kls/links/url.go

package links

import (
	"context"
	"log"
	"net/url"
	"time"

	"github.com/Masterminds/squirrel"
	"github.com/google/uuid"
	"github.com/jackc/pgx/v4"
	"github.com/jackc/pgx/v4/pgxpool"
)

var psql = squirrel.StatementBuilder.PlaceholderFormat(squirrel.Dollar)

type URL struct {
	ID        string
	URL       string
	NURL      string // normalized URL
	Short     string
	CreatedAt time.Time
}

func (u *URL) Timestamp() string {
	return u.CreatedAt.Format("2006-01-02 15:04")
}

func (u *URL) Store(ctx context.Context, db *pgxpool.Pool) error {
	stmt := psql.Insert("urls").
		Columns("id", "url", "nurl", "short", "created_at").
		Values(u.ID, u.URL, u.NURL, u.Short, u.CreatedAt)
	query, args, err := stmt.ToSql()
	if err != nil {
		return err
	}

	_, err = db.Exec(ctx, query, args...)
	return err
}

// Normalize cleans the URL to only the parts we care about.
func Normalize(u *url.URL) *url.URL {
	norm := &url.URL{
		Scheme:  u.Scheme,
		Host:    u.Host,
		Path:    u.Path,
		RawPath: u.RawPath,
	}
	return norm
}

func NormalizeString(s string) (string, error) {
	u, err := url.Parse(s)
	if err != nil {
		return "", err
	}

	u = Normalize(u)
	return u.String(), nil
}

// Clean should scrub out junk from the URL.
func Clean(u *url.URL, keepQuery bool) *url.URL {
	norm := &url.URL{
		Scheme:      u.Scheme,
		Host:        u.Host,
		Path:        u.Path,
		RawPath:     u.RawPath,
		Fragment:    u.Fragment,
		RawFragment: u.RawFragment,
	}

	if keepQuery {
		norm.RawQuery = u.RawQuery
	}
	return norm
}

func CleanString(s string, isRawURL bool) (string, error) {
	u, err := url.Parse(s)
	if err != nil {
		return "", err
	}

	u = Clean(u, isRawURL)
	return u.String(), nil
}

func New(u *url.URL) *URL {
	link := &URL{
		ID:        uuid.NewString(),
		URL:       u.String(),
		NURL:      Normalize(u).String(),
		Short:     GenCode(),
		CreatedAt: time.Now(),
	}

	return link
}

func FromString(s string) (*URL, error) {
	u, err := url.Parse(s)
	if err != nil {
		return nil, err
	}
	return New(u), nil
}

func Lookup(ctx context.Context, db *pgxpool.Pool, s string) (string, error) {
	u, err := url.Parse(s)
	if err != nil {
		return "", nil
	}

	u = Normalize(u)

	stmt := psql.Select("short").From("urls").
		Where(squirrel.Eq{"nurl": u.String()})
	query, args, err := stmt.ToSql()
	if err != nil {
		return "", err
	}

	row := db.QueryRow(ctx, query, args...)

	var short string
	err = row.Scan(&short)
	if err != nil {
		return "", err
	}

	return short, nil
}

func StoreURL(ctx context.Context, db *pgxpool.Pool, s string) (string, error) {
	short, err := Lookup(ctx, db, s)
	if err == nil {
		return short, nil
	}

	if err != pgx.ErrNoRows {
		return "", err
	}

	u, err := FromString(s)
	if err != nil {
		return "", err
	}

	err = u.Store(ctx, db)
	if err != nil {
		return "", err
	}

	return u.Short, nil
}

func RetrieveURL(ctx context.Context, db *pgxpool.Pool, short string) (string, error) {
	log.Printf("look up url for short code %s", short)
	stmt := psql.Select("url").From("urls").
		Where(squirrel.Eq{"short": short})
	query, args, err := stmt.ToSql()
	if err != nil {
		return "", err
	}

	row := db.QueryRow(ctx, query, args...)
	var url string

	err = row.Scan(&url)
	if err != nil {
		return "", err
	}

	return url, nil
}

func FetchAll(ctx context.Context, db *pgxpool.Pool) ([]*URL, error) {
	stmt := psql.Select("*").From("urls").OrderBy("created_at")
	query, args, err := stmt.ToSql()
	if err != nil {
		return nil, err
	}

	rows, err := db.Query(ctx, query, args...)
	if err != nil {
		return nil, err
	}

	var urls []*URL
	for rows.Next() {
		u := &URL{}
		err = rows.Scan(
			&u.ID,
			&u.URL,
			&u.NURL,
			&u.Short,
			&u.CreatedAt,
		)
		if err != nil {
			return nil, err
		}

		urls = append(urls, u)
	}

	return urls, nil
}