Replace the per-batch ClickHouse mutation loop with a single
ALTER TABLE ... DELETE, since ClickHouse DELETE is a heavy mutation
that rewrites data parts and per-batch mutations are pathologically
slow. Add deterministic unit tests covering ClickHouse DSN handling,
main-database rejection, TTL DDL generation, log ordering, request_id
backfill, and display id assignment.
Fall back to the OS hostname when NODE_NAME is not configured, so node
identity in audit/usage logs is populated automatically. In container and
Kubernetes deployments the hostname equals the container ID or Pod name,
which stays unique under autoscaling without any manual per-instance config.
resolveModelProvider in model-badge.tsx only covered mainstream providers
(OpenAI/Claude/Gemini/Grok/DeepSeek/Qwen/Doubao/Moonshot/Mistral/Meta/Cohere).
Models from Zhipu (glm-), MiniMax, Xiaomi MiMo, Baidu (ernie), iFlyTek
(spark), Tencent (hunyuan), Baichuan, InternLM, StepFun, and Yi rendered
as grey dots instead of their brand icons in the usage logs table.
Add keyword mappings for these providers, aligning coverage with the
classic frontend's getModelCategories. Icons resolved via @lobehub/icons.
Add an eye toggle in the flow section header that masks sensitive node
labels (users, tokens, nodes, groups, channels) in the Sankey while keeping
model names visible. Masking only rewrites display text; nodes stay distinct
via their key so graph structure, links, and highlighting are unaffected.
- Highlight full paths through a clicked node or link in the flow Sankey,
dimming unrelated nodes/links instead of removing them
- Disable VChart built-in emphasis to avoid crash, use custom highlight sets
- Initialize models filter dialog from currently applied filters so manual
time ranges are not overridden by preferences; auto-pick granularity by range
- Lift user charts time range/granularity/limit to dashboard as controlled state
* feat(dashboard): add traffic flow sankey chart
Add dashboard flow APIs and a Sankey-based flow view with user, optional API key, model, and channel layers.\n\nReuse the dashboard VChart palette, add precise link/node tooltips and interactions, and cover filtering, layer ordering, color stability, and error states with tests.
* feat: build flow chart from quota data
---------
Co-authored-by: CaIon <i@caion.me>
- Show 3-column card grid from xl breakpoint instead of 2xl
- Cap inline priority/weight width to avoid huge values stretching cards
- Collapse right-column grid to content-sized columns, removing wasted space
- Memoize channel columns, context value, upstream-update result, and ChannelCard
to avoid rebuilding/re-rendering all cards on unrelated state changes
Group #id/name and balance on the left and priority/weight plus
response/last-tested on the right. Simplify field labels by dropping
the CJK-unfriendly uppercase/letter-spacing, left-align the right
column, and shorten the "Last Tested" zh label.
Add an opt-in card view to the shared data-table stack (DataTablePage),
toggled via a segmented control in the toolbar with per-table localStorage
persistence. Cards render generically from column meta by default, with an
optional renderCard slot. Defaults to table-only so existing pages are
unchanged.
Wire it into the channels page with a bespoke ChannelCard that reuses every
column's cell renderer, preserving all table information and interactions
(selection, inline priority/weight, balance refresh, status, actions, tag
expand).
- add oxlint and oxfmt configuration for web/default quality checks
- replace ESLint and Prettier scripts with protected-header Oxc formatting
- update frontend workspace dependencies for oxlint and oxfmt
CaIon