forgejo/modules/indexer/issues/util.go

195 lines
5.7 KiB
Go
Raw 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 14:28:53 +08:00
// Copyright 2023 The Gitea Authors. All rights reserved.
// SPDX-License-Identifier: MIT
package issues
import (
"context"
"errors"
"fmt"
"forgejo.org/models/db"
issue_model "forgejo.org/models/issues"
"forgejo.org/modules/container"
"forgejo.org/modules/indexer/issues/internal"
"forgejo.org/modules/log"
"forgejo.org/modules/queue"
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 14:28:53 +08:00
)
// 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 14:28:53 +08:00
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,
Add issue number to the search index, rank number and title matches higher (#7956) (#7968) An attempt at solving #7956. This (and rebuilding the index) seems enough to ensure the issue *appears* among the results. However, I couldn't figure out from [bleve docs](https://github.com/blevesearch/bleve/blob/master/docs/scoring.md) how to affect the scoring based on specific fields, or whether that is possible at all. Disclaimer: I've never written Go before, sorry 😅 take it as a quick PoC more than anything. ### Tests - I added test coverage for Go changes... - [x] in their respective `*_test.go` for unit tests. - [ ] in the `tests/integration` directory if it involves interactions with a live Forgejo server. - I added test coverage for JavaScript changes... - [ ] in `web_src/js/*.test.js` if it can be unit tested. - [ ] in `tests/e2e/*.test.e2e.js` if it requires interactions with a live Forgejo server (see also the [developer guide for JavaScript testing](https://codeberg.org/forgejo/forgejo/src/branch/forgejo/tests/e2e/README.md#end-to-end-tests)). ### Documentation - [ ] I created a pull request [to the documentation](https://codeberg.org/forgejo/docs) to explain to Forgejo users how to use this change. - [x] I did not document these changes and I do not expect someone else to do it. ### Release notes - [ ] I do not want this change to show in the release notes. - [x] I want the title to show in the release notes with a link to this pull request. - [ ] I want the content of the `release-notes/<pull request number>.md` to be be used for the release notes instead of the title. <!--start release-notes-assistant--> ## Release notes <!--URL:https://codeberg.org/forgejo/forgejo--> - Features - [PR](https://codeberg.org/forgejo/forgejo/pulls/7968): <!--number 7968 --><!--line 0 --><!--description QWRkIGlzc3VlIG51bWJlciB0byB0aGUgc2VhcmNoIGluZGV4LCByYW5rIG51bWJlciBhbmQgdGl0bGUgbWF0Y2hlcyBoaWdoZXIgKCM3OTU2KQ==-->Add issue number to the search index, rank number and title matches higher (#7956)<!--description--> <!--end release-notes-assistant--> Reviewed-on: https://codeberg.org/forgejo/forgejo/pulls/7968 Reviewed-by: Shiny Nematoda <snematoda@noreply.codeberg.org> Co-authored-by: Danko Aleksejevs <danko@very.lv> Co-committed-by: Danko Aleksejevs <danko@very.lv>
2025-06-04 07:42:29 +02:00
Index: issue.Index,
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 14:28:53 +08:00
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,
ProjectColumnID: issue.ProjectColumnID(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 14:28:53 +08:00
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 14:28:53 +08:00
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 14:28:53 +08:00
}
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 14:28:53 +08:00
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 14:28:53 +08:00
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 14:28:53 +08:00
}
}