1. Async task model redirection (aligned with sync tasks):
- Integrate ModelMappedHelper in RelayTaskSubmit after model name
determination, populating OriginModelName / UpstreamModelName on RelayInfo.
- All task adaptors now send UpstreamModelName to upstream providers:
- Gemini & Vertex: BuildRequestURL uses UpstreamModelName.
- Doubao & Ali: BuildRequestBody conditionally overwrites body.Model.
- Vidu, Kling, Hailuo, Jimeng: convertToRequestPayload accepts RelayInfo
and unconditionally uses info.UpstreamModelName.
- Sora: BuildRequestBody parses JSON and multipart bodies to replace
the "model" field with UpstreamModelName.
- Frontend log visibility: LogTaskConsumption and taskBillingOther now
emit is_model_mapped / upstream_model_name in the "other" JSON field.
- Billing safety: RecalculateTaskQuotaByTokens reads model name from
BillingContext.OriginModelName (via taskModelName) instead of
task.Data["model"], preventing billing leaks from upstream model names.
2. Per-call billing (TaskPricePatches lifecycle):
- Rename TaskBillingContext.ModelName → OriginModelName; add PerCallBilling
bool field, populated from TaskPricePatches at submission time.
- settleTaskBillingOnComplete short-circuits when PerCallBilling is true,
skipping both adaptor adjustments and token-based recalculation.
- Remove ModelName from TaskSubmitResult; use relayInfo.OriginModelName
consistently in controller/relay.go for billing context and logging.
3. Multipart retry boundary mismatch fix:
- Root cause: after Sora (or OpenAI audio) rebuilds a multipart body with a
new boundary and overwrites c.Request.Header["Content-Type"], subsequent
calls to ParseMultipartFormReusable on retry would parse the cached
original body with the wrong boundary, causing "NextPart: EOF".
- Fix: ParseMultipartFormReusable now caches the original Content-Type in
gin context key "_original_multipart_ct" on first call and reuses it for
all subsequent parses, making multipart parsing retry-safe globally.
- Sora adaptor reverted to the standard pattern (direct header set/get),
which is now safe thanks to the root fix.
4. Tests:
- task_billing_test.go: update makeTask to use OriginModelName; add
PerCallBilling settlement tests (skip adaptor adjust, skip token recalc);
add non-per-call adaptor adjustment test with refund verification.
- Enhanced the RelayTask function to utilize a locked channel when available, allowing for better reuse during retries.
- Updated error handling to ensure proper context setup for the selected channel.
- Clarified comments in ResolveOriginTask regarding channel locking and retry behavior.
- Introduced a new field in TaskRelayInfo to store the locked channel object, improving type safety and reducing import cycles.
- Renamed RelayTask function to RelayTaskFetch for clarity.
- Updated routing in relay-router.go and video-router.go to use RelayTaskFetch for fetch operations.
- Enhanced error handling in RelayTaskFetch function.
- Adjusted task data conversion in TaskAdaptor to include task ID.
Add three billing lifecycle methods to the TaskAdaptor interface:
- EstimateBilling: compute OtherRatios from user request before pricing
- AdjustBillingOnSubmit: adjust ratios from upstream submit response
- AdjustBillingOnComplete: determine final quota at task terminal state
Introduce BaseBilling as embeddable no-op default for adaptors without
custom billing. Move Sora/Ali OtherRatios logic from shared validation
into per-adaptor EstimateBilling implementations.
Add TaskBillingContext to persist pricing params (model_price, group_ratio,
other_ratios) in task private data for async polling settlement.
Extract RecalculateTaskQuota as a general-purpose delta settlement
function and unify polling billing via settleTaskBillingOnComplete
(adaptor-first, then token-based fallback).
Restructure the task relay system for better separation of concerns:
- Extract task billing into service/task_billing.go with unified settlement flow
- Move task polling loop from controller to service/task_polling.go (supports Suno + video platforms)
- Split RelayTask into fetch/submit paths with dedicated retry logic (taskSubmitWithRetry)
- Add TaskDto, TaskResponse generics, and FetchReq to dto/task.go
- Add taskcommon/helpers.go for shared task adaptor utilities
- Remove controller/task_video.go (logic consolidated into service layer)
- Update all task adaptors (ali, doubao, gemini, hailuo, jimeng, kling, sora, suno, vertex, vidu)
- Simplify frontend task logs to use new TaskDto response format
* ci: create docker automation
* ✨ feat: add subscription billing system with admin management and user purchase flow
Implement a new subscription-based billing model alongside existing metered/per-request billing:
Backend:
- Add subscription plan models (SubscriptionPlan, SubscriptionPlanItem, UserSubscription, etc.)
- Implement CRUD APIs for subscription plan management (admin only)
- Add user subscription queries with support for multiple active/expired subscriptions
- Integrate payment gateways (Stripe, Creem, Epay) for subscription purchases
- Implement pre-consume and post-consume billing logic for subscription quota tracking
- Add billing preference settings (subscription_first, wallet_first, etc.)
- Enhance usage logs with subscription deduction details
Frontend - Admin:
- Add subscription management page with table view and drawer-based edit form
- Match UI/UX style with existing admin pages (redemption codes, users)
- Support enabling/disabling plans, configuring payment IDs, and model quotas
- Add user subscription binding modal in user management
Frontend - Wallet:
- Add subscription plans card with current subscription status display
- Show all subscriptions (active and expired) with remaining days/usage percentage
- Display purchasable plans with pricing cards following SaaS best practices
- Extract purchase modal to separate component matching payment confirm modal style
- Add skeleton loading states with active animation
- Implement billing preference selector in card header
- Handle payment gateway availability based on admin configuration
Frontend - Usage Logs:
- Display subscription deduction details in log entries
- Show step-by-step breakdown of subscription usage (pre-consumed, delta, final, remaining)
- Add subscription deduction tag for subscription-covered requests
* ✨ feat(admin): add user subscription management and refine UI/pagination
Add admin APIs to list/create/invalidate/delete user subscriptions
Add model helpers to fetch all user subscriptions (incl. expired) and support cancel/hard-delete
Wire new admin routes for user subscription operations
Replace “Bind subscription plan” entry with a dedicated User Subscriptions SideSheet in Users table
Use CardTable with responsive layout and working client-side pagination inside the SideSheet
Improve subscription purchase modal empty-gateway state with a Banner notice
* ✨ feat(admin): streamline subscription plan benefits editor with bulk actions
Restore the avatar/icon header for the “Model Benefits” section
Replace scattered controls with a compact toolbar-style workflow
Support multi-select add with a default quota for new items
Add row selection with bulk apply-to-selected / apply-to-all quota updates
Enable delete-selected to manage benefits faster and reduce mistakes
* ✨ fix(subscription): finalize payments, log billing, and clean up dead code
Complete subscription orders by creating a matching top-up record and writing billing logs
Add Epay return handler to verify and finalize browser callbacks
Require Stripe/Creem webhook configuration before starting subscription payments
Show subscription purchases in topup history with clearer labels/methods
Remove unused subscription helper, legacy Creem webhook struct, and unused topup fields
Simplify subscription self API payload to active/all lists only
* 🎨 style: format all code with gofmt and lint:fix
Apply consistent code formatting across the entire codebase using
gofmt and lint:fix tools. This ensures adherence to Go community
standards and improves code readability and maintainability.
Changes include:
- Run gofmt on all .go files to standardize formatting
- Apply lint:fix to automatically resolve linting issues
- Fix code style inconsistencies and formatting violations
No functional changes were made in this commit.
* ✨ feat(subscription): add quota reset periods and admin configuration
- Add reset period fields on subscription plans and user items
- Apply automatic quota resets during pre-consume based on plan schedule
- Expose reset-period configuration in the admin plan editor
- Display reset cadence in subscription cards and purchase modal
- Validate custom reset seconds on plan create/update
* ✨ feat(subscription): harden subscription billing with resets, idempotency, and production-grade stability
Add plan-level quota reset periods and display/reset cadence in admin/UI
Enforce natural reset alignment with background reset task and cleanup job
Make subscription pre-consume/refund idempotent with request-scoped records and retries
Use database time for consistent resets across multi-instance deployments
Harden payment callbacks with locking and idempotent order completion
Record subscription purchases in topup history and billing logs
Optimize subscription queries and add critical composite indexes
* ✨ feat(subscription): cache plan lookups and stabilize pre-consume
Introduce hybrid caches for subscription plans, items, and plan info with explicit
invalidation on admin updates. Streamline pre-consume transactions to reduce
redundant queries while preserving idempotency and reset logic.
* 🐛 fix(subscription): avoid pre-consume lookup noise
Use a RowsAffected check for the idempotency lookup so missing records
no longer surface as "record not found" errors while preserving behavior.
* 🔧 ci: Change workflow trigger to sub branch
Update the Docker image workflow to run on pushes to the sub branch instead of main.
* 💸 chore: Align subscription pricing display with global currency settings
Unify subscription price rendering to use the site-wide currency symbol/rate on the wallet and admin views.
Make subscription plan currency read-only in the editor and force USD on create/update to avoid drift.
Use global currency display type when creating Creem checkout payloads.
* 🔧 chore: Unify subscription plan status toggle with PATCH endpoint
Replace separate enable/disable flows with a single PATCH API that updates the enabled flag.
Update frontend hooks and table actions to call the unified endpoint and keep UI behavior consistent.
Introduce a minimal admin controller handler and route for the status update.
* ✨ feat: Add subscription limits and UI tags consistency
Add per-plan purchase limits with backend enforcement and UI disable states.
Expose limit configuration in admin plan editor and show limits in plan tables/cards.
Refine subscription UI tags with unified badge style and streamlined “My Subscriptions” layout.
* 🎨 style: tag color to white
* 🚀 refactor: Simplify subscription quota to total amount model
Remove per-model subscription items and switch to a single total quota per plan and user subscription. Update billing, reset, and logging flows to operate on total quota, and refactor admin/user UI to configure and display total quota consistently.
* 🚀 chore: Remove duplicate subscription usage percentage display
Keep the usage percentage shown only in the total quota line to avoid redundant “已用 0%” text while preserving remaining days in the summary.
* ✨ feat: Add subscription upgrade group with auto downgrade
* ✨ feat: Update subscription purchase modal display
Show total quota as currency with tooltip for raw quota, hide reset cycle when never, and display upgrade group when configured to match card display rules.
* ✨ feat: Extract quota conversion helpers to shared utils
Move quota display/conversion helpers into web/src/helpers/quota.js and update the subscription plan editor to import and use the shared utilities instead of inline functions.
* ✨ chore: Add upgrade group guidance in subscription editor
Add explanatory helper text under the upgrade group field to clarify automatic group upgrades, rollback conditions, and the expected delay before downgrading takes effect.
* 🔧 chore: remove unused Creem settings state
Drop the unused originInputs state and redundant updates to keep the Creem
settings form state minimal and easier to maintain.
* 🚀 chore: Remove useless action
* ✨ Add full i18n coverage for subscription-related UI across locales
* ✨ feat: harden subscription billing and improve UI consistency
Improve subscription payment safety and data integrity by handling user/URL lookup failures, fixing Stripe subscription mode, persisting quota reset fields, and correcting subscription delta accounting and DB timestamp casting. Refine the UI with stricter custom duration validation, accurate currency rounding, conditional Epay labeling, rollback on preference update failure, and shared subscription formatting helpers plus clearer component naming.
* 🔧 fix: make epay webhook and return flow subscription-aware
Ensure Epay webhook acknowledges success only after order completion, returning fail on processing errors to allow retries. Redirect subscription payment returns to the subscription page instead of top-up for correct user flow.
* 🚦 fix: guard epay return success on order completion
Redirect subscription return flow to failure when order completion fails, preventing false success states after payment verification.
* 🔧 fix: normalize epay error handling and webhook retries
Standardize SubscriptionRequestEpay error responses via ApiErrorMsg for a consistent schema.
Return "fail" on non-success trade statuses in the epay webhook to preserve retry behavior.
* 🧾 fix: persist epay orders before purchase
Create the subscription order before initiating epay payment and expire it if the provider call fails, preventing orphaned transactions and improving reconciliation.
* 🔧 fix: harden epay callbacks and billing fallbacks
Use POST and form parsing for epay notify/return routes, persist epay orders before provider calls with expiry on failure, and ensure notify handlers retry correctly.
Restrict subscription-first fallback to insufficient-subscription errors and log refund failures after retries to avoid silent quota drift.
* 🔧 fix: harden billing flow and sidebar settings
Add missing strings import for subscription fallback checks, log failed subscription refunds after retries, and extend sidebar module settings with a subscription management toggle plus translations.
* 🛡️ fix: fail fast on epay form parse errors
Handle ParseForm errors in epay notify/return handlers by returning fail or redirecting to failure, avoiding unsafe fallback to query parameters.
* ✨ fix: refine Japanese subscription status labels
Adjust Japanese UI wording for active-count labels to read more naturally and consistently.
* ✅ fix: standardize epay success response schema
Return subscription epay pay success responses via ApiSuccess to include the consistent success field and align with error schema.
* fix: channel affinity log styles
* fix: Issue with incorrect data storage when switching key sources
* feat: support not retrying after a single rule configuration fails
* fix: render channel affinity tooltip as multiline content
* feat: channel affinity cache hit
* fix: prevent ChannelAffinityUsageCacheModal infinite loading and hide data before fetch
* chore: format backend with gofmt and frontend with prettier/eslint autofix
Tighten oversized request handling across relay paths and make error matching reliable.
- Align `MAX_REQUEST_BODY_MB` fallback to `32` in request body reader and decompression middleware
- Stop ignoring `GetRequestBody` errors in relay retry paths; return consistent **413** on oversized bodies (400 for other read errors)
- Add `Unwrap()` to `types.NewAPIError` so `errors.Is/As` can match wrapped underlying errors
- `go test ./...` passes
Clamp request body size (including post-decompression) to avoid memory exhaustion caused by huge payloads/zip bombs, especially with large-context Claude requests. Add a configurable `MAX_REQUEST_BODY_MB` (default `32`) and document it.
- Enforce max request body size after gzip/br decompression via `http.MaxBytesReader`
- Add a secondary size guard in `common.GetRequestBody` and cache-safe handling
- Return **413 Request Entity Too Large** on oversized bodies in relay entry
- Avoid building large `TokenCountMeta.CombineText` when both token counting and sensitive check are disabled (use lightweight meta for pricing)
- Update READMEs (CN/EN/FR/JA) with `MAX_REQUEST_BODY_MB`
- Fix a handful of vet/formatting issues encountered during the change
- `go test ./...` passes
- Introduced new OpenAI text models in `common/model.go`.
- Added `IsOpenAITextModel` function to check for OpenAI text models.
- Refactored token estimation methods across various channels to use estimated prompt tokens instead of direct prompt token counts.
- Updated related functions and structures to accommodate the new token estimation approach, enhancing overall token management.
This commit refactors the logging mechanism across the application by replacing direct logger calls with a centralized logging approach using the `common` package. Key changes include:
- Replaced instances of `logger.SysLog` and `logger.FatalLog` with `common.SysLog` and `common.FatalLog` for consistent logging practices.
- Updated resource initialization error handling to utilize the new logging structure, enhancing maintainability and readability.
- Minor adjustments to improve code clarity and organization throughout various modules.
This change aims to streamline logging and improve the overall architecture of the codebase.
This commit introduces a major architectural refactoring to improve quota management, centralize logging, and streamline the relay handling logic.
Key changes:
- **Pre-consume Quota:** Implements a new mechanism to check and reserve user quota *before* making the request to the upstream provider. This ensures more accurate quota deduction and prevents users from exceeding their limits due to concurrent requests.
- **Unified Relay Handlers:** Refactors the relay logic to use generic handlers (e.g., `ChatHandler`, `ImageHandler`) instead of provider-specific implementations. This significantly reduces code duplication and simplifies adding new channels.
- **Centralized Logger:** A new dedicated `logger` package is introduced, and all system logging calls are migrated to use it, moving this responsibility out of the `common` package.
- **Code Reorganization:** DTOs are generalized (e.g., `dalle.go` -> `openai_image.go`) and utility code is moved to more appropriate packages (e.g., `common/http.go` -> `service/http.go`) for better code structure.
