pacnpal/Roo-Code
1
0
Fork 0
mirror of https://github.com/pacnpal/Roo-Code.git synced 2025-12-20 04:11:10 -05:00
Code Issues Packages Projects Releases Wiki Activity

Fix haiku prompt caching when tool prompt doesnt meet min required tokens

Browse Source
This commit is contained in:
Saoud Rizwan
2024-08-15 20:37:15 -04:00
parent 6989779dd6
commit a0f2a08f35
2 changed files with 14 additions and 26 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
src/api/anthropic.ts
View File

@@ -56,9 +56,7 @@ export class AnthropicHandler implements ApiHandler {
}
return message
}),
tools: tools.map((tool, index) =>
index === tools.length - 1 ? { ...tool, cache_control: { type: "ephemeral" } } : tool
),
tools, // cache breakpoints go from tools > system > messages, and since tools dont change, we can just set the breakpoint at the end of system (this avoids having to set a breakpoint at the end of tools which by itself does not meet min requirements for haiku caching)
tool_choice: { type: "auto" },
},
(() => {

36
webview-ui/src/components/ApiOptions.tsx
View File

@@ -215,45 +215,35 @@ const ModelInfoView = ({ modelInfo }: { modelInfo: ModelInfo }) => {
}).format(price)
}
const showPromptCachingPrices =
modelInfo.supportsPromptCache && modelInfo.cacheWritesPrice && modelInfo.cacheReadsPrice
return (
<p style={{ fontSize: "12px", marginTop: "2px", color: "var(--vscode-descriptionForeground)" }}>
<ModelInfoSupportsItem
isSupported={modelInfo.supportsPromptCache}
supportsLabel="Supports prompt caching"
doesNotSupportLabel="Does not support prompt caching"
/>{" "}
<VSCodeLink href="https://www.anthropic.com/news/prompt-caching" style={{ display: "inline" }}>
(what is this?)
</VSCodeLink>
<br />
<ModelInfoSupportsItem
isSupported={modelInfo.supportsImages}
supportsLabel="Supports images"
doesNotSupportLabel="Does not support images"
/>
<br />
<ModelInfoSupportsItem
isSupported={modelInfo.supportsPromptCache}
supportsLabel="Supports prompt caching"
doesNotSupportLabel="Does not support prompt caching"
/>
<br />
<span style={{ fontWeight: 500 }}>Max output:</span> {modelInfo.maxTokens.toLocaleString()} tokens
<br />
<span style={{ fontWeight: 500 }}>
{showPromptCachingPrices ? "Base input price:" : "Input price:"}
</span>{" "}
{formatPrice(modelInfo.inputPrice)} per million tokens
{showPromptCachingPrices && (
<span style={{ fontWeight: 500 }}>Input price:</span> {formatPrice(modelInfo.inputPrice)}/million tokens
{modelInfo.supportsPromptCache && modelInfo.cacheWritesPrice && modelInfo.cacheReadsPrice && (
<>
<br />
<span style={{ fontWeight: 500 }}>Prompt caching write price:</span>{" "}
{formatPrice(modelInfo.cacheWritesPrice || 0)} per million tokens
<span style={{ fontWeight: 500 }}>Cache writes price:</span>{" "}
{formatPrice(modelInfo.cacheWritesPrice || 0)}/million tokens
<br />
<span style={{ fontWeight: 500 }}>Prompt caching read price:</span>{" "}
{formatPrice(modelInfo.cacheReadsPrice || 0)} per million tokens
<span style={{ fontWeight: 500 }}>Cache reads price:</span>{" "}
{formatPrice(modelInfo.cacheReadsPrice || 0)}/million tokens
</>
)}
<br />
<span style={{ fontWeight: 500 }}>Output price:</span> {formatPrice(modelInfo.outputPrice)} per million
tokens
<span style={{ fontWeight: 500 }}>Output price:</span> {formatPrice(modelInfo.outputPrice)}/million tokens
</p>
)
}