import { SummarizeCompressor } from 'slimcontext';
import OpenAI from 'openai';
import {
SummarizeMessagesArgsSchema,
SlimContextError,
OpenAIError,
MCPToolResult,
SlimContextMessage,
} from '../types/index.js';
class OpenAISlimContextModel {
private openai: OpenAI;
private model: string;
constructor(apiKey: string, model: string = 'gpt-4o-mini') {
this.openai = new OpenAI({ apiKey });
this.model = model;
}
async invoke(messages: SlimContextMessage[]) {
try {
const openaiMessages = messages.map((msg) => ({
role:
msg.role === 'human'
? ('user' as const)
: msg.role === 'tool'
? ('assistant' as const)
: (msg.role as 'system' | 'user' | 'assistant'),
content: msg.content,
}));
const response = await this.openai.chat.completions.create({
model: this.model,
messages: openaiMessages,
temperature: 0.1,
max_tokens: 1000,
});
const content = response.choices[0]?.message?.content;
if (!content) {
throw new OpenAIError('No content returned from OpenAI API');
}
return { content };
} catch (error) {
if (error instanceof OpenAIError) {
throw error;
}
throw new OpenAIError(
`OpenAI API error: ${error instanceof Error ? error.message : 'Unknown error'}`,
error instanceof Error ? error : undefined
);
}
}
}
export async function handleSummarizeMessages(
args: unknown
): Promise<MCPToolResult> {
try {
const validatedArgs = SummarizeMessagesArgsSchema.parse(args);
const apiKey = validatedArgs.openaiApiKey || process.env.OPENAI_API_KEY;
if (!apiKey) {
throw new SlimContextError(
'OpenAI API key is required. Provide it via openaiApiKey parameter or OPENAI_API_KEY environment variable.'
);
}
const model = new OpenAISlimContextModel(apiKey, validatedArgs.openaiModel);
const summarizeConfig = {
model,
maxModelTokens: validatedArgs.maxModelTokens,
thresholdPercent: validatedArgs.thresholdPercent,
minRecentMessages: validatedArgs.minRecentMessages,
...(validatedArgs.customPrompt && { prompt: validatedArgs.customPrompt }),
};
const summarizeCompressor = new SummarizeCompressor(summarizeConfig);
const compressedMessages = await summarizeCompressor.compress(
validatedArgs.messages.map((msg) => ({
...msg,
role: msg.role === 'human' ? ('user' as const) : msg.role,
}))
);
const originalCount = validatedArgs.messages.length;
const compressedCount = compressedMessages.length;
const removedCount = originalCount - compressedCount;
const hasSummary = compressedMessages.some(
(msg) =>
msg.role === 'system' &&
!validatedArgs.messages.some((orig) => orig.content === msg.content)
);
return {
content: [
{
type: 'text',
text: JSON.stringify(
{
success: true,
original_message_count: originalCount,
compressed_message_count: compressedCount,
messages_removed: removedCount,
summary_generated: hasSummary,
compression_ratio: compressedCount / originalCount,
compressed_messages: compressedMessages,
},
null,
2
),
},
],
};
} catch (error) {
const errorMessage =
error instanceof Error ? error.message : 'Unknown error occurred';
const errorType =
error instanceof OpenAIError
? 'OpenAIError'
: error instanceof SlimContextError
? 'SlimContextError'
: 'UnknownError';
return {
content: [
{
type: 'text',
text: JSON.stringify(
{
success: false,
error: errorMessage,
error_type: errorType,
},
null,
2
),
},
],
};
}
}
