Added: downloading and returning wikipedia articles

This commit is contained in:
Oliwier Adamczyk
2025-10-04 23:14:41 +02:00
parent f542f01b49
commit 6df63dc4c1
26 changed files with 636 additions and 100 deletions

11
internal/article/dto.go Normal file
View File

@@ -0,0 +1,11 @@
package article
type ArticleDTO struct {
Uuid string
Title string
Content string
}
type ArticleQueryDTO struct {
Title string
}

12
internal/article/error.go Normal file
View File

@@ -0,0 +1,12 @@
package article
import "errors"
var (
ErrArticleDownloadFailed = errors.New("article: download failed")
ErrArticleQueryFailed = errors.New("article: article query failed")
ErrArticleCreateFailed = errors.New("article: create failed")
ErrArticleDeleteAllFailed = errors.New("article: failed to delete all articles")
ErrArticleTitleInvalidLength = errors.New("article: invalid article length")
)

View File

@@ -0,0 +1,9 @@
package article
import "database/sql"
type IArticleRepository interface {
CreateArticle(tx *sql.Tx, data ArticleCreateModel) error
GetArticlesByTitle(tx *sql.Tx, title string) ([]ArticleModel, error)
DeleteAllArticles(tx *sql.Tx) error
}

View File

@@ -0,0 +1,6 @@
package article
type IArticleService interface {
DownloadArticles() error
QueryArticles(ArticleQueryDTO) ([]ArticleDTO, error)
}

13
internal/article/model.go Normal file
View File

@@ -0,0 +1,13 @@
package article
type ArticleModel struct {
Uuid string
Title string
Content string
}
type ArticleCreateModel struct {
Uuid string
Title string
Content string
}

View File

@@ -0,0 +1,60 @@
package article
import (
"database/sql"
"fmt"
)
type ArticleRepository struct{}
func NewArticleRepository() IArticleRepository {
return &ArticleRepository{}
}
func (ArticleRepository) CreateArticle(tx *sql.Tx, data ArticleCreateModel) error {
query := `
INSERT INTO articles(uuid, title, content)
VALUES ($1, $2, $3);
`
_, err := tx.Exec(query, data.Uuid, data.Title, data.Content)
return err
}
func (ArticleRepository) GetArticlesByTitle(tx *sql.Tx, title string) ([]ArticleModel, error) {
fmt.Println(title, " ------------------")
query := `
SELECT uuid, title, content
FROM articles
WHERE title LIKE $1 || '%'
LIMIT 10;
`
rows, err := tx.Query(query, title)
if err != nil {
return nil, err
}
articles := []ArticleModel{}
for rows.Next() {
var a ArticleModel
err = rows.Scan(&a.Uuid, &a.Title, &a.Content)
if err != nil {
return nil, err
}
articles = append(articles, a)
}
fmt.Println(articles)
return articles, nil
}
func (ArticleRepository) DeleteAllArticles(tx *sql.Tx) error {
query := `DELETE FROM articles;`
_, err := tx.Exec(query)
return err
}

107
internal/article/service.go Normal file
View File

@@ -0,0 +1,107 @@
package article
import (
"log"
"scrap/internal/db"
"scrap/internal/wikipediadl"
"github.com/google/uuid"
)
type ArticleService struct {
txRepo db.ITxRepository
articleRepo IArticleRepository
}
func NewArticleService(
txRepo db.ITxRepository,
articleRepo IArticleRepository,
) IArticleService {
return &ArticleService{
txRepo: txRepo,
articleRepo: articleRepo,
}
}
func (a ArticleService) QueryArticles(data ArticleQueryDTO) ([]ArticleDTO, error) {
tx, err := a.txRepo.Begin()
if err != nil {
log.Println(err.Error())
return nil, err
}
defer a.txRepo.RollbackOnError(tx, &err)
articleTitleLength := len(data.Title)
if articleTitleLength < 1 || articleTitleLength > 255 {
return nil, ErrArticleTitleInvalidLength
}
articles, err := a.articleRepo.GetArticlesByTitle(tx, data.Title)
if err != nil {
log.Println(err.Error())
return nil, ErrArticleQueryFailed
}
articlesOut := make([]ArticleDTO, 0, len(articles))
for _, am := range articles {
a := ArticleDTO{
Uuid: am.Uuid,
Title: am.Title,
Content: am.Content,
}
articlesOut = append(articlesOut, a)
}
return articlesOut, nil
}
func (a ArticleService) DownloadArticles() error {
tx, err := a.txRepo.Begin()
if err != nil {
log.Println(err.Error())
return db.ErrTxBeginFailed
}
defer a.txRepo.RollbackOnError(tx, &err)
if err = a.articleRepo.DeleteAllArticles(tx); err != nil {
log.Println(err.Error())
return ErrArticleDeleteAllFailed
}
articleBundles, err := wikipediadl.FetchArticleBundles()
if err != nil {
log.Println(err.Error())
return ErrArticleDownloadFailed
}
for _, ab := range articleBundles {
articles, err := wikipediadl.ExtractArticles(ab)
if err != nil {
log.Println(err.Error())
return ErrArticleDownloadFailed
}
for _, article := range articles {
articleData := ArticleCreateModel{
Uuid: uuid.NewString(),
Title: article.Title,
Content: article.Revision.Text,
}
if err = a.articleRepo.CreateArticle(tx, articleData); err != nil {
log.Println(err.Error(), "tutaj ---------")
continue
}
}
}
if err = a.txRepo.Commit(tx); err != nil {
log.Println(err.Error())
return db.ErrTxCommitFailed
}
return nil
}