steve/go-extractor
0
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix: surface parsing errors instead of silently discarding them
All checks were successful
CI / vet (pull_request) Successful in 1m10s
Details
CI / build (pull_request) Successful in 1m21s
Details
CI / test (pull_request) Successful in 1m28s
Details

Browse Source 
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>
This commit is contained in:
Steve Dudenhoeffer
2026-02-15 16:31:56 +00:00
parent 7f24e97131
commit a9711ce904
4 changed files with 77 additions and 27 deletions
Download Patch File Download Diff File Expand all files Collapse all files

58
sites/wegmans/wegmans.go
View File

@@ -3,6 +3,7 @@ package wegmans
import (
"context"
"errors"
"fmt"
"log/slog"
"net/url"
"strconv"
@@ -54,7 +55,10 @@ func (c Config) GetItemPrice(ctx context.Context, b extractor.Browser, u *url.UR
return Item{}, ErrInvalidURL
}
id, _ := strconv.Atoi(a[2])
id, err := strconv.Atoi(a[2])
if err != nil {
return Item{}, fmt.Errorf("failed to parse product ID %q: %w", a[2], err)
}
if id == 0 {
return Item{}, ErrInvalidURL
@@ -67,7 +71,9 @@ func (c Config) GetItemPrice(ctx context.Context, b extractor.Browser, u *url.UR
defer extractor.DeferClose(doc)
timeout := 15 * time.Second
_ = doc.WaitForNetworkIdle(&timeout)
if err := doc.WaitForNetworkIdle(&timeout); err != nil {
slog.Warn("WaitForNetworkIdle failed", "err", err)
}
res := Item{
ID: id,
@@ -76,20 +82,29 @@ func (c Config) GetItemPrice(ctx context.Context, b extractor.Browser, u *url.UR
titles := doc.Select("h1[data-testid]")
if len(titles) != 0 {
res.Name, _ = titles[0].Text()
res.Name, err = titles[0].Text()
if err != nil {
slog.Warn("failed to get product name", "err", err)
}
}
prices := doc.Select("div.component--product-price:nth-child(1) > div:nth-child(1) > span:nth-child(1) > span:nth-child(2)")
slog.Info("prices", "len", len(prices))
if len(prices) != 0 {
priceStr, _ := prices[0].Text()
priceStr, err := prices[0].Text()
if err != nil {
return res, fmt.Errorf("failed to get price text: %w", err)
}
slog.Info("price", "0", prices[0], "text", priceStr)
priceStr = strings.ReplaceAll(priceStr, "$", "")
priceStr = strings.ReplaceAll(priceStr, ",", "")
// if there's a "/" in the price, then it's in the format of like "1.99/ea", so split it off
priceStr = strings.Split(priceStr, "/")[0]
price, _ := strconv.ParseFloat(priceStr, 64)
price, err := strconv.ParseFloat(priceStr, 64)
if err != nil {
return res, fmt.Errorf("failed to parse price %q: %w", priceStr, err)
}
slog.Info("price", "0", prices[0], "text", priceStr, "price", price)
res.Price = price
}
@@ -97,22 +112,29 @@ func (c Config) GetItemPrice(ctx context.Context, b extractor.Browser, u *url.UR
unitPrices := doc.Select(`div.component--product-price:nth-child(1) span.price-per-unit`)
if len(unitPrices) != 0 {
unitPriceStr, _ := unitPrices[0].Text()
unitPriceStr = strings.TrimSpace(unitPriceStr)
unitPriceStr = strings.ReplaceAll(unitPriceStr, "(", "")
unitPriceStr = strings.ReplaceAll(unitPriceStr, ")", "")
unitPriceStr = strings.ReplaceAll(unitPriceStr, "$", "")
unitPriceStr = strings.ReplaceAll(unitPriceStr, ",", "")
unitPriceStr, err := unitPrices[0].Text()
if err != nil {
slog.Warn("failed to get unit price text", "err", err)
} else {
unitPriceStr = strings.TrimSpace(unitPriceStr)
unitPriceStr = strings.ReplaceAll(unitPriceStr, "(", "")
unitPriceStr = strings.ReplaceAll(unitPriceStr, ")", "")
unitPriceStr = strings.ReplaceAll(unitPriceStr, "$", "")
unitPriceStr = strings.ReplaceAll(unitPriceStr, ",", "")
units := strings.Split(unitPriceStr, "/")
units := strings.Split(unitPriceStr, "/")
if len(units) > 1 {
res.Unit = strings.TrimSpace(units[1])
res.UnitPrice, _ = strconv.ParseFloat(units[0], 64)
if len(units) > 1 {
res.Unit = strings.TrimSpace(units[1])
res.UnitPrice, err = strconv.ParseFloat(units[0], 64)
if err != nil {
slog.Warn("failed to parse unit price", "text", units[0], "err", err)
}
// the unit might be like "lb.", so if it ends in a period, then just strip it off
if strings.HasSuffix(res.Unit, ".") {
res.Unit = strings.TrimSuffix(res.Unit, ".")
// the unit might be like "lb.", so if it ends in a period, then just strip it off
if strings.HasSuffix(res.Unit, ".") {
res.Unit = strings.TrimSuffix(res.Unit, ".")
}
}
}
}