Skip to content

Корректность ответов: честный finish_reason + точный usage.prompt_tokens #21

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
danscMax wants to merge 1 commit into
base: main
Choose a base branch
from
+10 −4
Open

Корректность ответов: честный finish_reason + точный usage.prompt_tokens #21

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
14 changes: 10 additions & 4 deletions server.js
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -722,7 +722,7 @@ function buildToolCallResponse(toolCall, model = 'deepseek-default', prompt = ''
};
}

function buildTextResponse(content, prompt, model = 'deepseek-default', reasoningContent = '') {
function buildTextResponse(content, prompt, model = 'deepseek-default', reasoningContent = '', finishReason = null) {
const message = { role: 'assistant', content };
if (reasoningContent) message.reasoning_content = reasoningContent;
return {
Expand All @@ -733,7 +733,9 @@ function buildTextResponse(content, prompt, model = 'deepseek-default', reasonin
choices: [{
index: 0,
message,
finish_reason: 'stop'
// Surface truncation: a 'length' finish lets length-aware clients re-request
// instead of silently treating a cut-off answer as a clean stop.
finish_reason: finishReason === 'length' ? 'length' : 'stop'
}],
usage: buildUsage(prompt, content, reasoningContent),
watermark: FORGETMEAI_WATERMARK
Expand Down Expand Up @@ -1266,6 +1268,10 @@ const server = http.createServer(async (req, res) => {
}

const { prompt, systemPrompt } = formatMessages(messages, tools);
// For usage accounting, count the CLIENT's original input — not the
// proxy-expanded fullPrompt (system + injected tools + history) — so
// prompt_tokens reflects what the caller actually sent.
const clientPromptText = messages.map(m => normalizeMessageContent(m.content)).join('\n');

const session = getOrCreateAgentSession(agentId);

Expand Down Expand Up @@ -1502,8 +1508,8 @@ const server = http.createServer(async (req, res) => {
storeHistory(agentId, prompt, fullContent, toolCall);

const openaiResponse = toolCall
? buildToolCallResponse(toolCall, requestedModel, fullPrompt, reasoningContent)
: buildTextResponse(fullContent, fullPrompt, requestedModel, reasoningContent);
? buildToolCallResponse(toolCall, requestedModel, clientPromptText, reasoningContent)
: buildTextResponse(fullContent, clientPromptText, requestedModel, reasoningContent, finishReason);

if (stream) {
if (apiMode === 'anthropic') {
Expand Down