dbos 1.4.0a1__py3-none-any.whl → 1.5.0a2__py3-none-any.whl

dbos/_admin_server.py

@@ -7,6 +7,8 @@ from functools import partial
 from http.server import BaseHTTPRequestHandler, ThreadingHTTPServer
 from typing import TYPE_CHECKING, Any, List, Optional, TypedDict
 
+from dbos._workflow_commands import garbage_collect, global_timeout
+
 from ._context import SetWorkflowID
 from ._error import DBOSException
 from ._logger import dbos_logger
@@ -20,6 +22,8 @@ _health_check_path = "/dbos-healthz"
 _workflow_recovery_path = "/dbos-workflow-recovery"
 _deactivate_path = "/deactivate"
 _workflow_queues_metadata_path = "/dbos-workflow-queues-metadata"
+_garbage_collect_path = "/dbos-garbage-collect"
+_global_timeout_path = "/dbos-global-timeout"
 # /workflows/:workflow_id/cancel
 # /workflows/:workflow_id/resume
 # /workflows/:workflow_id/restart
@@ -122,6 +126,23 @@ class AdminRequestHandler(BaseHTTPRequestHandler):
             self.send_response(200)
             self._end_headers()
             self.wfile.write(json.dumps(workflow_ids).encode("utf-8"))
+        elif self.path == _garbage_collect_path:
+            inputs = json.loads(post_data.decode("utf-8"))
+            cutoff_epoch_timestamp_ms = inputs.get("cutoff_epoch_timestamp_ms", None)
+            rows_threshold = inputs.get("rows_threshold", None)
+            garbage_collect(
+                self.dbos,
+                cutoff_epoch_timestamp_ms=cutoff_epoch_timestamp_ms,
+                rows_threshold=rows_threshold,
+            )
+            self.send_response(204)
+            self._end_headers()
+        elif self.path == _global_timeout_path:
+            inputs = json.loads(post_data.decode("utf-8"))
+            timeout_ms = inputs.get("timeout_ms", None)
+            global_timeout(self.dbos, timeout_ms)
+            self.send_response(204)
+            self._end_headers()
         else:
 
             restart_match = re.match(

dbos/_app_db.py

@@ -256,3 +256,21 @@ class ApplicationDatabase:
             )
 
             conn.execute(insert_stmt)
+
+    def garbage_collect(
+        self, cutoff_epoch_timestamp_ms: int, pending_workflow_ids: list[str]
+    ) -> None:
+        with self.engine.begin() as c:
+            delete_query = sa.delete(ApplicationSchema.transaction_outputs).where(
+                ApplicationSchema.transaction_outputs.c.created_at
+                < cutoff_epoch_timestamp_ms
+            )
+
+            if len(pending_workflow_ids) > 0:
+                delete_query = delete_query.where(
+                    ~ApplicationSchema.transaction_outputs.c.workflow_uuid.in_(
+                        pending_workflow_ids
+                    )
+                )
+
+            c.execute(delete_query)

dbos/_debug.py

@@ -15,11 +15,11 @@ class PythonModule:
 
 
 def debug_workflow(workflow_id: str, entrypoint: Union[str, PythonModule]) -> None:
-    # include the current directory (represented by empty string) in the search path
-    # if it not already included
-    if "" not in sys.path:
-        sys.path.insert(0, "")
     if isinstance(entrypoint, str):
+        # ensure the entrypoint parent directory is in sys.path
+        parent = str(Path(entrypoint).parent)
+        if parent not in sys.path:
+            sys.path.insert(0, parent)
         runpy.run_path(entrypoint)
     elif isinstance(entrypoint, PythonModule):
         runpy.run_module(entrypoint.module_name)

dbos/_sys_db.py

@@ -1852,6 +1852,62 @@ class SystemDatabase:
             dbos_logger.error(f"Error connecting to the DBOS system database: {e}")
             raise
 
+    def garbage_collect(
+        self, cutoff_epoch_timestamp_ms: Optional[int], rows_threshold: Optional[int]
+    ) -> Optional[tuple[int, list[str]]]:
+        if rows_threshold is not None:
+            with self.engine.begin() as c:
+                # Get the created_at timestamp of the rows_threshold newest row
+                result = c.execute(
+                    sa.select(SystemSchema.workflow_status.c.created_at)
+                    .order_by(SystemSchema.workflow_status.c.created_at.desc())
+                    .limit(1)
+                    .offset(rows_threshold - 1)
+                ).fetchone()
+
+                if result is not None:
+                    rows_based_cutoff = result[0]
+                    # Use the more restrictive cutoff (higher timestamp = more recent = more deletion)
+                    if (
+                        cutoff_epoch_timestamp_ms is None
+                        or rows_based_cutoff > cutoff_epoch_timestamp_ms
+                    ):
+                        cutoff_epoch_timestamp_ms = rows_based_cutoff
+
+        if cutoff_epoch_timestamp_ms is None:
+            return None
+
+        with self.engine.begin() as c:
+            # Delete all workflows older than cutoff that are NOT PENDING or ENQUEUED
+            c.execute(
+                sa.delete(SystemSchema.workflow_status)
+                .where(
+                    SystemSchema.workflow_status.c.created_at
+                    < cutoff_epoch_timestamp_ms
+                )
+                .where(
+                    ~SystemSchema.workflow_status.c.status.in_(
+                        [
+                            WorkflowStatusString.PENDING.value,
+                            WorkflowStatusString.ENQUEUED.value,
+                        ]
+                    )
+                )
+            )
+
+            # Then, get the IDs of all remaining old workflows
+            pending_enqueued_result = c.execute(
+                sa.select(SystemSchema.workflow_status.c.workflow_uuid).where(
+                    SystemSchema.workflow_status.c.created_at
+                    < cutoff_epoch_timestamp_ms
+                )
+            ).fetchall()
+
+            # Return the final cutoff and workflow IDs
+            return cutoff_epoch_timestamp_ms, [
+                row[0] for row in pending_enqueued_result
+            ]
+
 
 def reset_system_database(postgres_db_url: sa.URL, sysdb_name: str) -> None:
     try:

dbos/_workflow_commands.py

@@ -1,8 +1,9 @@
+import time
 import uuid
-from typing import List, Optional
+from datetime import datetime
+from typing import TYPE_CHECKING, List, Optional
 
 from dbos._context import get_local_dbos_context
-from dbos._error import DBOSException
 
 from ._app_db import ApplicationDatabase
 from ._sys_db import (
@@ -11,8 +12,12 @@ from ._sys_db import (
     StepInfo,
     SystemDatabase,
     WorkflowStatus,
+    WorkflowStatusString,
 )
 
+if TYPE_CHECKING:
+    from ._dbos import DBOS
+
 
 def list_workflows(
     sys_db: SystemDatabase,
@@ -118,3 +123,32 @@ def fork_workflow(
         application_version=application_version,
     )
     return forked_workflow_id
+
+
+def garbage_collect(
+    dbos: "DBOS",
+    cutoff_epoch_timestamp_ms: Optional[int],
+    rows_threshold: Optional[int],
+) -> None:
+    if cutoff_epoch_timestamp_ms is None and rows_threshold is None:
+        return
+    result = dbos._sys_db.garbage_collect(
+        cutoff_epoch_timestamp_ms=cutoff_epoch_timestamp_ms,
+        rows_threshold=rows_threshold,
+    )
+    if result is not None:
+        cutoff_epoch_timestamp_ms, pending_workflow_ids = result
+        dbos._app_db.garbage_collect(cutoff_epoch_timestamp_ms, pending_workflow_ids)
+
+
+def global_timeout(dbos: "DBOS", timeout_ms: int) -> None:
+    cutoff_epoch_timestamp_ms = int(time.time() * 1000) - timeout_ms
+    cutoff_iso = datetime.fromtimestamp(cutoff_epoch_timestamp_ms / 1000).isoformat()
+    for workflow in dbos.list_workflows(
+        status=WorkflowStatusString.PENDING.value, end_time=cutoff_iso
+    ):
+        dbos.cancel_workflow(workflow.workflow_id)
+    for workflow in dbos.list_workflows(
+        status=WorkflowStatusString.ENQUEUED.value, end_time=cutoff_iso
+    ):
+        dbos.cancel_workflow(workflow.workflow_id)

{dbos-1.4.0a1.dist-info → dbos-1.5.0a2.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: dbos
-Version: 1.4.0a1
+Version: 1.5.0a2
 Summary: Ultra-lightweight durable execution in Python
 Author-Email: "DBOS, Inc." <contact@dbos.dev>
 License: MIT

{dbos-1.4.0a1.dist-info → dbos-1.5.0a2.dist-info}/RECORD

@@ -1,11 +1,11 @@
-dbos-1.4.0a1.dist-info/METADATA,sha256=Snqbd6UknvRQFKSZJUL9tBkcEvfCGnRNXGH--L4nGPc,13267
-dbos-1.4.0a1.dist-info/WHEEL,sha256=tSfRZzRHthuv7vxpI4aehrdN9scLjk-dCJkPLzkHxGg,90
-dbos-1.4.0a1.dist-info/entry_points.txt,sha256=_QOQ3tVfEjtjBlr1jS4sHqHya9lI2aIEIWkz8dqYp14,58
-dbos-1.4.0a1.dist-info/licenses/LICENSE,sha256=VGZit_a5-kdw9WT6fY5jxAWVwGQzgLFyPWrcVVUhVNU,1067
+dbos-1.5.0a2.dist-info/METADATA,sha256=1_SM7qqqbdht1SS-0o1s9YSUSJ2ZIC2Hm7oyCZFU64w,13267
+dbos-1.5.0a2.dist-info/WHEEL,sha256=tSfRZzRHthuv7vxpI4aehrdN9scLjk-dCJkPLzkHxGg,90
+dbos-1.5.0a2.dist-info/entry_points.txt,sha256=_QOQ3tVfEjtjBlr1jS4sHqHya9lI2aIEIWkz8dqYp14,58
+dbos-1.5.0a2.dist-info/licenses/LICENSE,sha256=VGZit_a5-kdw9WT6fY5jxAWVwGQzgLFyPWrcVVUhVNU,1067
 dbos/__init__.py,sha256=NssPCubaBxdiKarOWa-wViz1hdJSkmBGcpLX_gQ4NeA,891
 dbos/__main__.py,sha256=G7Exn-MhGrVJVDbgNlpzhfh8WMX_72t3_oJaFT9Lmt8,653
-dbos/_admin_server.py,sha256=TWXi4drrzKFpKkUmEJpJkQBZxAtOalnhtYicEn2nDK0,10618
-dbos/_app_db.py,sha256=0PKqpxJ3EbIaak3Wl0lNl3hXvhBfz4EEHaCw1bUOvIM,9937
+dbos/_admin_server.py,sha256=SVk55SxT07OHi0wHt_VpQsBXOeuJL2017k7_YQI3oeg,11574
+dbos/_app_db.py,sha256=htblDPfqrpb_uZoFcvaud7cgQ-PDyn6Bn-cBidxdCTA,10603
 dbos/_classproperty.py,sha256=f0X-_BySzn3yFDRKB2JpCbLYQ9tLwt1XftfshvY7CBs,626
 dbos/_client.py,sha256=cQxw1Nbh_vKZ03lONt0EmUhwXBk3B3NczZrmfXXeefY,14667
 dbos/_conductor/conductor.py,sha256=o0IaZjwnZ2TOyHeP2H4iSX6UnXLXQ4uODvWAKD9hHMs,21703
@@ -15,7 +15,7 @@ dbos/_core.py,sha256=hvHKi31-3LG5yfWa-KhsnoFrXsV_eT-GeKIZFT4chx8,48533
 dbos/_croniter.py,sha256=XHAyUyibs_59sJQfSNWkP7rqQY6_XrlfuuCxk4jYqek,47559
 dbos/_dbos.py,sha256=GVx3NY59tKWW6nlAtH2PvX4Ne_eOHvY012MtXVK_FQA,47265
 dbos/_dbos_config.py,sha256=2CC1YR8lP9W-_NsMUMnTnW-v-70KN4XkbJEeNJ78RlQ,20373
-dbos/_debug.py,sha256=MNlQVZ6TscGCRQeEEL0VE8Uignvr6dPeDDDefS3xgIE,1823
+dbos/_debug.py,sha256=99j2SChWmCPAlZoDmjsJGe77tpU2LEa8E2TtLAnnh7o,1831
 dbos/_docker_pg_helper.py,sha256=tLJXWqZ4S-ExcaPnxg_i6cVxL6ZxrYlZjaGsklY-s2I,6115
 dbos/_error.py,sha256=q0OQJZTbR8FFHV9hEpAGpz9oWBT5L509zUhmyff7FJw,8500
 dbos/_event_loop.py,sha256=cvaFN9-II3MsHEOq8QoICc_8qSKrjikMlLfuhC3Y8Dk,2923
@@ -49,7 +49,7 @@ dbos/_schemas/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 dbos/_schemas/application_database.py,sha256=SypAS9l9EsaBHFn9FR8jmnqt01M74d9AF1AMa4m2hhI,1040
 dbos/_schemas/system_database.py,sha256=rbFKggONdvvbb45InvGz0TM6a7c-Ux9dcaL-h_7Z7pU,4438
 dbos/_serialization.py,sha256=bWuwhXSQcGmiazvhJHA5gwhrRWxtmFmcCFQSDJnqqkU,3666
-dbos/_sys_db.py,sha256=67z_K0aKH8M_oRs9c13zhp6skpT-sLAw8nYRBa3JM5w,77844
+dbos/_sys_db.py,sha256=now889o6Mlmcdopp8xF5_0LAE67KeVH9Vm-4svIqo5s,80170
 dbos/_templates/dbos-db-starter/README.md,sha256=GhxhBj42wjTt1fWEtwNriHbJuKb66Vzu89G4pxNHw2g,930
 dbos/_templates/dbos-db-starter/__package/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 dbos/_templates/dbos-db-starter/__package/main.py.dbos,sha256=aQnBPSSQpkB8ERfhf7gB7P9tsU6OPKhZscfeh0yiaD8,2702
@@ -62,11 +62,11 @@ dbos/_templates/dbos-db-starter/migrations/versions/2024_07_31_180642_init.py,sh
 dbos/_templates/dbos-db-starter/start_postgres_docker.py,sha256=lQVLlYO5YkhGPEgPqwGc7Y8uDKse9HsWv5fynJEFJHM,1681
 dbos/_tracer.py,sha256=yN6GRDKu_1p-EqtQLNarMocPfga2ZuqpzStzzSPYhzo,2732
 dbos/_utils.py,sha256=uywq1QrjMwy17btjxW4bES49povlQwYwYbvKwMT6C2U,1575
-dbos/_workflow_commands.py,sha256=UCpHWvCEXjVZtf5FNanFvtJpgUJDSI1EFBqQP0x_2A0,3346
+dbos/_workflow_commands.py,sha256=2_ubdzKNJZQ32oftiOGFo2JBsfU_koGC1giXRgwwexI,4539
 dbos/cli/_github_init.py,sha256=Y_bDF9gfO2jB1id4FV5h1oIxEJRWyqVjhb7bNEa5nQ0,3224
 dbos/cli/_template_init.py,sha256=7JBcpMqP1r2mfCnvWatu33z8ctEGHJarlZYKgB83cXE,2972
 dbos/cli/cli.py,sha256=EemOMqNpzSU2BQhAxV_e59pBRITDLwt49HF6W3uWBZg,20775
 dbos/dbos-config.schema.json,sha256=CjaspeYmOkx6Ip_pcxtmfXJTn_YGdSx_0pcPBF7KZmo,6060
 dbos/py.typed,sha256=QfzXT1Ktfk3Rj84akygc7_42z0lRpCq0Ilh8OXI6Zas,44
 version/__init__.py,sha256=L4sNxecRuqdtSFdpUGX3TtBi9KL3k7YsZVIvv-fv9-A,1678
-dbos-1.4.0a1.dist-info/RECORD,,
+dbos-1.5.0a2.dist-info/RECORD,,