fix: harden explore dataset collection

This commit is contained in:
2026-07-23 21:34:13 +08:00
parent 3749cf4543
commit 50d8246915
9 changed files with 254 additions and 87 deletions

View File

@@ -4,6 +4,7 @@ import (
"context"
"errors"
"net/url"
"os"
"testing"
"github.com/stretchr/testify/require"
@@ -98,6 +99,29 @@ func TestValidateFeedURLRejectsLocalAddressesAndCredentials(t *testing.T) {
require.NoError(t, validateFeedURL(parsed))
}
func TestLiveFeedCollectionFormats(t *testing.T) {
if os.Getenv("TEST_LIVE_FEEDS") != "1" {
t.Skip("set TEST_LIVE_FEEDS=1 to verify external feeds")
}
fetcher := NewHTTPFeedFetcher()
for _, test := range []struct {
name string
url string
wantFormat string
}{
{name: "JSON Feed", url: "https://bbg.buzzing.cc/feed.json", wantFormat: "json_feed"},
{name: "RSS XML", url: "https://rsshub.ktachibana.party/cls/hot", wantFormat: "rss"},
{name: "second RSS XML", url: "https://rsshub.ktachibana.party/jin10", wantFormat: "rss"},
} {
t.Run(test.name, func(t *testing.T) {
feed, err := fetcher.Fetch(context.Background(), test.url)
require.NoError(t, err)
require.Equal(t, test.wantFormat, feed.Format)
require.NotEmpty(t, feed.Items)
})
}
}
func TestSyncSourcesRecordsFeedFailure(t *testing.T) {
database := newDatasetTestDatabase(t)
user := createDatasetTestUser(t, database, "feed-owner@example.com")

View File

@@ -38,6 +38,10 @@ func TestDatasetAPIConnectsSourcesItemsAndSyncTasks(t *testing.T) {
require.NotEmpty(t, source.ID)
require.Equal(t, "Industry feed", source.Name)
require.Empty(t, source.IconURL)
var storedSource models.SaDatasetSource
require.NoError(t, database.Where("identity = ?", source.ID).First(&storedSource).Error)
require.Equal(t, owner.ID, storedSource.CreatedBy)
require.Equal(t, owner.Identity, storedSource.CreatedByIdentity)
createdItem := performDatasetRequest(t, ownerRouter, http.MethodPost, "/api/v1/dataset-items", map[string]any{
"sourceId": source.ID, "title": "A collected article", "summary": "Summary",
@@ -108,6 +112,10 @@ func newDatasetTestDatabase(t *testing.T) *gorm.DB {
&models.SaDatasetItem{},
&models.SaDatasetCron{},
))
require.NoError(t, database.Exec(`
CREATE UNIQUE INDEX idx_sa_dataset_item_source_external
ON sa_dataset_items (source_id, external_id)
`).Error)
return database
}

View File

@@ -10,6 +10,7 @@ import (
"time"
"gorm.io/gorm"
"gorm.io/gorm/clause"
"senlinai-agent/backend/internal/models"
)
@@ -326,24 +327,20 @@ func (s *Service) storeFeedItems(userID uint, source models.SaDatasetSource, ite
inserted := 0
err := s.db.Transaction(func(tx *gorm.DB) error {
for _, input := range items {
var count int64
if err := tx.Model(&models.SaDatasetItem{}).
Where("source_id = ? AND external_id = ?", source.ID, input.ExternalID).
Count(&count).Error; err != nil {
return err
}
if count > 0 {
continue
}
externalID := input.ExternalID
item := models.SaDatasetItem{
SourceID: source.ID, CreatedBy: userID, ExternalID: input.ExternalID,
SourceID: source.ID, CreatedBy: userID, ExternalID: &externalID,
Title: input.Title, Summary: input.Summary, Content: input.Content,
URL: input.URL, Status: "unread", PublishedAt: input.PublishedAt,
}
if err := tx.Create(&item).Error; err != nil {
return err
result := tx.Clauses(clause.OnConflict{
Columns: []clause.Column{{Name: "source_id"}, {Name: "external_id"}},
DoNothing: true,
}).Create(&item)
if result.Error != nil {
return result.Error
}
inserted++
inserted += int(result.RowsAffected)
}
return nil
})