69 lines
2.0 KiB
Go
69 lines
2.0 KiB
Go
package contentreviewser
|
|
|
|
import (
|
|
"91porn-server/common/log"
|
|
"91porn-server/models/v/contentreviewmod"
|
|
"91porn-server/models/v/mediacontentmod"
|
|
"91porn-server/models/v/mediamod"
|
|
|
|
"go.mongodb.org/mongo-driver/bson"
|
|
"go.mongodb.org/mongo-driver/bson/primitive"
|
|
)
|
|
|
|
// acgScanner ACG 扫描策略：每个 Media 可能产生 1 个 ACG_MEDIA + N 个 ACG_CONTENT 命中
|
|
// 仅匹配 title / summary / 章节名 / 章节内容（不检测 tag）
|
|
type acgScanner struct{}
|
|
|
|
func (s *acgScanner) CountTotal() (int64, error) {
|
|
return mediamod.CountForReview()
|
|
}
|
|
|
|
func (s *acgScanner) ScanBatch(taskID, lastID primitive.ObjectID, terms []string) (*ScanResult, error) {
|
|
batch, err := mediamod.FindForReviewBatch(lastID, scanBatchSize)
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
if len(batch) == 0 {
|
|
return &ScanResult{Done: true}, nil
|
|
}
|
|
|
|
issues := make([]*contentreviewmod.ReviewIssue, 0)
|
|
for _, m := range batch {
|
|
issues = append(issues, collectMediaIssues(taskID, m, terms)...)
|
|
}
|
|
|
|
return &ScanResult{
|
|
Issues: issues,
|
|
Checked: int64(len(batch)),
|
|
NextID: batch[len(batch)-1].ID,
|
|
}, nil
|
|
}
|
|
|
|
// collectMediaIssues 检测一个 Media 自身及其全部 MediaContent
|
|
func collectMediaIssues(taskID primitive.ObjectID, m *mediamod.Media, terms []string) []*contentreviewmod.ReviewIssue {
|
|
out := make([]*contentreviewmod.ReviewIssue, 0)
|
|
|
|
mb := newIssueBuilder(taskID, contentreviewmod.IssueTargetACGMedia, m.ID).
|
|
MatchTitle("标题", m.Title, terms).
|
|
MatchContent("内容", m.Summary, terms)
|
|
if mb.HasHits() {
|
|
out = append(out, mb.Build())
|
|
}
|
|
|
|
contents, err := mediacontentmod.QueryAllList(bson.M{"mediaId": m.ID, "isActive": true})
|
|
if err != nil {
|
|
log.Warn("query media contents fail", log.Any("mid", m.ID), log.E(err))
|
|
return out
|
|
}
|
|
for _, c := range contents {
|
|
cb := newIssueBuilder(taskID, contentreviewmod.IssueTargetACGContent, c.ID).
|
|
WithMediaID(m.ID).
|
|
MatchTitle("章节名", c.Name, terms).
|
|
MatchContent("章节内容", c.Text, terms)
|
|
if cb.HasHits() {
|
|
out = append(out, cb.Build())
|
|
}
|
|
}
|
|
return out
|
|
}
|