@juspay/neurolink
Version:
Universal AI Development Platform with working MCP integration, multi-provider support, voice (TTS/STT/realtime), and professional CLI. 58+ external MCP servers discoverable, multimodal file processing, RAG pipelines. Build, test, and deploy AI applicatio
240 lines • 8.28 kB
JavaScript
/**
* Centralized token usage extraction utilities
* Handles multiple provider formats and optional fields
*
* Consolidates token extraction logic from:
* - GenerationHandler.ts
* - analytics.ts
* - streamAnalytics.ts
*/
/**
* Extract input token count from various provider formats
* Priority: input > inputTokens > promptTokens
*/
export function extractInputTokens(usage) {
if (typeof usage.input === "number") {
return usage.input;
}
if (typeof usage.inputTokens === "number") {
return usage.inputTokens;
}
if (typeof usage.promptTokens === "number") {
return usage.promptTokens;
}
return 0;
}
/**
* Extract output token count from various provider formats
* Priority: output > outputTokens > completionTokens
*/
export function extractOutputTokens(usage) {
if (typeof usage.output === "number") {
return usage.output;
}
if (typeof usage.outputTokens === "number") {
return usage.outputTokens;
}
if (typeof usage.completionTokens === "number") {
return usage.completionTokens;
}
return 0;
}
/**
* Extract total token count from various provider formats
* Falls back to input + output if total is not provided
*/
export function extractTotalTokens(usage, input, output) {
if (typeof usage.total === "number") {
return usage.total;
}
if (typeof usage.totalTokens === "number") {
return usage.totalTokens;
}
return input + output;
}
/**
* Extract reasoning/thinking token count from various provider formats
* Supports: reasoningTokens, thinkingTokens, reasoning_tokens, reasoning
*/
export function extractReasoningTokens(usage) {
if (typeof usage.reasoningTokens === "number" && usage.reasoningTokens > 0) {
return usage.reasoningTokens;
}
if (typeof usage.thinkingTokens === "number" && usage.thinkingTokens > 0) {
return usage.thinkingTokens;
}
if (typeof usage.reasoning_tokens === "number" &&
usage.reasoning_tokens > 0) {
return usage.reasoning_tokens;
}
if (typeof usage.reasoning === "number" && usage.reasoning > 0) {
return usage.reasoning;
}
return undefined;
}
/**
* Extract cache creation token count from various provider formats
* Supports: cacheCreationInputTokens, cacheCreationTokens
*/
export function extractCacheCreationTokens(usage) {
if (typeof usage.cacheCreationInputTokens === "number" &&
usage.cacheCreationInputTokens > 0) {
return usage.cacheCreationInputTokens;
}
if (typeof usage.cacheCreationTokens === "number" &&
usage.cacheCreationTokens > 0) {
return usage.cacheCreationTokens;
}
return undefined;
}
/**
* Extract cache read token count from various provider formats
* Supports: cacheReadInputTokens, cacheReadTokens
*/
export function extractCacheReadTokens(usage) {
if (typeof usage.cacheReadInputTokens === "number" &&
usage.cacheReadInputTokens > 0) {
return usage.cacheReadInputTokens;
}
if (typeof usage.cacheReadTokens === "number" && usage.cacheReadTokens > 0) {
return usage.cacheReadTokens;
}
return undefined;
}
/**
* Calculate cache savings percentage
*
* This represents the percentage of input tokens served from cache.
* For Anthropic, cache read tokens cost 0.1x, so actual cost savings = cacheSavingsPercent * 0.9
* For other providers, cost savings may vary based on their cache pricing.
*
* @param cacheReadTokens Number of tokens read from cache
* @param inputTokens Number of non-cached input tokens
* @returns Percentage of tokens served from cache (0-100), or undefined if no cache usage
*/
export function calculateCacheSavingsPercent(cacheReadTokens, inputTokens) {
if (cacheReadTokens === undefined || cacheReadTokens <= 0) {
return undefined;
}
const totalInputWithCache = inputTokens + cacheReadTokens;
if (totalInputWithCache <= 0) {
return undefined;
}
return Math.round((cacheReadTokens / totalInputWithCache) * 100);
}
/**
* Extract token usage from various provider response formats
*
* Handles multiple input formats:
* - BaseProvider normalized format (input/output/total)
* - AI SDK format (inputTokens/outputTokens/totalTokens)
* - OpenAI/Mistral format (promptTokens/completionTokens)
* - Nested usage objects
*
* Also extracts optional fields:
* - Cache creation and read tokens (Anthropic-style)
* - Reasoning/thinking tokens (OpenAI o1, Anthropic, Google)
* - Cache savings percentage
*
* @param result Raw usage object from provider response
* @param options Extraction options
* @returns Normalized TokenUsage object
*/
export function extractTokenUsage(result, options = {}) {
const { calculateCacheSavings = true, missingOptionalBehavior = "undefined", } = options;
// Handle null/undefined input
if (!result) {
return { input: 0, output: 0, total: 0 };
}
// Handle nested usage object (some providers wrap usage in a usage property)
const usage = result.usage && typeof result.usage === "object" ? result.usage : result;
// Extract base token counts
const input = extractInputTokens(usage);
const output = extractOutputTokens(usage);
const total = extractTotalTokens(usage, input, output);
// Extract optional token fields
const reasoning = extractReasoningTokens(usage);
const cacheCreationTokens = extractCacheCreationTokens(usage);
const cacheReadTokens = extractCacheReadTokens(usage);
// Calculate cache savings if enabled
const cacheSavingsPercent = calculateCacheSavings
? calculateCacheSavingsPercent(cacheReadTokens, input)
: undefined;
// Build result object
const tokenUsage = {
input,
output,
total,
};
// Add optional fields based on behavior setting
if (missingOptionalBehavior === "zero") {
tokenUsage.cacheCreationTokens = cacheCreationTokens ?? 0;
tokenUsage.cacheReadTokens = cacheReadTokens ?? 0;
tokenUsage.reasoning = reasoning ?? 0;
tokenUsage.cacheSavingsPercent = cacheSavingsPercent ?? 0;
}
else {
// Only include optional fields if they have values
if (cacheCreationTokens !== undefined) {
tokenUsage.cacheCreationTokens = cacheCreationTokens;
}
if (cacheReadTokens !== undefined) {
tokenUsage.cacheReadTokens = cacheReadTokens;
}
if (reasoning !== undefined) {
tokenUsage.reasoning = reasoning;
}
if (cacheSavingsPercent !== undefined) {
tokenUsage.cacheSavingsPercent = cacheSavingsPercent;
}
}
return tokenUsage;
}
/**
* Create a default/empty TokenUsage object
* Useful for error handling and fallback scenarios
*/
export function createEmptyTokenUsage() {
return {
input: 0,
output: 0,
total: 0,
};
}
/**
* Check if a TokenUsage object has any non-zero values
*/
export function hasTokenUsage(usage) {
return usage.input > 0 || usage.output > 0 || usage.total > 0;
}
/**
* Merge two TokenUsage objects by summing their values
* Useful for aggregating usage across multiple calls
*/
export function mergeTokenUsage(a, b) {
const merged = {
input: a.input + b.input,
output: a.output + b.output,
total: a.total + b.total,
};
// Merge optional fields if present in either
const cacheCreationTokens = (a.cacheCreationTokens ?? 0) + (b.cacheCreationTokens ?? 0);
const cacheReadTokens = (a.cacheReadTokens ?? 0) + (b.cacheReadTokens ?? 0);
const reasoning = (a.reasoning ?? 0) + (b.reasoning ?? 0);
if (cacheCreationTokens > 0) {
merged.cacheCreationTokens = cacheCreationTokens;
}
if (cacheReadTokens > 0) {
merged.cacheReadTokens = cacheReadTokens;
}
if (reasoning > 0) {
merged.reasoning = reasoning;
}
// Recalculate cache savings for merged usage
const cacheSavingsPercent = calculateCacheSavingsPercent(merged.cacheReadTokens, merged.input);
if (cacheSavingsPercent !== undefined) {
merged.cacheSavingsPercent = cacheSavingsPercent;
}
return merged;
}
//# sourceMappingURL=tokenUtils.js.map