Steve Dudenhoeffer
a9711ce904
Return errors for required fields (ID, price) and log warnings for optional fields (title, description, unit price) across all site extractors instead of silently discarding them with _ =. Closes #24 Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
90 lines
1.8 KiB
Go
90 lines
1.8 KiB
Go
package aislegopher
|
|
|
|
import (
|
|
"context"
|
|
"errors"
|
|
"fmt"
|
|
"log/slog"
|
|
"net/url"
|
|
"strconv"
|
|
"strings"
|
|
|
|
"gitea.stevedudenhoeffer.com/steve/go-extractor"
|
|
)
|
|
|
|
type Config struct {
|
|
}
|
|
|
|
var DefaultConfig = Config{}
|
|
|
|
var (
|
|
ErrInvalidURL = errors.New("invalid url")
|
|
)
|
|
|
|
type Item struct {
|
|
ID int
|
|
Name string
|
|
Price float64
|
|
}
|
|
|
|
func GetItemFromURL(ctx context.Context, b extractor.Browser, u *url.URL) (Item, error) {
|
|
return DefaultConfig.GetItemFromURL(ctx, b, u)
|
|
}
|
|
|
|
func (c Config) GetItemFromURL(ctx context.Context, b extractor.Browser, u *url.URL) (Item, error) {
|
|
res := Item{}
|
|
|
|
// the url will be in the format of aislegopher.com/p/slug/id
|
|
// we need to parse the slug and id from the url
|
|
a := strings.Split(u.Path, "/")
|
|
if len(a) != 4 {
|
|
return res, ErrInvalidURL
|
|
}
|
|
|
|
if a[1] != "p" {
|
|
return res, ErrInvalidURL
|
|
}
|
|
|
|
if u.Host != "aislegopher.com" && u.Host != "www.aislegopher.com" {
|
|
return res, ErrInvalidURL
|
|
}
|
|
|
|
var err error
|
|
res.ID, err = strconv.Atoi(a[3])
|
|
if err != nil {
|
|
return res, fmt.Errorf("failed to parse product ID %q: %w", a[3], err)
|
|
}
|
|
|
|
doc, err := b.Open(ctx, u.String(), extractor.OpenPageOptions{})
|
|
if err != nil {
|
|
return res, fmt.Errorf("failed to open page: %w", err)
|
|
}
|
|
defer extractor.DeferClose(doc)
|
|
|
|
names := doc.Select(".h4")
|
|
|
|
if len(names) > 0 {
|
|
res.Name, err = names[0].Text()
|
|
if err != nil {
|
|
slog.Warn("failed to get product name", "err", err)
|
|
}
|
|
}
|
|
|
|
prices := doc.Select(".h2")
|
|
|
|
if len(prices) > 0 {
|
|
priceStr, err := prices[0].Text()
|
|
if err != nil {
|
|
return res, fmt.Errorf("failed to get price text: %w", err)
|
|
}
|
|
priceStr = strings.ReplaceAll(priceStr, "$", "")
|
|
priceStr = strings.TrimSpace(priceStr)
|
|
res.Price, err = strconv.ParseFloat(priceStr, 64)
|
|
if err != nil {
|
|
return res, fmt.Errorf("failed to parse price %q: %w", priceStr, err)
|
|
}
|
|
}
|
|
|
|
return res, nil
|
|
}
|