refactor: move channel model restriction from handler to scheduling phase

Move the model pricing restriction check from 8 handler entry points
to the account scheduling phase (SelectAccountForModelWithExclusions /
SelectAccountWithLoadAwareness), aligning restriction with billing:

- requested: check original request model against pricing list
- channel_mapped: check channel-mapped model against pricing list
- upstream: per-account check using account-mapped model

Handler layer now only resolves channel mapping (no restriction).
Scheduling layer performs pre-check for requested/channel_mapped,
and per-account filtering for upstream billing source.
This commit is contained in:
erio committed 2026-04-02 13:24:30 +08:00
1 parent 333ddfdb21
commit 513e653a4e
9 files changed
+100 -82

No files matched your search

+2 -6
View File
@@ -158,12 +158,8 @@ func (h *GatewayHandler) Messages(c *gin.Context) {
reqStream := parsedReq.Stream
reqLog = reqLog.With(zap.String("model", reqModel), zap.Bool("stream", reqStream))
// 解析渠道级模型映射 + 限制检查
channelMapping, restricted := h.gatewayService.ResolveChannelMappingAndRestrict(c.Request.Context(), apiKey.GroupID, reqModel)
if restricted {
h.errorResponse(c, http.StatusServiceUnavailable, "api_error", "The requested model is not available for this API key")
return
}
// 解析渠道级模型映射
channelMapping, _ := h.gatewayService.ResolveChannelMappingAndRestrict(c.Request.Context(), apiKey.GroupID, reqModel)
// 设置 max_tokens=1 + haiku 探测请求标识到 context 中
// 必须在 SetClaudeCodeClientContext 之前设置,因为 ClaudeCodeValidator 需要读取此标识进行绕过判断