self) -> None:
@@ -17,7 +19,7 @@ class RunEventPublisher:
     def subscribe(self, run_id: int) -> asyncio.Queue:
         if run_id not in self._subscribers:
             self._subscribers[run_id] = set()
-        queue: asyncio.Queue[Any] = asyncio.Queue()
+        queue: asyncio.Queue[Any] = asyncio.Queue(maxsize=_SUBSCRIBER_QUEUE_MAXSIZE)
         self._subscribers[run_id].add(queue)
         structured_log(
             logger,
@@ -51,6 +53,10 @@ class RunEventPublisher:
                     "runs.event_subscriber_queue_full",
                     run_id=run_id,
                 )
+                self._subscribers[run_id].discard(queue)
+
+    async def publish_run_complete(self, run_id: int) -> None:
+        await self.publish(run_id, "run_complete", {})
 
 
 run_events = RunEventPublisher()
@@ -62,8 +68,11 @@ async def event_generator(run_id: int) -> AsyncGenerator[str, None]:
         while True:
             # Wait for a new event
             message = await queue.get()
-            # Server-Sent Events format: "event: <type>\ndata: <json>\n\n"
             event_type = message["type"]
+            if event_type == "run_complete":
+                yield f"event: {event_type}\ndata: {{}}\n\n"
+                break
+            # Server-Sent Events format: "event: <type>\ndata: <json>\n\n"
             data_str = json.dumps(message["data"])
             yield f"event: {event_type}\ndata: {data_str}\n\n"
     except asyncio.CancelledError:

frontend/src/features/publications/components/PublicationTableRow.test.ts

@@ -13,7 +13,7 @@ function buildItem(overrides: Partial<PublicationItem> = {}): PublicationItem {
     citation_count: 42,
     pub_url: "https://example.com/pub",
     pdf_url: null,
-    pdf_status: null,
+    pdf_status: "untracked",
     is_read: false,
     is_favorite: false,
     is_new_in_latest_run: true,

frontend/src/features/publications/composables/usePublicationData.ts

@@ -1,4 +1,4 @@
-import { computed, ref, watch } from "vue";
+import { computed, onScopeDispose, ref, watch } from "vue";
 import { useRoute, useRouter } from "vue-router";
 import {
   listPublications,
@@ -271,6 +271,10 @@ export function usePublicationData() {
     }, 300);
   });
 
+  onScopeDispose(() => {
+    if (searchDebounceTimer !== null) clearTimeout(searchDebounceTimer);
+  });
+
   // --- Run-triggered refresh watcher ---
 
   watch(

frontend/src/features/settings/SettingsAdminPanel.vue

@@ -36,10 +36,8 @@ async function refreshForSection(): Promise<void> {
     return;
   }
   if (props.section === SECTION_REPAIRS) {
-    await Promise.all([
+    await usersRef.value?.load();
-      usersRef.value?.load(),
+    await repairsRef.value?.load();
-      repairsRef.value?.load(),
-    ]);
     return;
   }
   if (props.section === SECTION_PDF && pdfQueueRef.value) {

frontend/src/features/settings/components/AdminPdfQueueSection.vue

@@ -7,6 +7,7 @@ import AppHelpHint from "@/components/ui/AppHelpHint.vue";
 import AppRefreshButton from "@/components/ui/AppRefreshButton.vue";
 import AppSelect from "@/components/ui/AppSelect.vue";
 import AppTable from "@/components/ui/AppTable.vue";
+import RequestStateAlerts from "@/components/patterns/RequestStateAlerts.vue";
 import {
   listAdminPdfQueue,
   requeueAdminPdfLookup,
@@ -15,7 +16,7 @@ import {
 } from "@/features/admin_dbops";
 import { useRequestState } from "@/composables/useRequestState";
 
-const { clearAlerts, assignError, setSuccess } = useRequestState();
+const { errorMessage, errorRequestId, successMessage, clearAlerts, assignError, setSuccess } = useRequestState();
 
 const refreshingPdfQueue = ref(false);
 const requeueingPublicationId = ref<number | null>(null);
@@ -131,6 +132,16 @@ defineExpose({ load });
 </script>
 
 <template>
+  <section class="grid gap-4">
+    <RequestStateAlerts
+      :success-message="successMessage"
+      :error-message="errorMessage"
+      :error-request-id="errorRequestId"
+      success-title="PDF queue operation complete"
+      error-title="PDF queue operation failed"
+      @dismiss-success="successMessage = null"
+    />
+
     <AppCard class="space-y-3">
     <div class="flex flex-wrap items-center justify-between gap-2">
       <div class="flex items-center gap-1">
@@ -200,4 +211,5 @@ defineExpose({ load });
       </div>
     </div>
   </AppCard>
+  </section>
 </template>

frontend/src/features/settings/components/AdminRepairsSection.vue

@@ -9,6 +9,7 @@ import AppInput from "@/components/ui/AppInput.vue";
 import AppRefreshButton from "@/components/ui/AppRefreshButton.vue";
 import AppSelect from "@/components/ui/AppSelect.vue";
 import AppTable from "@/components/ui/AppTable.vue";
+import RequestStateAlerts from "@/components/patterns/RequestStateAlerts.vue";
 import {
   dropAllPublications,
   listAdminDbRepairJobs,
@@ -34,7 +35,7 @@ const DROP_PUBLICATIONS_CONFIRM_TEXT = "DROP ALL PUBLICATIONS";
 
 type RepairScopeMode = typeof SCOPE_SINGLE_USER | typeof SCOPE_ALL_USERS;
 
-const { clearAlerts, assignError, setSuccess } = useRequestState();
+const { errorMessage, errorRequestId, successMessage, clearAlerts, assignError, setSuccess } = useRequestState();
 
 const refreshingJobs = ref(false);
 const runningRepair = ref(false);
@@ -246,6 +247,15 @@ defineExpose({ load });
 
 <template>
   <section class="grid gap-4">
+    <RequestStateAlerts
+      :success-message="successMessage"
+      :error-message="errorMessage"
+      :error-request-id="errorRequestId"
+      success-title="Repair operation complete"
+      error-title="Repair operation failed"
+      @dismiss-success="successMessage = null"
+    />
+
     <AppCard class="space-y-3">
       <div class="flex items-center gap-1">
         <h2 class="text-lg font-semibold text-ink-primary">Publication Link Repair</h2>

frontend/src/pages/PublicationsPage.vue

@@ -335,6 +335,7 @@ async function onStartRun(): Promise<void> {
   pub.successMessage.value = null;
   pub.errorMessage.value = null;
   pub.errorRequestId.value = null;
+  try {
     const result = await runStatus.startManualCheck();
     if (result.kind === "started") { pub.successMessage.value = `Run #${result.runId} started.`; return; }
     if (result.kind === "already_running") {
@@ -343,6 +344,14 @@ async function onStartRun(): Promise<void> {
     }
     pub.errorMessage.value = result.message;
     pub.errorRequestId.value = result.requestId;
+  } catch (error) {
+    if (error instanceof ApiRequestError) {
+      pub.errorMessage.value = error.message;
+      pub.errorRequestId.value = error.requestId;
+    } else {
+      pub.errorMessage.value = "Unable to start manual run.";
+    }
+  }
 }
 
 // --- Lifecycle ---