paradizelost/paperless-ngx
1
0
Fork 0
mirror of https://github.com/paperless-ngx/paperless-ngx.git synced 2025-10-30 03:56:23 -05:00
Code Issues Packages Projects Releases Wiki Activity

Enhancement: support negative offset in scheduled workflows (#9746)

Browse Source
This commit is contained in:
shamoon
2025-05-11 13:04:46 -07:00
committed by GitHub
parent 73f0f1212d
commit 344cc70cd5
7 changed files with 227 additions and 97 deletions
Download Patch File Download Diff File Expand all files Collapse all files

22
src/documents/migrations/1066_alter_workflowtrigger_schedule_offset_days.py Normal file
View File

@@ -0,0 +1,22 @@
# Generated by Django 5.1.7 on 2025-04-15 19:18
from django.db import migrations
from django.db import models
class Migration(migrations.Migration):
dependencies = [
("documents", "1065_workflowaction_assign_custom_fields_values"),
]
operations = [
migrations.AlterField(
model_name="workflowtrigger",
name="schedule_offset_days",
field=models.IntegerField(
default=0,
help_text="The number of days to offset the schedule trigger by.",
verbose_name="schedule offset days",
),
),
]

2
src/documents/models.py
View File

@@ -1019,7 +1019,7 @@ class WorkflowTrigger(models.Model):
verbose_name=_("has this correspondent"),
)
schedule_offset_days = models.PositiveIntegerField(
schedule_offset_days = models.IntegerField(
_("schedule offset days"),
default=0,
help_text=_(

76
src/documents/tasks.py
View File

@@ -1,8 +1,8 @@
import datetime
import hashlib
import logging
import shutil
import uuid
from datetime import timedelta
from pathlib import Path
from tempfile import TemporaryDirectory
@@ -357,7 +357,7 @@ def empty_trash(doc_ids=None):
if doc_ids is not None
else Document.deleted_objects.filter(
deleted_at__lt=timezone.localtime(timezone.now())
- timedelta(
- datetime.timedelta(
days=settings.EMPTY_TRASH_DELAY,
),
)
@@ -397,6 +397,7 @@ def check_scheduled_workflows():
)
if scheduled_workflows.count() > 0:
logger.debug(f"Checking {len(scheduled_workflows)} scheduled workflows")
now = timezone.now()
for workflow in scheduled_workflows:
schedule_triggers = workflow.triggers.filter(
type=WorkflowTrigger.WorkflowTriggerType.SCHEDULED,
@@ -404,31 +405,60 @@ def check_scheduled_workflows():
trigger: WorkflowTrigger
for trigger in schedule_triggers:
documents = Document.objects.none()
offset_td = timedelta(days=trigger.schedule_offset_days)
offset_td = datetime.timedelta(days=-trigger.schedule_offset_days)
threshold = now - offset_td
logger.debug(
f"Checking trigger {trigger} with offset {offset_td} against field: {trigger.schedule_date_field}",
f"Trigger {trigger.id}: checking if (date + {offset_td}) <= now ({now})",
)
match trigger.schedule_date_field:
case WorkflowTrigger.ScheduleDateField.ADDED:
documents = Document.objects.filter(
added__lt=timezone.now() - offset_td,
)
documents = Document.objects.filter(added__lte=threshold)
case WorkflowTrigger.ScheduleDateField.CREATED:
documents = Document.objects.filter(
created__lt=timezone.now() - offset_td,
)
documents = Document.objects.filter(created__lte=threshold)
case WorkflowTrigger.ScheduleDateField.MODIFIED:
documents = Document.objects.filter(
modified__lt=timezone.now() - offset_td,
)
documents = Document.objects.filter(modified__lte=threshold)
case WorkflowTrigger.ScheduleDateField.CUSTOM_FIELD:
cf_instances = CustomFieldInstance.objects.filter(
field=trigger.schedule_date_custom_field,
value_date__lt=timezone.now() - offset_td,
)
documents = Document.objects.filter(
id__in=cf_instances.values_list("document", flat=True),
# cap earliest date to avoid massive scans
earliest_date = now - datetime.timedelta(days=365)
if offset_td.days < -365:
logger.warning(
f"Trigger {trigger.id} has large negative offset ({offset_td.days}), "
f"limiting earliest scan date to {earliest_date}",
)
cf_filter_kwargs = {
"field": trigger.schedule_date_custom_field,
"value_date__isnull": False,
"value_date__lte": threshold,
"value_date__gte": earliest_date,
}
recent_cf_instances = CustomFieldInstance.objects.filter(
**cf_filter_kwargs,
)
matched_ids = [
cfi.document_id
for cfi in recent_cf_instances
if cfi.value_date
and (
timezone.make_aware(
datetime.datetime.combine(
cfi.value_date,
datetime.time.min,
),
)
+ offset_td
<= now
)
]
documents = Document.objects.filter(id__in=matched_ids)
if documents.count() > 0:
logger.debug(
f"Found {documents.count()} documents for trigger {trigger}",
@@ -440,18 +470,18 @@ def check_scheduled_workflows():
workflow=workflow,
).order_by("-run_at")
if not trigger.schedule_is_recurring and workflow_runs.exists():
# schedule is non-recurring and the workflow has already been run
logger.debug(
f"Skipping document {document} for non-recurring workflow {workflow} as it has already been run",
)
continue
elif (
if (
trigger.schedule_is_recurring
and workflow_runs.exists()
and (
workflow_runs.last().run_at
> timezone.now()
- timedelta(
> now
- datetime.timedelta(
days=trigger.schedule_recurring_interval_days,
)
)

64
src/documents/tests/test_workflows.py
View File

@@ -1336,6 +1336,8 @@ class TestWorkflows(
GIVEN:
- Existing workflow with SCHEDULED trigger against the created field and action that assigns owner
- Existing doc that matches the trigger
- Workflow set to trigger at (now - offset) = now - 1 day
- Document created date is 2 days ago → trigger condition met
WHEN:
- Scheduled workflows are checked
THEN:
@@ -1359,7 +1361,7 @@ class TestWorkflows(
w.save()
now = timezone.localtime(timezone.now())
created = now - timedelta(weeks=520)
created = now - timedelta(days=2)
doc = Document.objects.create(
title="sample test",
correspondent=self.c,
@@ -1377,6 +1379,8 @@ class TestWorkflows(
GIVEN:
- Existing workflow with SCHEDULED trigger against the added field and action that assigns owner
- Existing doc that matches the trigger
- Workflow set to trigger at (now - offset) = now - 1 day
- Document added date is 365 days ago
WHEN:
- Scheduled workflows are checked
THEN:
@@ -1418,6 +1422,8 @@ class TestWorkflows(
GIVEN:
- Existing workflow with SCHEDULED trigger against the modified field and action that assigns owner
- Existing doc that matches the trigger
- Workflow set to trigger at (now - offset) = now - 1 day
- Document modified date is mocked as sufficiently in the past
WHEN:
- Scheduled workflows are checked
THEN:
@@ -1458,6 +1464,8 @@ class TestWorkflows(
GIVEN:
- Existing workflow with SCHEDULED trigger against a custom field and action that assigns owner
- Existing doc that matches the trigger
- Workflow set to trigger at (now - offset) = now - 1 day
- Custom field date is 2 days ago
WHEN:
- Scheduled workflows are checked
THEN:
@@ -1502,6 +1510,7 @@ class TestWorkflows(
GIVEN:
- Existing workflow with SCHEDULED trigger
- Existing doc that has already had the workflow run
- Document created 2 days ago, workflow offset = 1 day → trigger time = yesterday
WHEN:
- Scheduled workflows are checked
THEN:
@@ -1552,6 +1561,7 @@ class TestWorkflows(
GIVEN:
- Existing workflow with SCHEDULED trigger and recurring interval of 7 days
- Workflow run date is 6 days ago
- Document created 40 days ago, offset = 30 → trigger time = 10 days ago
WHEN:
- Scheduled workflows are checked
THEN:
@@ -1600,6 +1610,58 @@ class TestWorkflows(
doc.refresh_from_db()
self.assertIsNone(doc.owner)
def test_workflow_scheduled_trigger_negative_offset(self):
"""
GIVEN:
- Existing workflow with SCHEDULED trigger and negative offset of -7 days (so 7 days after date)
- Custom field date initially set to 5 days ago → trigger time = 2 days in future
- Then updated to 8 days ago → trigger time = 1 day ago
WHEN:
- Scheduled workflows are checked for document with custom field date 8 days in the past
THEN:
- Workflow runs and document owner is updated
"""
trigger = WorkflowTrigger.objects.create(
type=WorkflowTrigger.WorkflowTriggerType.SCHEDULED,
schedule_offset_days=-7,
schedule_date_field=WorkflowTrigger.ScheduleDateField.CUSTOM_FIELD,
schedule_date_custom_field=self.cf1,
)
action = WorkflowAction.objects.create(
assign_title="Doc assign owner",
assign_owner=self.user2,
)
w = Workflow.objects.create(
name="Workflow 1",
order=0,
)
w.triggers.add(trigger)
w.actions.add(action)
w.save()
doc = Document.objects.create(
title="sample test",
correspondent=self.c,
original_filename="sample.pdf",
)
cfi = CustomFieldInstance.objects.create(
document=doc,
field=self.cf1,
value_date=timezone.now() - timedelta(days=5),
)
tasks.check_scheduled_workflows()
doc.refresh_from_db()
self.assertIsNone(doc.owner) # has not triggered yet
cfi.value_date = timezone.now() - timedelta(days=8)
cfi.save()
tasks.check_scheduled_workflows()
doc.refresh_from_db()
self.assertEqual(doc.owner, self.user2)
def test_workflow_enabled_disabled(self):
trigger = WorkflowTrigger.objects.create(
type=WorkflowTrigger.WorkflowTriggerType.DOCUMENT_ADDED,