bypos-collector: 移除硬编码 sdogid、添加重试、补全 CSV 导出、增加测试

改进内容:
- sdogid 不再硬编码，优先从 $BYPOS_SDOGID 环境变量读取，其次 -sdogid 参数，
  也可在 Web 控制台输入框填写；三者都未设时启动警告、采集报错
- 网络错误自动重试（最多 3 次，指数退避 500ms/1s/2s）
- Web 控制台采集参数区增加 sdogid 输入框
- CSV 导出补全 retmsg 和 source 两列
- 新增 Go 单元测试（ean13Check、md5hex、sanitizeBarcodes、resolveSdogID 等）
- README 更新配置说明

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>

tools/bypos-collector/collect.go

@@ -20,12 +20,21 @@ import (
 )
 
 // ---- upstream config ----
-// sdogid is the 云店 account license id observed in the live request. It is
-// configurable so the tool is not tied to a single account.
-const defaultSdogID = "137966"
 
 const endpoint = "http://zc.bypos.net/byGoodsService/byMessage.asmx/GetGoodsinfo"
 
+// maxRetries is the number of retry attempts for transient network errors.
+const maxRetries = 3
+
+// resolveSdogID reads the account id from $BYPOS_SDOGID, falling back to the
+// explicit argument (CLI flag or Web UI input). Returns empty if neither set.
+func resolveSdogID(explicit string) string {
+	if v := os.Getenv("BYPOS_SDOGID"); v != "" {
+		return v
+	}
+	return explicit
+}
+
 var stringTagRe = regexp.MustCompile(`(?s)<string[^>]*>(.*)</string>`)
 
 // Product is the normalized record we persist (one JSON object per line).
@@ -87,8 +96,8 @@ func ean13Check(body string) (string, bool) {
 	return body + strconv.Itoa(chk), true
 }
 
-// lookup queries the upstream central library for one barcode.
-func (c *Collector) lookup(ctx context.Context, barcode string) (*Product, error) {
+// lookupOnce performs a single HTTP request to the upstream central library.
+func (c *Collector) lookupOnce(ctx context.Context, barcode string) (*Product, error) {
 	tsMs := strconv.FormatInt(time.Now().Unix()*1000, 10) // always ends in 000
 	sparm1 := md5hex(c.sdogID)
 	sparm2 := md5hex(barcode + tsMs)
@@ -147,6 +156,30 @@ func (c *Collector) lookup(ctx context.Context, barcode string) (*Product, error
 	return p, nil
 }
 
+// lookup queries the upstream with up to maxRetries retries on transient errors.
+func (c *Collector) lookup(ctx context.Context, barcode string) (*Product, error) {
+	var lastErr error
+	for attempt := 0; attempt <= maxRetries; attempt++ {
+		if ctx.Err() != nil {
+			return nil, ctx.Err()
+		}
+		p, err := c.lookupOnce(ctx, barcode)
+		if err == nil {
+			return p, nil
+		}
+		lastErr = err
+		if attempt < maxRetries {
+			backoff := time.Duration(1<<uint(attempt)) * 500 * time.Millisecond
+			select {
+			case <-time.After(backoff):
+			case <-ctx.Done():
+				return nil, ctx.Err()
+			}
+		}
+	}
+	return nil, lastErr
+}
+
 // ---- job / collector state ----
 
 type Stats struct {
@@ -187,9 +220,7 @@ type Collector struct {
 }
 
 func NewCollector(sdogID string) *Collector {
-	if sdogID == "" {
-		sdogID = defaultSdogID
-	}
+	sdogID = resolveSdogID(sdogID)
 	c := &Collector{
 		client: &http.Client{Timeout: 25 * time.Second},
 		sdogID: sdogID,
@@ -308,6 +339,9 @@ func (c *Collector) Start(req JobReq) error {
 	if req.SdogID != "" {
 		c.sdogID = req.SdogID
 	}
+	if c.sdogID == "" {
+		return fmt.Errorf("未指定 sdogid，请通过控制台输入框、环境变量 $BYPOS_SDOGID 或 -sdogid 参数配置")
+	}
 
 	// Build the list of barcodes to query.
 	var barcodes []string