feat(usage): support importing model pricing from models.dev (#4079)

* feat(usage): support importing model pricing from models.dev

Add an "Import from models.dev" button to the Add Pricing panel that
fetches https://models.dev/api.json, lists priced models sorted by
release date (newest 50 by default, full-text search across ~4800),
and bulk-imports the selected entries through the same
update_model_pricing command as manual entry.

- Normalize imported model IDs to match the backend's
  clean_model_id_for_pricing rules (strip vendor prefix, lowercase,
  truncate ':' suffix, map '@' to '-', drop the [1m] marker) so the
  stored rows actually match cost-attribution lookups
- Dedupe selections that collapse to the same model_id and report
  skipped duplicates in the success toast
- Invalidate usage queries on settled (not just success) so partial
  import failures still refresh the pricing list
- Keep ESC inside the picker's search input from closing the dialog
  and discarding the selection
- Add i18n keys for zh/en/zh-TW/ja and unit tests for the
  normalization, price formatting and flattening logic

Fixes #4017

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(usage): match scoped cost backfill against raw model aliases

The scoped backfill selected zero-cost rows via exact SQL string match,
but log columns store raw model strings (route prefixes, :free variants,
date suffixes), so alias rows were skipped until the next full backfill
on startup. Filter rows in Rust with the same model_pricing_candidates
normalization used by the pricing lookup; pricing decision logic is
untouched. Pre-existing gap from schema v11, surfaced by bulk import.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* refactor(usage): restrict models.dev pricing import to a single model

Each update_model_pricing call triggers a backfill pass that loads
every zero-cost usage row before filtering by model, so bulk-importing
N entries scaled as selectedModels x allZeroCostLogs full scans.
Re-importing pricing is rare, so drop the batch path instead of
optimizing it: the picker is now single-select, one import runs exactly
one update_model_pricing call and one backfill pass. This also removes
the normalized-ID dedup logic and the useImportModelPricing hook in
favor of the existing useUpdateModelPricing.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
This commit is contained in:
Xinyu Kuo
2026-06-16 22:01:56 +08:00
committed by GitHub
parent 81d6002ace
commit 0bb3b7515a
8 changed files with 778 additions and 16 deletions
@@ -0,0 +1,144 @@
import { describe, expect, it } from "vitest";
import {
flattenModels,
formatPrice,
normalizeModelIdForPricing,
} from "@/components/usage/ModelsDevPickerDialog";
describe("normalizeModelIdForPricing", () => {
it("keeps already-normalized ids unchanged", () => {
expect(normalizeModelIdForPricing("claude-opus-4-5")).toBe(
"claude-opus-4-5",
);
});
it("strips the vendor prefix before the last slash", () => {
expect(normalizeModelIdForPricing("z-ai/glm-4.7")).toBe("glm-4.7");
expect(normalizeModelIdForPricing("clarifai/main/models/mm-poly-8b")).toBe(
"mm-poly-8b",
);
});
it("lowercases the id", () => {
expect(normalizeModelIdForPricing("MiniMaxAI/MiniMax-M2.1")).toBe(
"minimax-m2.1",
);
});
it("truncates colon suffixes", () => {
expect(normalizeModelIdForPricing("claude-sonnet-4-thinking:8192")).toBe(
"claude-sonnet-4-thinking",
);
});
it("maps @ to -", () => {
expect(normalizeModelIdForPricing("claude-sonnet-4@20250514")).toBe(
"claude-sonnet-4-20250514",
);
});
it("strips the [1m] context marker", () => {
expect(normalizeModelIdForPricing("claude-sonnet-4-5[1m]")).toBe(
"claude-sonnet-4-5",
);
});
it("combines all rules", () => {
expect(normalizeModelIdForPricing("Vendor/Claude-Sonnet-4@2025:free")).toBe(
"claude-sonnet-4-2025",
);
});
});
describe("formatPrice", () => {
it("formats integers without a decimal point", () => {
expect(formatPrice(5)).toBe("5");
expect(formatPrice(25)).toBe("25");
});
it("trims trailing zeros", () => {
expect(formatPrice(0.5)).toBe("0.5");
expect(formatPrice(6.25)).toBe("6.25");
expect(formatPrice(1.0395)).toBe("1.0395");
});
it("keeps up to six decimal places", () => {
expect(formatPrice(0.000001)).toBe("0.000001");
expect(formatPrice(0.0000004)).toBe("0");
});
it("returns 0 for zero, negative and non-finite values", () => {
expect(formatPrice(0)).toBe("0");
expect(formatPrice(-1)).toBe("0");
expect(formatPrice(NaN)).toBe("0");
expect(formatPrice(Infinity)).toBe("0");
});
it("never produces exponent notation", () => {
// 后端 Decimal::from_str 不接受科学计数法
expect(formatPrice(1e-8)).toBe("0");
expect(formatPrice(1e21)).toBe("0");
for (const value of [5, 0.5, 0.000123, 123456.789]) {
expect(formatPrice(value)).toMatch(/^\d+(\.\d+)?$/);
}
});
});
describe("flattenModels", () => {
it("flattens providers, fills defaults and sorts by release date desc", () => {
const entries = flattenModels({
acme: {
id: "acme",
name: "Acme AI",
models: {
"old-model": {
id: "old-model",
name: "Old Model",
release_date: "2024-01-01",
cost: { input: 1, output: 2 },
},
"new-model": {
id: "new-model",
name: "New Model",
release_date: "2025-06-01",
cost: { input: 3, output: 6, cache_read: 0.3, cache_write: 3.75 },
},
"free-model": {
id: "free-model",
name: "No Cost Model",
},
},
},
bare: {
models: {
"Vendor/Some-Model:free": {
release_date: "2025-01",
cost: { input: 0.1 },
},
},
},
});
expect(entries.map((e) => e.key)).toEqual([
"acme/new-model",
"bare/Vendor/Some-Model:free",
"acme/old-model",
]);
const newModel = entries[0];
expect(newModel.normalizedId).toBe("new-model");
expect(newModel.cacheRead).toBe(0.3);
expect(newModel.cacheWrite).toBe(3.75);
// 没有 name 的 provider 用 id 兜底;缺失的成本字段补 0
const bareModel = entries[1];
expect(bareModel.providerName).toBe("bare");
expect(bareModel.normalizedId).toBe("some-model");
expect(bareModel.output).toBe(0);
expect(bareModel.cacheRead).toBe(0);
// 完全没有定价的模型被过滤
expect(entries.some((e) => e.modelId === "free-model")).toBe(false);
});
});