feat: harden governed runs and add dashboard integration

This commit is contained in:
2026-07-29 14:16:28 +02:00
parent 8c2b6035cc
commit 09c98087c5
10 changed files with 372 additions and 29 deletions
+3 -1
View File
@@ -67,7 +67,7 @@ def normalize_definition_scope(
return principal.tenant_id, "group", clean_id
if clean_type == "user":
if not clean_id:
clean_id = principal.membership_id or principal.account_id
clean_id = principal.account_id
if (
clean_id not in {principal.membership_id, principal.account_id}
and not administrative
@@ -75,6 +75,8 @@ def normalize_definition_scope(
raise PermissionError(
"Definitions can only be created for the current user."
)
if clean_id == principal.membership_id:
clean_id = principal.account_id
return principal.tenant_id, "user", clean_id
raise ValueError(
"Definition scope must be system, tenant, group, or user."
+12
View File
@@ -7,6 +7,7 @@ from govoplan_core.core.module_guards import (
persistent_table_uninstall_guard,
)
from govoplan_core.core.access import (
CAPABILITY_ACCESS_DIRECTORY,
CAPABILITY_AUTH_AUTOMATION_PRINCIPAL_PROVIDER,
)
from govoplan_core.core.dataflows import (
@@ -34,6 +35,7 @@ from govoplan_core.core.datasources import (
from govoplan_core.core.policy import (
CAPABILITY_POLICY_DEFINITION_GOVERNANCE,
)
from govoplan_core.core.views import ViewSurface
from govoplan_core.db.base import Base
from govoplan_dataflow.backend.db import models as dataflow_models
@@ -241,6 +243,7 @@ manifest = ModuleManifest(
"workflow",
),
optional_capabilities=(
CAPABILITY_ACCESS_DIRECTORY,
CAPABILITY_DATASOURCE_CATALOGUE,
CAPABILITY_DATASOURCE_LIFECYCLE,
CAPABILITY_DATASOURCE_PUBLICATION,
@@ -320,6 +323,15 @@ manifest = ModuleManifest(
order=72,
),
),
view_surfaces=(
ViewSurface(
id="dataflow.widget.pipelines",
module_id=MODULE_ID,
kind="section",
label="Dataflows widget",
order=70,
),
),
),
route_factory=_dataflow_router,
capability_factories={
+76 -2
View File
@@ -1,8 +1,12 @@
from __future__ import annotations
from fastapi import APIRouter, Depends, HTTPException, status
from fastapi import APIRouter, Depends, HTTPException, Query, status
from sqlalchemy.orm import Session
from govoplan_core.api.v1.schemas import (
ReferenceOptionListResponse,
ReferenceOptionResponse,
)
from govoplan_core.audit.logging import audit_event
from govoplan_core.auth import ApiPrincipal, get_api_principal, has_scope
from govoplan_core.core.automation import AutomationInvocation
@@ -21,6 +25,11 @@ from govoplan_core.core.datasources import (
datasource_catalogue,
datasource_lifecycle,
)
from govoplan_core.core.references import (
access_scope_reference_options,
access_scope_reference_provider_available,
validate_access_scope_reference,
)
from govoplan_core.core.tabular_sources import (
parse_tabular_csv,
)
@@ -264,6 +273,37 @@ def api_node_types(
return _node_library_response()
@router.get("/scope-targets", response_model=ReferenceOptionListResponse)
def api_scope_targets(
scope_type: str,
q: str = "",
selected: list[str] = Query(default=[]),
limit: int = Query(default=50, ge=1, le=200),
principal: ApiPrincipal = Depends(get_api_principal),
) -> ReferenceOptionListResponse:
_require_any_scope(principal, READ_SCOPE, WRITE_SCOPE, ADMIN_SCOPE)
registry = get_registry()
try:
options = access_scope_reference_options(
registry,
principal,
scope_type=scope_type,
query=q,
selected_values=selected,
limit=limit,
administrative=has_scope(principal, ADMIN_SCOPE),
)
except ValueError as exc:
raise HTTPException(
status_code=status.HTTP_422_UNPROCESSABLE_CONTENT,
detail=str(exc),
) from exc
return ReferenceOptionListResponse(
options=[ReferenceOptionResponse(**item.to_dict()) for item in options],
provider_available=access_scope_reference_provider_available(registry),
)
@router.get("/sources", response_model=TabularSourceListResponse)
def api_list_sources(
query: str = "",
@@ -417,6 +457,13 @@ def api_create_pipeline(
payload = payload.model_copy(
update={"scope_type": scope_type, "scope_id": scope_id}
)
scope_id = validate_access_scope_reference(
get_registry(),
tenant_id=tenant_id or principal.tenant_id,
scope_type=scope_type,
scope_id=scope_id,
)
payload = payload.model_copy(update={"scope_id": scope_id})
pipeline = create_pipeline(
session,
tenant_id=tenant_id or principal.tenant_id,
@@ -488,6 +535,26 @@ def api_update_pipeline(
registry=get_registry(),
action="edit",
)
tenant_id, scope_type, scope_id = normalize_definition_scope(
principal,
scope_type=payload.scope_type,
scope_id=payload.scope_id,
administrative=has_scope(principal, ADMIN_SCOPE),
)
scope_id = validate_access_scope_reference(
get_registry(),
tenant_id=tenant_id or principal.tenant_id,
scope_type=scope_type,
scope_id=scope_id,
preserve_existing=(
existing.scope_id
if existing.scope_type == scope_type
else None
),
)
payload = payload.model_copy(
update={"scope_type": scope_type, "scope_id": scope_id}
)
pipeline = update_pipeline(
session,
tenant_id=principal.tenant_id,
@@ -495,7 +562,7 @@ def api_update_pipeline(
actor_id=_actor_id(principal),
payload=payload,
)
except PermissionError as exc:
except (PermissionError, ValueError) as exc:
raise _governance_http_error(exc) from exc
except DataflowError as exc:
raise _http_error(exc) from exc
@@ -579,6 +646,13 @@ def api_derive_pipeline(
payload = payload.model_copy(
update={"scope_type": scope_type, "scope_id": scope_id}
)
scope_id = validate_access_scope_reference(
get_registry(),
tenant_id=tenant_id or principal.tenant_id,
scope_type=scope_type,
scope_id=scope_id,
)
payload = payload.model_copy(update={"scope_id": scope_id})
pipeline = derive_pipeline(
session,
tenant_id=tenant_id or principal.tenant_id,
+9 -1
View File
@@ -707,7 +707,15 @@ def _dispatch_delivery(
registry: object | None,
now: datetime,
) -> str:
trigger = session.get(DataflowTrigger, delivery.trigger_id)
# Serialize capacity checks per trigger. Delivery rows are independently
# leased with SKIP LOCKED, so without this lock two workers could each
# claim a different delivery and both observe the same available slot.
trigger = session.scalar(
select(DataflowTrigger)
.where(DataflowTrigger.id == delivery.trigger_id)
.with_for_update()
.execution_options(populate_existing=True)
)
pipeline = session.get(DataflowPipeline, delivery.pipeline_id)
if trigger is None or pipeline is None:
return _finish_delivery(