zulip/zerver/lib/presence.py

215 lines
7.4 KiB
Python
Raw Normal View History

import datetime
import time
from collections import defaultdict
from typing import Any, Dict, Mapping, Optional, Sequence, Set
presence: Rewrite the backend data model. This implements the core of the rewrite described in: For the backend data model for UserPresence to one that supports much more efficient queries and is more correct around handling of multiple clients. The main loss of functionality is that we no longer track which Client sent presence data (so we will no longer be able to say using UserPresence "the user was last online on their desktop 15 minutes ago, but was online with their phone 3 minutes ago"). If we consider that information important for the occasional investigation query, we have can construct that answer data via UserActivity already. It's not worth making Presence much more expensive/complex to support it. For slim_presence clients, this sends the same data format we sent before, albeit with less complexity involved in constructing it. Note that we at present will always send both last_active_time and last_connected_time; we may revisit that in the future. This commit doesn't include the finalizing migration, which drops the UserPresenceOld table. The way to deploy is to start the backfill migration with the server down and then start the server *without* the user_presence queue worker, to let the migration finish without having new data interfering with it. Once the migration is done, the queue worker can be started, leading to the presence data catching up to the current state as the queue worker goes over the queued up events and updating the UserPresence table. Co-authored-by: Mateusz Mandera <mateusz.mandera@zulip.com>
2020-06-11 16:03:47 +02:00
from django.conf import settings
from django.utils.timezone import now as timezone_now
from zerver.lib.timestamp import datetime_to_timestamp
from zerver.models import PushDeviceToken, Realm, UserPresence, UserProfile, query_for_ids
def get_presence_dicts_for_rows(
all_rows: Sequence[Mapping[str, Any]], mobile_user_ids: Set[int], slim_presence: bool
) -> Dict[str, Dict[str, Any]]:
if slim_presence:
# Stringify user_id here, since it's gonna be turned
# into a string anyway by JSON, and it keeps mypy happy.
get_user_key = lambda row: str(row["user_profile_id"])
get_user_presence_info = get_modern_user_presence_info
else:
get_user_key = lambda row: row["user_profile__email"]
get_user_presence_info = get_legacy_user_presence_info
user_statuses: Dict[str, Dict[str, Any]] = {}
presence: Rewrite the backend data model. This implements the core of the rewrite described in: For the backend data model for UserPresence to one that supports much more efficient queries and is more correct around handling of multiple clients. The main loss of functionality is that we no longer track which Client sent presence data (so we will no longer be able to say using UserPresence "the user was last online on their desktop 15 minutes ago, but was online with their phone 3 minutes ago"). If we consider that information important for the occasional investigation query, we have can construct that answer data via UserActivity already. It's not worth making Presence much more expensive/complex to support it. For slim_presence clients, this sends the same data format we sent before, albeit with less complexity involved in constructing it. Note that we at present will always send both last_active_time and last_connected_time; we may revisit that in the future. This commit doesn't include the finalizing migration, which drops the UserPresenceOld table. The way to deploy is to start the backfill migration with the server down and then start the server *without* the user_presence queue worker, to let the migration finish without having new data interfering with it. Once the migration is done, the queue worker can be started, leading to the presence data catching up to the current state as the queue worker goes over the queued up events and updating the UserPresence table. Co-authored-by: Mateusz Mandera <mateusz.mandera@zulip.com>
2020-06-11 16:03:47 +02:00
for presence_row in all_rows:
user_key = get_user_key(presence_row)
last_active_time = user_presence_datetime_with_date_joined_default(
presence_row["last_active_time"], presence_row["user_profile__date_joined"]
)
last_connected_time = user_presence_datetime_with_date_joined_default(
presence_row["last_connected_time"], presence_row["user_profile__date_joined"]
)
info = get_user_presence_info(
last_active_time,
last_connected_time,
)
user_statuses[user_key] = info
return user_statuses
def user_presence_datetime_with_date_joined_default(
dt: Optional[datetime.datetime], date_joined: datetime.datetime
) -> datetime.datetime:
"""
Our data models support UserPresence objects not having None
values for last_active_time/last_connected_time. The legacy API
however has always sent timestamps, so for backward
compatibility we cannot send such values through the API and need
to default to a sane datetime.
This helper functions expects to take a last_active_time or
last_connected_time value and the date_joined of the user, which
will serve as the default value if the first argument is None.
"""
if dt is None:
return date_joined
return dt
def get_modern_user_presence_info(
presence: Rewrite the backend data model. This implements the core of the rewrite described in: For the backend data model for UserPresence to one that supports much more efficient queries and is more correct around handling of multiple clients. The main loss of functionality is that we no longer track which Client sent presence data (so we will no longer be able to say using UserPresence "the user was last online on their desktop 15 minutes ago, but was online with their phone 3 minutes ago"). If we consider that information important for the occasional investigation query, we have can construct that answer data via UserActivity already. It's not worth making Presence much more expensive/complex to support it. For slim_presence clients, this sends the same data format we sent before, albeit with less complexity involved in constructing it. Note that we at present will always send both last_active_time and last_connected_time; we may revisit that in the future. This commit doesn't include the finalizing migration, which drops the UserPresenceOld table. The way to deploy is to start the backfill migration with the server down and then start the server *without* the user_presence queue worker, to let the migration finish without having new data interfering with it. Once the migration is done, the queue worker can be started, leading to the presence data catching up to the current state as the queue worker goes over the queued up events and updating the UserPresence table. Co-authored-by: Mateusz Mandera <mateusz.mandera@zulip.com>
2020-06-11 16:03:47 +02:00
last_active_time: datetime.datetime, last_connected_time: datetime.datetime
) -> Dict[str, Any]:
presence: Rewrite the backend data model. This implements the core of the rewrite described in: For the backend data model for UserPresence to one that supports much more efficient queries and is more correct around handling of multiple clients. The main loss of functionality is that we no longer track which Client sent presence data (so we will no longer be able to say using UserPresence "the user was last online on their desktop 15 minutes ago, but was online with their phone 3 minutes ago"). If we consider that information important for the occasional investigation query, we have can construct that answer data via UserActivity already. It's not worth making Presence much more expensive/complex to support it. For slim_presence clients, this sends the same data format we sent before, albeit with less complexity involved in constructing it. Note that we at present will always send both last_active_time and last_connected_time; we may revisit that in the future. This commit doesn't include the finalizing migration, which drops the UserPresenceOld table. The way to deploy is to start the backfill migration with the server down and then start the server *without* the user_presence queue worker, to let the migration finish without having new data interfering with it. Once the migration is done, the queue worker can be started, leading to the presence data catching up to the current state as the queue worker goes over the queued up events and updating the UserPresence table. Co-authored-by: Mateusz Mandera <mateusz.mandera@zulip.com>
2020-06-11 16:03:47 +02:00
# TODO: Do further bandwidth optimizations to this structure.
result = {}
presence: Rewrite the backend data model. This implements the core of the rewrite described in: For the backend data model for UserPresence to one that supports much more efficient queries and is more correct around handling of multiple clients. The main loss of functionality is that we no longer track which Client sent presence data (so we will no longer be able to say using UserPresence "the user was last online on their desktop 15 minutes ago, but was online with their phone 3 minutes ago"). If we consider that information important for the occasional investigation query, we have can construct that answer data via UserActivity already. It's not worth making Presence much more expensive/complex to support it. For slim_presence clients, this sends the same data format we sent before, albeit with less complexity involved in constructing it. Note that we at present will always send both last_active_time and last_connected_time; we may revisit that in the future. This commit doesn't include the finalizing migration, which drops the UserPresenceOld table. The way to deploy is to start the backfill migration with the server down and then start the server *without* the user_presence queue worker, to let the migration finish without having new data interfering with it. Once the migration is done, the queue worker can be started, leading to the presence data catching up to the current state as the queue worker goes over the queued up events and updating the UserPresence table. Co-authored-by: Mateusz Mandera <mateusz.mandera@zulip.com>
2020-06-11 16:03:47 +02:00
result["active_timestamp"] = datetime_to_timestamp(last_active_time)
result["idle_timestamp"] = datetime_to_timestamp(last_connected_time)
return result
def get_legacy_user_presence_info(
presence: Rewrite the backend data model. This implements the core of the rewrite described in: For the backend data model for UserPresence to one that supports much more efficient queries and is more correct around handling of multiple clients. The main loss of functionality is that we no longer track which Client sent presence data (so we will no longer be able to say using UserPresence "the user was last online on their desktop 15 minutes ago, but was online with their phone 3 minutes ago"). If we consider that information important for the occasional investigation query, we have can construct that answer data via UserActivity already. It's not worth making Presence much more expensive/complex to support it. For slim_presence clients, this sends the same data format we sent before, albeit with less complexity involved in constructing it. Note that we at present will always send both last_active_time and last_connected_time; we may revisit that in the future. This commit doesn't include the finalizing migration, which drops the UserPresenceOld table. The way to deploy is to start the backfill migration with the server down and then start the server *without* the user_presence queue worker, to let the migration finish without having new data interfering with it. Once the migration is done, the queue worker can be started, leading to the presence data catching up to the current state as the queue worker goes over the queued up events and updating the UserPresence table. Co-authored-by: Mateusz Mandera <mateusz.mandera@zulip.com>
2020-06-11 16:03:47 +02:00
last_active_time: datetime.datetime, last_connected_time: datetime.datetime
) -> Dict[str, Any]:
"""
Reformats the modern UserPresence data structure so that legacy
API clients can still access presence data.
We expect this code to remain mostly unchanged until we can delete it.
"""
presence: Rewrite the backend data model. This implements the core of the rewrite described in: For the backend data model for UserPresence to one that supports much more efficient queries and is more correct around handling of multiple clients. The main loss of functionality is that we no longer track which Client sent presence data (so we will no longer be able to say using UserPresence "the user was last online on their desktop 15 minutes ago, but was online with their phone 3 minutes ago"). If we consider that information important for the occasional investigation query, we have can construct that answer data via UserActivity already. It's not worth making Presence much more expensive/complex to support it. For slim_presence clients, this sends the same data format we sent before, albeit with less complexity involved in constructing it. Note that we at present will always send both last_active_time and last_connected_time; we may revisit that in the future. This commit doesn't include the finalizing migration, which drops the UserPresenceOld table. The way to deploy is to start the backfill migration with the server down and then start the server *without* the user_presence queue worker, to let the migration finish without having new data interfering with it. Once the migration is done, the queue worker can be started, leading to the presence data catching up to the current state as the queue worker goes over the queued up events and updating the UserPresence table. Co-authored-by: Mateusz Mandera <mateusz.mandera@zulip.com>
2020-06-11 16:03:47 +02:00
# Now we put things together in the legacy presence format with
# one client + an `aggregated` field.
#
# TODO: Look at whether we can drop to just the "aggregated" field
# if no clients look at the rest.
most_recent_info = format_legacy_presence_dict(last_active_time, last_connected_time)
result = {}
# The word "aggregated" here is possibly misleading.
# It's really just the most recent client's info.
result["aggregated"] = dict(
client=most_recent_info["client"],
status=most_recent_info["status"],
timestamp=most_recent_info["timestamp"],
)
result["website"] = most_recent_info
return result
def format_legacy_presence_dict(
last_active_time: datetime.datetime, last_connected_time: datetime.datetime
) -> Dict[str, Any]:
presence: Rewrite the backend data model. This implements the core of the rewrite described in: For the backend data model for UserPresence to one that supports much more efficient queries and is more correct around handling of multiple clients. The main loss of functionality is that we no longer track which Client sent presence data (so we will no longer be able to say using UserPresence "the user was last online on their desktop 15 minutes ago, but was online with their phone 3 minutes ago"). If we consider that information important for the occasional investigation query, we have can construct that answer data via UserActivity already. It's not worth making Presence much more expensive/complex to support it. For slim_presence clients, this sends the same data format we sent before, albeit with less complexity involved in constructing it. Note that we at present will always send both last_active_time and last_connected_time; we may revisit that in the future. This commit doesn't include the finalizing migration, which drops the UserPresenceOld table. The way to deploy is to start the backfill migration with the server down and then start the server *without* the user_presence queue worker, to let the migration finish without having new data interfering with it. Once the migration is done, the queue worker can be started, leading to the presence data catching up to the current state as the queue worker goes over the queued up events and updating the UserPresence table. Co-authored-by: Mateusz Mandera <mateusz.mandera@zulip.com>
2020-06-11 16:03:47 +02:00
"""
This function assumes it's being called right after the presence object was updated,
and is not meant to be used on old presence data.
"""
if (
last_active_time
presence: Rewrite the backend data model. This implements the core of the rewrite described in: For the backend data model for UserPresence to one that supports much more efficient queries and is more correct around handling of multiple clients. The main loss of functionality is that we no longer track which Client sent presence data (so we will no longer be able to say using UserPresence "the user was last online on their desktop 15 minutes ago, but was online with their phone 3 minutes ago"). If we consider that information important for the occasional investigation query, we have can construct that answer data via UserActivity already. It's not worth making Presence much more expensive/complex to support it. For slim_presence clients, this sends the same data format we sent before, albeit with less complexity involved in constructing it. Note that we at present will always send both last_active_time and last_connected_time; we may revisit that in the future. This commit doesn't include the finalizing migration, which drops the UserPresenceOld table. The way to deploy is to start the backfill migration with the server down and then start the server *without* the user_presence queue worker, to let the migration finish without having new data interfering with it. Once the migration is done, the queue worker can be started, leading to the presence data catching up to the current state as the queue worker goes over the queued up events and updating the UserPresence table. Co-authored-by: Mateusz Mandera <mateusz.mandera@zulip.com>
2020-06-11 16:03:47 +02:00
+ datetime.timedelta(seconds=settings.PRESENCE_LEGACY_EVENT_OFFSET_FOR_ACTIVITY_SECONDS)
>= last_connected_time
presence: Rewrite the backend data model. This implements the core of the rewrite described in: For the backend data model for UserPresence to one that supports much more efficient queries and is more correct around handling of multiple clients. The main loss of functionality is that we no longer track which Client sent presence data (so we will no longer be able to say using UserPresence "the user was last online on their desktop 15 minutes ago, but was online with their phone 3 minutes ago"). If we consider that information important for the occasional investigation query, we have can construct that answer data via UserActivity already. It's not worth making Presence much more expensive/complex to support it. For slim_presence clients, this sends the same data format we sent before, albeit with less complexity involved in constructing it. Note that we at present will always send both last_active_time and last_connected_time; we may revisit that in the future. This commit doesn't include the finalizing migration, which drops the UserPresenceOld table. The way to deploy is to start the backfill migration with the server down and then start the server *without* the user_presence queue worker, to let the migration finish without having new data interfering with it. Once the migration is done, the queue worker can be started, leading to the presence data catching up to the current state as the queue worker goes over the queued up events and updating the UserPresence table. Co-authored-by: Mateusz Mandera <mateusz.mandera@zulip.com>
2020-06-11 16:03:47 +02:00
):
status = UserPresence.LEGACY_STATUS_ACTIVE
timestamp = datetime_to_timestamp(last_active_time)
presence: Rewrite the backend data model. This implements the core of the rewrite described in: For the backend data model for UserPresence to one that supports much more efficient queries and is more correct around handling of multiple clients. The main loss of functionality is that we no longer track which Client sent presence data (so we will no longer be able to say using UserPresence "the user was last online on their desktop 15 minutes ago, but was online with their phone 3 minutes ago"). If we consider that information important for the occasional investigation query, we have can construct that answer data via UserActivity already. It's not worth making Presence much more expensive/complex to support it. For slim_presence clients, this sends the same data format we sent before, albeit with less complexity involved in constructing it. Note that we at present will always send both last_active_time and last_connected_time; we may revisit that in the future. This commit doesn't include the finalizing migration, which drops the UserPresenceOld table. The way to deploy is to start the backfill migration with the server down and then start the server *without* the user_presence queue worker, to let the migration finish without having new data interfering with it. Once the migration is done, the queue worker can be started, leading to the presence data catching up to the current state as the queue worker goes over the queued up events and updating the UserPresence table. Co-authored-by: Mateusz Mandera <mateusz.mandera@zulip.com>
2020-06-11 16:03:47 +02:00
else:
status = UserPresence.LEGACY_STATUS_IDLE
timestamp = datetime_to_timestamp(last_connected_time)
# This field was never used by clients of the legacy API, so we
# just set it to a fixed value for API format compatibility.
pushable = False
presence: Rewrite the backend data model. This implements the core of the rewrite described in: For the backend data model for UserPresence to one that supports much more efficient queries and is more correct around handling of multiple clients. The main loss of functionality is that we no longer track which Client sent presence data (so we will no longer be able to say using UserPresence "the user was last online on their desktop 15 minutes ago, but was online with their phone 3 minutes ago"). If we consider that information important for the occasional investigation query, we have can construct that answer data via UserActivity already. It's not worth making Presence much more expensive/complex to support it. For slim_presence clients, this sends the same data format we sent before, albeit with less complexity involved in constructing it. Note that we at present will always send both last_active_time and last_connected_time; we may revisit that in the future. This commit doesn't include the finalizing migration, which drops the UserPresenceOld table. The way to deploy is to start the backfill migration with the server down and then start the server *without* the user_presence queue worker, to let the migration finish without having new data interfering with it. Once the migration is done, the queue worker can be started, leading to the presence data catching up to the current state as the queue worker goes over the queued up events and updating the UserPresence table. Co-authored-by: Mateusz Mandera <mateusz.mandera@zulip.com>
2020-06-11 16:03:47 +02:00
return dict(client="website", status=status, timestamp=timestamp, pushable=pushable)
presence: Rewrite the backend data model. This implements the core of the rewrite described in: For the backend data model for UserPresence to one that supports much more efficient queries and is more correct around handling of multiple clients. The main loss of functionality is that we no longer track which Client sent presence data (so we will no longer be able to say using UserPresence "the user was last online on their desktop 15 minutes ago, but was online with their phone 3 minutes ago"). If we consider that information important for the occasional investigation query, we have can construct that answer data via UserActivity already. It's not worth making Presence much more expensive/complex to support it. For slim_presence clients, this sends the same data format we sent before, albeit with less complexity involved in constructing it. Note that we at present will always send both last_active_time and last_connected_time; we may revisit that in the future. This commit doesn't include the finalizing migration, which drops the UserPresenceOld table. The way to deploy is to start the backfill migration with the server down and then start the server *without* the user_presence queue worker, to let the migration finish without having new data interfering with it. Once the migration is done, the queue worker can be started, leading to the presence data catching up to the current state as the queue worker goes over the queued up events and updating the UserPresence table. Co-authored-by: Mateusz Mandera <mateusz.mandera@zulip.com>
2020-06-11 16:03:47 +02:00
def get_presence_for_user(
user_profile_id: int, slim_presence: bool = False
) -> Dict[str, Dict[str, Any]]:
query = UserPresence.objects.filter(user_profile_id=user_profile_id).values(
presence: Rewrite the backend data model. This implements the core of the rewrite described in: For the backend data model for UserPresence to one that supports much more efficient queries and is more correct around handling of multiple clients. The main loss of functionality is that we no longer track which Client sent presence data (so we will no longer be able to say using UserPresence "the user was last online on their desktop 15 minutes ago, but was online with their phone 3 minutes ago"). If we consider that information important for the occasional investigation query, we have can construct that answer data via UserActivity already. It's not worth making Presence much more expensive/complex to support it. For slim_presence clients, this sends the same data format we sent before, albeit with less complexity involved in constructing it. Note that we at present will always send both last_active_time and last_connected_time; we may revisit that in the future. This commit doesn't include the finalizing migration, which drops the UserPresenceOld table. The way to deploy is to start the backfill migration with the server down and then start the server *without* the user_presence queue worker, to let the migration finish without having new data interfering with it. Once the migration is done, the queue worker can be started, leading to the presence data catching up to the current state as the queue worker goes over the queued up events and updating the UserPresence table. Co-authored-by: Mateusz Mandera <mateusz.mandera@zulip.com>
2020-06-11 16:03:47 +02:00
"last_active_time",
"last_connected_time",
"user_profile__email",
"user_profile_id",
"user_profile__enable_offline_push_notifications",
"user_profile__date_joined",
)
presence_rows = list(query)
python: Convert assignment type annotations to Python 3.6 style. This commit was split by tabbott; this piece covers the vast majority of files in Zulip, but excludes scripts/, tools/, and puppet/ to help ensure we at least show the right error messages for Xenial systems. We can likely further refine the remaining pieces with some testing. Generated by com2ann, with whitespace fixes and various manual fixes for runtime issues: - invoiced_through: Optional[LicenseLedger] = models.ForeignKey( + invoiced_through: Optional["LicenseLedger"] = models.ForeignKey( -_apns_client: Optional[APNsClient] = None +_apns_client: Optional["APNsClient"] = None - notifications_stream: Optional[Stream] = models.ForeignKey('Stream', related_name='+', null=True, blank=True, on_delete=CASCADE) - signup_notifications_stream: Optional[Stream] = models.ForeignKey('Stream', related_name='+', null=True, blank=True, on_delete=CASCADE) + notifications_stream: Optional["Stream"] = models.ForeignKey('Stream', related_name='+', null=True, blank=True, on_delete=CASCADE) + signup_notifications_stream: Optional["Stream"] = models.ForeignKey('Stream', related_name='+', null=True, blank=True, on_delete=CASCADE) - author: Optional[UserProfile] = models.ForeignKey('UserProfile', blank=True, null=True, on_delete=CASCADE) + author: Optional["UserProfile"] = models.ForeignKey('UserProfile', blank=True, null=True, on_delete=CASCADE) - bot_owner: Optional[UserProfile] = models.ForeignKey('self', null=True, on_delete=models.SET_NULL) + bot_owner: Optional["UserProfile"] = models.ForeignKey('self', null=True, on_delete=models.SET_NULL) - default_sending_stream: Optional[Stream] = models.ForeignKey('zerver.Stream', null=True, related_name='+', on_delete=CASCADE) - default_events_register_stream: Optional[Stream] = models.ForeignKey('zerver.Stream', null=True, related_name='+', on_delete=CASCADE) + default_sending_stream: Optional["Stream"] = models.ForeignKey('zerver.Stream', null=True, related_name='+', on_delete=CASCADE) + default_events_register_stream: Optional["Stream"] = models.ForeignKey('zerver.Stream', null=True, related_name='+', on_delete=CASCADE) -descriptors_by_handler_id: Dict[int, ClientDescriptor] = {} +descriptors_by_handler_id: Dict[int, "ClientDescriptor"] = {} -worker_classes: Dict[str, Type[QueueProcessingWorker]] = {} -queues: Dict[str, Dict[str, Type[QueueProcessingWorker]]] = {} +worker_classes: Dict[str, Type["QueueProcessingWorker"]] = {} +queues: Dict[str, Dict[str, Type["QueueProcessingWorker"]]] = {} -AUTH_LDAP_REVERSE_EMAIL_SEARCH: Optional[LDAPSearch] = None +AUTH_LDAP_REVERSE_EMAIL_SEARCH: Optional["LDAPSearch"] = None Signed-off-by: Anders Kaseorg <anders@zulipchat.com>
2020-04-22 01:09:50 +02:00
mobile_user_ids: Set[int] = set()
if PushDeviceToken.objects.filter(user_id=user_profile_id).exists(): # nocoverage
# TODO: Add a test, though this is low priority, since we don't use mobile_user_ids yet.
mobile_user_ids.add(user_profile_id)
return get_presence_dicts_for_rows(presence_rows, mobile_user_ids, slim_presence)
def get_presence_dict_by_realm(
realm_id: int, slim_presence: bool = False
) -> Dict[str, Dict[str, Any]]:
two_weeks_ago = timezone_now() - datetime.timedelta(weeks=2)
query = UserPresence.objects.filter(
realm_id=realm_id,
presence: Rewrite the backend data model. This implements the core of the rewrite described in: For the backend data model for UserPresence to one that supports much more efficient queries and is more correct around handling of multiple clients. The main loss of functionality is that we no longer track which Client sent presence data (so we will no longer be able to say using UserPresence "the user was last online on their desktop 15 minutes ago, but was online with their phone 3 minutes ago"). If we consider that information important for the occasional investigation query, we have can construct that answer data via UserActivity already. It's not worth making Presence much more expensive/complex to support it. For slim_presence clients, this sends the same data format we sent before, albeit with less complexity involved in constructing it. Note that we at present will always send both last_active_time and last_connected_time; we may revisit that in the future. This commit doesn't include the finalizing migration, which drops the UserPresenceOld table. The way to deploy is to start the backfill migration with the server down and then start the server *without* the user_presence queue worker, to let the migration finish without having new data interfering with it. Once the migration is done, the queue worker can be started, leading to the presence data catching up to the current state as the queue worker goes over the queued up events and updating the UserPresence table. Co-authored-by: Mateusz Mandera <mateusz.mandera@zulip.com>
2020-06-11 16:03:47 +02:00
last_connected_time__gte=two_weeks_ago,
user_profile__is_active=True,
user_profile__is_bot=False,
).values(
presence: Rewrite the backend data model. This implements the core of the rewrite described in: For the backend data model for UserPresence to one that supports much more efficient queries and is more correct around handling of multiple clients. The main loss of functionality is that we no longer track which Client sent presence data (so we will no longer be able to say using UserPresence "the user was last online on their desktop 15 minutes ago, but was online with their phone 3 minutes ago"). If we consider that information important for the occasional investigation query, we have can construct that answer data via UserActivity already. It's not worth making Presence much more expensive/complex to support it. For slim_presence clients, this sends the same data format we sent before, albeit with less complexity involved in constructing it. Note that we at present will always send both last_active_time and last_connected_time; we may revisit that in the future. This commit doesn't include the finalizing migration, which drops the UserPresenceOld table. The way to deploy is to start the backfill migration with the server down and then start the server *without* the user_presence queue worker, to let the migration finish without having new data interfering with it. Once the migration is done, the queue worker can be started, leading to the presence data catching up to the current state as the queue worker goes over the queued up events and updating the UserPresence table. Co-authored-by: Mateusz Mandera <mateusz.mandera@zulip.com>
2020-06-11 16:03:47 +02:00
"last_active_time",
"last_connected_time",
"user_profile__email",
"user_profile_id",
"user_profile__enable_offline_push_notifications",
"user_profile__date_joined",
)
presence_rows = list(query)
mobile_query = PushDeviceToken.objects.distinct("user_id").values_list(
"user_id",
flat=True,
)
user_profile_ids = [presence_row["user_profile_id"] for presence_row in presence_rows]
if len(user_profile_ids) == 0:
# This conditional is necessary because query_for_ids
# throws an exception if passed an empty list.
#
# It's not clear this condition is actually possible,
# though, because it shouldn't be possible to end up with
# a realm with 0 active users.
return {}
mobile_query_ids = query_for_ids(
query=mobile_query,
user_ids=user_profile_ids,
field="user_id",
)
mobile_user_ids = set(mobile_query_ids)
return get_presence_dicts_for_rows(presence_rows, mobile_user_ids, slim_presence)
def get_presences_for_realm(
realm: Realm, slim_presence: bool
) -> Dict[str, Dict[str, Dict[str, Any]]]:
if realm.presence_disabled:
# Return an empty dict if presence is disabled in this realm
return defaultdict(dict)
return get_presence_dict_by_realm(realm.id, slim_presence)
def get_presence_response(
requesting_user_profile: UserProfile, slim_presence: bool
) -> Dict[str, Any]:
realm = requesting_user_profile.realm
server_timestamp = time.time()
presences = get_presences_for_realm(realm, slim_presence)
return dict(presences=presences, server_timestamp=server_timestamp)