Register a DoubaoVideo asset adapter that forwards the Ark-compatible
asset API (POST /api/v1/volcengine/asset?Action=xx) to the official
/openApi/portrait endpoint with a Volcengine V4 HMAC-SHA256 signature
(canonical request over content-type/host/x-content-sha256/x-date,
date/cn-beijing/ark/request scope). The AK/SK pair and optional
project code are stored in channel_asset_credentials (pool_id holds
the project code), so asset credentials stay separate from the video
Bearer key; the channel form now accepts them for type 54.
DoubaoVideo joins the seedance asset family so a user's asset upload
channel and video task channel stay the same (asset:// references
pass through to the upstream). Both resolution chains (asset proxy
resolver and video asset family matcher) now require the asset
credential to be configured before a type-54 channel can serve,
bind, or be auto-matched for assets, falling through to compatible
channels otherwise. The /api/v3 native-path whitelist simplifies
back to the family list now that 54 is a member.
Verified end-to-end against the real upstream: CreateAssetGroup,
CreateAsset (real video upload, Processing -> Active moderation),
ListAssets/GetAsset, and a video generation task referencing the
asset via asset://video reference (succeeded, 432900 tokens).
Co-Authored-By: ZCode <noreply@anthropic.com>
Add KlingAiping usage-token settlement from upstream final_unit_deduction and duration fallback handling.
Normalize model endpoint config persistence and convert model pricing values between stored USD and display currency in the editor.
Co-Authored-By: Codex <noreply@anthropic.com>
Add candidate user selection and explicit_ids batch creation flow for overseas migration. Also add cancellable batches to release selected-user locks and extend backend, frontend, and e2e coverage for the new flow.
Co-Authored-By: Codex <noreply@anthropic.com>
Add unit tests for group ratio, cross-group retry, and price calculation:
- Group ratio CRUD and special ratio override (12 tests)
- HandleGroupRatio with auto-group and special ratios (5 tests)
- ModelPriceHelper with group ratio in price/ratio modes (3 tests)
- Channel selection priority, groups, weighted LB (6 tests)
- RetryParam and AutoGroup service logic (7 tests)
Also restore the group pricing card in model detail SideSheet.
gpt-5.4 returns `arguments` as a JSON object for some tools (e.g.
apply_patch, tool_search), but ResponsesOutput.Arguments was typed as
string, causing json.Unmarshal to fail on every such SSE event.
The failure silently dropped all output_item events and the
response.completed event from the forwarded stream, so:
- the client never received tool call content
- usage/token counts could not be extracted (billed as 0)
- client reported "stream closed before response.completed"
Fix: change Arguments to json.RawMessage and add GetArguments() which
normalises both forms to a plain string (unescapes JSON strings,
returns raw bytes for objects).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
fengsilin
winter