mirrors/ArchiveBox
1
0
Fork 0
mirror of https://github.com/ArchiveBox/ArchiveBox.git synced 2025-05-13 22:54:27 -04:00
Code Issues Projects Releases Packages Wiki Activity Actions

working state machine flow yay

Browse source
This commit is contained in:
Nick Sweeting 2024-11-18 04:25:45 -08:00
parent 9adfe0e2e6
commit eb53145e4e
No known key found for this signature in database
7 changed files with 107 additions and 52 deletions
Download patch file Download diff file Expand all files Collapse all files

39
archivebox/core/statemachines.py
View file

@ -29,9 +29,9 @@ class SnapshotMachine(StateMachine, strict_states=True):
# Tick Event
tick = (
queued.to.itself(unless='can_start', internal=True) |
queued.to.itself(unless='can_start') |
queued.to(started, cond='can_start') |
started.to.itself(unless='is_finished', internal=True) |
started.to.itself(unless='is_finished') |
started.to(sealed, cond='is_finished')
)
@ -40,7 +40,7 @@ class SnapshotMachine(StateMachine, strict_states=True):
super().__init__(snapshot, *args, **kwargs)
def can_start(self) -> bool:
can_start = bool(self.snapshot.url and (self.snapshot.retry_at < timezone.now()))
can_start = bool(self.snapshot.url)
if not can_start:
print(f'SnapshotMachine[{self.snapshot.ABID}].can_start() False: {self.snapshot.url} {self.snapshot.retry_at} {timezone.now()}')
return can_start
@ -63,24 +63,34 @@ class SnapshotMachine(StateMachine, strict_states=True):
@queued.enter
def enter_queued(self):
print(f'SnapshotMachine[{self.snapshot.ABID}].on_queued(): snapshot.retry_at = now()')
self.snapshot.status = Snapshot.StatusChoices.QUEUED
self.snapshot.retry_at = timezone.now()
self.snapshot.save()
self.snapshot.update_for_workers(
retry_at=timezone.now(),
status=Snapshot.StatusChoices.QUEUED,
)
@started.enter
def enter_started(self):
print(f'SnapshotMachine[{self.snapshot.ABID}].on_started(): snapshot.create_pending_archiveresults() + snapshot.bump_retry_at(+60s)')
self.snapshot.bump_retry_at(seconds=30) # if failed, wait 10s before retrying
self.snapshot.save()
# lock the snapshot while we create the pending archiveresults
self.snapshot.update_for_workers(
retry_at=timezone.now() + timedelta(seconds=30), # if failed, wait 30s before retrying
)
# create the pending archiveresults
self.snapshot.create_pending_archiveresults()
self.snapshot.status = Snapshot.StatusChoices.STARTED
# unlock the snapshot after we're done creating the pending archiveresults + set status = started
self.snapshot.update_for_workers(
retry_at=timezone.now() + timedelta(seconds=5), # wait 5s before checking it again
status=Snapshot.StatusChoices.STARTED,
)
@sealed.enter
def enter_sealed(self):
print(f'SnapshotMachine[{self.snapshot.ABID}].on_sealed(): snapshot.retry_at=None')
self.snapshot.status = Snapshot.StatusChoices.SEALED
self.snapshot.retry_at = None
self.snapshot.save()
self.snapshot.update_for_workers(
retry_at=None,
status=Snapshot.StatusChoices.SEALED,
)
class SnapshotWorker(ActorType[Snapshot]):
@ -136,7 +146,10 @@ class ArchiveResultMachine(StateMachine, strict_states=True):
super().__init__(archiveresult, *args, **kwargs)
def can_start(self) -> bool:
return self.archiveresult.snapshot and (self.archiveresult.retry_at < timezone.now())
can_start = bool(self.archiveresult.snapshot.url)
if not can_start:
print(f'ArchiveResultMachine[{self.archiveresult.ABID}].can_start() False: {self.archiveresult.snapshot.url} {self.archiveresult.retry_at} {timezone.now()}')
return can_start
def is_succeeded(self) -> bool:
if self.archiveresult.output and 'err' not in self.archiveresult.output.lower():