Files
zulip/static/js/unread.js
N-Shar-ma 0e55b2aed9 unread: Show unread count for only 1:1 human PMs for right sidebar icon.
Uptil now, the right sidebar user list unread count included bot and group
PMs which sometimes resulted in the confusing state of the user list icon
indicating unread messages but on expanding the user list, no username
had a counter beside it, since the list only has individual human users.

Now this right sidebar unread count too excludes bots and groups, so the
unread count and the user list are consistent, without any ghost counts.
2023-02-07 15:38:32 -08:00

858 lines
27 KiB
JavaScript

import * as blueslip from "./blueslip";
import {FoldDict} from "./fold_dict";
import * as message_store from "./message_store";
import * as people from "./people";
import * as recent_topics_ui from "./recent_topics_ui";
import * as recent_topics_util from "./recent_topics_util";
import * as settings_config from "./settings_config";
import * as stream_data from "./stream_data";
import * as sub_store from "./sub_store";
import {user_settings} from "./user_settings";
import * as user_topics from "./user_topics";
import * as util from "./util";
// The unread module tracks the message IDs and locations of the
// user's unread messages. The tracking is initialized with
// server-provided data of the total set of unread messages in the
// user's history via page_params.unread_msgs (well, it cuts off at
// MAX_UNREAD_MESSAGES unreads for performance reasons). As a result,
// it can contain many thousands of messages that we don't have full
// data for in `message_store`, so we cannot in general look these
// messages up there.
// See https://zulip.readthedocs.io/en/latest/subsystems/pointer.html
// for more details on how this system is designed.
export let messages_read_in_narrow = false;
export function set_messages_read_in_narrow(value) {
messages_read_in_narrow = value;
}
export let old_unreads_missing = false;
export function clear_old_unreads_missing() {
old_unreads_missing = false;
}
export const unread_mentions_counter = new Set();
const unread_messages = new Set();
// Map with keys of the form "{stream_id}:{topic.toLowerCase()}" and
// values being Sets of message IDs for unread messages mentioning the
// user within that topic. Use `recent_topics_util.get_topic_key` to
// calculate keys.
//
// Functionally a cache; see clear_and_populate_unread_mention_topics
// for how we can refresh it efficiently.
export const unread_mention_topics = new Map();
class Bucketer {
// Maps item_id => bucket_key for items present in a bucket.
reverse_lookup = new Map();
constructor(options) {
this.key_to_bucket = new options.KeyDict();
this.make_bucket = options.make_bucket;
}
clear() {
this.key_to_bucket.clear();
this.reverse_lookup.clear();
}
add(opts) {
const bucket_key = opts.bucket_key;
const item_id = opts.item_id;
const add_callback = opts.add_callback;
let bucket = this.key_to_bucket.get(bucket_key);
if (!bucket) {
bucket = this.make_bucket();
this.key_to_bucket.set(bucket_key, bucket);
}
if (add_callback) {
add_callback(bucket, item_id);
} else {
bucket.add(item_id);
}
this.reverse_lookup.set(item_id, bucket_key);
}
delete(item_id) {
const bucket_key = this.reverse_lookup.get(item_id);
if (bucket_key) {
const bucket = this.get_bucket(bucket_key);
bucket.delete(item_id);
this.reverse_lookup.delete(item_id);
}
}
get_bucket(bucket_key) {
return this.key_to_bucket.get(bucket_key);
}
keys() {
return this.key_to_bucket.keys();
}
values() {
return this.key_to_bucket.values();
}
[Symbol.iterator]() {
return this.key_to_bucket[Symbol.iterator]();
}
}
class UnreadPMCounter {
bucketer = new Bucketer({
KeyDict: Map,
make_bucket: () => new Set(),
});
clear() {
this.bucketer.clear();
}
set_pms(pms) {
for (const obj of pms) {
const user_ids_string = obj.other_user_id.toString();
this.set_message_ids(user_ids_string, obj.unread_message_ids);
}
}
set_huddles(huddles) {
for (const obj of huddles) {
const user_ids_string = people.pm_lookup_key(obj.user_ids_string);
this.set_message_ids(user_ids_string, obj.unread_message_ids);
}
}
set_message_ids(user_ids_string, unread_message_ids) {
for (const msg_id of unread_message_ids) {
this.bucketer.add({
bucket_key: user_ids_string,
item_id: msg_id,
});
}
}
add({message_id, user_ids_string}) {
if (user_ids_string) {
this.bucketer.add({
bucket_key: user_ids_string,
item_id: message_id,
});
}
}
delete(message_id) {
this.bucketer.delete(message_id);
}
get_counts() {
const pm_dict = new Map(); // Hash by user_ids_string -> count
let total_count = 0;
let right_sidebar_count = 0;
for (const [user_ids_string, id_set] of this.bucketer) {
const count = id_set.size;
pm_dict.set(user_ids_string, count);
const user_ids = people.user_ids_string_to_ids_array(user_ids_string);
const is_with_one_human =
user_ids.length === 1 && !people.get_by_user_id(user_ids[0]).is_bot;
if (is_with_one_human) {
right_sidebar_count += count;
}
total_count += count;
}
return {
total_count,
pm_dict,
right_sidebar_count,
};
}
num_unread(user_ids_string) {
if (!user_ids_string) {
return 0;
}
const bucket = this.bucketer.get_bucket(user_ids_string);
if (!bucket) {
return 0;
}
return bucket.size;
}
get_msg_ids() {
const ids = [];
for (const id_set of this.bucketer.values()) {
for (const id of id_set) {
ids.push(id);
}
}
return util.sorted_ids(ids);
}
get_msg_ids_for_user_ids_string(user_ids_string) {
if (!user_ids_string) {
return [];
}
const bucket = this.bucketer.get_bucket(user_ids_string);
if (!bucket) {
return [];
}
const ids = Array.from(bucket);
return util.sorted_ids(ids);
}
}
const unread_pm_counter = new UnreadPMCounter();
function make_per_stream_bucketer() {
return new Bucketer({
KeyDict: FoldDict, // bucket keys are topics
make_bucket: () => new Set(),
});
}
class UnreadTopicCounter {
bucketer = new Bucketer({
KeyDict: Map, // bucket keys are stream_ids
make_bucket: make_per_stream_bucketer,
});
clear() {
this.bucketer.clear();
}
set_streams(objs) {
for (const obj of objs) {
const stream_id = obj.stream_id;
const topic = obj.topic;
const unread_message_ids = obj.unread_message_ids;
for (const message_id of unread_message_ids) {
this.add({message_id, stream_id, topic});
}
}
}
add({message_id, stream_id, topic}) {
this.bucketer.add({
bucket_key: stream_id,
item_id: message_id,
add_callback(per_stream_bucketer) {
per_stream_bucketer.add({
bucket_key: topic,
item_id: message_id,
});
},
});
}
delete(msg_id) {
this.bucketer.delete(msg_id);
}
get_counts() {
const res = {};
res.stream_unread_messages = 0;
res.stream_count = new Map(); // hash by stream_id -> count
for (const [stream_id, per_stream_bucketer] of this.bucketer) {
// We track unread counts for streams that may be currently
// unsubscribed. Since users may re-subscribe, we don't
// completely throw away the data. But we do ignore it here,
// so that callers have a view of the **current** world.
const sub = sub_store.get(stream_id);
if (!sub || !stream_data.is_subscribed(stream_id)) {
continue;
}
let stream_count = 0;
for (const [topic, msgs] of per_stream_bucketer) {
const topic_count = msgs.size;
if (!user_topics.is_topic_muted(stream_id, topic)) {
stream_count += topic_count;
}
}
res.stream_count.set(stream_id, stream_count);
if (!stream_data.is_muted(stream_id)) {
res.stream_unread_messages += stream_count;
}
}
return res;
}
get_missing_topics(opts) {
/* Clients have essentially complete unread data, but
* stream_topic_history.is_complete_for_stream_id() can be
* false. In that situation, this function helps ensure that
* we include all topics with unread messages in data that.
*
* It will return all topics in the provided stream with a
* nonzero unread count that are not already present in the
* topic_dict parameter.
*/
const stream_id = opts.stream_id;
const topic_dict = opts.topic_dict;
const per_stream_bucketer = this.bucketer.get_bucket(stream_id);
if (!per_stream_bucketer) {
return [];
}
let topic_names = Array.from(per_stream_bucketer.keys());
/* Include topics that have at least one unread. It would likely
* be better design for buckets to be deleted when emptied. */
topic_names = topic_names.filter((topic_name) => {
const messages = Array.from(per_stream_bucketer.get_bucket(topic_name));
return messages.length > 0;
});
/* And aren't already present in topic_dict. */
topic_names = topic_names.filter((topic_name) => !topic_dict.has(topic_name));
const result = topic_names.map((topic_name) => {
const msgs = per_stream_bucketer.get_bucket(topic_name);
return {
pretty_name: topic_name,
message_id: Math.max(...Array.from(msgs)),
};
});
return result;
}
get_stream_count(stream_id) {
let stream_count = 0;
const per_stream_bucketer = this.bucketer.get_bucket(stream_id);
if (!per_stream_bucketer) {
return 0;
}
const sub = sub_store.get(stream_id);
for (const [topic, msgs] of per_stream_bucketer) {
if (sub && !user_topics.is_topic_muted(stream_id, topic)) {
stream_count += msgs.size;
}
}
return stream_count;
}
get(stream_id, topic) {
const per_stream_bucketer = this.bucketer.get_bucket(stream_id);
if (!per_stream_bucketer) {
return 0;
}
const topic_bucket = per_stream_bucketer.get_bucket(topic);
if (!topic_bucket) {
return 0;
}
return topic_bucket.size;
}
get_msg_ids_for_stream(stream_id) {
const per_stream_bucketer = this.bucketer.get_bucket(stream_id);
if (!per_stream_bucketer) {
return [];
}
const ids = [];
const sub = sub_store.get(stream_id);
for (const [topic, id_set] of per_stream_bucketer) {
if (sub && !user_topics.is_topic_muted(stream_id, topic)) {
for (const id of id_set) {
ids.push(id);
}
}
}
return util.sorted_ids(ids);
}
get_msg_ids_for_topic(stream_id, topic) {
const per_stream_bucketer = this.bucketer.get_bucket(stream_id);
if (!per_stream_bucketer) {
return [];
}
const topic_bucket = per_stream_bucketer.get_bucket(topic);
if (!topic_bucket) {
return [];
}
const ids = Array.from(topic_bucket);
return util.sorted_ids(ids);
}
get_streams_with_unread_mentions() {
const streams_with_mentions = new Set();
// Collect the set of streams containing at least one mention.
// We can do this efficiently, since unread_mentions_counter
// contains all unread message IDs, and we use stream_ids as
// bucket keys in our outer bucketer.
for (const message_id of unread_mentions_counter) {
const stream_id = this.bucketer.reverse_lookup.get(message_id);
streams_with_mentions.add(stream_id);
}
return streams_with_mentions;
}
topic_has_any_unread(stream_id, topic) {
const per_stream_bucketer = this.bucketer.get_bucket(stream_id);
if (!per_stream_bucketer) {
return false;
}
const id_set = per_stream_bucketer.get_bucket(topic);
if (!id_set) {
return false;
}
return id_set.size !== 0;
}
get_topics_with_unread_mentions(stream_id) {
// Returns the set of lower cased topics with unread mentions
// in the given stream.
const result = new Set();
const per_stream_bucketer = this.bucketer.get_bucket(stream_id);
if (!per_stream_bucketer) {
return result;
}
for (const message_id of unread_mentions_counter) {
// Because bucket keys in per_stream_bucketer are topics,
// we can just directly use reverse_lookup to find the
// topic in this stream containing a given unread message
// ID. If it's not in this stream, we'll get undefined.
const topic_match = per_stream_bucketer.reverse_lookup.get(message_id);
if (topic_match !== undefined) {
// Important: We lower-case topics here before adding them
// to this set, to support case-insensitive checks.
result.add(topic_match.toLowerCase());
}
}
return result;
}
}
const unread_topic_counter = new UnreadTopicCounter();
function add_message_to_unread_mention_topics(message_id) {
const message = message_store.get(message_id);
if (message.type !== "stream") {
return;
}
const topic_key = recent_topics_util.get_topic_key(message.stream_id, message.topic);
if (unread_mention_topics.has(topic_key)) {
unread_mention_topics.get(topic_key).add(message_id);
}
unread_mention_topics.set(topic_key, new Set([message_id]));
}
function remove_message_from_unread_mention_topics(message_id) {
const stream_id = unread_topic_counter.bucketer.reverse_lookup.get(message_id);
if (!stream_id) {
// Private messages and messages that were already not unread
// exit here.
return;
}
const per_stream_bucketer = unread_topic_counter.bucketer.get_bucket(stream_id);
if (!per_stream_bucketer) {
blueslip.error(`Could not find per_stream_bucketer for ${message_id}.`);
return;
}
const topic = per_stream_bucketer.reverse_lookup.get(message_id);
const topic_key = recent_topics_util.get_topic_key(stream_id, topic);
if (unread_mention_topics.has(topic_key)) {
unread_mention_topics.get(topic_key).delete(message_id);
}
}
export function clear_and_populate_unread_mention_topics() {
// The unread_mention_topics is an important data structure for
// efficiently querying whether a given stream/topic pair contains
// unread mentions.
//
// It is effectively a cache, since it can be reconstructed from
// unread_mentions_counter (IDs for all unread mentions) and
// unread_topic_counter (Streams/topics for all unread messages).
//
// Since this function runs in O(unread mentions) time, we can use
// it in topic editing code paths where it might be onerous to
// write custom live-update code; but we should avoid calling it
// in loops.
unread_mention_topics.clear();
for (const message_id of unread_mentions_counter) {
const stream_id = unread_topic_counter.bucketer.reverse_lookup.get(message_id);
if (!stream_id) {
continue;
}
const per_stream_bucketer = unread_topic_counter.bucketer.get_bucket(stream_id);
const topic = per_stream_bucketer.reverse_lookup.get(message_id);
const topic_key = recent_topics_util.get_topic_key(stream_id, topic);
if (unread_mention_topics.has(topic_key)) {
unread_mention_topics.get(topic_key).add(message_id);
}
unread_mention_topics.set(topic_key, new Set([message_id]));
}
}
export function message_unread(message) {
if (message === undefined) {
return false;
}
return message.unread;
}
export function get_read_message_ids(message_ids) {
return message_ids.filter((message_id) => !unread_messages.has(message_id));
}
export function get_unread_message_ids(message_ids) {
return message_ids.filter((message_id) => unread_messages.has(message_id));
}
export function get_unread_messages(messages) {
return messages.filter((message) => unread_messages.has(message.id));
}
export function get_unread_message_count() {
return unread_messages.size;
}
export function update_unread_topics(msg, event) {
const new_topic = util.get_edit_event_topic(event);
const {new_stream_id} = event;
if (new_topic === undefined && new_stream_id === undefined) {
return;
}
if (!unread_messages.has(msg.id)) {
return;
}
unread_topic_counter.delete(msg.id);
unread_topic_counter.add({
message_id: msg.id,
stream_id: new_stream_id || msg.stream_id,
topic: new_topic || msg.topic,
});
}
export function process_loaded_messages(messages, expect_no_new_unreads = false) {
// Process a set of messages that we have full copies of from the
// server for whether any are unread but not tracked as such by
// our data structures. This can occur due to old_unreads_missing,
// changes in muting configuration, innocent races, or potentially bugs.
//
// Returns whether there were any new unread messages; in that
// case, the caller will need to trigger a rerender of UI
// displaying unread counts.
let any_untracked_unread_messages = false;
for (const message of messages) {
if (message.unread) {
if (unread_messages.has(message.id)) {
// If we're already tracking this message as unread, there's nothing to do.
continue;
}
if (expect_no_new_unreads && !old_unreads_missing) {
// This may happen due to races, where someone narrows
// to a view and the message_fetch request returns
// before server_events system delivers the message to
// the client.
blueslip.log(`New unread ${message.id} discovered in process_loaded_messages.`);
}
const user_ids_string =
message.type === "private" ? people.pm_reply_user_string(message) : undefined;
process_unread_message({
id: message.id,
mentioned: message.mentioned,
mentioned_me_directly: message.mentioned_me_directly,
stream_id: message.stream_id,
topic: message.topic,
type: message.type,
unread: true,
user_ids_string,
});
any_untracked_unread_messages = true;
}
}
return any_untracked_unread_messages;
}
export function process_unread_message(message) {
// The `message` here just needs to require certain fields. For example,
// the "message" may actually be constructed from a Zulip event that doesn't
// include fields like "content". The caller must verify that the message
// is actually unread--we don't defend against that.
unread_messages.add(message.id);
if (message.type === "private") {
unread_pm_counter.add({
message_id: message.id,
user_ids_string: message.user_ids_string,
});
}
if (message.type === "stream") {
unread_topic_counter.add({
message_id: message.id,
stream_id: message.stream_id,
topic: message.topic,
});
}
update_message_for_mention(message);
}
export function update_message_for_mention(message, content_edited = false) {
if (!message.unread) {
unread_mentions_counter.delete(message.id);
remove_message_from_unread_mention_topics(message.id);
return;
}
const is_unmuted_mention =
message.type === "stream" &&
message.mentioned &&
!user_topics.is_topic_muted(message.stream_id, message.topic);
if (is_unmuted_mention || message.mentioned_me_directly) {
unread_mentions_counter.add(message.id);
add_message_to_unread_mention_topics(message.id);
} else {
unread_mentions_counter.delete(message.id);
remove_message_from_unread_mention_topics(message.id);
}
if (content_edited && message.type === "stream") {
// We only need to update recent topics here if this was a content change in an unread
// mention, since in other cases recent topics gets rerendered by other functions.
const topic_key = recent_topics_util.get_topic_key(message.stream_id, message.topic);
recent_topics_ui.inplace_rerender(topic_key);
}
}
export function mark_as_read(message_id) {
// We don't need to check anything about the message, since all
// the following methods are cheap and work fine even if message_id
// was never set to unread.
unread_pm_counter.delete(message_id);
// Important: This function uses `unread_topic_counter` to look up
// the stream/topic for this previously unread message, so much
// happen before the message is removed from that data structure.
remove_message_from_unread_mention_topics(message_id);
unread_topic_counter.delete(message_id);
unread_mentions_counter.delete(message_id);
unread_messages.delete(message_id);
const message = message_store.get(message_id);
if (message) {
message.unread = false;
}
}
export function declare_bankruptcy() {
// Only used in tests.
unread_pm_counter.clear();
unread_topic_counter.clear();
unread_mentions_counter.clear();
unread_messages.clear();
unread_mention_topics.clear();
}
export function get_counts() {
const res = {};
// Return a data structure with various counts. This function should be
// pretty cheap, even if you don't care about all the counts, and you
// should strive to keep it free of side effects on globals or DOM.
res.private_message_count = 0;
res.mentioned_message_count = unread_mentions_counter.size;
// This sets stream_count, topic_count, and home_unread_messages
const topic_res = unread_topic_counter.get_counts();
const streams_with_mentions = unread_topic_counter.get_streams_with_unread_mentions();
res.home_unread_messages = topic_res.stream_unread_messages;
res.stream_count = topic_res.stream_count;
res.streams_with_mentions = Array.from(streams_with_mentions);
const pm_res = unread_pm_counter.get_counts();
res.pm_count = pm_res.pm_dict;
res.private_message_count = pm_res.total_count;
res.right_sidebar_private_message_count = pm_res.right_sidebar_count;
res.home_unread_messages += pm_res.total_count;
return res;
}
// Saves us from calling to get_counts() when we can avoid it.
export function calculate_notifiable_count(res) {
let new_message_count = 0;
const only_show_notifiable =
user_settings.desktop_icon_count_display ===
settings_config.desktop_icon_count_display_values.notifiable.code;
const no_notifications =
user_settings.desktop_icon_count_display ===
settings_config.desktop_icon_count_display_values.none.code;
if (only_show_notifiable) {
// DESKTOP_ICON_COUNT_DISPLAY_NOTIFIABLE
new_message_count = res.mentioned_message_count + res.private_message_count;
} else if (no_notifications) {
// DESKTOP_ICON_COUNT_DISPLAY_NONE
new_message_count = 0;
} else {
// DESKTOP_ICON_COUNT_DISPLAY_MESSAGES
new_message_count = res.home_unread_messages;
}
return new_message_count;
}
export function get_notifiable_count() {
const res = get_counts();
return calculate_notifiable_count(res);
}
export function num_unread_for_stream(stream_id) {
return unread_topic_counter.get_stream_count(stream_id);
}
export function num_unread_for_topic(stream_id, topic_name) {
return unread_topic_counter.get(stream_id, topic_name);
}
export function stream_has_any_unread_mentions(stream_id) {
// This function is somewhat inefficient and thus should not be
// called in loops, since runs in O(total unread mentions) time.
const streams_with_mentions = unread_topic_counter.get_streams_with_unread_mentions();
return streams_with_mentions.has(stream_id);
}
export function topic_has_any_unread_mentions(stream_id, topic) {
// Because this function is called in a loop for every displayed
// Recent Topics row, it's important for it to run in O(1) time.
const topic_key = stream_id + ":" + topic.toLowerCase();
return unread_mention_topics.get(topic_key) && unread_mention_topics.get(topic_key).size > 0;
}
export function topic_has_any_unread(stream_id, topic) {
return unread_topic_counter.topic_has_any_unread(stream_id, topic);
}
export function get_topics_with_unread_mentions(stream_id) {
return unread_topic_counter.get_topics_with_unread_mentions(stream_id);
}
export function num_unread_for_user_ids_string(user_ids_string) {
return unread_pm_counter.num_unread(user_ids_string);
}
export function get_msg_ids_for_stream(stream_id) {
return unread_topic_counter.get_msg_ids_for_stream(stream_id);
}
export function get_msg_ids_for_topic(stream_id, topic_name) {
return unread_topic_counter.get_msg_ids_for_topic(stream_id, topic_name);
}
export function get_msg_ids_for_user_ids_string(user_ids_string) {
return unread_pm_counter.get_msg_ids_for_user_ids_string(user_ids_string);
}
export function get_msg_ids_for_private() {
return unread_pm_counter.get_msg_ids();
}
export function get_msg_ids_for_mentions() {
const ids = Array.from(unread_mentions_counter);
return util.sorted_ids(ids);
}
export function get_all_msg_ids() {
const ids = Array.from(unread_messages);
return util.sorted_ids(ids);
}
export function get_missing_topics(opts) {
return unread_topic_counter.get_missing_topics(opts);
}
export function get_msg_ids_for_starred() {
// This is here for API consistency sake--we never
// have unread starred messages. (Some day we may ironically
// want to make starring the same as mark-as-unread, but
// for now starring === reading.)
return [];
}
export function initialize(params) {
const unread_msgs = params.unread_msgs;
old_unreads_missing = unread_msgs.old_unreads_missing;
unread_pm_counter.set_huddles(unread_msgs.huddles);
unread_pm_counter.set_pms(unread_msgs.pms);
unread_topic_counter.set_streams(unread_msgs.streams);
for (const message_id of unread_msgs.mentions) {
unread_mentions_counter.add(message_id);
}
clear_and_populate_unread_mention_topics();
for (const obj of unread_msgs.huddles) {
for (const message_id of obj.unread_message_ids) {
unread_messages.add(message_id);
}
}
for (const obj of unread_msgs.pms) {
for (const message_id of obj.unread_message_ids) {
unread_messages.add(message_id);
}
}
for (const obj of unread_msgs.streams) {
for (const message_id of obj.unread_message_ids) {
unread_messages.add(message_id);
}
}
for (const message_id of unread_msgs.mentions) {
unread_messages.add(message_id);
}
}