feat: enhance text protocol conversion and advanced custom routing (#5825)

* refactor: consolidate relay protocol converters

* refactor relayconvert text converters

* feat: refine relay converters and advanced custom routing

* refactor: enhance logging and add thought signature handling for Gemini requests

* refactor: enhance channel cache and pricing endpoint handling for advanced custom models

* feat: preserve billing usage semantics

* feat: add protocol-aware billing usage

* Delete useless files

* chore: update action versions in workflow files

* chore: update Docker action versions in workflow files

* fix: harden billing usage settlement and hot-path route matching

- estimate Gemini completion tokens locally when billable usageMetadata is
  prompt-only but output content was received (e.g. client aborts the stream
  before the final chunk), and rebuild the attached billing_usage as estimated
  so settlement does not bill zero output tokens
- guard NewClaudeMessagesBillingUsage against all-zero ClaudeUsage, matching
  the OpenAI/Gemini constructors, so a zero billing_usage cannot override a
  non-zero top-level usage during settlement
- cache compiled advanced-custom route model regexes; they run on the request
  hot path and were recompiled per request
- move the effectiveBillingUsage remap to PostTextConsumeQuota only, and
  document that calculateTextQuotaSummary expects remapped usage
- document the updatePricingLock -> channelSyncLock lock ordering that
  InitChannelCache/CacheUpdateChannel rely on, and the aux-struct pitfall in
  GeminiChatResponse.UnmarshalJSON

service/text_quota.go

@@ -175,6 +175,9 @@ func composeTieredTextQuota(relayInfo *relaycommon.RelayInfo, summary textQuotaS
 	return total
 }
 
+// calculateTextQuotaSummary expects a usage already remapped by
+// effectiveBillingUsage; PostTextConsumeQuota performs that remap once and shares
+// the result with tiered billing, affinity observation and logging.
 func calculateTextQuotaSummary(ctx *gin.Context, relayInfo *relaycommon.RelayInfo, usage *dto.Usage) textQuotaSummary {
 	summary := textQuotaSummary{
 		ModelName:            relayInfo.OriginModelName,
@@ -335,15 +338,16 @@ func usageSemanticFromUsage(relayInfo *relaycommon.RelayInfo, usage *dto.Usage)
 
 func PostTextConsumeQuota(ctx *gin.Context, relayInfo *relaycommon.RelayInfo, usage *dto.Usage, extraContent []string) {
 	originUsage := usage
+	billingUsage := effectiveBillingUsage(usage)
 	if usage == nil {
 		extraContent = append(extraContent, "上游无计费信息")
 	}
 	if originUsage != nil {
-		ObserveChannelAffinityUsageCacheByRelayFormat(ctx, usage, relayInfo.GetFinalRequestRelayFormat())
+		ObserveChannelAffinityUsageCacheByRelayFormat(ctx, billingUsage, relayInfo.GetFinalRequestRelayFormat())
 	}
 
 	adminRejectReason := common.GetContextKeyString(ctx, constant.ContextKeyAdminRejectReason)
-	summary := calculateTextQuotaSummary(ctx, relayInfo, usage)
+	summary := calculateTextQuotaSummary(ctx, relayInfo, billingUsage)
 
 	var tieredResult *billingexpr.TieredResult
 	tieredBillingApplied := false
@@ -352,7 +356,7 @@ func PostTextConsumeQuota(ctx *gin.Context, relayInfo *relaycommon.RelayInfo, us
 		if snap := relayInfo.TieredBillingSnapshot; snap != nil {
 			tieredUsedVars = billingexpr.UsedVars(snap.ExprString)
 		}
-		tieredOk, tieredQuota, tieredRes := TryTieredSettle(relayInfo, BuildTieredTokenParams(usage, summary.IsClaudeUsageSemantic, tieredUsedVars))
+		tieredOk, tieredQuota, tieredRes := TryTieredSettle(relayInfo, BuildTieredTokenParams(billingUsage, summary.IsClaudeUsageSemantic, tieredUsedVars))
 		if tieredOk {
 			tieredBillingApplied = true
 			tieredResult = tieredRes
@@ -412,6 +416,7 @@ func PostTextConsumeQuota(ctx *gin.Context, relayInfo *relaycommon.RelayInfo, us
 	} else {
 		other = GenerateTextOtherInfo(ctx, relayInfo, summary.ModelRatio, summary.GroupRatio, summary.CompletionRatio, summary.CacheTokens, summary.CacheRatio, summary.ModelPrice, relayInfo.PriceData.GroupRatioInfo.GroupSpecialRatio)
 	}
+	appendUsageBillingPathForLog(other, common.GetContextKeyBool(ctx, constant.ContextKeyLocalCountTokens), originUsage)
 	if adminRejectReason != "" {
 		other["reject_reason"] = adminRejectReason
 	}
@@ -462,12 +467,12 @@ func PostTextConsumeQuota(ctx *gin.Context, relayInfo *relaycommon.RelayInfo, us
 		// to cache_creation_tokens.
 		other["cache_write_tokens"] = cacheWriteTokens
 	}
-	if relayInfo.GetFinalRequestRelayFormat() != types.RelayFormatClaude && usage != nil && usage.UsageSource != "" && usage.InputTokens > 0 {
+	if relayInfo.GetFinalRequestRelayFormat() != types.RelayFormatClaude && billingUsage != nil && billingUsage.UsageSource != "" && billingUsage.InputTokens > 0 {
 		// input_tokens_total: explicit normalized total input used by the usage log UI.
 		// Only write this field when upstream/current conversion has already provided a
 		// reliable total input value and tagged the usage source. Do not infer it from
 		// prompt/cache fields here, otherwise old upstream payloads may be double-counted.
-		other["input_tokens_total"] = usage.InputTokens
+		other["input_tokens_total"] = billingUsage.InputTokens
 	}
 	if tieredBillingApplied {
 		InjectTieredBillingInfo(other, relayInfo, tieredResult)