Files
govoplan-scheduling/src/govoplan_scheduling/backend/dsar_provider.py
T

892 lines
30 KiB
Python

from __future__ import annotations
from collections.abc import Mapping, Sequence
from datetime import datetime, timezone
from sqlalchemy import func, or_
from sqlalchemy.orm import Session
from govoplan_core.core.dsar import (
DsarErasureActionRef,
DsarExecutionResultRef,
DsarRecordRef,
DsarSubjectRef,
dsar_capability_name,
)
from govoplan_scheduling.backend.db.models import (
SchedulingCandidateSlot,
SchedulingNotification,
SchedulingParticipant,
SchedulingRequest,
)
SCHEDULING_DSAR_CAPABILITY = dsar_capability_name("scheduling")
_MAX_RECORDS = 5_000
_TERMINAL_REQUEST_STATUSES = frozenset(
{"decided", "handed_off", "cancelled", "archived"}
)
_TERMINAL_NOTIFICATION_STATUSES = frozenset({"sent", "failed", "skipped", "cancelled"})
class SchedulingDsarProvider:
provider_id = "scheduling"
module_id = "scheduling"
def search_subject(
self,
session: object,
*,
tenant_id: str,
subject: DsarSubjectRef,
) -> Sequence[DsarRecordRef]:
db = _session(session)
references = _scheduling_references(subject)
membership_ids = _membership_ids(subject)
respondent_ids = _respondent_ids(subject)
account_ids = _account_ids(subject)
email = _subject_email(subject)
if not any(
(
references,
membership_ids,
respondent_ids,
account_ids,
email,
)
):
return ()
participants = _matching_participants(
db,
tenant_id=tenant_id,
participant_id=references.get("participant"),
respondent_ids=respondent_ids,
account_ids=account_ids,
email=email,
)
participant_ids = {row.id for row in participants}
request_ids = {row.request_id for row in participants}
if references.get("request"):
request_ids.add(references["request"])
requests = _matching_requests(
db,
tenant_id=tenant_id,
request_ids=request_ids,
organizer_ids=membership_ids,
)
requests_by_id = {row.id: row for row in requests}
request_ids = set(requests_by_id)
notifications = _matching_notifications(
db,
tenant_id=tenant_id,
request_ids=request_ids,
participant_ids=participant_ids,
respondent_ids=respondent_ids,
email=email,
)
notifications_by_participant: dict[str, list[SchedulingNotification]] = {}
for notification in notifications:
if notification.participant_id:
notifications_by_participant.setdefault(
notification.participant_id, []
).append(notification)
records: list[DsarRecordRef] = []
def append(record: DsarRecordRef) -> None:
if len(records) >= _MAX_RECORDS:
raise ValueError(
"Scheduling DSAR match limit exceeded; narrow the subject selectors."
)
records.append(record)
for request in requests:
participant_context = request.id in {row.request_id for row in participants}
organizer_context = bool(
request.organizer_user_id
and request.organizer_user_id in membership_ids
)
immutable = _request_is_evidence(request)
append(
_record(
"scheduling_request",
request.id,
"scheduling_request",
request.title,
{
"match_fields": (
["organizer_user_id"] if organizer_context else []
),
"participant_context": participant_context,
"title": request.title,
"description": request.description,
"location": request.location,
"timezone": request.timezone,
"status": request.status,
"deadline_at": _iso(request.deadline_at),
"allow_external_participants": request.allow_external_participants,
"allow_participant_updates": request.allow_participant_updates,
"result_visibility": request.result_visibility,
"participant_visibility": request.participant_visibility,
"notify_on_answers": request.notify_on_answers,
"single_choice": request.single_choice,
"max_participants_per_option": request.max_participants_per_option,
"allow_maybe": request.allow_maybe,
"allow_comments": request.allow_comments,
"participant_email_required": request.participant_email_required,
"calendar_integration_enabled": request.calendar_integration_enabled,
"calendar_freebusy_enabled": request.calendar_freebusy_enabled,
"calendar_hold_enabled": request.calendar_hold_enabled,
"create_calendar_event_on_decision": request.create_calendar_event_on_decision,
"handed_off_at": _iso(request.handed_off_at),
"cancelled_at": _iso(request.cancelled_at),
"deleted_at": _iso(request.deleted_at),
},
observed_at=request.updated_at,
immutable=immutable,
retention_reason=(
"Decided, handed-off, cancelled, archived, or deleted scheduling state is institutional decision and coordination evidence."
if immutable
else None
),
source_path=f"/scheduling?request={request.id}",
)
)
for slot in _candidate_slots(
db,
tenant_id=tenant_id,
request_ids=request_ids,
):
request = requests_by_id[slot.request_id]
immutable = _request_is_evidence(request) or slot.deleted_at is not None
append(
_record(
"scheduling_candidate_slot",
slot.id,
"scheduling_candidate_context",
slot.label,
{
"request_id": slot.request_id,
"label": slot.label,
"description": slot.description,
"start_at": _iso(slot.start_at),
"end_at": _iso(slot.end_at),
"timezone": slot.timezone,
"location": slot.location,
"position": slot.position,
"freebusy_checked_at": _iso(slot.freebusy_checked_at),
"freebusy_status": slot.freebusy_status,
"deleted_at": _iso(slot.deleted_at),
},
observed_at=slot.updated_at,
immutable=immutable,
retention_reason=(
"Candidate timing retained with terminal Scheduling decision evidence."
if immutable
else None
),
source_path=f"/scheduling?request={slot.request_id}",
)
)
for participant in participants:
request = requests_by_id.get(participant.request_id)
if request is None:
continue
matching_fields = _participant_matching_fields(
participant,
participant_id=references.get("participant"),
respondent_ids=respondent_ids,
account_ids=account_ids,
email=email,
)
immutable = _request_is_evidence(request) or _participant_is_evidence(
participant,
notifications_by_participant.get(participant.id, ()),
)
erasable_without_coordination = not immutable and _participant_is_unengaged(
participant,
notifications_by_participant.get(participant.id, ()),
)
append(
_record(
"scheduling_participant",
participant.id,
"scheduling_participation",
participant.display_name or "Scheduling participant",
{
"match_fields": matching_fields,
"request_id": participant.request_id,
"display_name": participant.display_name,
"email": _normalized_email(participant.email),
"participant_type": participant.participant_type,
"required": participant.required,
"status": participant.status,
"account_bound_at": _iso(participant.account_bound_at),
"last_invited_at": _iso(participant.last_invited_at),
"responded_at": _iso(participant.responded_at),
"response_comment": _bounded_text(participant.response_comment),
"deleted_at": _iso(participant.deleted_at),
"erasable_without_coordination": erasable_without_coordination,
},
observed_at=participant.updated_at,
immutable=immutable,
retention_reason=(
"Responded, notified, removed, or terminal participant state is retained with Poll and Scheduling decision evidence."
if immutable
else None
),
source_path=f"/scheduling?request={participant.request_id}",
)
)
for notification in notifications:
immutable = bool(
notification.sent_at
or notification.status in _TERMINAL_NOTIFICATION_STATUSES
)
append(
_record(
"scheduling_notification",
notification.id,
"scheduling_notification_evidence",
f"Scheduling {notification.event_kind} notification",
{
"match_fields": _notification_matching_fields(
notification,
participant_ids=participant_ids,
respondent_ids=respondent_ids,
email=email,
),
"request_id": notification.request_id,
"participant_id": (
notification.participant_id
if notification.participant_id in participant_ids
else None
),
"event_kind": notification.event_kind,
"channel": notification.channel,
"recipient": _matching_recipient(
notification.recipient,
respondent_ids=respondent_ids,
email=email,
),
"status": notification.status,
"sent_at": _iso(notification.sent_at),
},
observed_at=notification.updated_at,
immutable=immutable,
retention_reason=(
"Terminal Scheduling notification state is delivery and participant-access evidence."
if immutable
else None
),
)
)
return tuple(records)
def plan_erasure(
self,
session: object,
*,
tenant_id: str,
subject: DsarSubjectRef,
records: Sequence[DsarRecordRef],
) -> Sequence[DsarErasureActionRef]:
del session
actions: list[DsarErasureActionRef] = []
for record in records:
if (
record.provider_id != self.provider_id
or record.module_id != self.module_id
):
raise ValueError("Scheduling DSAR received a foreign provider record.")
actions.append(
_action(
f"scheduling:{'retain' if record.immutable_evidence else 'review'}:{record.resource_type}:{record.resource_id}",
"retain" if record.immutable_evidence else "manual_review",
record,
f"{'Retain' if record.immutable_evidence else 'Review'} {record.title}",
record.retention_reason
or "Scheduling data can overlap Poll responses, Calendar effects, shared participants, and institutional decisions; review it through the owning lifecycle controls.",
executable=False,
)
)
if (
record.resource_type == "scheduling_participant"
and record.data.get("erasable_without_coordination") is True
and record.data.get("match_fields")
):
actions.append(
_action(
f"scheduling:anonymize:scheduling_participant:{record.resource_id}",
"anonymize",
record,
"Anonymize unengaged Scheduling participant",
"A participant without invitation, response, self-enrollment, notification, or terminal decision evidence can be removed without changing Poll or Calendar state.",
executable=True,
irreversible=True,
metadata={"tenant_id": tenant_id},
)
)
action_ids = [action.action_id for action in actions]
if len(action_ids) != len(set(action_ids)):
raise ValueError("Scheduling DSAR produced duplicate action ids.")
return tuple(actions)
def execute_erasure(
self,
session: object,
*,
tenant_id: str,
subject: DsarSubjectRef,
actions: Sequence[DsarErasureActionRef],
request_id: str,
) -> Sequence[DsarExecutionResultRef]:
db = _session(session)
results: list[DsarExecutionResultRef] = []
for action in actions:
if (
action.provider_id != self.provider_id
or action.module_id != self.module_id
or action.metadata.get("tenant_id") != tenant_id
or not action.action_id.startswith(
"scheduling:anonymize:scheduling_participant:"
)
):
results.append(
_blocked(action, "The Scheduling DSAR action is stale or invalid.")
)
continue
results.append(
_anonymize_unengaged_participant(
db,
tenant_id=tenant_id,
subject=subject,
action=action,
request_id=request_id,
)
)
db.flush()
return tuple(results)
def _matching_participants(
session: Session,
*,
tenant_id: str,
participant_id: str | None,
respondent_ids: set[str],
account_ids: set[str],
email: str | None,
) -> list[SchedulingParticipant]:
conditions = []
if participant_id:
conditions.append(SchedulingParticipant.id == participant_id)
if respondent_ids:
conditions.append(SchedulingParticipant.respondent_id.in_(respondent_ids))
if account_ids:
conditions.append(SchedulingParticipant.bound_account_id.in_(account_ids))
if email:
conditions.append(func.lower(SchedulingParticipant.email) == email)
if not conditions:
return []
candidates = _bounded_rows(
session.query(SchedulingParticipant)
.filter(SchedulingParticipant.tenant_id == tenant_id, or_(*conditions))
.order_by(SchedulingParticipant.id)
)
return [
row
for row in candidates
if _participant_matching_fields(
row,
participant_id=participant_id,
respondent_ids=respondent_ids,
account_ids=account_ids,
email=email,
)
]
def _matching_requests(
session: Session,
*,
tenant_id: str,
request_ids: set[str],
organizer_ids: set[str],
) -> list[SchedulingRequest]:
conditions = []
if request_ids:
conditions.append(SchedulingRequest.id.in_(request_ids))
if organizer_ids:
conditions.append(SchedulingRequest.organizer_user_id.in_(organizer_ids))
if not conditions:
return []
return _bounded_rows(
session.query(SchedulingRequest)
.filter(SchedulingRequest.tenant_id == tenant_id, or_(*conditions))
.order_by(SchedulingRequest.id)
)
def _candidate_slots(
session: Session,
*,
tenant_id: str,
request_ids: set[str],
) -> list[SchedulingCandidateSlot]:
if not request_ids:
return []
return _bounded_rows(
session.query(SchedulingCandidateSlot)
.filter(
SchedulingCandidateSlot.tenant_id == tenant_id,
SchedulingCandidateSlot.request_id.in_(request_ids),
)
.order_by(SchedulingCandidateSlot.request_id, SchedulingCandidateSlot.position)
)
def _matching_notifications(
session: Session,
*,
tenant_id: str,
request_ids: set[str],
participant_ids: set[str],
respondent_ids: set[str],
email: str | None,
) -> list[SchedulingNotification]:
if not request_ids:
return []
conditions = []
if participant_ids:
conditions.append(SchedulingNotification.participant_id.in_(participant_ids))
if respondent_ids:
conditions.append(SchedulingNotification.recipient.in_(respondent_ids))
if email:
conditions.append(func.lower(SchedulingNotification.recipient) == email)
if not conditions:
return []
candidates = _bounded_rows(
session.query(SchedulingNotification)
.filter(
SchedulingNotification.tenant_id == tenant_id,
SchedulingNotification.request_id.in_(request_ids),
or_(*conditions),
)
.order_by(SchedulingNotification.id)
)
return [
row
for row in candidates
if _notification_matching_fields(
row,
participant_ids=participant_ids,
respondent_ids=respondent_ids,
email=email,
)
]
def _participant_matching_fields(
row: SchedulingParticipant,
*,
participant_id: str | None,
respondent_ids: set[str],
account_ids: set[str],
email: str | None,
) -> list[str]:
fields = []
if participant_id and row.id == participant_id:
fields.append("id")
if row.respondent_id and row.respondent_id in respondent_ids:
fields.append("respondent_id")
if row.bound_account_id and row.bound_account_id in account_ids:
fields.append("bound_account_id")
if email and _normalized_email(row.email) == email:
fields.append("email")
return fields
def _notification_matching_fields(
row: SchedulingNotification,
*,
participant_ids: set[str],
respondent_ids: set[str],
email: str | None,
) -> list[str]:
fields = []
if row.participant_id and row.participant_id in participant_ids:
fields.append("participant_id")
recipient = str(row.recipient or "").strip()
if recipient in respondent_ids:
fields.append("recipient")
if email and _normalized_email(recipient) == email:
fields.append("recipient")
return list(dict.fromkeys(fields))
def _matching_recipient(
value: object,
*,
respondent_ids: set[str],
email: str | None,
) -> str | None:
candidate = str(value or "").strip()
if candidate in respondent_ids:
return candidate
if email and _normalized_email(candidate) == email:
return email
return None
def _request_is_evidence(row: SchedulingRequest) -> bool:
return bool(
row.status in _TERMINAL_REQUEST_STATUSES
or row.handed_off_at
or row.cancelled_at
or row.deleted_at
)
def _participant_is_evidence(
row: SchedulingParticipant,
notifications: Sequence[SchedulingNotification],
) -> bool:
return bool(
row.poll_invitation_id
or row.responded_at
or row.response_comment
or row.self_enrollment_link_id
or row.self_enrollment_proof_hash
or row.status in {"responded", "removed"}
or notifications
)
def _participant_is_unengaged(
row: SchedulingParticipant,
notifications: Sequence[SchedulingNotification],
) -> bool:
return bool(
row.deleted_at is None
and row.status in {"draft", "invited"}
and row.poll_invitation_id is None
and row.responded_at is None
and not row.response_comment
and row.self_enrollment_link_id is None
and row.self_enrollment_proof_hash is None
and not notifications
)
def _anonymize_unengaged_participant(
session: Session,
*,
tenant_id: str,
subject: DsarSubjectRef,
action: DsarErasureActionRef,
request_id: str,
) -> DsarExecutionResultRef:
row = (
session.query(SchedulingParticipant)
.filter(
SchedulingParticipant.id == action.resource_id,
SchedulingParticipant.tenant_id == tenant_id,
)
.with_for_update()
.one_or_none()
)
if row is None:
return _result(
action,
"unchanged",
"The Scheduling participant was already absent.",
{"request_id": request_id},
)
if _participant_is_anonymized(row):
return _result(
action,
"unchanged",
"The Scheduling participant was already anonymized.",
{"request_id": request_id},
)
if not _participant_matches_subject(row, subject):
return _blocked(
action, "The Scheduling participant identity changed after planning."
)
request = (
session.query(SchedulingRequest)
.filter(
SchedulingRequest.id == row.request_id,
SchedulingRequest.tenant_id == tenant_id,
)
.with_for_update()
.one_or_none()
)
if request is None or _request_is_evidence(request):
return _blocked(
action,
"The Scheduling request is absent or became retained decision evidence.",
)
notifications = _participant_notifications(session, row)
if not _participant_is_unengaged(row, notifications):
return _blocked(
action,
"The participant gained invitation, response, enrollment, or notification evidence after planning.",
)
now = datetime.now(timezone.utc)
row.respondent_id = None
row.display_name = None
row.email = None
row.status = "removed"
row.poll_invitation_id = None
row.participation_gateway = None
row.self_enrollment_link_id = None
row.self_enrollment_proof_hash = None
row.bound_account_id = None
row.account_bound_at = None
row.last_invited_at = None
row.responded_at = None
row.response_comment = None
row.deleted_at = now
row.metadata_ = None
return _result(
action,
"executed",
"The unengaged Scheduling participant was anonymized and retired.",
{"request_id": request_id},
)
def _participant_notifications(
session: Session,
row: SchedulingParticipant,
) -> list[SchedulingNotification]:
conditions = [SchedulingNotification.participant_id == row.id]
email = _normalized_email(row.email)
if email:
conditions.append(func.lower(SchedulingNotification.recipient) == email)
return _bounded_rows(
session.query(SchedulingNotification)
.filter(
SchedulingNotification.tenant_id == row.tenant_id,
SchedulingNotification.request_id == row.request_id,
or_(*conditions),
)
.order_by(SchedulingNotification.id)
)
def _participant_matches_subject(
row: SchedulingParticipant,
subject: DsarSubjectRef,
) -> bool:
return bool(
_participant_matching_fields(
row,
participant_id=_scheduling_references(subject).get("participant"),
respondent_ids=_respondent_ids(subject),
account_ids=_account_ids(subject),
email=_subject_email(subject),
)
)
def _participant_is_anonymized(row: SchedulingParticipant) -> bool:
return bool(
row.deleted_at is not None
and row.status == "removed"
and row.respondent_id is None
and row.display_name is None
and row.email is None
and row.poll_invitation_id is None
and row.self_enrollment_link_id is None
and row.self_enrollment_proof_hash is None
and row.bound_account_id is None
and row.response_comment is None
and row.metadata_ is None
)
def _scheduling_references(subject: DsarSubjectRef) -> dict[str, str]:
aliases = {
"scheduling.request": "request",
"scheduling.request_id": "request",
"scheduling.participant": "participant",
"scheduling.participant_id": "participant",
}
references: dict[str, str] = {}
for key, target in aliases.items():
value = str(subject.external_references.get(key) or "").strip()
if value and target not in references:
references[target] = value
return references
def _membership_ids(subject: DsarSubjectRef) -> set[str]:
values = [subject.membership_id]
values.extend(
subject.external_references.get(key)
for key in (
"scheduling.user",
"scheduling.membership",
"access.membership",
"membership_id",
)
)
return _identifiers(values)
def _respondent_ids(subject: DsarSubjectRef) -> set[str]:
values = [subject.membership_id, subject.identity_id, subject.account_id]
values.extend(
subject.external_references.get(key)
for key in (
"scheduling.respondent",
"poll.respondent",
"access.membership",
"membership_id",
"identity_id",
"account_id",
)
)
return _identifiers(values)
def _account_ids(subject: DsarSubjectRef) -> set[str]:
values = [subject.account_id]
values.extend(
subject.external_references.get(key)
for key in ("scheduling.account", "access.account", "account_id")
)
return _identifiers(values)
def _identifiers(values: Sequence[object]) -> set[str]:
return {text for value in values if (text := str(value or "").strip())}
def _subject_email(subject: DsarSubjectRef) -> str | None:
candidates = [subject.email]
candidates.extend(
subject.external_references.get(key)
for key in ("scheduling.email", "scheduling.participant_email")
)
normalized = {
email for value in candidates if (email := _normalized_email(value)) is not None
}
return normalized.pop() if len(normalized) == 1 else None
def _normalized_email(value: object) -> str | None:
if not isinstance(value, str):
return None
normalized = value.strip().casefold()
return normalized or None
def _bounded_text(value: str | None) -> str | None:
return value[:2_000] if value else None
def _record(
resource_type: str,
resource_id: str,
category: str,
title: str,
data: Mapping[str, object],
*,
observed_at: datetime | None = None,
immutable: bool = False,
retention_reason: str | None = None,
source_path: str | None = None,
) -> DsarRecordRef:
return DsarRecordRef(
provider_id="scheduling",
module_id="scheduling",
resource_type=resource_type,
resource_id=resource_id,
category=category,
title=title,
data=data,
observed_at=observed_at,
immutable_evidence=immutable,
retention_reason=retention_reason,
source_path=source_path,
)
def _action(
action_id: str,
kind: str,
record: DsarRecordRef,
title: str,
rationale: str,
*,
executable: bool,
irreversible: bool = False,
metadata: Mapping[str, object] | None = None,
) -> DsarErasureActionRef:
return DsarErasureActionRef(
action_id=action_id,
provider_id="scheduling",
module_id="scheduling",
kind=kind, # type: ignore[arg-type]
resource_type=record.resource_type,
resource_id=record.resource_id,
title=title,
rationale=rationale,
executable=executable,
irreversible=irreversible,
metadata=metadata or {},
)
def _result(
action: DsarErasureActionRef,
status: str,
summary: str,
evidence: Mapping[str, object] | None = None,
) -> DsarExecutionResultRef:
return DsarExecutionResultRef(
action_id=action.action_id,
status=status, # type: ignore[arg-type]
summary=summary,
evidence=evidence or {},
)
def _blocked(action: DsarErasureActionRef, summary: str) -> DsarExecutionResultRef:
return _result(action, "blocked", summary)
def _session(value: object) -> Session:
if not isinstance(value, Session):
raise TypeError("Scheduling DSAR provider requires a SQLAlchemy session.")
return value
def _bounded_rows(query: object) -> list[object]:
rows = query.limit(_MAX_RECORDS + 1).all() # type: ignore[attr-defined]
if len(rows) > _MAX_RECORDS:
raise ValueError(
"Scheduling DSAR match limit exceeded; narrow the subject selectors."
)
return rows
def _iso(value: datetime | None) -> str | None:
if value is None:
return None
if value.tzinfo is None:
value = value.replace(tzinfo=timezone.utc)
return value.isoformat()
__all__ = ["SCHEDULING_DSAR_CAPABILITY", "SchedulingDsarProvider"]