Fix openai cache tracking and cost estimates #2616
Conversation
Set correct cacheReadsPrice (cached input price) for gpt-4.1, gpt-4.1 mini, and gpt-4.1 nano based on official OpenAI pricing. No changes to cacheWritesPrice as per current OpenAI documentation. This ensures prompt caching costs are accurately reflected for these models in cost calculations.
|
dosubot
bot
added
the
size:L
![ellipsis-dev[bot]](https://avatars.githubusercontent.com/in/64358?s=60&v=4){kind=small}
| // Check usage data fields but use toBeCloseTo for floating point comparison | ||
| expect(results[1].type).toBe("usage") | ||
| // Use type assertion to avoid TypeScript errors | ||
| expect((results[1] as any).inputTokens).toBe(10) |
There was a problem hiding this comment.
Consider defining proper result types to eliminate repeated 'as any' assertions, enhancing maintainability.
| // Check usage data fields but use toBeCloseTo for floating point comparison | ||
| expect(results[3].type).toBe("usage") | ||
| // Use type assertion to avoid TypeScript errors | ||
| expect((results[3] as any).inputTokens).toBe(10) |
There was a problem hiding this comment.
Consider replacing the any cast with a more specific type or unknown to improve type safety in these assertions.
| // Check usage data fields but use toBeCloseTo for floating point comparison | ||
| expect(results[1].type).toBe("usage") | ||
| // Use type assertion to avoid TypeScript errors | ||
| expect((results[1] as any).inputTokens).toBe(10) |
There was a problem hiding this comment.
Repeated type assertions for usage properties appear in multiple tests. Consider extracting these assertions into a helper function to reduce duplication and improve maintainability.
Important
Enhances OpenAI cache tracking and cost estimation by updating model handling and tests in
openai-native.tsandopenai-native.test.ts.OpenAiNativeHandlerinopenai-native.tsto include cache read tokens in cost calculation usingcalculateApiCostOpenAI.cacheReadsPricetoopenAiNativeModelsinapi.tsfor cost estimation.openai-native.test.tsto verify individual response fields and usage data, includingtotalCost.handleStreamResponseand related functions inopenai-native.tsto pass model information for accurate cost calculation.This description was created by
for 1c814a9. It will automatically update as commits are pushed.