feat: upgrade to .NET 6, refactor everything
This commit is contained in:
spiral
@@ -1,17 +1,10 @@
|
||||
using System;
|
||||
using System.Collections.Concurrent;
|
||||
using System.Collections.Generic;
|
||||
using System.Linq;
|
||||
using System.Net.WebSockets;
|
||||
using System.Threading;
|
||||
using System.Threading.Tasks;
|
||||
|
||||
using App.Metrics;
|
||||
|
||||
using Autofac;
|
||||
|
||||
using Myriad.Cache;
|
||||
using Myriad.Extensions;
|
||||
using Myriad.Gateway;
|
||||
using Myriad.Rest;
|
||||
using Myriad.Types;
|
||||
@@ -25,267 +18,261 @@ using Sentry;
|
||||
using Serilog;
|
||||
using Serilog.Context;
|
||||
|
||||
namespace PluralKit.Bot
|
||||
namespace PluralKit.Bot;
|
||||
|
||||
public class Bot
|
||||
{
|
||||
public class Bot
|
||||
private readonly IDiscordCache _cache;
|
||||
|
||||
private readonly Cluster _cluster;
|
||||
private readonly PeriodicStatCollector _collector;
|
||||
private readonly CommandMessageService _commandMessageService;
|
||||
private readonly BotConfig _config;
|
||||
private readonly ErrorMessageService _errorMessageService;
|
||||
private readonly ILogger _logger;
|
||||
private readonly IMetrics _metrics;
|
||||
private readonly DiscordApiClient _rest;
|
||||
private readonly ILifetimeScope _services;
|
||||
|
||||
private bool _hasReceivedReady;
|
||||
private Timer _periodicTask; // Never read, just kept here for GC reasons
|
||||
|
||||
public Bot(ILifetimeScope services, ILogger logger, PeriodicStatCollector collector, IMetrics metrics,
|
||||
BotConfig config,
|
||||
ErrorMessageService errorMessageService, CommandMessageService commandMessageService,
|
||||
Cluster cluster, DiscordApiClient rest, IDiscordCache cache)
|
||||
{
|
||||
_logger = logger.ForContext<Bot>();
|
||||
_services = services;
|
||||
_collector = collector;
|
||||
_metrics = metrics;
|
||||
_config = config;
|
||||
_errorMessageService = errorMessageService;
|
||||
_commandMessageService = commandMessageService;
|
||||
_cluster = cluster;
|
||||
_rest = rest;
|
||||
_cache = cache;
|
||||
}
|
||||
|
||||
private readonly Cluster _cluster;
|
||||
private readonly DiscordApiClient _rest;
|
||||
private readonly ILogger _logger;
|
||||
private readonly ILifetimeScope _services;
|
||||
private readonly PeriodicStatCollector _collector;
|
||||
private readonly IMetrics _metrics;
|
||||
private readonly BotConfig _config;
|
||||
private readonly ErrorMessageService _errorMessageService;
|
||||
private readonly CommandMessageService _commandMessageService;
|
||||
private readonly IDiscordCache _cache;
|
||||
public void Init()
|
||||
{
|
||||
_cluster.EventReceived += OnEventReceived;
|
||||
|
||||
private bool _hasReceivedReady = false;
|
||||
private Timer _periodicTask; // Never read, just kept here for GC reasons
|
||||
// Init the shard stuff
|
||||
_services.Resolve<ShardInfoService>().Init();
|
||||
|
||||
public Bot(ILifetimeScope services, ILogger logger, PeriodicStatCollector collector, IMetrics metrics, BotConfig config,
|
||||
ErrorMessageService errorMessageService, CommandMessageService commandMessageService, Cluster cluster, DiscordApiClient rest, IDiscordCache cache)
|
||||
// Not awaited, just needs to run in the background
|
||||
// Trying our best to run it at whole minute boundaries (xx:00), with ~250ms buffer
|
||||
// This *probably* doesn't matter in practice but I jut think it's neat, y'know.
|
||||
var timeNow = SystemClock.Instance.GetCurrentInstant();
|
||||
var timeTillNextWholeMinute = TimeSpan.FromMilliseconds(60000 - timeNow.ToUnixTimeMilliseconds() % 60000 + 250);
|
||||
_periodicTask = new Timer(_ =>
|
||||
{
|
||||
_logger = logger.ForContext<Bot>();
|
||||
_services = services;
|
||||
_collector = collector;
|
||||
_metrics = metrics;
|
||||
_config = config;
|
||||
_errorMessageService = errorMessageService;
|
||||
_commandMessageService = commandMessageService;
|
||||
_cluster = cluster;
|
||||
_rest = rest;
|
||||
_cache = cache;
|
||||
}
|
||||
var __ = UpdatePeriodic();
|
||||
}, null, timeTillNextWholeMinute, TimeSpan.FromMinutes(1));
|
||||
}
|
||||
|
||||
public void Init()
|
||||
{
|
||||
_cluster.EventReceived += OnEventReceived;
|
||||
private async Task OnEventReceived(Shard shard, IGatewayEvent evt)
|
||||
{
|
||||
await _cache.TryUpdateSelfMember(shard, evt);
|
||||
await _cache.HandleGatewayEvent(evt);
|
||||
|
||||
// Init the shard stuff
|
||||
_services.Resolve<ShardInfoService>().Init();
|
||||
// HandleEvent takes a type parameter, automatically inferred by the event type
|
||||
// It will then look up an IEventHandler<TypeOfEvent> in the DI container and call that object's handler method
|
||||
// For registering new ones, see Modules.cs
|
||||
if (evt is MessageCreateEvent mc)
|
||||
await HandleEvent(shard, mc);
|
||||
if (evt is MessageUpdateEvent mu)
|
||||
await HandleEvent(shard, mu);
|
||||
if (evt is MessageDeleteEvent md)
|
||||
await HandleEvent(shard, md);
|
||||
if (evt is MessageDeleteBulkEvent mdb)
|
||||
await HandleEvent(shard, mdb);
|
||||
if (evt is MessageReactionAddEvent mra)
|
||||
await HandleEvent(shard, mra);
|
||||
if (evt is InteractionCreateEvent ic)
|
||||
await HandleEvent(shard, ic);
|
||||
|
||||
// Not awaited, just needs to run in the background
|
||||
// Trying our best to run it at whole minute boundaries (xx:00), with ~250ms buffer
|
||||
// This *probably* doesn't matter in practice but I jut think it's neat, y'know.
|
||||
var timeNow = SystemClock.Instance.GetCurrentInstant();
|
||||
var timeTillNextWholeMinute = TimeSpan.FromMilliseconds(60000 - timeNow.ToUnixTimeMilliseconds() % 60000 + 250);
|
||||
_periodicTask = new Timer(_ =>
|
||||
{
|
||||
var __ = UpdatePeriodic();
|
||||
}, null, timeTillNextWholeMinute, TimeSpan.FromMinutes(1));
|
||||
}
|
||||
// Update shard status for shards immediately on connect
|
||||
if (evt is ReadyEvent re)
|
||||
await HandleReady(shard, re);
|
||||
if (evt is ResumedEvent)
|
||||
await HandleResumed(shard);
|
||||
}
|
||||
|
||||
private async Task OnEventReceived(Shard shard, IGatewayEvent evt)
|
||||
{
|
||||
await _cache.TryUpdateSelfMember(shard, evt);
|
||||
await _cache.HandleGatewayEvent(evt);
|
||||
private Task HandleResumed(Shard shard) => UpdateBotStatus(shard);
|
||||
|
||||
// HandleEvent takes a type parameter, automatically inferred by the event type
|
||||
// It will then look up an IEventHandler<TypeOfEvent> in the DI container and call that object's handler method
|
||||
// For registering new ones, see Modules.cs
|
||||
if (evt is MessageCreateEvent mc)
|
||||
await HandleEvent(shard, mc);
|
||||
if (evt is MessageUpdateEvent mu)
|
||||
await HandleEvent(shard, mu);
|
||||
if (evt is MessageDeleteEvent md)
|
||||
await HandleEvent(shard, md);
|
||||
if (evt is MessageDeleteBulkEvent mdb)
|
||||
await HandleEvent(shard, mdb);
|
||||
if (evt is MessageReactionAddEvent mra)
|
||||
await HandleEvent(shard, mra);
|
||||
if (evt is InteractionCreateEvent ic)
|
||||
await HandleEvent(shard, ic);
|
||||
private Task HandleReady(Shard shard, ReadyEvent _)
|
||||
{
|
||||
_hasReceivedReady = true;
|
||||
return UpdateBotStatus(shard);
|
||||
}
|
||||
|
||||
// Update shard status for shards immediately on connect
|
||||
if (evt is ReadyEvent re)
|
||||
await HandleReady(shard, re);
|
||||
if (evt is ResumedEvent)
|
||||
await HandleResumed(shard);
|
||||
}
|
||||
public async Task Shutdown()
|
||||
{
|
||||
// This will stop the timer and prevent any subsequent invocations
|
||||
await _periodicTask.DisposeAsync();
|
||||
|
||||
private Task HandleResumed(Shard shard)
|
||||
{
|
||||
return UpdateBotStatus(shard);
|
||||
}
|
||||
|
||||
private Task HandleReady(Shard shard, ReadyEvent _)
|
||||
{
|
||||
_hasReceivedReady = true;
|
||||
return UpdateBotStatus(shard);
|
||||
}
|
||||
|
||||
public async Task Shutdown()
|
||||
{
|
||||
// This will stop the timer and prevent any subsequent invocations
|
||||
await _periodicTask.DisposeAsync();
|
||||
|
||||
// Send users a lil status message
|
||||
// We're not actually properly disconnecting from the gateway (lol) so it'll linger for a few minutes
|
||||
// Should be plenty of time for the bot to connect again next startup and set the real status
|
||||
if (_hasReceivedReady)
|
||||
{
|
||||
await Task.WhenAll(_cluster.Shards.Values.Select(shard =>
|
||||
shard.UpdateStatus(new GatewayStatusUpdate
|
||||
// Send users a lil status message
|
||||
// We're not actually properly disconnecting from the gateway (lol) so it'll linger for a few minutes
|
||||
// Should be plenty of time for the bot to connect again next startup and set the real status
|
||||
if (_hasReceivedReady)
|
||||
await Task.WhenAll(_cluster.Shards.Values.Select(shard =>
|
||||
shard.UpdateStatus(new GatewayStatusUpdate
|
||||
{
|
||||
Activities = new[]
|
||||
{
|
||||
Activities = new[]
|
||||
{
|
||||
new ActivityPartial
|
||||
{
|
||||
Name = "Restarting... (please wait)",
|
||||
Type = ActivityType.Game
|
||||
}
|
||||
},
|
||||
Status = GatewayStatusUpdate.UserStatus.Idle
|
||||
})));
|
||||
}
|
||||
}
|
||||
new ActivityPartial {Name = "Restarting... (please wait)", Type = ActivityType.Game}
|
||||
},
|
||||
Status = GatewayStatusUpdate.UserStatus.Idle
|
||||
})));
|
||||
}
|
||||
|
||||
private Task HandleEvent<T>(Shard shard, T evt) where T : IGatewayEvent
|
||||
private Task HandleEvent<T>(Shard shard, T evt) where T : IGatewayEvent
|
||||
{
|
||||
// We don't want to stall the event pipeline, so we'll "fork" inside here
|
||||
var _ = HandleEventInner();
|
||||
return Task.CompletedTask;
|
||||
|
||||
async Task HandleEventInner()
|
||||
{
|
||||
// We don't want to stall the event pipeline, so we'll "fork" inside here
|
||||
var _ = HandleEventInner();
|
||||
return Task.CompletedTask;
|
||||
await Task.Yield();
|
||||
|
||||
async Task HandleEventInner()
|
||||
await using var serviceScope = _services.BeginLifetimeScope();
|
||||
|
||||
// Find an event handler that can handle the type of event (<T>) we're given
|
||||
IEventHandler<T> handler;
|
||||
try
|
||||
{
|
||||
await Task.Yield();
|
||||
|
||||
await using var serviceScope = _services.BeginLifetimeScope();
|
||||
|
||||
// Find an event handler that can handle the type of event (<T>) we're given
|
||||
IEventHandler<T> handler;
|
||||
try
|
||||
{
|
||||
handler = serviceScope.Resolve<IEventHandler<T>>();
|
||||
}
|
||||
catch (Exception e)
|
||||
{
|
||||
_logger.Error(e, "Error instantiating handler class");
|
||||
return;
|
||||
}
|
||||
|
||||
try
|
||||
{
|
||||
var queue = serviceScope.ResolveOptional<HandlerQueue<T>>();
|
||||
|
||||
using var _ = LogContext.PushProperty("EventId", Guid.NewGuid());
|
||||
using var __ = LogContext.Push(await serviceScope.Resolve<SerilogGatewayEnricherFactory>().GetEnricher(shard, evt));
|
||||
_logger.Verbose("Received gateway event: {@Event}", evt);
|
||||
|
||||
// Also, find a Sentry enricher for the event type (if one is present), and ask it to put some event data in the Sentry scope
|
||||
var sentryEnricher = serviceScope.ResolveOptional<ISentryEnricher<T>>();
|
||||
sentryEnricher?.Enrich(serviceScope.Resolve<Scope>(), shard, evt);
|
||||
|
||||
using var timer = _metrics.Measure.Timer.Time(BotMetrics.EventsHandled,
|
||||
new MetricTags("event", typeof(T).Name.Replace("Event", "")));
|
||||
|
||||
// Delegate to the queue to see if it wants to handle this event
|
||||
// the TryHandle call returns true if it's handled the event
|
||||
// Usually it won't, so just pass it on to the main handler
|
||||
if (queue == null || !await queue.TryHandle(evt))
|
||||
await handler.Handle(shard, evt);
|
||||
}
|
||||
catch (Exception exc)
|
||||
{
|
||||
await HandleError(shard, handler, evt, serviceScope, exc);
|
||||
}
|
||||
handler = serviceScope.Resolve<IEventHandler<T>>();
|
||||
}
|
||||
}
|
||||
|
||||
private async Task HandleError<T>(Shard shard, IEventHandler<T> handler, T evt, ILifetimeScope serviceScope, Exception exc)
|
||||
where T : IGatewayEvent
|
||||
{
|
||||
_metrics.Measure.Meter.Mark(BotMetrics.BotErrors, exc.GetType().FullName);
|
||||
|
||||
// Make this beforehand so we can access the event ID for logging
|
||||
var sentryEvent = new SentryEvent(exc);
|
||||
|
||||
// If the event is us responding to our own error messages, don't bother logging
|
||||
if (evt is MessageCreateEvent mc && mc.Author.Id == shard.User?.Id)
|
||||
catch (Exception e)
|
||||
{
|
||||
_logger.Error(e, "Error instantiating handler class");
|
||||
return;
|
||||
|
||||
var shouldReport = exc.IsOurProblem();
|
||||
if (shouldReport)
|
||||
{
|
||||
// only log exceptions if they're our problem
|
||||
_logger.Error(exc, "Exception in event handler: {SentryEventId}", sentryEvent.EventId);
|
||||
|
||||
// Report error to Sentry
|
||||
// This will just no-op if there's no URL set
|
||||
var sentryScope = serviceScope.Resolve<Scope>();
|
||||
|
||||
// Add some specific info about Discord error responses, as a breadcrumb
|
||||
// TODO: headers to dict
|
||||
// if (exc is BadRequestException bre)
|
||||
// sentryScope.AddBreadcrumb(bre.Response, "response.error", data: new Dictionary<string, string>(bre.Response.Headers));
|
||||
// if (exc is NotFoundException nfe)
|
||||
// sentryScope.AddBreadcrumb(nfe.Response, "response.error", data: new Dictionary<string, string>(nfe.Response.Headers));
|
||||
// if (exc is UnauthorizedException ue)
|
||||
// sentryScope.AddBreadcrumb(ue.Response, "response.error", data: new Dictionary<string, string>(ue.Response.Headers));
|
||||
|
||||
SentrySdk.CaptureEvent(sentryEvent, sentryScope);
|
||||
|
||||
// most of these errors aren't useful...
|
||||
if (_config.DisableErrorReporting)
|
||||
return;
|
||||
|
||||
// Once we've sent it to Sentry, report it to the user (if we have permission to)
|
||||
var reportChannel = handler.ErrorChannelFor(evt);
|
||||
if (reportChannel == null)
|
||||
return;
|
||||
|
||||
var botPerms = await _cache.PermissionsIn(reportChannel.Value);
|
||||
if (botPerms.HasFlag(PermissionSet.SendMessages | PermissionSet.EmbedLinks))
|
||||
await _errorMessageService.SendErrorMessage(reportChannel.Value, sentryEvent.EventId.ToString());
|
||||
}
|
||||
}
|
||||
|
||||
private async Task UpdatePeriodic()
|
||||
{
|
||||
_logger.Debug("Running once-per-minute scheduled tasks");
|
||||
|
||||
await UpdateBotStatus();
|
||||
|
||||
// Collect some stats, submit them to the metrics backend
|
||||
await _collector.CollectStats();
|
||||
await Task.WhenAll(((IMetricsRoot)_metrics).ReportRunner.RunAllAsync());
|
||||
_logger.Debug("Submitted metrics to backend");
|
||||
}
|
||||
|
||||
private async Task UpdateBotStatus(Shard specificShard = null)
|
||||
{
|
||||
// If we're not on any shards, don't bother (this happens if the periodic timer fires before the first Ready)
|
||||
if (!_hasReceivedReady) return;
|
||||
|
||||
var totalGuilds = await _cache.GetAllGuilds().CountAsync();
|
||||
|
||||
try // DiscordClient may throw an exception if the socket is closed (e.g just after OP 7 received)
|
||||
try
|
||||
{
|
||||
Task UpdateStatus(Shard shard) =>
|
||||
shard.UpdateStatus(new GatewayStatusUpdate
|
||||
{
|
||||
Activities = new[]
|
||||
{
|
||||
new ActivityPartial
|
||||
{
|
||||
Name = $"pk;help | in {totalGuilds:N0} servers | shard #{shard.ShardId}",
|
||||
Type = ActivityType.Game,
|
||||
Url = "https://pluralkit.me/"
|
||||
}
|
||||
}
|
||||
});
|
||||
var queue = serviceScope.ResolveOptional<HandlerQueue<T>>();
|
||||
|
||||
if (specificShard != null)
|
||||
await UpdateStatus(specificShard);
|
||||
else // Run shard updates concurrently
|
||||
await Task.WhenAll(_cluster.Shards.Values.Select(UpdateStatus));
|
||||
using var _ = LogContext.PushProperty("EventId", Guid.NewGuid());
|
||||
using var __ = LogContext.Push(await serviceScope.Resolve<SerilogGatewayEnricherFactory>()
|
||||
.GetEnricher(shard, evt));
|
||||
_logger.Verbose("Received gateway event: {@Event}", evt);
|
||||
|
||||
// Also, find a Sentry enricher for the event type (if one is present), and ask it to put some event data in the Sentry scope
|
||||
var sentryEnricher = serviceScope.ResolveOptional<ISentryEnricher<T>>();
|
||||
sentryEnricher?.Enrich(serviceScope.Resolve<Scope>(), shard, evt);
|
||||
|
||||
using var timer = _metrics.Measure.Timer.Time(BotMetrics.EventsHandled,
|
||||
new MetricTags("event", typeof(T).Name.Replace("Event", "")));
|
||||
|
||||
// Delegate to the queue to see if it wants to handle this event
|
||||
// the TryHandle call returns true if it's handled the event
|
||||
// Usually it won't, so just pass it on to the main handler
|
||||
if (queue == null || !await queue.TryHandle(evt))
|
||||
await handler.Handle(shard, evt);
|
||||
}
|
||||
catch (WebSocketException)
|
||||
catch (Exception exc)
|
||||
{
|
||||
// TODO: this still thrown?
|
||||
await HandleError(shard, handler, evt, serviceScope, exc);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
private async Task HandleError<T>(Shard shard, IEventHandler<T> handler, T evt, ILifetimeScope serviceScope,
|
||||
Exception exc)
|
||||
where T : IGatewayEvent
|
||||
{
|
||||
_metrics.Measure.Meter.Mark(BotMetrics.BotErrors, exc.GetType().FullName);
|
||||
|
||||
// Make this beforehand so we can access the event ID for logging
|
||||
var sentryEvent = new SentryEvent(exc);
|
||||
|
||||
// If the event is us responding to our own error messages, don't bother logging
|
||||
if (evt is MessageCreateEvent mc && mc.Author.Id == shard.User?.Id)
|
||||
return;
|
||||
|
||||
var shouldReport = exc.IsOurProblem();
|
||||
if (shouldReport)
|
||||
{
|
||||
// only log exceptions if they're our problem
|
||||
_logger.Error(exc, "Exception in event handler: {SentryEventId}", sentryEvent.EventId);
|
||||
|
||||
// Report error to Sentry
|
||||
// This will just no-op if there's no URL set
|
||||
var sentryScope = serviceScope.Resolve<Scope>();
|
||||
|
||||
// Add some specific info about Discord error responses, as a breadcrumb
|
||||
// TODO: headers to dict
|
||||
// if (exc is BadRequestException bre)
|
||||
// sentryScope.AddBreadcrumb(bre.Response, "response.error", data: new Dictionary<string, string>(bre.Response.Headers));
|
||||
// if (exc is NotFoundException nfe)
|
||||
// sentryScope.AddBreadcrumb(nfe.Response, "response.error", data: new Dictionary<string, string>(nfe.Response.Headers));
|
||||
// if (exc is UnauthorizedException ue)
|
||||
// sentryScope.AddBreadcrumb(ue.Response, "response.error", data: new Dictionary<string, string>(ue.Response.Headers));
|
||||
|
||||
SentrySdk.CaptureEvent(sentryEvent, sentryScope);
|
||||
|
||||
// most of these errors aren't useful...
|
||||
if (_config.DisableErrorReporting)
|
||||
return;
|
||||
|
||||
// Once we've sent it to Sentry, report it to the user (if we have permission to)
|
||||
var reportChannel = handler.ErrorChannelFor(evt);
|
||||
if (reportChannel == null)
|
||||
return;
|
||||
|
||||
var botPerms = await _cache.PermissionsIn(reportChannel.Value);
|
||||
if (botPerms.HasFlag(PermissionSet.SendMessages | PermissionSet.EmbedLinks))
|
||||
await _errorMessageService.SendErrorMessage(reportChannel.Value, sentryEvent.EventId.ToString());
|
||||
}
|
||||
}
|
||||
|
||||
private async Task UpdatePeriodic()
|
||||
{
|
||||
_logger.Debug("Running once-per-minute scheduled tasks");
|
||||
|
||||
await UpdateBotStatus();
|
||||
|
||||
// Collect some stats, submit them to the metrics backend
|
||||
await _collector.CollectStats();
|
||||
await Task.WhenAll(((IMetricsRoot)_metrics).ReportRunner.RunAllAsync());
|
||||
_logger.Debug("Submitted metrics to backend");
|
||||
}
|
||||
|
||||
private async Task UpdateBotStatus(Shard specificShard = null)
|
||||
{
|
||||
// If we're not on any shards, don't bother (this happens if the periodic timer fires before the first Ready)
|
||||
if (!_hasReceivedReady) return;
|
||||
|
||||
var totalGuilds = await _cache.GetAllGuilds().CountAsync();
|
||||
|
||||
try // DiscordClient may throw an exception if the socket is closed (e.g just after OP 7 received)
|
||||
{
|
||||
Task UpdateStatus(Shard shard) =>
|
||||
shard.UpdateStatus(new GatewayStatusUpdate
|
||||
{
|
||||
Activities = new[]
|
||||
{
|
||||
new ActivityPartial
|
||||
{
|
||||
Name = $"pk;help | in {totalGuilds:N0} servers | shard #{shard.ShardId}",
|
||||
Type = ActivityType.Game,
|
||||
Url = "https://pluralkit.me/"
|
||||
}
|
||||
}
|
||||
});
|
||||
|
||||
if (specificShard != null)
|
||||
await UpdateStatus(specificShard);
|
||||
else // Run shard updates concurrently
|
||||
await Task.WhenAll(_cluster.Shards.Values.Select(UpdateStatus));
|
||||
}
|
||||
catch (WebSocketException)
|
||||
{
|
||||
// TODO: this still thrown?
|
||||
}
|
||||
}
|
||||
}
|
||||
Reference in New Issue
Block a user