forgejo/modules/indexer/issues/util.go

194 lines
5.7 KiB
Go
Raw Permalink Normal View History

Refactor and enhance issue indexer to support both searching, filtering and paging (#26012) Fix #24662. Replace #24822 and #25708 (although it has been merged) ## Background In the past, Gitea supported issue searching with a keyword and conditions in a less efficient way. It worked by searching for issues with the keyword and obtaining limited IDs (as it is heavy to get all) on the indexer (bleve/elasticsearch/meilisearch), and then querying with conditions on the database to find a subset of the found IDs. This is why the results could be incomplete. To solve this issue, we need to store all fields that could be used as conditions in the indexer and support both keyword and additional conditions when searching with the indexer. ## Major changes - Redefine `IndexerData` to include all fields that could be used as filter conditions. - Refactor `Search(ctx context.Context, kw string, repoIDs []int64, limit, start int, state string)` to `Search(ctx context.Context, options *SearchOptions)`, so it supports more conditions now. - Change the data type stored in `issueIndexerQueue`. Use `IndexerMetadata` instead of `IndexerData` in case the data has been updated while it is in the queue. This also reduces the storage size of the queue. - Enhance searching with Bleve/Elasticsearch/Meilisearch, make them fully support `SearchOptions`. Also, update the data versions. - Keep most logic of database indexer, but remove `issues.SearchIssueIDsByKeyword` in `models` to avoid confusion where is the entry point to search issues. - Start a Meilisearch instance to test it in unit tests. - Add unit tests with almost full coverage to test Bleve/Elasticsearch/Meilisearch indexer. --------- Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
2023-07-31 06:28:53 +00:00
// Copyright 2023 The Gitea Authors. All rights reserved.
// SPDX-License-Identifier: MIT
package issues
import (
"context"
"errors"
"fmt"
"code.gitea.io/gitea/models/db"
issue_model "code.gitea.io/gitea/models/issues"
"code.gitea.io/gitea/modules/container"
"code.gitea.io/gitea/modules/indexer/issues/internal"
"code.gitea.io/gitea/modules/log"
"code.gitea.io/gitea/modules/queue"
)
// getIssueIndexerData returns the indexer data of an issue and a bool value indicating whether the issue exists.
func getIssueIndexerData(ctx context.Context, issueID int64) (*internal.IndexerData, bool, error) {
issue, err := issue_model.GetIssueByID(ctx, issueID)
if err != nil {
if issue_model.IsErrIssueNotExist(err) {
return nil, false, nil
}
return nil, false, err
}
// FIXME: what if users want to search for a review comment of a pull request?
// The comment type is CommentTypeCode or CommentTypeReview.
// But LoadDiscussComments only loads CommentTypeComment.
if err := issue.LoadDiscussComments(ctx); err != nil {
return nil, false, err
}
comments := make([]string, 0, len(issue.Comments))
for _, comment := range issue.Comments {
if comment.Content != "" {
// what ever the comment type is, index the content if it is not empty.
comments = append(comments, comment.Content)
}
}
if err := issue.LoadAttributes(ctx); err != nil {
return nil, false, err
}
labels := make([]int64, 0, len(issue.Labels))
for _, label := range issue.Labels {
labels = append(labels, label.ID)
}
mentionIDs, err := issue_model.GetIssueMentionIDs(ctx, issueID)
if err != nil {
return nil, false, err
}
var (
reviewedIDs []int64
reviewRequestedIDs []int64
)
{
reviews, err := issue_model.FindReviews(ctx, issue_model.FindReviewOptions{
ListOptions: db.ListOptionsAll,
Refactor and enhance issue indexer to support both searching, filtering and paging (#26012) Fix #24662. Replace #24822 and #25708 (although it has been merged) ## Background In the past, Gitea supported issue searching with a keyword and conditions in a less efficient way. It worked by searching for issues with the keyword and obtaining limited IDs (as it is heavy to get all) on the indexer (bleve/elasticsearch/meilisearch), and then querying with conditions on the database to find a subset of the found IDs. This is why the results could be incomplete. To solve this issue, we need to store all fields that could be used as conditions in the indexer and support both keyword and additional conditions when searching with the indexer. ## Major changes - Redefine `IndexerData` to include all fields that could be used as filter conditions. - Refactor `Search(ctx context.Context, kw string, repoIDs []int64, limit, start int, state string)` to `Search(ctx context.Context, options *SearchOptions)`, so it supports more conditions now. - Change the data type stored in `issueIndexerQueue`. Use `IndexerMetadata` instead of `IndexerData` in case the data has been updated while it is in the queue. This also reduces the storage size of the queue. - Enhance searching with Bleve/Elasticsearch/Meilisearch, make them fully support `SearchOptions`. Also, update the data versions. - Keep most logic of database indexer, but remove `issues.SearchIssueIDsByKeyword` in `models` to avoid confusion where is the entry point to search issues. - Start a Meilisearch instance to test it in unit tests. - Add unit tests with almost full coverage to test Bleve/Elasticsearch/Meilisearch indexer. --------- Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
2023-07-31 06:28:53 +00:00
IssueID: issueID,
OfficialOnly: false,
})
if err != nil {
return nil, false, err
}
reviewedIDsSet := make(container.Set[int64], len(reviews))
reviewRequestedIDsSet := make(container.Set[int64], len(reviews))
for _, review := range reviews {
if review.Type == issue_model.ReviewTypeRequest {
reviewRequestedIDsSet.Add(review.ReviewerID)
} else {
reviewedIDsSet.Add(review.ReviewerID)
}
}
reviewedIDs = reviewedIDsSet.Values()
reviewRequestedIDs = reviewRequestedIDsSet.Values()
}
subscriberIDs, err := issue_model.GetIssueWatchersIDs(ctx, issue.ID, true)
if err != nil {
return nil, false, err
}
var projectID int64
if issue.Project != nil {
projectID = issue.Project.ID
}
return &internal.IndexerData{
ID: issue.ID,
RepoID: issue.RepoID,
IsPublic: !issue.Repo.IsPrivate,
Title: issue.Title,
Content: issue.Content,
Comments: comments,
IsPull: issue.IsPull,
IsClosed: issue.IsClosed,
LabelIDs: labels,
NoLabel: len(labels) == 0,
MilestoneID: issue.MilestoneID,
ProjectID: projectID,
ProjectBoardID: issue.ProjectBoardID(ctx),
Refactor and enhance issue indexer to support both searching, filtering and paging (#26012) Fix #24662. Replace #24822 and #25708 (although it has been merged) ## Background In the past, Gitea supported issue searching with a keyword and conditions in a less efficient way. It worked by searching for issues with the keyword and obtaining limited IDs (as it is heavy to get all) on the indexer (bleve/elasticsearch/meilisearch), and then querying with conditions on the database to find a subset of the found IDs. This is why the results could be incomplete. To solve this issue, we need to store all fields that could be used as conditions in the indexer and support both keyword and additional conditions when searching with the indexer. ## Major changes - Redefine `IndexerData` to include all fields that could be used as filter conditions. - Refactor `Search(ctx context.Context, kw string, repoIDs []int64, limit, start int, state string)` to `Search(ctx context.Context, options *SearchOptions)`, so it supports more conditions now. - Change the data type stored in `issueIndexerQueue`. Use `IndexerMetadata` instead of `IndexerData` in case the data has been updated while it is in the queue. This also reduces the storage size of the queue. - Enhance searching with Bleve/Elasticsearch/Meilisearch, make them fully support `SearchOptions`. Also, update the data versions. - Keep most logic of database indexer, but remove `issues.SearchIssueIDsByKeyword` in `models` to avoid confusion where is the entry point to search issues. - Start a Meilisearch instance to test it in unit tests. - Add unit tests with almost full coverage to test Bleve/Elasticsearch/Meilisearch indexer. --------- Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
2023-07-31 06:28:53 +00:00
PosterID: issue.PosterID,
AssigneeID: issue.AssigneeID,
MentionIDs: mentionIDs,
ReviewedIDs: reviewedIDs,
ReviewRequestedIDs: reviewRequestedIDs,
SubscriberIDs: subscriberIDs,
UpdatedUnix: issue.UpdatedUnix,
CreatedUnix: issue.CreatedUnix,
DeadlineUnix: issue.DeadlineUnix,
CommentCount: int64(len(issue.Comments)),
}, true, nil
}
func updateRepoIndexer(ctx context.Context, repoID int64) error {
ids, err := issue_model.GetIssueIDsByRepoID(ctx, repoID)
if err != nil {
return fmt.Errorf("issue_model.GetIssueIDsByRepoID: %w", err)
}
for _, id := range ids {
if err := updateIssueIndexer(ctx, id); err != nil {
Refactor and enhance issue indexer to support both searching, filtering and paging (#26012) Fix #24662. Replace #24822 and #25708 (although it has been merged) ## Background In the past, Gitea supported issue searching with a keyword and conditions in a less efficient way. It worked by searching for issues with the keyword and obtaining limited IDs (as it is heavy to get all) on the indexer (bleve/elasticsearch/meilisearch), and then querying with conditions on the database to find a subset of the found IDs. This is why the results could be incomplete. To solve this issue, we need to store all fields that could be used as conditions in the indexer and support both keyword and additional conditions when searching with the indexer. ## Major changes - Redefine `IndexerData` to include all fields that could be used as filter conditions. - Refactor `Search(ctx context.Context, kw string, repoIDs []int64, limit, start int, state string)` to `Search(ctx context.Context, options *SearchOptions)`, so it supports more conditions now. - Change the data type stored in `issueIndexerQueue`. Use `IndexerMetadata` instead of `IndexerData` in case the data has been updated while it is in the queue. This also reduces the storage size of the queue. - Enhance searching with Bleve/Elasticsearch/Meilisearch, make them fully support `SearchOptions`. Also, update the data versions. - Keep most logic of database indexer, but remove `issues.SearchIssueIDsByKeyword` in `models` to avoid confusion where is the entry point to search issues. - Start a Meilisearch instance to test it in unit tests. - Add unit tests with almost full coverage to test Bleve/Elasticsearch/Meilisearch indexer. --------- Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
2023-07-31 06:28:53 +00:00
return err
}
}
return nil
}
func updateIssueIndexer(ctx context.Context, issueID int64) error {
return pushIssueIndexerQueue(ctx, &IndexerMetadata{ID: issueID})
Refactor and enhance issue indexer to support both searching, filtering and paging (#26012) Fix #24662. Replace #24822 and #25708 (although it has been merged) ## Background In the past, Gitea supported issue searching with a keyword and conditions in a less efficient way. It worked by searching for issues with the keyword and obtaining limited IDs (as it is heavy to get all) on the indexer (bleve/elasticsearch/meilisearch), and then querying with conditions on the database to find a subset of the found IDs. This is why the results could be incomplete. To solve this issue, we need to store all fields that could be used as conditions in the indexer and support both keyword and additional conditions when searching with the indexer. ## Major changes - Redefine `IndexerData` to include all fields that could be used as filter conditions. - Refactor `Search(ctx context.Context, kw string, repoIDs []int64, limit, start int, state string)` to `Search(ctx context.Context, options *SearchOptions)`, so it supports more conditions now. - Change the data type stored in `issueIndexerQueue`. Use `IndexerMetadata` instead of `IndexerData` in case the data has been updated while it is in the queue. This also reduces the storage size of the queue. - Enhance searching with Bleve/Elasticsearch/Meilisearch, make them fully support `SearchOptions`. Also, update the data versions. - Keep most logic of database indexer, but remove `issues.SearchIssueIDsByKeyword` in `models` to avoid confusion where is the entry point to search issues. - Start a Meilisearch instance to test it in unit tests. - Add unit tests with almost full coverage to test Bleve/Elasticsearch/Meilisearch indexer. --------- Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
2023-07-31 06:28:53 +00:00
}
func deleteRepoIssueIndexer(ctx context.Context, repoID int64) error {
var ids []int64
ids, err := issue_model.GetIssueIDsByRepoID(ctx, repoID)
if err != nil {
return fmt.Errorf("issue_model.GetIssueIDsByRepoID: %w", err)
}
if len(ids) == 0 {
return nil
}
return pushIssueIndexerQueue(ctx, &IndexerMetadata{
Refactor and enhance issue indexer to support both searching, filtering and paging (#26012) Fix #24662. Replace #24822 and #25708 (although it has been merged) ## Background In the past, Gitea supported issue searching with a keyword and conditions in a less efficient way. It worked by searching for issues with the keyword and obtaining limited IDs (as it is heavy to get all) on the indexer (bleve/elasticsearch/meilisearch), and then querying with conditions on the database to find a subset of the found IDs. This is why the results could be incomplete. To solve this issue, we need to store all fields that could be used as conditions in the indexer and support both keyword and additional conditions when searching with the indexer. ## Major changes - Redefine `IndexerData` to include all fields that could be used as filter conditions. - Refactor `Search(ctx context.Context, kw string, repoIDs []int64, limit, start int, state string)` to `Search(ctx context.Context, options *SearchOptions)`, so it supports more conditions now. - Change the data type stored in `issueIndexerQueue`. Use `IndexerMetadata` instead of `IndexerData` in case the data has been updated while it is in the queue. This also reduces the storage size of the queue. - Enhance searching with Bleve/Elasticsearch/Meilisearch, make them fully support `SearchOptions`. Also, update the data versions. - Keep most logic of database indexer, but remove `issues.SearchIssueIDsByKeyword` in `models` to avoid confusion where is the entry point to search issues. - Start a Meilisearch instance to test it in unit tests. - Add unit tests with almost full coverage to test Bleve/Elasticsearch/Meilisearch indexer. --------- Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
2023-07-31 06:28:53 +00:00
IDs: ids,
IsDelete: true,
})
}
type keepRetryKey struct{}
// contextWithKeepRetry returns a context with a key indicating that the indexer should keep retrying.
// Please note that it's for background tasks only, and it should not be used for user requests, or it may cause blocking.
func contextWithKeepRetry(ctx context.Context) context.Context {
return context.WithValue(ctx, keepRetryKey{}, true)
}
func pushIssueIndexerQueue(ctx context.Context, data *IndexerMetadata) error {
Refactor and enhance issue indexer to support both searching, filtering and paging (#26012) Fix #24662. Replace #24822 and #25708 (although it has been merged) ## Background In the past, Gitea supported issue searching with a keyword and conditions in a less efficient way. It worked by searching for issues with the keyword and obtaining limited IDs (as it is heavy to get all) on the indexer (bleve/elasticsearch/meilisearch), and then querying with conditions on the database to find a subset of the found IDs. This is why the results could be incomplete. To solve this issue, we need to store all fields that could be used as conditions in the indexer and support both keyword and additional conditions when searching with the indexer. ## Major changes - Redefine `IndexerData` to include all fields that could be used as filter conditions. - Refactor `Search(ctx context.Context, kw string, repoIDs []int64, limit, start int, state string)` to `Search(ctx context.Context, options *SearchOptions)`, so it supports more conditions now. - Change the data type stored in `issueIndexerQueue`. Use `IndexerMetadata` instead of `IndexerData` in case the data has been updated while it is in the queue. This also reduces the storage size of the queue. - Enhance searching with Bleve/Elasticsearch/Meilisearch, make them fully support `SearchOptions`. Also, update the data versions. - Keep most logic of database indexer, but remove `issues.SearchIssueIDsByKeyword` in `models` to avoid confusion where is the entry point to search issues. - Start a Meilisearch instance to test it in unit tests. - Add unit tests with almost full coverage to test Bleve/Elasticsearch/Meilisearch indexer. --------- Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
2023-07-31 06:28:53 +00:00
if issueIndexerQueue == nil {
// Some unit tests will trigger indexing, but the queue is not initialized.
// It's OK to ignore it, but log a warning message in case it's not a unit test.
log.Warn("Trying to push %+v to issue indexer queue, but the queue is not initialized, it's OK if it's a unit test", data)
return nil
}
for {
select {
case <-ctx.Done():
return ctx.Err()
default:
}
err := issueIndexerQueue.Push(data)
if errors.Is(err, queue.ErrAlreadyInQueue) {
return nil
}
if errors.Is(err, context.DeadlineExceeded) { // the queue is full
log.Warn("It seems that issue indexer is slow and the queue is full. Please check the issue indexer or increase the queue size.")
if ctx.Value(keepRetryKey{}) == nil {
return err
}
// It will be better to increase the queue size instead of retrying, but users may ignore the previous warning message.
// However, even it retries, it may still cause index loss when there's a deadline in the context.
log.Debug("Retry to push %+v to issue indexer queue", data)
continue
}
return err
Refactor and enhance issue indexer to support both searching, filtering and paging (#26012) Fix #24662. Replace #24822 and #25708 (although it has been merged) ## Background In the past, Gitea supported issue searching with a keyword and conditions in a less efficient way. It worked by searching for issues with the keyword and obtaining limited IDs (as it is heavy to get all) on the indexer (bleve/elasticsearch/meilisearch), and then querying with conditions on the database to find a subset of the found IDs. This is why the results could be incomplete. To solve this issue, we need to store all fields that could be used as conditions in the indexer and support both keyword and additional conditions when searching with the indexer. ## Major changes - Redefine `IndexerData` to include all fields that could be used as filter conditions. - Refactor `Search(ctx context.Context, kw string, repoIDs []int64, limit, start int, state string)` to `Search(ctx context.Context, options *SearchOptions)`, so it supports more conditions now. - Change the data type stored in `issueIndexerQueue`. Use `IndexerMetadata` instead of `IndexerData` in case the data has been updated while it is in the queue. This also reduces the storage size of the queue. - Enhance searching with Bleve/Elasticsearch/Meilisearch, make them fully support `SearchOptions`. Also, update the data versions. - Keep most logic of database indexer, but remove `issues.SearchIssueIDsByKeyword` in `models` to avoid confusion where is the entry point to search issues. - Start a Meilisearch instance to test it in unit tests. - Add unit tests with almost full coverage to test Bleve/Elasticsearch/Meilisearch indexer. --------- Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
2023-07-31 06:28:53 +00:00
}
}