fix(tagging): end the batch job after its cleanup and report pre-pass progress (#379)

The job read as ended while it still held the model gate, so a quick
restart could be refused with model-busy. Pre-pass chunks now publish
progress, and a cancel between chunks keeps exact counts.

Part of #379
This commit is contained in:
Vito Sansevero
2026-10-27 05:49:40 -07:00
parent cfa54dc8ae
commit bb4925c42d
2 changed files with 80 additions and 20 deletions
+62
View File
@@ -372,6 +372,68 @@ class TestBatchLifecycle(BatchTestCase):
self.assertEqual(self.tagger.calls, [])
self.assertIsNone(model_gate.holder())
async def test_cancel_between_prepass_chunks_keeps_counts(self):
for n in range(3):
self.image(f"{n}.png", (n, n, n))
await self.run_batch()
seen = []
real = jobs_module._prepass_chunk
def chunk(db, items, skip_marked, max_file):
result = real(db, items, skip_marked, max_file)
job = self.api._jobs.active("autotag")
seen.append(job.processed)
job.cancelled = True # cancel after the first chunk
return result
with (
patch.object(jobs_module, "PREPASS_CHUNK", 1),
patch.object(jobs_module, "_prepass_chunk", side_effect=chunk),
):
job = await self.run_batch()
self.assertEqual(job.state, "cancelled")
self.assertEqual(job.counters["skipped"], 1)
self.assertEqual((job.processed, job.total), (1, 3))
self.assertEqual(seen, [0])
async def test_the_job_ends_only_after_its_cleanup(self):
self.image("a.png", (1, 1, 1))
PromptManagerConfig.SAFETY_BACKUP = True
states = []
async def prune(job, _db):
states.append((job.state, model_gate.holder() is not None))
with (
patch("py.api.maintenance.take_safety_backup", return_value="db.backup_x"),
patch.object(self.api, "_prune_backup_pool", side_effect=prune),
):
job = await self.run_batch()
self.assertEqual(job.state, "finished")
self.assertEqual(states, [("running", True)])
self.assertIsNone(model_gate.holder())
async def test_prepass_publishes_its_progress(self):
for n in range(2):
self.image(f"{n}.png", (n, n, n))
await self.run_batch()
published = []
real = type(self.api._jobs).start
def start(registry, *args, **kwargs):
job = real(registry, *args, **kwargs)
original = job.publish
job.publish = lambda payload: (
published.append(dict(payload)),
original(payload),
)
return job
with patch.object(type(self.api._jobs), "start", start):
await self.run_batch()
self.assertIn({"type": "progress", "progress": 100}, published)
self.assertEqual(published[-1], {"type": "finished", "progress": 100})
async def test_restore_mid_run_fails_the_job(self):
self.image("a.png", (1, 1, 1))
self.image("b.png", (2, 2, 2))