支持导入定价文档并在统计页估算费用。
把用量与单价对齐，方便对照各模型实际开销。 Co-authored-by: Cursor <cursoragent@cursor.com>
This commit is contained in:
Cursor
@@ -0,0 +1,176 @@
|
||||
import type { TokenUsageRecord } from "./token-store.js";
|
||||
import {
|
||||
findModelRate,
|
||||
resolveRateCatalog,
|
||||
type ResolvedRate,
|
||||
} from "./pricing.js";
|
||||
import type { PeakWindow, RateBand } from "./pricing-parse.js";
|
||||
|
||||
export type Money = {
|
||||
amount: number;
|
||||
currency: string;
|
||||
};
|
||||
|
||||
export type RecordCost = {
|
||||
cost: Money | null;
|
||||
unmatched: boolean;
|
||||
sourceName?: string;
|
||||
};
|
||||
|
||||
export type ModelCostRow = {
|
||||
model: string;
|
||||
calls: number;
|
||||
promptTokens: number;
|
||||
completionTokens: number;
|
||||
totalTokens: number;
|
||||
cachedTokens: number;
|
||||
cacheMissTokens: number;
|
||||
cost: Money | null;
|
||||
unmatched: boolean;
|
||||
sourceName?: string;
|
||||
sourceUrl?: string;
|
||||
};
|
||||
|
||||
export type CostReport = {
|
||||
total: Money | null;
|
||||
byCurrency: Record<string, number>;
|
||||
unpricedCalls: number;
|
||||
pricedCalls: number;
|
||||
byModel: ModelCostRow[];
|
||||
};
|
||||
|
||||
export function isPeakAt(
|
||||
at: Date,
|
||||
windows: PeakWindow[] | undefined,
|
||||
weekdays: number[] | undefined,
|
||||
): boolean {
|
||||
if (!windows?.length) return false;
|
||||
if (weekdays?.length && !weekdays.includes(at.getUTCDay())) return false;
|
||||
const minutes = at.getUTCHours() * 60 + at.getUTCMinutes();
|
||||
return windows.some((w) => minutes >= w.startMinutes && minutes < w.endMinutes);
|
||||
}
|
||||
|
||||
export function splitPromptTokens(record: TokenUsageRecord): {
|
||||
cached: number;
|
||||
miss: number;
|
||||
output: number;
|
||||
} {
|
||||
const cached = Math.max(0, record.cachedTokens ?? 0);
|
||||
let miss = record.cacheMissTokens;
|
||||
if (miss == null) {
|
||||
miss = Math.max(0, record.promptTokens - cached);
|
||||
}
|
||||
const covered = cached + miss;
|
||||
if (covered < record.promptTokens) {
|
||||
miss += record.promptTokens - covered;
|
||||
}
|
||||
return { cached, miss, output: Math.max(0, record.completionTokens) };
|
||||
}
|
||||
|
||||
function bandPrice(band: RateBand | undefined, peak: boolean): number {
|
||||
if (!band) return 0;
|
||||
if (peak && band.peak != null) return band.peak;
|
||||
return band.offPeak;
|
||||
}
|
||||
|
||||
export function costForRecord(
|
||||
record: TokenUsageRecord,
|
||||
rate: ResolvedRate,
|
||||
): Money {
|
||||
const at = new Date(record.at);
|
||||
const peak = isPeakAt(at, rate.peakWindows, rate.peakWeekdays);
|
||||
const parts = splitPromptTokens(record);
|
||||
const hitPrice = rate.inputCacheHit
|
||||
? bandPrice(rate.inputCacheHit, peak)
|
||||
: bandPrice(rate.inputCacheMiss, peak);
|
||||
const missPrice = bandPrice(rate.inputCacheMiss, peak);
|
||||
const outPrice = bandPrice(rate.output, peak);
|
||||
const unit = rate.unitTokens || 1_000_000;
|
||||
const amount =
|
||||
(parts.cached * hitPrice + parts.miss * missPrice + parts.output * outPrice) /
|
||||
unit;
|
||||
return { amount: Math.round(amount * 1e10) / 1e10, currency: rate.currency };
|
||||
}
|
||||
|
||||
export function attachRecordCost(
|
||||
record: TokenUsageRecord,
|
||||
catalog = resolveRateCatalog(),
|
||||
): RecordCost {
|
||||
const rate = findModelRate(record.model || "", catalog);
|
||||
if (!rate) return { cost: null, unmatched: true };
|
||||
return {
|
||||
cost: costForRecord(record, rate),
|
||||
unmatched: false,
|
||||
sourceName: rate.sourceName,
|
||||
};
|
||||
}
|
||||
|
||||
export function buildCostReport(
|
||||
records: TokenUsageRecord[],
|
||||
catalog = resolveRateCatalog(),
|
||||
): CostReport {
|
||||
const byModel = new Map<string, ModelCostRow>();
|
||||
const byCurrency: Record<string, number> = {};
|
||||
let unpricedCalls = 0;
|
||||
let pricedCalls = 0;
|
||||
|
||||
for (const record of records) {
|
||||
const model = record.model?.trim() || "(未填型号)";
|
||||
const rate = findModelRate(record.model || "", catalog);
|
||||
const priced = rate ? costForRecord(record, rate) : null;
|
||||
if (priced) {
|
||||
pricedCalls += 1;
|
||||
byCurrency[priced.currency] = (byCurrency[priced.currency] ?? 0) + priced.amount;
|
||||
} else {
|
||||
unpricedCalls += 1;
|
||||
}
|
||||
|
||||
const prev = byModel.get(model) ?? {
|
||||
model,
|
||||
calls: 0,
|
||||
promptTokens: 0,
|
||||
completionTokens: 0,
|
||||
totalTokens: 0,
|
||||
cachedTokens: 0,
|
||||
cacheMissTokens: 0,
|
||||
cost: priced ? { amount: 0, currency: priced.currency } : null,
|
||||
unmatched: !rate,
|
||||
sourceName: rate?.sourceName,
|
||||
sourceUrl: rate?.sourceUrl,
|
||||
};
|
||||
prev.calls += 1;
|
||||
prev.promptTokens += record.promptTokens;
|
||||
prev.completionTokens += record.completionTokens;
|
||||
prev.totalTokens += record.totalTokens;
|
||||
prev.cachedTokens += record.cachedTokens ?? 0;
|
||||
prev.cacheMissTokens += record.cacheMissTokens ?? 0;
|
||||
if (priced) {
|
||||
if (!prev.cost) prev.cost = { amount: 0, currency: priced.currency };
|
||||
if (prev.cost.currency === priced.currency) {
|
||||
prev.cost.amount += priced.amount;
|
||||
}
|
||||
prev.unmatched = false;
|
||||
}
|
||||
byModel.set(model, prev);
|
||||
}
|
||||
|
||||
const currencies = Object.keys(byCurrency);
|
||||
const total =
|
||||
currencies.length === 1
|
||||
? { amount: byCurrency[currencies[0]], currency: currencies[0] }
|
||||
: null;
|
||||
|
||||
const rows = [...byModel.values()].sort((a, b) => {
|
||||
const ac = a.cost?.amount ?? -1;
|
||||
const bc = b.cost?.amount ?? -1;
|
||||
return bc - ac;
|
||||
});
|
||||
|
||||
return {
|
||||
total,
|
||||
byCurrency,
|
||||
unpricedCalls,
|
||||
pricedCalls,
|
||||
byModel: rows,
|
||||
};
|
||||
}
|
||||
@@ -0,0 +1,615 @@
|
||||
/** 从官方价目文档（HTML / Markdown / JSON）抽出按型号的单价。 */
|
||||
|
||||
export type PeakWindow = {
|
||||
/** 距 UTC 当天 00:00 的分钟数，含起点不含终点 */
|
||||
startMinutes: number;
|
||||
endMinutes: number;
|
||||
};
|
||||
|
||||
export type RateBand = {
|
||||
offPeak: number;
|
||||
peak?: number;
|
||||
};
|
||||
|
||||
export type ModelRate = {
|
||||
model: string;
|
||||
aliases: string[];
|
||||
currency: string;
|
||||
unitTokens: number;
|
||||
inputCacheHit?: RateBand;
|
||||
inputCacheMiss: RateBand;
|
||||
output: RateBand;
|
||||
peakWindows?: PeakWindow[];
|
||||
/** JS getUTCDay：0 日 … 6 六。省略则每天都可能高峰 */
|
||||
peakWeekdays?: number[];
|
||||
};
|
||||
|
||||
export type ParsePricingResult = {
|
||||
title?: string;
|
||||
rates: ModelRate[];
|
||||
warnings: string[];
|
||||
};
|
||||
|
||||
const UNIT_1M = 1_000_000;
|
||||
|
||||
const MODEL_STOP = new Set(
|
||||
[
|
||||
"model",
|
||||
"models",
|
||||
"pricing",
|
||||
"features",
|
||||
"peak",
|
||||
"off-peak",
|
||||
"offpeak",
|
||||
"input",
|
||||
"output",
|
||||
"tokens",
|
||||
"token",
|
||||
"cache",
|
||||
"hit",
|
||||
"miss",
|
||||
"thinking",
|
||||
"maximum",
|
||||
"vision",
|
||||
"usd",
|
||||
"cny",
|
||||
"base",
|
||||
"url",
|
||||
"openai",
|
||||
"anthropic",
|
||||
"format",
|
||||
"version",
|
||||
"concurrency",
|
||||
"limit",
|
||||
].map((s) => s.toLowerCase()),
|
||||
);
|
||||
|
||||
export function parsePricingDocument(
|
||||
text: string,
|
||||
contentType = "text/html",
|
||||
): ParsePricingResult {
|
||||
const trimmed = text.trim();
|
||||
const type = contentType.toLowerCase();
|
||||
if (type.includes("json") || (trimmed.startsWith("{") && trimmed.includes("pricing"))) {
|
||||
const jsonResult = parsePricingJson(trimmed);
|
||||
if (jsonResult.rates.length) return jsonResult;
|
||||
}
|
||||
const html = /<table[\s>]/i.test(trimmed) ? parsePricingHtml(trimmed) : { rates: [], warnings: [] as string[], title: pageTitle(trimmed) };
|
||||
if (html.rates.length) {
|
||||
return decorate(html, trimmed);
|
||||
}
|
||||
const md = parsePricingMarkdown(stripTags(trimmed));
|
||||
if (md.rates.length) return decorate(md, trimmed);
|
||||
return {
|
||||
title: pageTitle(trimmed),
|
||||
rates: [],
|
||||
warnings: ["未能从文档解析出型号价格"],
|
||||
};
|
||||
}
|
||||
|
||||
function decorate(result: ParsePricingResult, source: string): ParsePricingResult {
|
||||
const peak = parsePeakSchedule(source);
|
||||
const aliases = parseLegacyAliases(source, result.rates.map((r) => r.model));
|
||||
const rates = result.rates.map((r) => ({
|
||||
...r,
|
||||
aliases: unique([...(r.aliases ?? []), ...(aliases.get(r.model.toLowerCase()) ?? [])]),
|
||||
...(peak
|
||||
? { peakWindows: peak.windows, peakWeekdays: peak.weekdays }
|
||||
: {}),
|
||||
}));
|
||||
return {
|
||||
title: result.title ?? pageTitle(source),
|
||||
rates,
|
||||
warnings: result.warnings,
|
||||
};
|
||||
}
|
||||
|
||||
function parsePricingHtml(html: string): ParsePricingResult {
|
||||
const tables = extractHtmlTables(html).map(expandTableGrid);
|
||||
return extractFromGrids(tables, html);
|
||||
}
|
||||
|
||||
function parsePricingMarkdown(text: string): ParsePricingResult {
|
||||
const tables = extractMarkdownTables(text);
|
||||
return extractFromGrids(tables, text);
|
||||
}
|
||||
|
||||
function extractFromGrids(grids: string[][][], source: string): ParsePricingResult {
|
||||
const warnings: string[] = [];
|
||||
const rates: ModelRate[] = [];
|
||||
const currencyHint = detectCurrency(source);
|
||||
const unit = detectUnitTokens(source);
|
||||
|
||||
for (const grid of grids) {
|
||||
const fromMatrix = extractDeepSeekStyle(grid, currencyHint, unit);
|
||||
if (fromMatrix.length) {
|
||||
rates.push(...fromMatrix);
|
||||
continue;
|
||||
}
|
||||
const generic = extractGenericHeaderTable(grid, currencyHint, unit);
|
||||
rates.push(...generic);
|
||||
}
|
||||
|
||||
const uniqueRates = mergeRatesByModel(rates);
|
||||
if (!uniqueRates.length) {
|
||||
warnings.push("表格里没有识别到单价");
|
||||
}
|
||||
return { title: pageTitle(source), rates: uniqueRates, warnings };
|
||||
}
|
||||
|
||||
function extractDeepSeekStyle(
|
||||
grid: string[][],
|
||||
currencyHint: string,
|
||||
unit: number,
|
||||
): ModelRate[] {
|
||||
let models: string[] = [];
|
||||
for (const row of grid) {
|
||||
const idx = row.findIndex((c) => /^model$/i.test(cleanCell(c)));
|
||||
if (idx < 0) continue;
|
||||
const found = row
|
||||
.slice(idx + 1)
|
||||
.map(cleanCell)
|
||||
.filter((c) => looksLikeModelId(c));
|
||||
if (found.length) {
|
||||
models = unique(found);
|
||||
break;
|
||||
}
|
||||
}
|
||||
if (!models.length) return [];
|
||||
|
||||
const hit: Partial<Record<string, number[]>> = {};
|
||||
const miss: Partial<Record<string, number[]>> = {};
|
||||
const output: Partial<Record<string, number[]>> = {};
|
||||
let currency = currencyHint;
|
||||
|
||||
for (const row of grid) {
|
||||
const cells = row.map(cleanCell);
|
||||
const joined = cells.join(" ").toLowerCase();
|
||||
const band = bandOf(joined);
|
||||
if (!band) continue;
|
||||
const money = parseMoneyCells(row);
|
||||
if (money.length < models.length) continue;
|
||||
const slice = money.slice(-models.length);
|
||||
currency = slice[0]?.currency ?? currency;
|
||||
const amounts = slice.map((m) => m.amount);
|
||||
if (/cache\s*hit/.test(joined) || /缓存命中/.test(joined)) {
|
||||
hit[band] = amounts;
|
||||
} else if (/cache\s*miss/.test(joined) || /缓存未命中/.test(joined)) {
|
||||
miss[band] = amounts;
|
||||
} else if (/output/.test(joined) || /输出/.test(joined)) {
|
||||
output[band] = amounts;
|
||||
}
|
||||
}
|
||||
|
||||
if (!miss.offPeak && !output.offPeak) return [];
|
||||
|
||||
return models.map((model, i) => {
|
||||
const missOff = miss.offPeak?.[i] ?? miss.peak?.[i];
|
||||
const outOff = output.offPeak?.[i] ?? output.peak?.[i];
|
||||
const missPeak = miss.peak?.[i];
|
||||
const outPeak = output.peak?.[i];
|
||||
const hitOff = hit.offPeak?.[i];
|
||||
const hitPeak = hit.peak?.[i];
|
||||
const inputMiss = missOff ?? missPeak;
|
||||
const out = outOff ?? outPeak;
|
||||
if (inputMiss == null || out == null) {
|
||||
return null;
|
||||
}
|
||||
const rate: ModelRate = {
|
||||
model,
|
||||
aliases: [],
|
||||
currency,
|
||||
unitTokens: unit,
|
||||
inputCacheMiss: {
|
||||
offPeak: inputMiss,
|
||||
...(missPeak != null && missPeak !== inputMiss ? { peak: missPeak } : {}),
|
||||
},
|
||||
output: {
|
||||
offPeak: out,
|
||||
...(outPeak != null && outPeak !== out ? { peak: outPeak } : {}),
|
||||
},
|
||||
};
|
||||
if (hitOff != null || hitPeak != null) {
|
||||
const off = hitOff ?? hitPeak!;
|
||||
rate.inputCacheHit = {
|
||||
offPeak: off,
|
||||
...(hitPeak != null && hitPeak !== off ? { peak: hitPeak } : {}),
|
||||
};
|
||||
}
|
||||
return rate;
|
||||
}).filter((r): r is ModelRate => r != null);
|
||||
}
|
||||
|
||||
function extractGenericHeaderTable(
|
||||
grid: string[][],
|
||||
currencyHint: string,
|
||||
unit: number,
|
||||
): ModelRate[] {
|
||||
if (grid.length < 2) return [];
|
||||
const header = grid[0].map(cleanCell);
|
||||
const kinds = header.map(headerKind);
|
||||
const modelCol = kinds.indexOf("model");
|
||||
const inputCol = kinds.indexOf("input");
|
||||
const cachedCol = kinds.indexOf("cached");
|
||||
const outputCol = kinds.indexOf("output");
|
||||
if (modelCol < 0 || inputCol < 0 || outputCol < 0) return [];
|
||||
|
||||
const rates: ModelRate[] = [];
|
||||
for (const row of grid.slice(1)) {
|
||||
const model = cleanCell(row[modelCol] ?? "");
|
||||
if (!looksLikeModelId(model)) continue;
|
||||
const input = parsePriceCell(row[inputCol] ?? "", currencyHint);
|
||||
const output = parsePriceCell(row[outputCol] ?? "", currencyHint);
|
||||
if (!input || !output) continue;
|
||||
const cached =
|
||||
cachedCol >= 0 ? parsePriceCell(row[cachedCol] ?? "", currencyHint) : null;
|
||||
const currency = input.currency ?? output.currency ?? currencyHint;
|
||||
const rate: ModelRate = {
|
||||
model,
|
||||
aliases: [],
|
||||
currency,
|
||||
unitTokens: unit,
|
||||
inputCacheMiss: { offPeak: input.amount },
|
||||
output: { offPeak: output.amount },
|
||||
};
|
||||
if (cached) rate.inputCacheHit = { offPeak: cached.amount };
|
||||
rates.push(rate);
|
||||
}
|
||||
return rates;
|
||||
}
|
||||
|
||||
function headerKind(cell: string): "model" | "input" | "cached" | "output" | "other" {
|
||||
const t = cell.toLowerCase().replace(/\s+/g, " ").trim();
|
||||
if (/模型名称|^(model|型号|模型)$/i.test(t) || t === "id") return "model";
|
||||
if (/缓存存储|cache(?:d)?\s*input\s*storage/.test(t)) return "other";
|
||||
if (/cache\s*hit|cached(?:\s+input)?$|缓存命中/.test(t)) return "cached";
|
||||
if (/output|completion|输出/.test(t)) return "output";
|
||||
if (/input|prompt|cache\s*miss|输入/.test(t)) return "input";
|
||||
return "other";
|
||||
}
|
||||
|
||||
function bandOf(joined: string): "offPeak" | "peak" | null {
|
||||
if (/off[- ]?peak|闲时|非高峰|低峰/.test(joined)) return "offPeak";
|
||||
if (/(^|[^a-z])peak([^a-z]|$)|高峰/.test(joined)) return "peak";
|
||||
return null;
|
||||
}
|
||||
|
||||
export function parsePeakSchedule(text: string): {
|
||||
windows: PeakWindow[];
|
||||
weekdays: number[];
|
||||
} | null {
|
||||
const plain = stripTags(text).replace(/\s+/g, " ");
|
||||
const windows: PeakWindow[] = [];
|
||||
const utcRange =
|
||||
/peak hours are\s+(\d{1,2}:\d{2})\s*[-–—]\s*(\d{1,2}:\d{2})\s+and\s+(\d{1,2}:\d{2})\s*[-–—]\s*(\d{1,2}:\d{2})\s*utc/i.exec(
|
||||
plain,
|
||||
) ??
|
||||
/高峰.*?(\d{1,2}:\d{2})\s*[-–—]\s*(\d{1,2}:\d{2}).*?(\d{1,2}:\d{2})\s*[-–—]\s*(\d{1,2}:\d{2}).*utc/i.exec(
|
||||
plain,
|
||||
);
|
||||
if (utcRange) {
|
||||
windows.push(
|
||||
{ startMinutes: parseHhMm(utcRange[1]), endMinutes: parseHhMm(utcRange[2]) },
|
||||
{ startMinutes: parseHhMm(utcRange[3]), endMinutes: parseHhMm(utcRange[4]) },
|
||||
);
|
||||
}
|
||||
if (!windows.length) return null;
|
||||
const weekdays = /monday through friday|周一至周五|星期一至星期五/i.test(plain)
|
||||
? [1, 2, 3, 4, 5]
|
||||
: [0, 1, 2, 3, 4, 5, 6];
|
||||
return { windows, weekdays };
|
||||
}
|
||||
|
||||
export function parseLegacyAliases(
|
||||
text: string,
|
||||
models: string[],
|
||||
): Map<string, string[]> {
|
||||
const map = new Map<string, string[]>();
|
||||
const codes = [
|
||||
...text.matchAll(/<code\b[^>]*>([^<]+)<\/code>/gi),
|
||||
...text.matchAll(/`([a-zA-Z][a-zA-Z0-9._:+/-]{1,80})`/g),
|
||||
].map((m) => cleanCell(m[1]));
|
||||
const plain = stripTags(text).replace(/\s+/g, " ");
|
||||
const flash = models.find((m) => /flash/i.test(m));
|
||||
if (
|
||||
flash &&
|
||||
/legacy names[\s\S]{0,800}billed at the flash price/i.test(plain)
|
||||
) {
|
||||
const legacy = codes.filter(
|
||||
(c) => /flash/i.test(c) && c.toLowerCase() !== flash.toLowerCase(),
|
||||
);
|
||||
if (legacy.length) map.set(flash.toLowerCase(), unique(legacy));
|
||||
}
|
||||
return map;
|
||||
}
|
||||
|
||||
function parsePricingJson(text: string): ParsePricingResult {
|
||||
try {
|
||||
const data = JSON.parse(text) as unknown;
|
||||
const rates: ModelRate[] = [];
|
||||
const rows = jsonModelRows(data);
|
||||
for (const row of rows) {
|
||||
const model = String(row.id ?? row.model ?? "").trim();
|
||||
if (!model) continue;
|
||||
const pricing = (row.pricing ?? row) as Record<string, unknown>;
|
||||
const prompt = jsonPrice(pricing.prompt ?? pricing.input ?? pricing.inputCacheMiss);
|
||||
const completion = jsonPrice(pricing.completion ?? pricing.output);
|
||||
if (prompt == null || completion == null) continue;
|
||||
const cached = jsonPrice(pricing.input_cache_read ?? pricing.cached ?? pricing.inputCacheHit);
|
||||
// OpenRouter 等接口用「每个 token」；官方价目表通常是每 1M。
|
||||
const perToken = prompt > 0 && prompt < 1e-4;
|
||||
const rate: ModelRate = {
|
||||
model: model.includes("/") ? model.split("/").pop()! : model,
|
||||
aliases: model.includes("/") ? [model] : [],
|
||||
currency: "USD",
|
||||
unitTokens: UNIT_1M,
|
||||
inputCacheMiss: { offPeak: perToken ? prompt * UNIT_1M : prompt },
|
||||
output: { offPeak: perToken ? completion * UNIT_1M : completion },
|
||||
};
|
||||
if (cached != null) {
|
||||
rate.inputCacheHit = { offPeak: perToken ? cached * UNIT_1M : cached };
|
||||
}
|
||||
rates.push(rate);
|
||||
}
|
||||
return { rates: mergeRatesByModel(rates), warnings: rates.length ? [] : ["JSON 中没有单价"] };
|
||||
} catch {
|
||||
return { rates: [], warnings: ["JSON 解析失败"] };
|
||||
}
|
||||
}
|
||||
|
||||
function jsonModelRows(data: unknown): Array<Record<string, unknown>> {
|
||||
if (Array.isArray(data)) return data.filter(isRecord);
|
||||
if (!isRecord(data)) return [];
|
||||
if (Array.isArray(data.models)) return data.models.filter(isRecord);
|
||||
if (Array.isArray(data.data)) return data.data.filter(isRecord);
|
||||
if (Array.isArray(data.rates)) return data.rates.filter(isRecord);
|
||||
return [];
|
||||
}
|
||||
|
||||
function jsonPrice(value: unknown): number | null {
|
||||
if (typeof value === "number" && Number.isFinite(value)) return value;
|
||||
if (typeof value === "string") {
|
||||
const n = Number(value);
|
||||
return Number.isFinite(n) ? n : null;
|
||||
}
|
||||
return null;
|
||||
}
|
||||
|
||||
function isRecord(v: unknown): v is Record<string, unknown> {
|
||||
return typeof v === "object" && v != null && !Array.isArray(v);
|
||||
}
|
||||
|
||||
type RawCell = { text: string; colspan: number; rowspan: number };
|
||||
|
||||
function extractHtmlTables(html: string): RawCell[][][] {
|
||||
const tables: RawCell[][][] = [];
|
||||
const re = /<table\b[^>]*>([\s\S]*?)<\/table>/gi;
|
||||
let m: RegExpExecArray | null;
|
||||
while ((m = re.exec(html))) {
|
||||
tables.push(parseHtmlTableRows(m[1]));
|
||||
}
|
||||
return tables;
|
||||
}
|
||||
|
||||
function parseHtmlTableRows(inner: string): RawCell[][] {
|
||||
const rows: RawCell[][] = [];
|
||||
const trRe = /<tr\b[^>]*>([\s\S]*?)<\/tr>/gi;
|
||||
let tr: RegExpExecArray | null;
|
||||
while ((tr = trRe.exec(inner))) {
|
||||
const cells: RawCell[] = [];
|
||||
const tdRe = /<t[dh]\b([^>]*)>([\s\S]*?)<\/t[dh]>/gi;
|
||||
let td: RegExpExecArray | null;
|
||||
while ((td = tdRe.exec(tr[1]))) {
|
||||
const attrs = td[1];
|
||||
cells.push({
|
||||
text: cellText(td[2]),
|
||||
colspan: attrInt(attrs, "colspan") || 1,
|
||||
rowspan: attrInt(attrs, "rowspan") || 1,
|
||||
});
|
||||
}
|
||||
if (cells.length) rows.push(cells);
|
||||
}
|
||||
return rows;
|
||||
}
|
||||
|
||||
function expandTableGrid(rows: RawCell[][]): string[][] {
|
||||
const occupier: Array<{ left: number; text: string } | undefined> = [];
|
||||
const grid: string[][] = [];
|
||||
for (const row of rows) {
|
||||
const out: string[] = [];
|
||||
let col = 0;
|
||||
let i = 0;
|
||||
while (i < row.length || occupier.some((o, idx) => idx >= col && o && o.left > 0)) {
|
||||
const occ = occupier[col];
|
||||
if (occ && occ.left > 0) {
|
||||
out[col] = occ.text;
|
||||
occ.left -= 1;
|
||||
col += 1;
|
||||
continue;
|
||||
}
|
||||
if (i >= row.length) break;
|
||||
const cell = row[i++];
|
||||
for (let k = 0; k < cell.colspan; k++) {
|
||||
while (occupier[col]?.left) {
|
||||
out[col] = occupier[col]!.text;
|
||||
occupier[col]!.left -= 1;
|
||||
col += 1;
|
||||
}
|
||||
out[col] = cell.text;
|
||||
if (cell.rowspan > 1) {
|
||||
occupier[col] = { left: cell.rowspan - 1, text: cell.text };
|
||||
}
|
||||
col += 1;
|
||||
}
|
||||
}
|
||||
grid.push(out);
|
||||
}
|
||||
return grid;
|
||||
}
|
||||
|
||||
function extractMarkdownTables(text: string): string[][][] {
|
||||
const lines = text.split(/\r?\n/);
|
||||
const tables: string[][][] = [];
|
||||
let current: string[][] = [];
|
||||
const flush = () => {
|
||||
if (current.length >= 2) tables.push(current);
|
||||
current = [];
|
||||
};
|
||||
for (const line of lines) {
|
||||
if (!/^\s*\|/.test(line)) {
|
||||
flush();
|
||||
continue;
|
||||
}
|
||||
if (/^\s*\|?\s*:?-{3,}/.test(line)) continue;
|
||||
const cells = line
|
||||
.replace(/^\s*\|/, "")
|
||||
.replace(/\|\s*$/, "")
|
||||
.split("|")
|
||||
.map((c) => cleanCell(c));
|
||||
current.push(cells);
|
||||
}
|
||||
flush();
|
||||
return tables;
|
||||
}
|
||||
|
||||
function attrInt(attrs: string, name: string): number {
|
||||
const m = new RegExp(`${name}\\s*=\\s*["']?(\\d+)`, "i").exec(attrs);
|
||||
return m ? Number(m[1]) : 0;
|
||||
}
|
||||
|
||||
function cellText(raw: string): string {
|
||||
return cleanCell(
|
||||
raw
|
||||
.replace(/<sup\b[^>]*>[\s\S]*?<\/sup>/gi, "")
|
||||
.replace(/<br\s*\/?>/gi, " ")
|
||||
.replace(/<\/(p|div|tr|h\d)>/gi, " "),
|
||||
);
|
||||
}
|
||||
|
||||
export function cleanCell(raw: string): string {
|
||||
return decodeEntities(stripTags(raw))
|
||||
.replace(/\u00a0/g, " ")
|
||||
.replace(/\s+/g, " ")
|
||||
.trim();
|
||||
}
|
||||
|
||||
function stripTags(html: string): string {
|
||||
return html.replace(/<script[\s\S]*?<\/script>/gi, " ").replace(/<[^>]+>/g, " ");
|
||||
}
|
||||
|
||||
function decodeEntities(s: string): string {
|
||||
return s
|
||||
.replace(/&/g, "&")
|
||||
.replace(/</g, "<")
|
||||
.replace(/>/g, ">")
|
||||
.replace(/"/g, '"')
|
||||
.replace(/'/g, "'")
|
||||
.replace(/ /g, " ");
|
||||
}
|
||||
|
||||
function pageTitle(text: string): string | undefined {
|
||||
const m = /<title[^>]*>([\s\S]*?)<\/title>/i.exec(text);
|
||||
if (!m) return undefined;
|
||||
return decodeEntities(m[1]).replace(/\s+/g, " ").trim() || undefined;
|
||||
}
|
||||
|
||||
function looksLikeModelId(s: string): boolean {
|
||||
const t = s.trim();
|
||||
if (t.length < 2 || t.length > 80) return false;
|
||||
if (MODEL_STOP.has(t.toLowerCase())) return false;
|
||||
if (!/^[a-zA-Z][a-zA-Z0-9._:+/-]*$/.test(t)) return false;
|
||||
if (/^(https?:|www\.)/i.test(t)) return false;
|
||||
return /[-/._]/.test(t) || /^(gpt|o[1-9]|claude|gemini|deepseek|qwen|glm|mistral|llama)/i.test(t);
|
||||
}
|
||||
|
||||
function parseMoneyCells(row: string[]): Array<{ amount: number; currency: string }> {
|
||||
const out: Array<{ amount: number; currency: string }> = [];
|
||||
for (const cell of row) {
|
||||
const m = parseMoneyCell(cell);
|
||||
if (m) out.push(m);
|
||||
}
|
||||
return out;
|
||||
}
|
||||
|
||||
function parsePriceCell(
|
||||
cell: string,
|
||||
currencyHint: string,
|
||||
): { amount: number; currency: string } | null {
|
||||
const t = cleanCell(cell);
|
||||
if (!t) return null;
|
||||
if (/^(?:[-–—\\/]|不支持|暂不支持)$/i.test(t)) return null;
|
||||
if (/^(free|免费)$/i.test(t)) return { amount: 0, currency: currencyHint };
|
||||
const money = parseMoneyCell(t);
|
||||
if (money) return money;
|
||||
return parseBareNumber(t, currencyHint);
|
||||
}
|
||||
|
||||
function parseMoneyCell(cell: string): { amount: number; currency: string } | null {
|
||||
const t = cleanCell(cell);
|
||||
const m =
|
||||
/^[\$]\s*([0-9]+(?:\.[0-9]+)?)\b/.exec(t) ||
|
||||
/^[¥￥]\s*([0-9]+(?:\.[0-9]+)?)\b/.exec(t) ||
|
||||
/^([0-9]+(?:\.[0-9]+)?)\s*元/.exec(t);
|
||||
if (!m) return null;
|
||||
const amount = Number(m[1]);
|
||||
if (!Number.isFinite(amount)) return null;
|
||||
const currency = t.includes("¥") || t.includes("￥") || t.includes("元") ? "CNY" : "USD";
|
||||
return { amount, currency };
|
||||
}
|
||||
|
||||
function parseBareNumber(
|
||||
cell: string,
|
||||
currencyHint = "USD",
|
||||
): { amount: number; currency: string } | null {
|
||||
const t = cleanCell(cell);
|
||||
if (!/^[0-9]+(?:\.[0-9]+)?$/.test(t)) return null;
|
||||
const amount = Number(t);
|
||||
return Number.isFinite(amount) ? { amount, currency: currencyHint } : null;
|
||||
}
|
||||
|
||||
function detectCurrency(text: string): string {
|
||||
if (/元\s*[/／]\s*百万|元\/百万|人民币/.test(text)) return "CNY";
|
||||
if (/\$\s*[0-9]|prices are in usd/i.test(text)) return "USD";
|
||||
if (/[¥￥]|CNY/.test(text)) return "CNY";
|
||||
return "USD";
|
||||
}
|
||||
|
||||
function detectUnitTokens(text: string): number {
|
||||
if (/1\s*[Mm]\s*tokens|per\s*1M|每\s*100\s*万|百万\s*token/i.test(text)) return UNIT_1M;
|
||||
return UNIT_1M;
|
||||
}
|
||||
|
||||
function parseHhMm(s: string): number {
|
||||
const [h, m] = s.split(":").map(Number);
|
||||
return h * 60 + (m || 0);
|
||||
}
|
||||
|
||||
function unique(xs: string[]): string[] {
|
||||
const seen = new Set<string>();
|
||||
const out: string[] = [];
|
||||
for (const x of xs) {
|
||||
const k = x.toLowerCase();
|
||||
if (seen.has(k)) continue;
|
||||
seen.add(k);
|
||||
out.push(x);
|
||||
}
|
||||
return out;
|
||||
}
|
||||
|
||||
function mergeRatesByModel(rates: ModelRate[]): ModelRate[] {
|
||||
const map = new Map<string, ModelRate>();
|
||||
for (const r of rates) {
|
||||
const key = r.model.toLowerCase();
|
||||
const prev = map.get(key);
|
||||
if (!prev) {
|
||||
map.set(key, { ...r, aliases: unique(r.aliases) });
|
||||
continue;
|
||||
}
|
||||
map.set(key, {
|
||||
...prev,
|
||||
...r,
|
||||
aliases: unique([...prev.aliases, ...r.aliases]),
|
||||
});
|
||||
}
|
||||
return [...map.values()];
|
||||
}
|
||||
@@ -0,0 +1,275 @@
|
||||
import { readFileSync, writeFileSync } from "node:fs";
|
||||
import path from "node:path";
|
||||
import { randomUUID } from "node:crypto";
|
||||
import { ensureUserDataDirs, getUserDataDir } from "../config/user-data-dir.js";
|
||||
import {
|
||||
parsePricingDocument,
|
||||
type ModelRate,
|
||||
} from "./pricing-parse.js";
|
||||
|
||||
export type PricingSource = {
|
||||
id: string;
|
||||
name: string;
|
||||
url: string;
|
||||
createdAt: string;
|
||||
fetchedAt?: string;
|
||||
fetchError?: string;
|
||||
title?: string;
|
||||
rates: ModelRate[];
|
||||
};
|
||||
|
||||
type PricingFile = {
|
||||
version: 1;
|
||||
sources: PricingSource[];
|
||||
};
|
||||
|
||||
const FILE_NAME = "pricing-sources.json";
|
||||
const FETCH_TIMEOUT_MS = 20_000;
|
||||
const MAX_BYTES = 2_000_000;
|
||||
|
||||
export const SUGGESTED_PRICING_DOCS: Array<{ name: string; url: string }> = [
|
||||
{
|
||||
name: "DeepSeek Models & Pricing",
|
||||
url: "https://api-docs.deepseek.com/quick_start/pricing",
|
||||
},
|
||||
{
|
||||
name: "DeepSeek 模型与价格",
|
||||
url: "https://api-docs.deepseek.com/zh-cn/quick_start/pricing",
|
||||
},
|
||||
{
|
||||
name: "智谱 GLM API 定价",
|
||||
url: "https://docs.bigmodel.cn/cn/guide/start/pricing",
|
||||
},
|
||||
{
|
||||
name: "Z.AI GLM Pricing",
|
||||
url: "https://docs.z.ai/guides/overview/pricing",
|
||||
},
|
||||
];
|
||||
|
||||
function filePath(): string {
|
||||
return path.join(getUserDataDir(), FILE_NAME);
|
||||
}
|
||||
|
||||
function emptyFile(): PricingFile {
|
||||
return { version: 1, sources: [] };
|
||||
}
|
||||
|
||||
function readFile(): PricingFile {
|
||||
ensureUserDataDirs();
|
||||
try {
|
||||
const raw = readFileSync(filePath(), "utf8");
|
||||
const parsed = JSON.parse(raw) as PricingFile;
|
||||
if (parsed.version !== 1 || !Array.isArray(parsed.sources)) return emptyFile();
|
||||
return parsed;
|
||||
} catch {
|
||||
return emptyFile();
|
||||
}
|
||||
}
|
||||
|
||||
function writeFile(data: PricingFile): void {
|
||||
ensureUserDataDirs();
|
||||
writeFileSync(filePath(), JSON.stringify(data, null, 2), "utf8");
|
||||
}
|
||||
|
||||
export function listPricingSources(): PricingSource[] {
|
||||
return readFile().sources;
|
||||
}
|
||||
|
||||
export function getPricingSource(id: string): PricingSource | null {
|
||||
return listPricingSources().find((s) => s.id === id) ?? null;
|
||||
}
|
||||
|
||||
export function normalizePricingUrl(raw: string): string {
|
||||
const trimmed = raw.trim();
|
||||
let url: URL;
|
||||
try {
|
||||
url = new URL(trimmed);
|
||||
} catch {
|
||||
throw new Error("不是合法 URL");
|
||||
}
|
||||
if (url.protocol !== "http:" && url.protocol !== "https:") {
|
||||
throw new Error("只支持 http/https 官方文档");
|
||||
}
|
||||
url.hash = "";
|
||||
if (url.pathname.length > 1) {
|
||||
url.pathname = url.pathname.replace(/\/+$/, "");
|
||||
}
|
||||
return url.toString();
|
||||
}
|
||||
|
||||
export async function addPricingSource(input: {
|
||||
url: string;
|
||||
name?: string;
|
||||
}): Promise<PricingSource> {
|
||||
const url = normalizePricingUrl(input.url);
|
||||
const data = readFile();
|
||||
const existing = data.sources.find((s) => s.url === url);
|
||||
if (existing) {
|
||||
return refreshPricingSource(existing.id);
|
||||
}
|
||||
const source: PricingSource = {
|
||||
id: randomUUID(),
|
||||
name: input.name?.trim() || hostLabel(url),
|
||||
url,
|
||||
createdAt: new Date().toISOString(),
|
||||
rates: [],
|
||||
};
|
||||
data.sources.push(source);
|
||||
writeFile(data);
|
||||
return refreshPricingSource(source.id);
|
||||
}
|
||||
|
||||
export function deletePricingSource(id: string): boolean {
|
||||
const data = readFile();
|
||||
const next = data.sources.filter((s) => s.id !== id);
|
||||
if (next.length === data.sources.length) return false;
|
||||
writeFile({ version: 1, sources: next });
|
||||
return true;
|
||||
}
|
||||
|
||||
export async function refreshPricingSource(id: string): Promise<PricingSource> {
|
||||
const data = readFile();
|
||||
const source = data.sources.find((s) => s.id === id);
|
||||
if (!source) throw new Error("价目来源不存在");
|
||||
try {
|
||||
const parsed = await fetchAndParsePricing(source.url);
|
||||
source.fetchedAt = new Date().toISOString();
|
||||
source.title = parsed.title;
|
||||
source.rates = parsed.rates;
|
||||
if (parsed.title && looksGenericName(source.name)) {
|
||||
source.name = shortenTitle(parsed.title);
|
||||
}
|
||||
source.fetchError = parsed.rates.length
|
||||
? undefined
|
||||
: parsed.warnings[0] || "未能从文档解析出型号价格";
|
||||
} catch (err) {
|
||||
source.fetchedAt = new Date().toISOString();
|
||||
source.fetchError = err instanceof Error ? err.message : String(err);
|
||||
}
|
||||
writeFile(data);
|
||||
return source;
|
||||
}
|
||||
|
||||
export type FetchedDoc = { contentType: string; text: string };
|
||||
|
||||
export function markdownDocFallbacks(url: string): string[] {
|
||||
try {
|
||||
const parsed = new URL(url);
|
||||
if (parsed.protocol !== "http:" && parsed.protocol !== "https:") return [];
|
||||
if (parsed.pathname.endsWith(".md")) return [];
|
||||
const docPath = parsed.pathname.replace(/\/+$/, "") || "/";
|
||||
parsed.hash = "";
|
||||
parsed.search = "";
|
||||
parsed.pathname = `${docPath}.md`;
|
||||
return [parsed.toString()];
|
||||
} catch {
|
||||
return [];
|
||||
}
|
||||
}
|
||||
|
||||
export async function fetchAndParsePricing(url: string): Promise<{
|
||||
title?: string;
|
||||
rates: ModelRate[];
|
||||
warnings: string[];
|
||||
}> {
|
||||
const candidates = [url, ...markdownDocFallbacks(url)];
|
||||
let lastError = "未能从文档解析出型号价格";
|
||||
let lastTitle: string | undefined;
|
||||
for (const candidate of candidates) {
|
||||
try {
|
||||
const doc = await fetchPricingDocument(candidate);
|
||||
const parsed = parsePricingDocument(doc.text, doc.contentType);
|
||||
lastTitle = parsed.title ?? lastTitle;
|
||||
if (parsed.rates.length) return parsed;
|
||||
lastError = parsed.warnings[0] || lastError;
|
||||
} catch (err) {
|
||||
lastError = err instanceof Error ? err.message : String(err);
|
||||
}
|
||||
}
|
||||
return { title: lastTitle, rates: [], warnings: [lastError] };
|
||||
}
|
||||
|
||||
export async function fetchPricingDocument(url: string): Promise<FetchedDoc> {
|
||||
const res = await fetch(url, {
|
||||
redirect: "follow",
|
||||
headers: {
|
||||
Accept: "text/html, text/markdown, application/json, text/plain;q=0.9",
|
||||
"User-Agent": "WritingAgent/0.1 (local token cost)",
|
||||
},
|
||||
signal: AbortSignal.timeout(FETCH_TIMEOUT_MS),
|
||||
});
|
||||
if (!res.ok) {
|
||||
throw new Error(`拉取失败 HTTP ${res.status}`);
|
||||
}
|
||||
const buf = Buffer.from(await res.arrayBuffer());
|
||||
if (buf.length > MAX_BYTES) {
|
||||
throw new Error("文档过大");
|
||||
}
|
||||
const contentType = res.headers.get("content-type") ?? "text/html";
|
||||
return { contentType, text: buf.toString("utf8") };
|
||||
}
|
||||
|
||||
export type ResolvedRate = ModelRate & {
|
||||
sourceId: string;
|
||||
sourceName: string;
|
||||
sourceUrl: string;
|
||||
};
|
||||
|
||||
export function resolveRateCatalog(sources = listPricingSources()): {
|
||||
byKey: Map<string, ResolvedRate>;
|
||||
rates: ResolvedRate[];
|
||||
} {
|
||||
const byKey = new Map<string, ResolvedRate>();
|
||||
for (const source of sources) {
|
||||
for (const rate of source.rates) {
|
||||
const resolved: ResolvedRate = {
|
||||
...rate,
|
||||
sourceId: source.id,
|
||||
sourceName: source.name,
|
||||
sourceUrl: source.url,
|
||||
};
|
||||
byKey.set(rate.model.toLowerCase(), resolved);
|
||||
for (const alias of rate.aliases) {
|
||||
byKey.set(alias.toLowerCase(), resolved);
|
||||
}
|
||||
}
|
||||
}
|
||||
const seen = new Set<string>();
|
||||
const rates: ResolvedRate[] = [];
|
||||
for (const rate of byKey.values()) {
|
||||
const id = `${rate.sourceId}:${rate.model.toLowerCase()}`;
|
||||
if (seen.has(id)) continue;
|
||||
seen.add(id);
|
||||
rates.push(rate);
|
||||
}
|
||||
return { byKey, rates };
|
||||
}
|
||||
|
||||
export function findModelRate(
|
||||
model: string,
|
||||
catalog = resolveRateCatalog(),
|
||||
): ResolvedRate | undefined {
|
||||
const key = model.trim().toLowerCase();
|
||||
if (!key) return undefined;
|
||||
const exact = catalog.byKey.get(key);
|
||||
if (exact) return exact;
|
||||
const tail = key.split("/").pop();
|
||||
if (tail && tail !== key) return catalog.byKey.get(tail);
|
||||
return undefined;
|
||||
}
|
||||
|
||||
function hostLabel(url: string): string {
|
||||
try {
|
||||
return new URL(url).hostname.replace(/^www\./, "");
|
||||
} catch {
|
||||
return url;
|
||||
}
|
||||
}
|
||||
|
||||
function looksGenericName(name: string): boolean {
|
||||
return /^(https?:|www\.|[a-z0-9.-]+\.[a-z]{2,}$)/i.test(name);
|
||||
}
|
||||
|
||||
function shortenTitle(title: string): string {
|
||||
return title.replace(/\s*[·|]\s*.*$/, "").trim() || title;
|
||||
}
|
||||
@@ -129,7 +129,7 @@ export function readRecords(query: TokenStatsQuery = {}): TokenUsageRecord[] {
|
||||
const lines = readFileSync(file, "utf8").split(/\r?\n/).filter(Boolean);
|
||||
for (let j = lines.length - 1; j >= 0 && records.length < limit; j--) {
|
||||
try {
|
||||
const record = JSON.parse(lines[j]) as TokenUsageRecord;
|
||||
const record = JSON.parse(lines[j].trim()) as TokenUsageRecord;
|
||||
if (query.bookId && record.bookId !== query.bookId) continue;
|
||||
if (query.orchestratorId && record.orchestratorId !== query.orchestratorId) {
|
||||
continue;
|
||||
@@ -171,7 +171,10 @@ function accumulateCaller(
|
||||
}
|
||||
|
||||
export function summarizeTokenUsage(query: TokenStatsQuery = {}): TokenStatsSummary {
|
||||
const records = readRecords({ ...query, limit: query.limit ?? 2000 });
|
||||
return summarizeRecords(readRecords({ ...query, limit: query.limit ?? 2000 }));
|
||||
}
|
||||
|
||||
export function summarizeRecords(records: TokenUsageRecord[]): TokenStatsSummary {
|
||||
const summary: TokenStatsSummary = {
|
||||
totalCalls: records.length,
|
||||
promptTokens: 0,
|
||||
|
||||
Reference in New Issue
Block a user