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>
This commit is contained in:
Jason Song 2023-07-31 14:28:53 +08:00 committed by GitHub
parent aba9096999
commit 1e76a824bc
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
37 changed files with 2965 additions and 861 deletions

View file

@ -132,74 +132,73 @@ func SearchIssues(ctx *context.APIContext) {
isClosed = util.OptionalBoolFalse
}
// find repos user can access (for issue search)
opts := &repo_model.SearchRepoOptions{
Private: false,
AllPublic: true,
TopicOnly: false,
Collaborate: util.OptionalBoolNone,
// This needs to be a column that is not nil in fixtures or
// MySQL will return different results when sorting by null in some cases
OrderBy: db.SearchOrderByAlphabetically,
Actor: ctx.Doer,
}
if ctx.IsSigned {
opts.Private = true
opts.AllLimited = true
}
if ctx.FormString("owner") != "" {
owner, err := user_model.GetUserByName(ctx, ctx.FormString("owner"))
if err != nil {
if user_model.IsErrUserNotExist(err) {
ctx.Error(http.StatusBadRequest, "Owner not found", err)
} else {
ctx.Error(http.StatusInternalServerError, "GetUserByName", err)
var (
repoIDs []int64
allPublic bool
)
{
// find repos user can access (for issue search)
opts := &repo_model.SearchRepoOptions{
Private: false,
AllPublic: true,
TopicOnly: false,
Collaborate: util.OptionalBoolNone,
// This needs to be a column that is not nil in fixtures or
// MySQL will return different results when sorting by null in some cases
OrderBy: db.SearchOrderByAlphabetically,
Actor: ctx.Doer,
}
if ctx.IsSigned {
opts.Private = true
opts.AllLimited = true
}
if ctx.FormString("owner") != "" {
owner, err := user_model.GetUserByName(ctx, ctx.FormString("owner"))
if err != nil {
if user_model.IsErrUserNotExist(err) {
ctx.Error(http.StatusBadRequest, "Owner not found", err)
} else {
ctx.Error(http.StatusInternalServerError, "GetUserByName", err)
}
return
}
return
opts.OwnerID = owner.ID
opts.AllLimited = false
opts.AllPublic = false
opts.Collaborate = util.OptionalBoolFalse
}
opts.OwnerID = owner.ID
opts.AllLimited = false
opts.AllPublic = false
opts.Collaborate = util.OptionalBoolFalse
}
if ctx.FormString("team") != "" {
if ctx.FormString("owner") == "" {
ctx.Error(http.StatusBadRequest, "", "Owner organisation is required for filtering on team")
return
}
team, err := organization.GetTeam(ctx, opts.OwnerID, ctx.FormString("team"))
if err != nil {
if organization.IsErrTeamNotExist(err) {
ctx.Error(http.StatusBadRequest, "Team not found", err)
} else {
ctx.Error(http.StatusInternalServerError, "GetUserByName", err)
if ctx.FormString("team") != "" {
if ctx.FormString("owner") == "" {
ctx.Error(http.StatusBadRequest, "", "Owner organisation is required for filtering on team")
return
}
team, err := organization.GetTeam(ctx, opts.OwnerID, ctx.FormString("team"))
if err != nil {
if organization.IsErrTeamNotExist(err) {
ctx.Error(http.StatusBadRequest, "Team not found", err)
} else {
ctx.Error(http.StatusInternalServerError, "GetUserByName", err)
}
return
}
opts.TeamID = team.ID
}
if opts.AllPublic {
allPublic = true
opts.AllPublic = false // set it false to avoid returning too many repos, we could filter by indexer
}
repoIDs, _, err = repo_model.SearchRepositoryIDs(opts)
if err != nil {
ctx.Error(http.StatusInternalServerError, "SearchRepositoryIDs", err)
return
}
opts.TeamID = team.ID
}
repoCond := repo_model.SearchRepositoryCondition(opts)
repoIDs, _, err := repo_model.SearchRepositoryIDs(opts)
if err != nil {
ctx.Error(http.StatusInternalServerError, "SearchRepositoryIDs", err)
return
}
var issues []*issues_model.Issue
var filteredCount int64
keyword := ctx.FormTrim("q")
if strings.IndexByte(keyword, 0) >= 0 {
keyword = ""
}
var issueIDs []int64
if len(keyword) > 0 && len(repoIDs) > 0 {
if issueIDs, err = issue_indexer.SearchIssuesByKeyword(ctx, repoIDs, keyword, ctx.FormString("state")); err != nil {
ctx.Error(http.StatusInternalServerError, "SearchIssuesByKeyword", err)
return
}
}
var isPull util.OptionalBool
switch ctx.FormString("type") {
@ -211,16 +210,33 @@ func SearchIssues(ctx *context.APIContext) {
isPull = util.OptionalBoolNone
}
labels := ctx.FormTrim("labels")
var includedLabelNames []string
if len(labels) > 0 {
includedLabelNames = strings.Split(labels, ",")
var includedAnyLabels []int64
{
labels := ctx.FormTrim("labels")
var includedLabelNames []string
if len(labels) > 0 {
includedLabelNames = strings.Split(labels, ",")
}
includedAnyLabels, err = issues_model.GetLabelIDsByNames(ctx, includedLabelNames)
if err != nil {
ctx.Error(http.StatusInternalServerError, "GetLabelIDsByNames", err)
return
}
}
milestones := ctx.FormTrim("milestones")
var includedMilestones []string
if len(milestones) > 0 {
includedMilestones = strings.Split(milestones, ",")
var includedMilestones []int64
{
milestones := ctx.FormTrim("milestones")
var includedMilestoneNames []string
if len(milestones) > 0 {
includedMilestoneNames = strings.Split(milestones, ",")
}
includedMilestones, err = issues_model.GetMilestoneIDsByNames(ctx, includedMilestoneNames)
if err != nil {
ctx.Error(http.StatusInternalServerError, "GetMilestoneIDsByNames", err)
return
}
}
// this api is also used in UI,
@ -232,64 +248,64 @@ func SearchIssues(ctx *context.APIContext) {
limit = setting.API.MaxResponseItems
}
// Only fetch the issues if we either don't have a keyword or the search returned issues
// This would otherwise return all issues if no issues were found by the search.
if len(keyword) == 0 || len(issueIDs) > 0 || len(includedLabelNames) > 0 || len(includedMilestones) > 0 {
issuesOpt := &issues_model.IssuesOptions{
ListOptions: db.ListOptions{
Page: ctx.FormInt("page"),
PageSize: limit,
},
RepoCond: repoCond,
IsClosed: isClosed,
IssueIDs: issueIDs,
IncludedLabelNames: includedLabelNames,
IncludeMilestones: includedMilestones,
SortType: "priorityrepo",
PriorityRepoID: ctx.FormInt64("priority_repo_id"),
IsPull: isPull,
UpdatedBeforeUnix: before,
UpdatedAfterUnix: since,
}
searchOpt := &issue_indexer.SearchOptions{
Paginator: &db.ListOptions{
PageSize: limit,
Page: ctx.FormInt("page"),
},
Keyword: keyword,
RepoIDs: repoIDs,
AllPublic: allPublic,
IsPull: isPull,
IsClosed: isClosed,
IncludedAnyLabelIDs: includedAnyLabels,
MilestoneIDs: includedMilestones,
SortBy: issue_indexer.SortByCreatedDesc,
}
ctxUserID := int64(0)
if ctx.IsSigned {
ctxUserID = ctx.Doer.ID
}
if since != 0 {
searchOpt.UpdatedAfterUnix = &since
}
if before != 0 {
searchOpt.UpdatedBeforeUnix = &before
}
// Filter for: Created by User, Assigned to User, Mentioning User, Review of User Requested
if ctx.IsSigned {
ctxUserID := ctx.Doer.ID
if ctx.FormBool("created") {
issuesOpt.PosterID = ctxUserID
searchOpt.PosterID = &ctxUserID
}
if ctx.FormBool("assigned") {
issuesOpt.AssigneeID = ctxUserID
searchOpt.AssigneeID = &ctxUserID
}
if ctx.FormBool("mentioned") {
issuesOpt.MentionedID = ctxUserID
searchOpt.MentionID = &ctxUserID
}
if ctx.FormBool("review_requested") {
issuesOpt.ReviewRequestedID = ctxUserID
searchOpt.ReviewRequestedID = &ctxUserID
}
if ctx.FormBool("reviewed") {
issuesOpt.ReviewedID = ctxUserID
}
if issues, err = issues_model.Issues(ctx, issuesOpt); err != nil {
ctx.Error(http.StatusInternalServerError, "Issues", err)
return
}
issuesOpt.ListOptions = db.ListOptions{
Page: -1,
}
if filteredCount, err = issues_model.CountIssues(ctx, issuesOpt); err != nil {
ctx.Error(http.StatusInternalServerError, "CountIssues", err)
return
searchOpt.ReviewedID = &ctxUserID
}
}
ctx.SetLinkHeader(int(filteredCount), limit)
ctx.SetTotalCountHeader(filteredCount)
// FIXME: It's unsupported to sort by priority repo when searching by indexer,
// it's indeed an regression, but I think it is worth to support filtering by indexer first.
_ = ctx.FormInt64("priority_repo_id")
ids, total, err := issue_indexer.SearchIssues(ctx, searchOpt)
if err != nil {
ctx.Error(http.StatusInternalServerError, "SearchIssues", err)
return
}
issues, err := issues_model.GetIssuesByIDs(ctx, ids, true)
if err != nil {
ctx.Error(http.StatusInternalServerError, "FindIssuesByIDs", err)
return
}
ctx.SetLinkHeader(int(total), limit)
ctx.SetTotalCountHeader(total)
ctx.JSON(http.StatusOK, convert.ToAPIIssueList(ctx, issues))
}
@ -384,23 +400,12 @@ func ListIssues(ctx *context.APIContext) {
isClosed = util.OptionalBoolFalse
}
var issues []*issues_model.Issue
var filteredCount int64
keyword := ctx.FormTrim("q")
if strings.IndexByte(keyword, 0) >= 0 {
keyword = ""
}
var issueIDs []int64
var labelIDs []int64
if len(keyword) > 0 {
issueIDs, err = issue_indexer.SearchIssuesByKeyword(ctx, []int64{ctx.Repo.Repository.ID}, keyword, ctx.FormString("state"))
if err != nil {
ctx.Error(http.StatusInternalServerError, "SearchIssuesByKeyword", err)
return
}
}
var labelIDs []int64
if splitted := strings.Split(ctx.FormString("labels"), ","); len(splitted) > 0 {
labelIDs, err = issues_model.GetLabelIDsInRepoByNames(ctx.Repo.Repository.ID, splitted)
if err != nil {
@ -465,40 +470,61 @@ func ListIssues(ctx *context.APIContext) {
return
}
// Only fetch the issues if we either don't have a keyword or the search returned issues
// This would otherwise return all issues if no issues were found by the search.
if len(keyword) == 0 || len(issueIDs) > 0 || len(labelIDs) > 0 {
issuesOpt := &issues_model.IssuesOptions{
ListOptions: listOptions,
RepoIDs: []int64{ctx.Repo.Repository.ID},
IsClosed: isClosed,
IssueIDs: issueIDs,
LabelIDs: labelIDs,
MilestoneIDs: mileIDs,
IsPull: isPull,
UpdatedBeforeUnix: before,
UpdatedAfterUnix: since,
PosterID: createdByID,
AssigneeID: assignedByID,
MentionedID: mentionedByID,
}
if issues, err = issues_model.Issues(ctx, issuesOpt); err != nil {
ctx.Error(http.StatusInternalServerError, "Issues", err)
return
}
issuesOpt.ListOptions = db.ListOptions{
Page: -1,
}
if filteredCount, err = issues_model.CountIssues(ctx, issuesOpt); err != nil {
ctx.Error(http.StatusInternalServerError, "CountIssues", err)
return
searchOpt := &issue_indexer.SearchOptions{
Paginator: &listOptions,
Keyword: keyword,
RepoIDs: []int64{ctx.Repo.Repository.ID},
IsPull: isPull,
IsClosed: isClosed,
SortBy: issue_indexer.SortByCreatedDesc,
}
if since != 0 {
searchOpt.UpdatedAfterUnix = &since
}
if before != 0 {
searchOpt.UpdatedBeforeUnix = &before
}
if len(labelIDs) == 1 && labelIDs[0] == 0 {
searchOpt.NoLabelOnly = true
} else {
for _, labelID := range labelIDs {
if labelID > 0 {
searchOpt.IncludedLabelIDs = append(searchOpt.IncludedLabelIDs, labelID)
} else {
searchOpt.ExcludedLabelIDs = append(searchOpt.ExcludedLabelIDs, -labelID)
}
}
}
ctx.SetLinkHeader(int(filteredCount), listOptions.PageSize)
ctx.SetTotalCountHeader(filteredCount)
if len(mileIDs) == 1 && mileIDs[0] == db.NoConditionID {
searchOpt.MilestoneIDs = []int64{0}
} else {
searchOpt.MilestoneIDs = mileIDs
}
if createdByID > 0 {
searchOpt.PosterID = &createdByID
}
if assignedByID > 0 {
searchOpt.AssigneeID = &assignedByID
}
if mentionedByID > 0 {
searchOpt.MentionID = &mentionedByID
}
ids, total, err := issue_indexer.SearchIssues(ctx, searchOpt)
if err != nil {
ctx.Error(http.StatusInternalServerError, "SearchIssues", err)
return
}
issues, err := issues_model.GetIssuesByIDs(ctx, ids, true)
if err != nil {
ctx.Error(http.StatusInternalServerError, "FindIssuesByIDs", err)
return
}
ctx.SetLinkHeader(int(total), listOptions.PageSize)
ctx.SetTotalCountHeader(total)
ctx.JSON(http.StatusOK, convert.ToAPIIssueList(ctx, issues))
}

View file

@ -309,7 +309,7 @@ func prepareForReplaceOrAdd(ctx *context.APIContext, form api.IssueLabelsOption)
return nil, nil, err
}
labels, err := issues_model.GetLabelsByIDs(form.Labels)
labels, err := issues_model.GetLabelsByIDs(form.Labels, "id", "repo_id", "org_id")
if err != nil {
ctx.Error(http.StatusInternalServerError, "GetLabelsByIDs", err)
return nil, nil, err