mirror of
1
Fork 0
gotosocial/internal/db/bundb/conversation.go

566 lines
18 KiB
Go
Raw Normal View History

[feature] Conversations API (#3013) * Implement conversations API * Sort and page conversations by last status ID * Appease linter * Fix deleting conversations and statuses * Refactor to make migrations automatic * Lint * Update tests post-merge * Fixes from live-fire testing * Linter caught a format problem * Refactor tests, fix cache * Negative test for non-DMs * Run conversations advanced migration on testrig startup as well as regular server startup * Document (lack of) side effects of API method for deleting a conversation * Make not-found check less nested for readability * Rename PutConversation to UpsertConversation * Use util.Ptr instead of IIFE * Reduce cache used by conversations * Remove unnecessary TableExpr/ColumnExpr * Use struct tags for both unique constraints on Conversation * Make it clear how paging with GetDirectStatusIDsBatch should be used * Let conversation paging skip conversations it can't render * Use Bun NewDropTable * Convert delete raw query to Bun * Convert update raw query to Bun * Convert latestConversationStatusesTempTable raw query partially to Bun * Convert conversationStatusesTempTable raw query partially to Bun * Rename field used to store result of MaxDirectStatusID * Move advanced migrations to their own tiny processor * Catch up util function name with main * Remove json.… wrappers * Remove redundant check * Combine error checks * Replace map with slice of structs * Address processor/type converter comments - Add context info for errors - Extract some common processor code into shared methods - Move conversation eligibility check ahead of populating conversation * Add error context when dropping temp tables
2024-07-23 21:44:31 +02:00
// GoToSocial
// Copyright (C) GoToSocial Authors admin@gotosocial.org
// SPDX-License-Identifier: AGPL-3.0-or-later
//
// This program is free software: you can redistribute it and/or modify
// it under the terms of the GNU Affero General Public License as published by
// the Free Software Foundation, either version 3 of the License, or
// (at your option) any later version.
//
// This program is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
// GNU Affero General Public License for more details.
//
// You should have received a copy of the GNU Affero General Public License
// along with this program. If not, see <http://www.gnu.org/licenses/>.
package bundb
import (
"context"
"errors"
"slices"
"time"
[feature] Conversations API (#3013) * Implement conversations API * Sort and page conversations by last status ID * Appease linter * Fix deleting conversations and statuses * Refactor to make migrations automatic * Lint * Update tests post-merge * Fixes from live-fire testing * Linter caught a format problem * Refactor tests, fix cache * Negative test for non-DMs * Run conversations advanced migration on testrig startup as well as regular server startup * Document (lack of) side effects of API method for deleting a conversation * Make not-found check less nested for readability * Rename PutConversation to UpsertConversation * Use util.Ptr instead of IIFE * Reduce cache used by conversations * Remove unnecessary TableExpr/ColumnExpr * Use struct tags for both unique constraints on Conversation * Make it clear how paging with GetDirectStatusIDsBatch should be used * Let conversation paging skip conversations it can't render * Use Bun NewDropTable * Convert delete raw query to Bun * Convert update raw query to Bun * Convert latestConversationStatusesTempTable raw query partially to Bun * Convert conversationStatusesTempTable raw query partially to Bun * Rename field used to store result of MaxDirectStatusID * Move advanced migrations to their own tiny processor * Catch up util function name with main * Remove json.… wrappers * Remove redundant check * Combine error checks * Replace map with slice of structs * Address processor/type converter comments - Add context info for errors - Extract some common processor code into shared methods - Move conversation eligibility check ahead of populating conversation * Add error context when dropping temp tables
2024-07-23 21:44:31 +02:00
"github.com/superseriousbusiness/gotosocial/internal/db"
"github.com/superseriousbusiness/gotosocial/internal/gtscontext"
"github.com/superseriousbusiness/gotosocial/internal/gtserror"
"github.com/superseriousbusiness/gotosocial/internal/gtsmodel"
"github.com/superseriousbusiness/gotosocial/internal/id"
"github.com/superseriousbusiness/gotosocial/internal/log"
"github.com/superseriousbusiness/gotosocial/internal/paging"
"github.com/superseriousbusiness/gotosocial/internal/state"
"github.com/superseriousbusiness/gotosocial/internal/util"
"github.com/uptrace/bun"
"github.com/uptrace/bun/dialect"
)
type conversationDB struct {
db *bun.DB
state *state.State
}
func (c *conversationDB) GetConversationByID(ctx context.Context, id string) (*gtsmodel.Conversation, error) {
return c.getConversation(
ctx,
"ID",
func(conversation *gtsmodel.Conversation) error {
return c.db.
NewSelect().
Model(conversation).
Where("? = ?", bun.Ident("id"), id).
Scan(ctx)
},
id,
)
}
func (c *conversationDB) GetConversationByThreadAndAccountIDs(ctx context.Context, threadID string, accountID string, otherAccountIDs []string) (*gtsmodel.Conversation, error) {
otherAccountsKey := gtsmodel.ConversationOtherAccountsKey(otherAccountIDs)
return c.getConversation(
ctx,
"ThreadID,AccountID,OtherAccountsKey",
func(conversation *gtsmodel.Conversation) error {
return c.db.
NewSelect().
Model(conversation).
Where("? = ?", bun.Ident("thread_id"), threadID).
Where("? = ?", bun.Ident("account_id"), accountID).
Where("? = ?", bun.Ident("other_accounts_key"), otherAccountsKey).
Scan(ctx)
},
threadID,
accountID,
otherAccountsKey,
)
}
func (c *conversationDB) getConversation(
ctx context.Context,
lookup string,
dbQuery func(conversation *gtsmodel.Conversation) error,
keyParts ...any,
) (*gtsmodel.Conversation, error) {
// Fetch conversation from cache with loader callback
conversation, err := c.state.Caches.DB.Conversation.LoadOne(lookup, func() (*gtsmodel.Conversation, error) {
[feature] Conversations API (#3013) * Implement conversations API * Sort and page conversations by last status ID * Appease linter * Fix deleting conversations and statuses * Refactor to make migrations automatic * Lint * Update tests post-merge * Fixes from live-fire testing * Linter caught a format problem * Refactor tests, fix cache * Negative test for non-DMs * Run conversations advanced migration on testrig startup as well as regular server startup * Document (lack of) side effects of API method for deleting a conversation * Make not-found check less nested for readability * Rename PutConversation to UpsertConversation * Use util.Ptr instead of IIFE * Reduce cache used by conversations * Remove unnecessary TableExpr/ColumnExpr * Use struct tags for both unique constraints on Conversation * Make it clear how paging with GetDirectStatusIDsBatch should be used * Let conversation paging skip conversations it can't render * Use Bun NewDropTable * Convert delete raw query to Bun * Convert update raw query to Bun * Convert latestConversationStatusesTempTable raw query partially to Bun * Convert conversationStatusesTempTable raw query partially to Bun * Rename field used to store result of MaxDirectStatusID * Move advanced migrations to their own tiny processor * Catch up util function name with main * Remove json.… wrappers * Remove redundant check * Combine error checks * Replace map with slice of structs * Address processor/type converter comments - Add context info for errors - Extract some common processor code into shared methods - Move conversation eligibility check ahead of populating conversation * Add error context when dropping temp tables
2024-07-23 21:44:31 +02:00
var conversation gtsmodel.Conversation
// Not cached! Perform database query
if err := dbQuery(&conversation); err != nil {
return nil, err
}
return &conversation, nil
}, keyParts...)
if err != nil {
// already processe
return nil, err
}
if gtscontext.Barebones(ctx) {
// Only a barebones model was requested.
return conversation, nil
}
if err := c.populateConversation(ctx, conversation); err != nil {
return nil, err
}
return conversation, nil
}
func (c *conversationDB) populateConversation(ctx context.Context, conversation *gtsmodel.Conversation) error {
var (
errs gtserror.MultiError
err error
)
if conversation.Account == nil {
conversation.Account, err = c.state.DB.GetAccountByID(
gtscontext.SetBarebones(ctx),
conversation.AccountID,
)
if err != nil {
errs.Appendf("error populating conversation owner account: %w", err)
}
}
if conversation.OtherAccounts == nil {
conversation.OtherAccounts, err = c.state.DB.GetAccountsByIDs(
gtscontext.SetBarebones(ctx),
conversation.OtherAccountIDs,
)
if err != nil {
errs.Appendf("error populating other conversation accounts: %w", err)
}
}
if conversation.LastStatus == nil && conversation.LastStatusID != "" {
conversation.LastStatus, err = c.state.DB.GetStatusByID(
gtscontext.SetBarebones(ctx),
conversation.LastStatusID,
)
if err != nil {
errs.Appendf("error populating conversation last status: %w", err)
}
}
return errs.Combine()
}
func (c *conversationDB) GetConversationsByOwnerAccountID(ctx context.Context, accountID string, page *paging.Page) ([]*gtsmodel.Conversation, error) {
conversationLastStatusIDs, err := c.getAccountConversationLastStatusIDs(ctx, accountID, page)
if err != nil {
return nil, err
}
return c.getConversationsByLastStatusIDs(ctx, accountID, conversationLastStatusIDs)
}
func (c *conversationDB) getAccountConversationLastStatusIDs(ctx context.Context, accountID string, page *paging.Page) ([]string, error) {
return loadPagedIDs(&c.state.Caches.DB.ConversationLastStatusIDs, accountID, page, func() ([]string, error) {
[feature] Conversations API (#3013) * Implement conversations API * Sort and page conversations by last status ID * Appease linter * Fix deleting conversations and statuses * Refactor to make migrations automatic * Lint * Update tests post-merge * Fixes from live-fire testing * Linter caught a format problem * Refactor tests, fix cache * Negative test for non-DMs * Run conversations advanced migration on testrig startup as well as regular server startup * Document (lack of) side effects of API method for deleting a conversation * Make not-found check less nested for readability * Rename PutConversation to UpsertConversation * Use util.Ptr instead of IIFE * Reduce cache used by conversations * Remove unnecessary TableExpr/ColumnExpr * Use struct tags for both unique constraints on Conversation * Make it clear how paging with GetDirectStatusIDsBatch should be used * Let conversation paging skip conversations it can't render * Use Bun NewDropTable * Convert delete raw query to Bun * Convert update raw query to Bun * Convert latestConversationStatusesTempTable raw query partially to Bun * Convert conversationStatusesTempTable raw query partially to Bun * Rename field used to store result of MaxDirectStatusID * Move advanced migrations to their own tiny processor * Catch up util function name with main * Remove json.… wrappers * Remove redundant check * Combine error checks * Replace map with slice of structs * Address processor/type converter comments - Add context info for errors - Extract some common processor code into shared methods - Move conversation eligibility check ahead of populating conversation * Add error context when dropping temp tables
2024-07-23 21:44:31 +02:00
var conversationLastStatusIDs []string
// Conversation last status IDs not in cache. Perform DB query.
if _, err := c.db.
NewSelect().
Model((*gtsmodel.Conversation)(nil)).
Column("last_status_id").
Where("? = ?", bun.Ident("account_id"), accountID).
OrderExpr("? DESC", bun.Ident("last_status_id")).
Exec(ctx, &conversationLastStatusIDs); // nocollapse
err != nil && !errors.Is(err, db.ErrNoEntries) {
return nil, err
}
return conversationLastStatusIDs, nil
})
}
func (c *conversationDB) getConversationsByLastStatusIDs(
ctx context.Context,
accountID string,
conversationLastStatusIDs []string,
) ([]*gtsmodel.Conversation, error) {
// Load all conversation IDs via cache loader callbacks.
conversations, err := c.state.Caches.DB.Conversation.LoadIDs2Part(
[feature] Conversations API (#3013) * Implement conversations API * Sort and page conversations by last status ID * Appease linter * Fix deleting conversations and statuses * Refactor to make migrations automatic * Lint * Update tests post-merge * Fixes from live-fire testing * Linter caught a format problem * Refactor tests, fix cache * Negative test for non-DMs * Run conversations advanced migration on testrig startup as well as regular server startup * Document (lack of) side effects of API method for deleting a conversation * Make not-found check less nested for readability * Rename PutConversation to UpsertConversation * Use util.Ptr instead of IIFE * Reduce cache used by conversations * Remove unnecessary TableExpr/ColumnExpr * Use struct tags for both unique constraints on Conversation * Make it clear how paging with GetDirectStatusIDsBatch should be used * Let conversation paging skip conversations it can't render * Use Bun NewDropTable * Convert delete raw query to Bun * Convert update raw query to Bun * Convert latestConversationStatusesTempTable raw query partially to Bun * Convert conversationStatusesTempTable raw query partially to Bun * Rename field used to store result of MaxDirectStatusID * Move advanced migrations to their own tiny processor * Catch up util function name with main * Remove json.… wrappers * Remove redundant check * Combine error checks * Replace map with slice of structs * Address processor/type converter comments - Add context info for errors - Extract some common processor code into shared methods - Move conversation eligibility check ahead of populating conversation * Add error context when dropping temp tables
2024-07-23 21:44:31 +02:00
"AccountID,LastStatusID",
accountID,
conversationLastStatusIDs,
func(accountID string, uncached []string) ([]*gtsmodel.Conversation, error) {
// Preallocate expected length of uncached conversations.
[performance] cache more database calls, reduce required database calls overall (#3290) * improvements to caching for lists and relationship to accounts / follows * fix nil panic in AddToList() * ensure list related caches are correctly invalidated * ensure returned ID lists are ordered correctly * bump go-structr to v0.8.9 (returns early if zero uncached keys to be loaded) * remove zero checks in uncached key load functions (go-structr now handles this) * fix issues after rebase on upstream/main * update the expected return order of CSV exports (since list entries are now down by entry creation date) * rename some funcs, allow deleting list entries for multiple follow IDs at a time, fix up more tests * use returning statements on delete to get cache invalidation info * fixes to recent database delete changes * fix broken list entries delete sql * remove unused db function * update remainder of delete functions to behave in similar way, some other small tweaks * fix delete user sql, allow returning on err no entries * uncomment + fix list database tests * update remaining list tests * update envparsing test * add comments to each specific key being invalidated * add more cache invalidation explanatory comments * whoops; actually delete poll votes from database in the DeletePollByID() func * remove added but-commented-out field * improved comment regarding paging being disabled * make cache invalidation comments match what's actually happening * fix up delete query comments to match what is happening * rename function to read a bit better * don't use ErrNoEntries on delete when not needed (it's only needed for a RETURNING call) * update function name in test * move list exclusivity check to AFTER eligibility check. use log.Panic() instead of panic() * use the poll_id column in poll_votes for selecting votes in poll ID * fix function name
2024-09-16 18:46:09 +02:00
conversations := make([]*gtsmodel.Conversation, 0, len(uncached))
[feature] Conversations API (#3013) * Implement conversations API * Sort and page conversations by last status ID * Appease linter * Fix deleting conversations and statuses * Refactor to make migrations automatic * Lint * Update tests post-merge * Fixes from live-fire testing * Linter caught a format problem * Refactor tests, fix cache * Negative test for non-DMs * Run conversations advanced migration on testrig startup as well as regular server startup * Document (lack of) side effects of API method for deleting a conversation * Make not-found check less nested for readability * Rename PutConversation to UpsertConversation * Use util.Ptr instead of IIFE * Reduce cache used by conversations * Remove unnecessary TableExpr/ColumnExpr * Use struct tags for both unique constraints on Conversation * Make it clear how paging with GetDirectStatusIDsBatch should be used * Let conversation paging skip conversations it can't render * Use Bun NewDropTable * Convert delete raw query to Bun * Convert update raw query to Bun * Convert latestConversationStatusesTempTable raw query partially to Bun * Convert conversationStatusesTempTable raw query partially to Bun * Rename field used to store result of MaxDirectStatusID * Move advanced migrations to their own tiny processor * Catch up util function name with main * Remove json.… wrappers * Remove redundant check * Combine error checks * Replace map with slice of structs * Address processor/type converter comments - Add context info for errors - Extract some common processor code into shared methods - Move conversation eligibility check ahead of populating conversation * Add error context when dropping temp tables
2024-07-23 21:44:31 +02:00
// Perform database query scanning the remaining (uncached) IDs.
if err := c.db.NewSelect().
Model(&conversations).
Where("? = ?", bun.Ident("account_id"), accountID).
Where("? IN (?)", bun.Ident("last_status_id"), bun.In(uncached)).
Scan(ctx); err != nil {
return nil, err
}
return conversations, nil
},
)
if err != nil {
return nil, err
}
// Reorder the conversations by their last status IDs to ensure correct order.
getID := func(b *gtsmodel.Conversation) string { return b.ID }
util.OrderBy(conversations, conversationLastStatusIDs, getID)
if gtscontext.Barebones(ctx) {
// no need to fully populate.
return conversations, nil
}
// Populate all loaded conversations, removing those we fail to populate.
conversations = slices.DeleteFunc(conversations, func(conversation *gtsmodel.Conversation) bool {
if err := c.populateConversation(ctx, conversation); err != nil {
log.Errorf(ctx, "error populating conversation %s: %v", conversation.ID, err)
return true
}
return false
})
return conversations, nil
}
func (c *conversationDB) UpsertConversation(ctx context.Context, conversation *gtsmodel.Conversation, columns ...string) error {
// If we're updating by column, ensure "updated_at" is included.
if len(columns) > 0 {
columns = append(columns, "updated_at")
}
return c.state.Caches.DB.Conversation.Store(conversation, func() error {
[feature] Conversations API (#3013) * Implement conversations API * Sort and page conversations by last status ID * Appease linter * Fix deleting conversations and statuses * Refactor to make migrations automatic * Lint * Update tests post-merge * Fixes from live-fire testing * Linter caught a format problem * Refactor tests, fix cache * Negative test for non-DMs * Run conversations advanced migration on testrig startup as well as regular server startup * Document (lack of) side effects of API method for deleting a conversation * Make not-found check less nested for readability * Rename PutConversation to UpsertConversation * Use util.Ptr instead of IIFE * Reduce cache used by conversations * Remove unnecessary TableExpr/ColumnExpr * Use struct tags for both unique constraints on Conversation * Make it clear how paging with GetDirectStatusIDsBatch should be used * Let conversation paging skip conversations it can't render * Use Bun NewDropTable * Convert delete raw query to Bun * Convert update raw query to Bun * Convert latestConversationStatusesTempTable raw query partially to Bun * Convert conversationStatusesTempTable raw query partially to Bun * Rename field used to store result of MaxDirectStatusID * Move advanced migrations to their own tiny processor * Catch up util function name with main * Remove json.… wrappers * Remove redundant check * Combine error checks * Replace map with slice of structs * Address processor/type converter comments - Add context info for errors - Extract some common processor code into shared methods - Move conversation eligibility check ahead of populating conversation * Add error context when dropping temp tables
2024-07-23 21:44:31 +02:00
_, err := NewUpsert(c.db).
Model(conversation).
Constraint("id").
Column(columns...).
Exec(ctx)
return err
})
}
func (c *conversationDB) LinkConversationToStatus(ctx context.Context, conversationID string, statusID string) error {
conversationToStatus := &gtsmodel.ConversationToStatus{
ConversationID: conversationID,
StatusID: statusID,
}
if _, err := c.db.NewInsert().
Model(conversationToStatus).
Exec(ctx); // nocollapse
err != nil {
return err
}
return nil
}
func (c *conversationDB) DeleteConversationByID(ctx context.Context, id string) error {
[performance] cache more database calls, reduce required database calls overall (#3290) * improvements to caching for lists and relationship to accounts / follows * fix nil panic in AddToList() * ensure list related caches are correctly invalidated * ensure returned ID lists are ordered correctly * bump go-structr to v0.8.9 (returns early if zero uncached keys to be loaded) * remove zero checks in uncached key load functions (go-structr now handles this) * fix issues after rebase on upstream/main * update the expected return order of CSV exports (since list entries are now down by entry creation date) * rename some funcs, allow deleting list entries for multiple follow IDs at a time, fix up more tests * use returning statements on delete to get cache invalidation info * fixes to recent database delete changes * fix broken list entries delete sql * remove unused db function * update remainder of delete functions to behave in similar way, some other small tweaks * fix delete user sql, allow returning on err no entries * uncomment + fix list database tests * update remaining list tests * update envparsing test * add comments to each specific key being invalidated * add more cache invalidation explanatory comments * whoops; actually delete poll votes from database in the DeletePollByID() func * remove added but-commented-out field * improved comment regarding paging being disabled * make cache invalidation comments match what's actually happening * fix up delete query comments to match what is happening * rename function to read a bit better * don't use ErrNoEntries on delete when not needed (it's only needed for a RETURNING call) * update function name in test * move list exclusivity check to AFTER eligibility check. use log.Panic() instead of panic() * use the poll_id column in poll_votes for selecting votes in poll ID * fix function name
2024-09-16 18:46:09 +02:00
// Gather necessary fields from
// deleted for cache invaliation.
var deleted gtsmodel.Conversation
deleted.ID = id
// Delete conversation from DB.
if _, err := c.db.NewDelete().
Model(&deleted).
Where("? = ?", bun.Ident("id"), id).
Returning("?", bun.Ident("account_id")).
Exec(ctx); err != nil &&
!errors.Is(err, db.ErrNoEntries) {
[feature] Conversations API (#3013) * Implement conversations API * Sort and page conversations by last status ID * Appease linter * Fix deleting conversations and statuses * Refactor to make migrations automatic * Lint * Update tests post-merge * Fixes from live-fire testing * Linter caught a format problem * Refactor tests, fix cache * Negative test for non-DMs * Run conversations advanced migration on testrig startup as well as regular server startup * Document (lack of) side effects of API method for deleting a conversation * Make not-found check less nested for readability * Rename PutConversation to UpsertConversation * Use util.Ptr instead of IIFE * Reduce cache used by conversations * Remove unnecessary TableExpr/ColumnExpr * Use struct tags for both unique constraints on Conversation * Make it clear how paging with GetDirectStatusIDsBatch should be used * Let conversation paging skip conversations it can't render * Use Bun NewDropTable * Convert delete raw query to Bun * Convert update raw query to Bun * Convert latestConversationStatusesTempTable raw query partially to Bun * Convert conversationStatusesTempTable raw query partially to Bun * Rename field used to store result of MaxDirectStatusID * Move advanced migrations to their own tiny processor * Catch up util function name with main * Remove json.… wrappers * Remove redundant check * Combine error checks * Replace map with slice of structs * Address processor/type converter comments - Add context info for errors - Extract some common processor code into shared methods - Move conversation eligibility check ahead of populating conversation * Add error context when dropping temp tables
2024-07-23 21:44:31 +02:00
return err
}
[performance] cache more database calls, reduce required database calls overall (#3290) * improvements to caching for lists and relationship to accounts / follows * fix nil panic in AddToList() * ensure list related caches are correctly invalidated * ensure returned ID lists are ordered correctly * bump go-structr to v0.8.9 (returns early if zero uncached keys to be loaded) * remove zero checks in uncached key load functions (go-structr now handles this) * fix issues after rebase on upstream/main * update the expected return order of CSV exports (since list entries are now down by entry creation date) * rename some funcs, allow deleting list entries for multiple follow IDs at a time, fix up more tests * use returning statements on delete to get cache invalidation info * fixes to recent database delete changes * fix broken list entries delete sql * remove unused db function * update remainder of delete functions to behave in similar way, some other small tweaks * fix delete user sql, allow returning on err no entries * uncomment + fix list database tests * update remaining list tests * update envparsing test * add comments to each specific key being invalidated * add more cache invalidation explanatory comments * whoops; actually delete poll votes from database in the DeletePollByID() func * remove added but-commented-out field * improved comment regarding paging being disabled * make cache invalidation comments match what's actually happening * fix up delete query comments to match what is happening * rename function to read a bit better * don't use ErrNoEntries on delete when not needed (it's only needed for a RETURNING call) * update function name in test * move list exclusivity check to AFTER eligibility check. use log.Panic() instead of panic() * use the poll_id column in poll_votes for selecting votes in poll ID * fix function name
2024-09-16 18:46:09 +02:00
// Invalidate cached conversation by ID,
// manually invalidate hook in case not cached.
c.state.Caches.DB.Conversation.Invalidate("ID", id)
c.state.Caches.OnInvalidateConversation(&deleted)
[feature] Conversations API (#3013) * Implement conversations API * Sort and page conversations by last status ID * Appease linter * Fix deleting conversations and statuses * Refactor to make migrations automatic * Lint * Update tests post-merge * Fixes from live-fire testing * Linter caught a format problem * Refactor tests, fix cache * Negative test for non-DMs * Run conversations advanced migration on testrig startup as well as regular server startup * Document (lack of) side effects of API method for deleting a conversation * Make not-found check less nested for readability * Rename PutConversation to UpsertConversation * Use util.Ptr instead of IIFE * Reduce cache used by conversations * Remove unnecessary TableExpr/ColumnExpr * Use struct tags for both unique constraints on Conversation * Make it clear how paging with GetDirectStatusIDsBatch should be used * Let conversation paging skip conversations it can't render * Use Bun NewDropTable * Convert delete raw query to Bun * Convert update raw query to Bun * Convert latestConversationStatusesTempTable raw query partially to Bun * Convert conversationStatusesTempTable raw query partially to Bun * Rename field used to store result of MaxDirectStatusID * Move advanced migrations to their own tiny processor * Catch up util function name with main * Remove json.… wrappers * Remove redundant check * Combine error checks * Replace map with slice of structs * Address processor/type converter comments - Add context info for errors - Extract some common processor code into shared methods - Move conversation eligibility check ahead of populating conversation * Add error context when dropping temp tables
2024-07-23 21:44:31 +02:00
[performance] cache more database calls, reduce required database calls overall (#3290) * improvements to caching for lists and relationship to accounts / follows * fix nil panic in AddToList() * ensure list related caches are correctly invalidated * ensure returned ID lists are ordered correctly * bump go-structr to v0.8.9 (returns early if zero uncached keys to be loaded) * remove zero checks in uncached key load functions (go-structr now handles this) * fix issues after rebase on upstream/main * update the expected return order of CSV exports (since list entries are now down by entry creation date) * rename some funcs, allow deleting list entries for multiple follow IDs at a time, fix up more tests * use returning statements on delete to get cache invalidation info * fixes to recent database delete changes * fix broken list entries delete sql * remove unused db function * update remainder of delete functions to behave in similar way, some other small tweaks * fix delete user sql, allow returning on err no entries * uncomment + fix list database tests * update remaining list tests * update envparsing test * add comments to each specific key being invalidated * add more cache invalidation explanatory comments * whoops; actually delete poll votes from database in the DeletePollByID() func * remove added but-commented-out field * improved comment regarding paging being disabled * make cache invalidation comments match what's actually happening * fix up delete query comments to match what is happening * rename function to read a bit better * don't use ErrNoEntries on delete when not needed (it's only needed for a RETURNING call) * update function name in test * move list exclusivity check to AFTER eligibility check. use log.Panic() instead of panic() * use the poll_id column in poll_votes for selecting votes in poll ID * fix function name
2024-09-16 18:46:09 +02:00
return nil
[feature] Conversations API (#3013) * Implement conversations API * Sort and page conversations by last status ID * Appease linter * Fix deleting conversations and statuses * Refactor to make migrations automatic * Lint * Update tests post-merge * Fixes from live-fire testing * Linter caught a format problem * Refactor tests, fix cache * Negative test for non-DMs * Run conversations advanced migration on testrig startup as well as regular server startup * Document (lack of) side effects of API method for deleting a conversation * Make not-found check less nested for readability * Rename PutConversation to UpsertConversation * Use util.Ptr instead of IIFE * Reduce cache used by conversations * Remove unnecessary TableExpr/ColumnExpr * Use struct tags for both unique constraints on Conversation * Make it clear how paging with GetDirectStatusIDsBatch should be used * Let conversation paging skip conversations it can't render * Use Bun NewDropTable * Convert delete raw query to Bun * Convert update raw query to Bun * Convert latestConversationStatusesTempTable raw query partially to Bun * Convert conversationStatusesTempTable raw query partially to Bun * Rename field used to store result of MaxDirectStatusID * Move advanced migrations to their own tiny processor * Catch up util function name with main * Remove json.… wrappers * Remove redundant check * Combine error checks * Replace map with slice of structs * Address processor/type converter comments - Add context info for errors - Extract some common processor code into shared methods - Move conversation eligibility check ahead of populating conversation * Add error context when dropping temp tables
2024-07-23 21:44:31 +02:00
}
func (c *conversationDB) DeleteConversationsByOwnerAccountID(ctx context.Context, accountID string) error {
defer func() {
// Invalidate any cached conversations and conversation IDs owned by this account on return.
// Conversation invalidate hooks only invalidate the conversation ID cache,
// so we don't need to load all conversations into the cache to run invalidation hooks,
// as with some other object types (blocks, for example).
c.state.Caches.DB.Conversation.Invalidate("AccountID", accountID)
[feature] Conversations API (#3013) * Implement conversations API * Sort and page conversations by last status ID * Appease linter * Fix deleting conversations and statuses * Refactor to make migrations automatic * Lint * Update tests post-merge * Fixes from live-fire testing * Linter caught a format problem * Refactor tests, fix cache * Negative test for non-DMs * Run conversations advanced migration on testrig startup as well as regular server startup * Document (lack of) side effects of API method for deleting a conversation * Make not-found check less nested for readability * Rename PutConversation to UpsertConversation * Use util.Ptr instead of IIFE * Reduce cache used by conversations * Remove unnecessary TableExpr/ColumnExpr * Use struct tags for both unique constraints on Conversation * Make it clear how paging with GetDirectStatusIDsBatch should be used * Let conversation paging skip conversations it can't render * Use Bun NewDropTable * Convert delete raw query to Bun * Convert update raw query to Bun * Convert latestConversationStatusesTempTable raw query partially to Bun * Convert conversationStatusesTempTable raw query partially to Bun * Rename field used to store result of MaxDirectStatusID * Move advanced migrations to their own tiny processor * Catch up util function name with main * Remove json.… wrappers * Remove redundant check * Combine error checks * Replace map with slice of structs * Address processor/type converter comments - Add context info for errors - Extract some common processor code into shared methods - Move conversation eligibility check ahead of populating conversation * Add error context when dropping temp tables
2024-07-23 21:44:31 +02:00
// In case there were no cached conversations,
// explicitly invalidate the user's conversation last status ID cache.
c.state.Caches.DB.ConversationLastStatusIDs.Invalidate(accountID)
[feature] Conversations API (#3013) * Implement conversations API * Sort and page conversations by last status ID * Appease linter * Fix deleting conversations and statuses * Refactor to make migrations automatic * Lint * Update tests post-merge * Fixes from live-fire testing * Linter caught a format problem * Refactor tests, fix cache * Negative test for non-DMs * Run conversations advanced migration on testrig startup as well as regular server startup * Document (lack of) side effects of API method for deleting a conversation * Make not-found check less nested for readability * Rename PutConversation to UpsertConversation * Use util.Ptr instead of IIFE * Reduce cache used by conversations * Remove unnecessary TableExpr/ColumnExpr * Use struct tags for both unique constraints on Conversation * Make it clear how paging with GetDirectStatusIDsBatch should be used * Let conversation paging skip conversations it can't render * Use Bun NewDropTable * Convert delete raw query to Bun * Convert update raw query to Bun * Convert latestConversationStatusesTempTable raw query partially to Bun * Convert conversationStatusesTempTable raw query partially to Bun * Rename field used to store result of MaxDirectStatusID * Move advanced migrations to their own tiny processor * Catch up util function name with main * Remove json.… wrappers * Remove redundant check * Combine error checks * Replace map with slice of structs * Address processor/type converter comments - Add context info for errors - Extract some common processor code into shared methods - Move conversation eligibility check ahead of populating conversation * Add error context when dropping temp tables
2024-07-23 21:44:31 +02:00
}()
return c.db.RunInTx(ctx, nil, func(ctx context.Context, tx bun.Tx) error {
// Delete conversations matching the account ID.
deletedConversationIDs := []string{}
if err := tx.NewDelete().
Model((*gtsmodel.Conversation)(nil)).
Where("? = ?", bun.Ident("account_id"), accountID).
Returning("?", bun.Ident("id")).
Scan(ctx, &deletedConversationIDs); // nocollapse
err != nil {
return gtserror.Newf("error deleting conversations for account %s: %w", accountID, err)
}
if len(deletedConversationIDs) == 0 {
// Nothing
// to delete.
return nil
}
// Delete any conversation-to-status links
// matching the deleted conversation IDs.
[feature] Conversations API (#3013) * Implement conversations API * Sort and page conversations by last status ID * Appease linter * Fix deleting conversations and statuses * Refactor to make migrations automatic * Lint * Update tests post-merge * Fixes from live-fire testing * Linter caught a format problem * Refactor tests, fix cache * Negative test for non-DMs * Run conversations advanced migration on testrig startup as well as regular server startup * Document (lack of) side effects of API method for deleting a conversation * Make not-found check less nested for readability * Rename PutConversation to UpsertConversation * Use util.Ptr instead of IIFE * Reduce cache used by conversations * Remove unnecessary TableExpr/ColumnExpr * Use struct tags for both unique constraints on Conversation * Make it clear how paging with GetDirectStatusIDsBatch should be used * Let conversation paging skip conversations it can't render * Use Bun NewDropTable * Convert delete raw query to Bun * Convert update raw query to Bun * Convert latestConversationStatusesTempTable raw query partially to Bun * Convert conversationStatusesTempTable raw query partially to Bun * Rename field used to store result of MaxDirectStatusID * Move advanced migrations to their own tiny processor * Catch up util function name with main * Remove json.… wrappers * Remove redundant check * Combine error checks * Replace map with slice of structs * Address processor/type converter comments - Add context info for errors - Extract some common processor code into shared methods - Move conversation eligibility check ahead of populating conversation * Add error context when dropping temp tables
2024-07-23 21:44:31 +02:00
if _, err := tx.NewDelete().
Model((*gtsmodel.ConversationToStatus)(nil)).
Where("? IN (?)", bun.Ident("conversation_id"), bun.In(deletedConversationIDs)).
Exec(ctx); // nocollapse
err != nil {
return gtserror.Newf("error deleting conversation-to-status links for account %s: %w", accountID, err)
}
return nil
})
}
func (c *conversationDB) DeleteStatusFromConversations(ctx context.Context, statusID string) error {
var (
updatedConversationIDs = []string{}
deletedConversationIDs = []string{}
[feature] Conversations API (#3013) * Implement conversations API * Sort and page conversations by last status ID * Appease linter * Fix deleting conversations and statuses * Refactor to make migrations automatic * Lint * Update tests post-merge * Fixes from live-fire testing * Linter caught a format problem * Refactor tests, fix cache * Negative test for non-DMs * Run conversations advanced migration on testrig startup as well as regular server startup * Document (lack of) side effects of API method for deleting a conversation * Make not-found check less nested for readability * Rename PutConversation to UpsertConversation * Use util.Ptr instead of IIFE * Reduce cache used by conversations * Remove unnecessary TableExpr/ColumnExpr * Use struct tags for both unique constraints on Conversation * Make it clear how paging with GetDirectStatusIDsBatch should be used * Let conversation paging skip conversations it can't render * Use Bun NewDropTable * Convert delete raw query to Bun * Convert update raw query to Bun * Convert latestConversationStatusesTempTable raw query partially to Bun * Convert conversationStatusesTempTable raw query partially to Bun * Rename field used to store result of MaxDirectStatusID * Move advanced migrations to their own tiny processor * Catch up util function name with main * Remove json.… wrappers * Remove redundant check * Combine error checks * Replace map with slice of structs * Address processor/type converter comments - Add context info for errors - Extract some common processor code into shared methods - Move conversation eligibility check ahead of populating conversation * Add error context when dropping temp tables
2024-07-23 21:44:31 +02:00
// Method of creating + dropping temp
// tables differs depending on driver.
tmpQ string
)
if c.db.Dialect().Name() == dialect.PG {
// On Postgres, we can instruct PG to clean
// up temp tables on commit, so we can just
// use any connection from the pool without
// caring what happens to it when we're done.
tmpQ = "CREATE TEMPORARY TABLE ? ON COMMIT DROP AS (?)"
} else {
// On SQLite, we can't instruct SQLite to drop
// temp tables on commit, and we can't manually
// drop temp tables without triggering a bug.
// So we leave the temp tables alone, in the
// knowledge they'll be cleaned up when this
// connection gets recycled (in max 5min).
tmpQ = "CREATE TEMPORARY TABLE ? AS ?"
}
[feature] Conversations API (#3013) * Implement conversations API * Sort and page conversations by last status ID * Appease linter * Fix deleting conversations and statuses * Refactor to make migrations automatic * Lint * Update tests post-merge * Fixes from live-fire testing * Linter caught a format problem * Refactor tests, fix cache * Negative test for non-DMs * Run conversations advanced migration on testrig startup as well as regular server startup * Document (lack of) side effects of API method for deleting a conversation * Make not-found check less nested for readability * Rename PutConversation to UpsertConversation * Use util.Ptr instead of IIFE * Reduce cache used by conversations * Remove unnecessary TableExpr/ColumnExpr * Use struct tags for both unique constraints on Conversation * Make it clear how paging with GetDirectStatusIDsBatch should be used * Let conversation paging skip conversations it can't render * Use Bun NewDropTable * Convert delete raw query to Bun * Convert update raw query to Bun * Convert latestConversationStatusesTempTable raw query partially to Bun * Convert conversationStatusesTempTable raw query partially to Bun * Rename field used to store result of MaxDirectStatusID * Move advanced migrations to their own tiny processor * Catch up util function name with main * Remove json.… wrappers * Remove redundant check * Combine error checks * Replace map with slice of structs * Address processor/type converter comments - Add context info for errors - Extract some common processor code into shared methods - Move conversation eligibility check ahead of populating conversation * Add error context when dropping temp tables
2024-07-23 21:44:31 +02:00
if err := c.db.RunInTx(ctx, nil, func(ctx context.Context, tx bun.Tx) error {
// First delete this status from
// conversation-to-status links.
_, err := tx.
NewDelete().
Table("conversation_to_statuses").
[feature] Conversations API (#3013) * Implement conversations API * Sort and page conversations by last status ID * Appease linter * Fix deleting conversations and statuses * Refactor to make migrations automatic * Lint * Update tests post-merge * Fixes from live-fire testing * Linter caught a format problem * Refactor tests, fix cache * Negative test for non-DMs * Run conversations advanced migration on testrig startup as well as regular server startup * Document (lack of) side effects of API method for deleting a conversation * Make not-found check less nested for readability * Rename PutConversation to UpsertConversation * Use util.Ptr instead of IIFE * Reduce cache used by conversations * Remove unnecessary TableExpr/ColumnExpr * Use struct tags for both unique constraints on Conversation * Make it clear how paging with GetDirectStatusIDsBatch should be used * Let conversation paging skip conversations it can't render * Use Bun NewDropTable * Convert delete raw query to Bun * Convert update raw query to Bun * Convert latestConversationStatusesTempTable raw query partially to Bun * Convert conversationStatusesTempTable raw query partially to Bun * Rename field used to store result of MaxDirectStatusID * Move advanced migrations to their own tiny processor * Catch up util function name with main * Remove json.… wrappers * Remove redundant check * Combine error checks * Replace map with slice of structs * Address processor/type converter comments - Add context info for errors - Extract some common processor code into shared methods - Move conversation eligibility check ahead of populating conversation * Add error context when dropping temp tables
2024-07-23 21:44:31 +02:00
Where("? = ?", bun.Ident("status_id"), statusID).
Exec(ctx)
if err != nil {
return gtserror.Newf(
"error deleting conversation-to-status links while deleting status %s: %w",
statusID, err,
)
[feature] Conversations API (#3013) * Implement conversations API * Sort and page conversations by last status ID * Appease linter * Fix deleting conversations and statuses * Refactor to make migrations automatic * Lint * Update tests post-merge * Fixes from live-fire testing * Linter caught a format problem * Refactor tests, fix cache * Negative test for non-DMs * Run conversations advanced migration on testrig startup as well as regular server startup * Document (lack of) side effects of API method for deleting a conversation * Make not-found check less nested for readability * Rename PutConversation to UpsertConversation * Use util.Ptr instead of IIFE * Reduce cache used by conversations * Remove unnecessary TableExpr/ColumnExpr * Use struct tags for both unique constraints on Conversation * Make it clear how paging with GetDirectStatusIDsBatch should be used * Let conversation paging skip conversations it can't render * Use Bun NewDropTable * Convert delete raw query to Bun * Convert update raw query to Bun * Convert latestConversationStatusesTempTable raw query partially to Bun * Convert conversationStatusesTempTable raw query partially to Bun * Rename field used to store result of MaxDirectStatusID * Move advanced migrations to their own tiny processor * Catch up util function name with main * Remove json.… wrappers * Remove redundant check * Combine error checks * Replace map with slice of structs * Address processor/type converter comments - Add context info for errors - Extract some common processor code into shared methods - Move conversation eligibility check ahead of populating conversation * Add error context when dropping temp tables
2024-07-23 21:44:31 +02:00
}
// Note: Bun doesn't currently support `CREATE TABLE … AS SELECT …`
// so we need to use raw queries to create temporary tables.
// Create a temporary table containing all statuses other than
// the deleted status, in each conversation for which the deleted
// status is the last status, if there are such statuses.
//
// This will produce a query like:
//
// CREATE TEMPORARY TABLE "conversation_statuses_01J78T2AR0YCZ4YR12WSCZ608S"
// AS (
// SELECT
// "conversations"."id" AS "conversation_id",
// "conversation_to_statuses"."status_id" AS "id",
// "statuses"."created_at"
// FROM
// "conversations"
// LEFT JOIN "conversation_to_statuses" ON (
// "conversations"."id" = "conversation_to_statuses"."conversation_id"
// )
// AND (
// "conversation_to_statuses"."status_id" != '01J78T2BQ4TN5S2XSC9VNQ5GBS'
// )
// LEFT JOIN "statuses" ON (
// "conversation_to_statuses"."status_id" = "statuses"."id"
// )
// WHERE
// (
// "conversations"."last_status_id" = '01J78T2BQ4TN5S2XSC9VNQ5GBS'
// )
// )
conversationStatusesTmp := "conversation_statuses_" + id.NewULID()
conversationStatusesTmpQ := tx.NewRaw(
tmpQ,
bun.Ident(conversationStatusesTmp),
[feature] Conversations API (#3013) * Implement conversations API * Sort and page conversations by last status ID * Appease linter * Fix deleting conversations and statuses * Refactor to make migrations automatic * Lint * Update tests post-merge * Fixes from live-fire testing * Linter caught a format problem * Refactor tests, fix cache * Negative test for non-DMs * Run conversations advanced migration on testrig startup as well as regular server startup * Document (lack of) side effects of API method for deleting a conversation * Make not-found check less nested for readability * Rename PutConversation to UpsertConversation * Use util.Ptr instead of IIFE * Reduce cache used by conversations * Remove unnecessary TableExpr/ColumnExpr * Use struct tags for both unique constraints on Conversation * Make it clear how paging with GetDirectStatusIDsBatch should be used * Let conversation paging skip conversations it can't render * Use Bun NewDropTable * Convert delete raw query to Bun * Convert update raw query to Bun * Convert latestConversationStatusesTempTable raw query partially to Bun * Convert conversationStatusesTempTable raw query partially to Bun * Rename field used to store result of MaxDirectStatusID * Move advanced migrations to their own tiny processor * Catch up util function name with main * Remove json.… wrappers * Remove redundant check * Combine error checks * Replace map with slice of structs * Address processor/type converter comments - Add context info for errors - Extract some common processor code into shared methods - Move conversation eligibility check ahead of populating conversation * Add error context when dropping temp tables
2024-07-23 21:44:31 +02:00
tx.NewSelect().
ColumnExpr(
"? AS ?",
bun.Ident("conversations.id"),
bun.Ident("conversation_id"),
).
ColumnExpr(
"? AS ?",
bun.Ident("conversation_to_statuses.status_id"),
bun.Ident("id"),
).
Column("statuses.created_at").
Table("conversations").
Join("LEFT JOIN ?", bun.Ident("conversation_to_statuses")).
JoinOn(
"? = ?",
bun.Ident("conversations.id"),
bun.Ident("conversation_to_statuses.conversation_id"),
).
JoinOn(
"? != ?",
bun.Ident("conversation_to_statuses.status_id"),
statusID,
).
Join("LEFT JOIN ?", bun.Ident("statuses")).
JoinOn(
"? = ?",
bun.Ident("conversation_to_statuses.status_id"),
bun.Ident("statuses.id"),
).
Where(
"? = ?",
bun.Ident("conversations.last_status_id"),
statusID,
),
)
_, err = conversationStatusesTmpQ.Exec(ctx)
if err != nil {
return gtserror.Newf(
"error creating temp table %s while deleting status %s: %w",
conversationStatusesTmp, statusID, err,
)
[feature] Conversations API (#3013) * Implement conversations API * Sort and page conversations by last status ID * Appease linter * Fix deleting conversations and statuses * Refactor to make migrations automatic * Lint * Update tests post-merge * Fixes from live-fire testing * Linter caught a format problem * Refactor tests, fix cache * Negative test for non-DMs * Run conversations advanced migration on testrig startup as well as regular server startup * Document (lack of) side effects of API method for deleting a conversation * Make not-found check less nested for readability * Rename PutConversation to UpsertConversation * Use util.Ptr instead of IIFE * Reduce cache used by conversations * Remove unnecessary TableExpr/ColumnExpr * Use struct tags for both unique constraints on Conversation * Make it clear how paging with GetDirectStatusIDsBatch should be used * Let conversation paging skip conversations it can't render * Use Bun NewDropTable * Convert delete raw query to Bun * Convert update raw query to Bun * Convert latestConversationStatusesTempTable raw query partially to Bun * Convert conversationStatusesTempTable raw query partially to Bun * Rename field used to store result of MaxDirectStatusID * Move advanced migrations to their own tiny processor * Catch up util function name with main * Remove json.… wrappers * Remove redundant check * Combine error checks * Replace map with slice of structs * Address processor/type converter comments - Add context info for errors - Extract some common processor code into shared methods - Move conversation eligibility check ahead of populating conversation * Add error context when dropping temp tables
2024-07-23 21:44:31 +02:00
}
// Create a temporary table with the most recently created
// status in each conversation for which the deleted status
// is the last status, if there is such a status.
//
// This will produce a query like:
//
// CREATE TEMPORARY TABLE "latest_conversation_statuses_01J78T2AR0E46SJSH6C7NRZ7MR"
// AS (
// SELECT
// "conversation_statuses"."conversation_id",
// "conversation_statuses"."id"
// FROM
// "conversation_statuses_01J78T2AR0YCZ4YR12WSCZ608S" AS "conversation_statuses"
// LEFT JOIN "conversation_statuses_01J78T2AR0YCZ4YR12WSCZ608S" AS "later_statuses" ON (
// "conversation_statuses"."conversation_id" = "later_statuses"."conversation_id"
// )
// AND (
// "later_statuses"."created_at" > "conversation_statuses"."created_at"
// )
// WHERE
// ("later_statuses"."id" IS NULL)
// )
latestConversationStatusesTmp := "latest_conversation_statuses_" + id.NewULID()
latestConversationStatusesTmpQ := tx.NewRaw(
tmpQ,
bun.Ident(latestConversationStatusesTmp),
[feature] Conversations API (#3013) * Implement conversations API * Sort and page conversations by last status ID * Appease linter * Fix deleting conversations and statuses * Refactor to make migrations automatic * Lint * Update tests post-merge * Fixes from live-fire testing * Linter caught a format problem * Refactor tests, fix cache * Negative test for non-DMs * Run conversations advanced migration on testrig startup as well as regular server startup * Document (lack of) side effects of API method for deleting a conversation * Make not-found check less nested for readability * Rename PutConversation to UpsertConversation * Use util.Ptr instead of IIFE * Reduce cache used by conversations * Remove unnecessary TableExpr/ColumnExpr * Use struct tags for both unique constraints on Conversation * Make it clear how paging with GetDirectStatusIDsBatch should be used * Let conversation paging skip conversations it can't render * Use Bun NewDropTable * Convert delete raw query to Bun * Convert update raw query to Bun * Convert latestConversationStatusesTempTable raw query partially to Bun * Convert conversationStatusesTempTable raw query partially to Bun * Rename field used to store result of MaxDirectStatusID * Move advanced migrations to their own tiny processor * Catch up util function name with main * Remove json.… wrappers * Remove redundant check * Combine error checks * Replace map with slice of structs * Address processor/type converter comments - Add context info for errors - Extract some common processor code into shared methods - Move conversation eligibility check ahead of populating conversation * Add error context when dropping temp tables
2024-07-23 21:44:31 +02:00
tx.NewSelect().
Column(
"conversation_statuses.conversation_id",
"conversation_statuses.id",
).
TableExpr(
"? AS ?",
bun.Ident(conversationStatusesTmp),
[feature] Conversations API (#3013) * Implement conversations API * Sort and page conversations by last status ID * Appease linter * Fix deleting conversations and statuses * Refactor to make migrations automatic * Lint * Update tests post-merge * Fixes from live-fire testing * Linter caught a format problem * Refactor tests, fix cache * Negative test for non-DMs * Run conversations advanced migration on testrig startup as well as regular server startup * Document (lack of) side effects of API method for deleting a conversation * Make not-found check less nested for readability * Rename PutConversation to UpsertConversation * Use util.Ptr instead of IIFE * Reduce cache used by conversations * Remove unnecessary TableExpr/ColumnExpr * Use struct tags for both unique constraints on Conversation * Make it clear how paging with GetDirectStatusIDsBatch should be used * Let conversation paging skip conversations it can't render * Use Bun NewDropTable * Convert delete raw query to Bun * Convert update raw query to Bun * Convert latestConversationStatusesTempTable raw query partially to Bun * Convert conversationStatusesTempTable raw query partially to Bun * Rename field used to store result of MaxDirectStatusID * Move advanced migrations to their own tiny processor * Catch up util function name with main * Remove json.… wrappers * Remove redundant check * Combine error checks * Replace map with slice of structs * Address processor/type converter comments - Add context info for errors - Extract some common processor code into shared methods - Move conversation eligibility check ahead of populating conversation * Add error context when dropping temp tables
2024-07-23 21:44:31 +02:00
bun.Ident("conversation_statuses"),
).
Join(
"LEFT JOIN ? AS ?",
bun.Ident(conversationStatusesTmp),
[feature] Conversations API (#3013) * Implement conversations API * Sort and page conversations by last status ID * Appease linter * Fix deleting conversations and statuses * Refactor to make migrations automatic * Lint * Update tests post-merge * Fixes from live-fire testing * Linter caught a format problem * Refactor tests, fix cache * Negative test for non-DMs * Run conversations advanced migration on testrig startup as well as regular server startup * Document (lack of) side effects of API method for deleting a conversation * Make not-found check less nested for readability * Rename PutConversation to UpsertConversation * Use util.Ptr instead of IIFE * Reduce cache used by conversations * Remove unnecessary TableExpr/ColumnExpr * Use struct tags for both unique constraints on Conversation * Make it clear how paging with GetDirectStatusIDsBatch should be used * Let conversation paging skip conversations it can't render * Use Bun NewDropTable * Convert delete raw query to Bun * Convert update raw query to Bun * Convert latestConversationStatusesTempTable raw query partially to Bun * Convert conversationStatusesTempTable raw query partially to Bun * Rename field used to store result of MaxDirectStatusID * Move advanced migrations to their own tiny processor * Catch up util function name with main * Remove json.… wrappers * Remove redundant check * Combine error checks * Replace map with slice of structs * Address processor/type converter comments - Add context info for errors - Extract some common processor code into shared methods - Move conversation eligibility check ahead of populating conversation * Add error context when dropping temp tables
2024-07-23 21:44:31 +02:00
bun.Ident("later_statuses"),
).
JoinOn(
"? = ?",
bun.Ident("conversation_statuses.conversation_id"),
bun.Ident("later_statuses.conversation_id"),
).
JoinOn(
"? > ?",
bun.Ident("later_statuses.created_at"),
bun.Ident("conversation_statuses.created_at"),
).
Where("? IS NULL", bun.Ident("later_statuses.id")),
)
_, err = latestConversationStatusesTmpQ.Exec(ctx)
if err != nil {
return gtserror.Newf(
"error creating temp table %s while deleting status %s: %w",
conversationStatusesTmp, statusID, err,
)
[feature] Conversations API (#3013) * Implement conversations API * Sort and page conversations by last status ID * Appease linter * Fix deleting conversations and statuses * Refactor to make migrations automatic * Lint * Update tests post-merge * Fixes from live-fire testing * Linter caught a format problem * Refactor tests, fix cache * Negative test for non-DMs * Run conversations advanced migration on testrig startup as well as regular server startup * Document (lack of) side effects of API method for deleting a conversation * Make not-found check less nested for readability * Rename PutConversation to UpsertConversation * Use util.Ptr instead of IIFE * Reduce cache used by conversations * Remove unnecessary TableExpr/ColumnExpr * Use struct tags for both unique constraints on Conversation * Make it clear how paging with GetDirectStatusIDsBatch should be used * Let conversation paging skip conversations it can't render * Use Bun NewDropTable * Convert delete raw query to Bun * Convert update raw query to Bun * Convert latestConversationStatusesTempTable raw query partially to Bun * Convert conversationStatusesTempTable raw query partially to Bun * Rename field used to store result of MaxDirectStatusID * Move advanced migrations to their own tiny processor * Catch up util function name with main * Remove json.… wrappers * Remove redundant check * Combine error checks * Replace map with slice of structs * Address processor/type converter comments - Add context info for errors - Extract some common processor code into shared methods - Move conversation eligibility check ahead of populating conversation * Add error context when dropping temp tables
2024-07-23 21:44:31 +02:00
}
// For every conversation where the given status was the last one,
// reset its last status to the most recently created in the
// conversation other than that one, if there is such a status.
[feature] Conversations API (#3013) * Implement conversations API * Sort and page conversations by last status ID * Appease linter * Fix deleting conversations and statuses * Refactor to make migrations automatic * Lint * Update tests post-merge * Fixes from live-fire testing * Linter caught a format problem * Refactor tests, fix cache * Negative test for non-DMs * Run conversations advanced migration on testrig startup as well as regular server startup * Document (lack of) side effects of API method for deleting a conversation * Make not-found check less nested for readability * Rename PutConversation to UpsertConversation * Use util.Ptr instead of IIFE * Reduce cache used by conversations * Remove unnecessary TableExpr/ColumnExpr * Use struct tags for both unique constraints on Conversation * Make it clear how paging with GetDirectStatusIDsBatch should be used * Let conversation paging skip conversations it can't render * Use Bun NewDropTable * Convert delete raw query to Bun * Convert update raw query to Bun * Convert latestConversationStatusesTempTable raw query partially to Bun * Convert conversationStatusesTempTable raw query partially to Bun * Rename field used to store result of MaxDirectStatusID * Move advanced migrations to their own tiny processor * Catch up util function name with main * Remove json.… wrappers * Remove redundant check * Combine error checks * Replace map with slice of structs * Address processor/type converter comments - Add context info for errors - Extract some common processor code into shared methods - Move conversation eligibility check ahead of populating conversation * Add error context when dropping temp tables
2024-07-23 21:44:31 +02:00
// Return conversation IDs for invalidation.
updateQ := tx.NewUpdate().
Table("conversations").
TableExpr("? AS ?", bun.Ident(latestConversationStatusesTmp), bun.Ident("latest_conversation_statuses")).
Set("? = ?", bun.Ident("last_status_id"), bun.Ident("latest_conversation_statuses.id")).
Set("? = ?", bun.Ident("updated_at"), time.Now()).
Where("? = ?", bun.Ident("conversations.id"), bun.Ident("latest_conversation_statuses.conversation_id")).
[feature] Conversations API (#3013) * Implement conversations API * Sort and page conversations by last status ID * Appease linter * Fix deleting conversations and statuses * Refactor to make migrations automatic * Lint * Update tests post-merge * Fixes from live-fire testing * Linter caught a format problem * Refactor tests, fix cache * Negative test for non-DMs * Run conversations advanced migration on testrig startup as well as regular server startup * Document (lack of) side effects of API method for deleting a conversation * Make not-found check less nested for readability * Rename PutConversation to UpsertConversation * Use util.Ptr instead of IIFE * Reduce cache used by conversations * Remove unnecessary TableExpr/ColumnExpr * Use struct tags for both unique constraints on Conversation * Make it clear how paging with GetDirectStatusIDsBatch should be used * Let conversation paging skip conversations it can't render * Use Bun NewDropTable * Convert delete raw query to Bun * Convert update raw query to Bun * Convert latestConversationStatusesTempTable raw query partially to Bun * Convert conversationStatusesTempTable raw query partially to Bun * Rename field used to store result of MaxDirectStatusID * Move advanced migrations to their own tiny processor * Catch up util function name with main * Remove json.… wrappers * Remove redundant check * Combine error checks * Replace map with slice of structs * Address processor/type converter comments - Add context info for errors - Extract some common processor code into shared methods - Move conversation eligibility check ahead of populating conversation * Add error context when dropping temp tables
2024-07-23 21:44:31 +02:00
Where("? IS NOT NULL", bun.Ident("latest_conversation_statuses.id")).
Returning("?", bun.Ident("conversations.id"))
_, err = updateQ.Exec(ctx, &updatedConversationIDs)
if err != nil {
return gtserror.Newf(
"error rolling back last status for conversation while deleting status %s: %w",
statusID, err,
)
[feature] Conversations API (#3013) * Implement conversations API * Sort and page conversations by last status ID * Appease linter * Fix deleting conversations and statuses * Refactor to make migrations automatic * Lint * Update tests post-merge * Fixes from live-fire testing * Linter caught a format problem * Refactor tests, fix cache * Negative test for non-DMs * Run conversations advanced migration on testrig startup as well as regular server startup * Document (lack of) side effects of API method for deleting a conversation * Make not-found check less nested for readability * Rename PutConversation to UpsertConversation * Use util.Ptr instead of IIFE * Reduce cache used by conversations * Remove unnecessary TableExpr/ColumnExpr * Use struct tags for both unique constraints on Conversation * Make it clear how paging with GetDirectStatusIDsBatch should be used * Let conversation paging skip conversations it can't render * Use Bun NewDropTable * Convert delete raw query to Bun * Convert update raw query to Bun * Convert latestConversationStatusesTempTable raw query partially to Bun * Convert conversationStatusesTempTable raw query partially to Bun * Rename field used to store result of MaxDirectStatusID * Move advanced migrations to their own tiny processor * Catch up util function name with main * Remove json.… wrappers * Remove redundant check * Combine error checks * Replace map with slice of structs * Address processor/type converter comments - Add context info for errors - Extract some common processor code into shared methods - Move conversation eligibility check ahead of populating conversation * Add error context when dropping temp tables
2024-07-23 21:44:31 +02:00
}
// If there is no such status,
// just delete the conversation.
// Return IDs for invalidation.
_, err = tx.
NewDelete().
Table("conversations").
[feature] Conversations API (#3013) * Implement conversations API * Sort and page conversations by last status ID * Appease linter * Fix deleting conversations and statuses * Refactor to make migrations automatic * Lint * Update tests post-merge * Fixes from live-fire testing * Linter caught a format problem * Refactor tests, fix cache * Negative test for non-DMs * Run conversations advanced migration on testrig startup as well as regular server startup * Document (lack of) side effects of API method for deleting a conversation * Make not-found check less nested for readability * Rename PutConversation to UpsertConversation * Use util.Ptr instead of IIFE * Reduce cache used by conversations * Remove unnecessary TableExpr/ColumnExpr * Use struct tags for both unique constraints on Conversation * Make it clear how paging with GetDirectStatusIDsBatch should be used * Let conversation paging skip conversations it can't render * Use Bun NewDropTable * Convert delete raw query to Bun * Convert update raw query to Bun * Convert latestConversationStatusesTempTable raw query partially to Bun * Convert conversationStatusesTempTable raw query partially to Bun * Rename field used to store result of MaxDirectStatusID * Move advanced migrations to their own tiny processor * Catch up util function name with main * Remove json.… wrappers * Remove redundant check * Combine error checks * Replace map with slice of structs * Address processor/type converter comments - Add context info for errors - Extract some common processor code into shared methods - Move conversation eligibility check ahead of populating conversation * Add error context when dropping temp tables
2024-07-23 21:44:31 +02:00
Where(
"? IN (?)",
bun.Ident("id"),
tx.NewSelect().
Table(latestConversationStatusesTmp).
[feature] Conversations API (#3013) * Implement conversations API * Sort and page conversations by last status ID * Appease linter * Fix deleting conversations and statuses * Refactor to make migrations automatic * Lint * Update tests post-merge * Fixes from live-fire testing * Linter caught a format problem * Refactor tests, fix cache * Negative test for non-DMs * Run conversations advanced migration on testrig startup as well as regular server startup * Document (lack of) side effects of API method for deleting a conversation * Make not-found check less nested for readability * Rename PutConversation to UpsertConversation * Use util.Ptr instead of IIFE * Reduce cache used by conversations * Remove unnecessary TableExpr/ColumnExpr * Use struct tags for both unique constraints on Conversation * Make it clear how paging with GetDirectStatusIDsBatch should be used * Let conversation paging skip conversations it can't render * Use Bun NewDropTable * Convert delete raw query to Bun * Convert update raw query to Bun * Convert latestConversationStatusesTempTable raw query partially to Bun * Convert conversationStatusesTempTable raw query partially to Bun * Rename field used to store result of MaxDirectStatusID * Move advanced migrations to their own tiny processor * Catch up util function name with main * Remove json.… wrappers * Remove redundant check * Combine error checks * Replace map with slice of structs * Address processor/type converter comments - Add context info for errors - Extract some common processor code into shared methods - Move conversation eligibility check ahead of populating conversation * Add error context when dropping temp tables
2024-07-23 21:44:31 +02:00
Column("conversation_id").
Where("? IS NULL", bun.Ident("id")),
).
Returning("?", bun.Ident("id")).
Exec(ctx, &deletedConversationIDs)
if err != nil {
return gtserror.Newf(
"error deleting conversation while deleting status %s: %w",
statusID, err,
)
[feature] Conversations API (#3013) * Implement conversations API * Sort and page conversations by last status ID * Appease linter * Fix deleting conversations and statuses * Refactor to make migrations automatic * Lint * Update tests post-merge * Fixes from live-fire testing * Linter caught a format problem * Refactor tests, fix cache * Negative test for non-DMs * Run conversations advanced migration on testrig startup as well as regular server startup * Document (lack of) side effects of API method for deleting a conversation * Make not-found check less nested for readability * Rename PutConversation to UpsertConversation * Use util.Ptr instead of IIFE * Reduce cache used by conversations * Remove unnecessary TableExpr/ColumnExpr * Use struct tags for both unique constraints on Conversation * Make it clear how paging with GetDirectStatusIDsBatch should be used * Let conversation paging skip conversations it can't render * Use Bun NewDropTable * Convert delete raw query to Bun * Convert update raw query to Bun * Convert latestConversationStatusesTempTable raw query partially to Bun * Convert conversationStatusesTempTable raw query partially to Bun * Rename field used to store result of MaxDirectStatusID * Move advanced migrations to their own tiny processor * Catch up util function name with main * Remove json.… wrappers * Remove redundant check * Combine error checks * Replace map with slice of structs * Address processor/type converter comments - Add context info for errors - Extract some common processor code into shared methods - Move conversation eligibility check ahead of populating conversation * Add error context when dropping temp tables
2024-07-23 21:44:31 +02:00
}
return nil
}); err != nil {
return err
}
// Invalidate cache entries.
[feature] Conversations API (#3013) * Implement conversations API * Sort and page conversations by last status ID * Appease linter * Fix deleting conversations and statuses * Refactor to make migrations automatic * Lint * Update tests post-merge * Fixes from live-fire testing * Linter caught a format problem * Refactor tests, fix cache * Negative test for non-DMs * Run conversations advanced migration on testrig startup as well as regular server startup * Document (lack of) side effects of API method for deleting a conversation * Make not-found check less nested for readability * Rename PutConversation to UpsertConversation * Use util.Ptr instead of IIFE * Reduce cache used by conversations * Remove unnecessary TableExpr/ColumnExpr * Use struct tags for both unique constraints on Conversation * Make it clear how paging with GetDirectStatusIDsBatch should be used * Let conversation paging skip conversations it can't render * Use Bun NewDropTable * Convert delete raw query to Bun * Convert update raw query to Bun * Convert latestConversationStatusesTempTable raw query partially to Bun * Convert conversationStatusesTempTable raw query partially to Bun * Rename field used to store result of MaxDirectStatusID * Move advanced migrations to their own tiny processor * Catch up util function name with main * Remove json.… wrappers * Remove redundant check * Combine error checks * Replace map with slice of structs * Address processor/type converter comments - Add context info for errors - Extract some common processor code into shared methods - Move conversation eligibility check ahead of populating conversation * Add error context when dropping temp tables
2024-07-23 21:44:31 +02:00
updatedConversationIDs = append(updatedConversationIDs, deletedConversationIDs...)
updatedConversationIDs = util.Deduplicate(updatedConversationIDs)
c.state.Caches.DB.Conversation.InvalidateIDs("ID", updatedConversationIDs)
[feature] Conversations API (#3013) * Implement conversations API * Sort and page conversations by last status ID * Appease linter * Fix deleting conversations and statuses * Refactor to make migrations automatic * Lint * Update tests post-merge * Fixes from live-fire testing * Linter caught a format problem * Refactor tests, fix cache * Negative test for non-DMs * Run conversations advanced migration on testrig startup as well as regular server startup * Document (lack of) side effects of API method for deleting a conversation * Make not-found check less nested for readability * Rename PutConversation to UpsertConversation * Use util.Ptr instead of IIFE * Reduce cache used by conversations * Remove unnecessary TableExpr/ColumnExpr * Use struct tags for both unique constraints on Conversation * Make it clear how paging with GetDirectStatusIDsBatch should be used * Let conversation paging skip conversations it can't render * Use Bun NewDropTable * Convert delete raw query to Bun * Convert update raw query to Bun * Convert latestConversationStatusesTempTable raw query partially to Bun * Convert conversationStatusesTempTable raw query partially to Bun * Rename field used to store result of MaxDirectStatusID * Move advanced migrations to their own tiny processor * Catch up util function name with main * Remove json.… wrappers * Remove redundant check * Combine error checks * Replace map with slice of structs * Address processor/type converter comments - Add context info for errors - Extract some common processor code into shared methods - Move conversation eligibility check ahead of populating conversation * Add error context when dropping temp tables
2024-07-23 21:44:31 +02:00
return nil
}