openlit
Version:
OpenTelemetry-native Auto instrumentation library for monitoring LLM Applications, facilitating the integration of observability into your GenAI-driven projects
372 lines • 18.1 kB
JavaScript
;
var __createBinding = (this && this.__createBinding) || (Object.create ? (function(o, m, k, k2) {
if (k2 === undefined) k2 = k;
var desc = Object.getOwnPropertyDescriptor(m, k);
if (!desc || ("get" in desc ? !m.__esModule : desc.writable || desc.configurable)) {
desc = { enumerable: true, get: function() { return m[k]; } };
}
Object.defineProperty(o, k2, desc);
}) : (function(o, m, k, k2) {
if (k2 === undefined) k2 = k;
o[k2] = m[k];
}));
var __setModuleDefault = (this && this.__setModuleDefault) || (Object.create ? (function(o, v) {
Object.defineProperty(o, "default", { enumerable: true, value: v });
}) : function(o, v) {
o["default"] = v;
});
var __importStar = (this && this.__importStar) || (function () {
var ownKeys = function(o) {
ownKeys = Object.getOwnPropertyNames || function (o) {
var ar = [];
for (var k in o) if (Object.prototype.hasOwnProperty.call(o, k)) ar[ar.length] = k;
return ar;
};
return ownKeys(o);
};
return function (mod) {
if (mod && mod.__esModule) return mod;
var result = {};
if (mod != null) for (var k = ownKeys(mod), i = 0; i < k.length; i++) if (k[i] !== "default") __createBinding(result, mod, k[i]);
__setModuleDefault(result, mod);
return result;
};
})();
var __importDefault = (this && this.__importDefault) || function (mod) {
return (mod && mod.__esModule) ? mod : { "default": mod };
};
Object.defineProperty(exports, "__esModule", { value: true });
const api_1 = require("@opentelemetry/api");
const config_1 = __importDefault(require("../../config"));
const helpers_1 = __importStar(require("../../helpers"));
const semantic_convention_1 = __importDefault(require("../../semantic-convention"));
const base_wrapper_1 = __importDefault(require("../base-wrapper"));
const FINISH_REASON_MAP = {
end_turn: 'stop',
max_tokens: 'length',
stop_sequence: 'stop',
tool_use: 'tool_call',
};
function mapFinishReason(reason) {
return FINISH_REASON_MAP[reason] || reason || 'stop';
}
function spanCreationAttrs(operationName, requestModel) {
return {
[semantic_convention_1.default.GEN_AI_OPERATION]: operationName,
[semantic_convention_1.default.GEN_AI_PROVIDER_NAME_OTEL]: AnthropicWrapper.aiSystem,
[semantic_convention_1.default.GEN_AI_REQUEST_MODEL]: requestModel,
[semantic_convention_1.default.SERVER_ADDRESS]: AnthropicWrapper.serverAddress,
[semantic_convention_1.default.SERVER_PORT]: AnthropicWrapper.serverPort,
};
}
class AnthropicWrapper extends base_wrapper_1.default {
static _patchMessageCreate(tracer) {
const genAIEndpoint = 'anthropic.resources.messages';
return (originalMethod) => {
return async function (...args) {
if ((0, helpers_1.isFrameworkLlmActive)())
return originalMethod.apply(this, args);
const requestModel = args[0]?.model || 'claude-3-5-sonnet-latest';
const spanName = `${semantic_convention_1.default.GEN_AI_OPERATION_TYPE_CHAT} ${requestModel}`;
const effectiveCtx = (0, helpers_1.getFrameworkParentContext)() ?? api_1.context.active();
const span = tracer.startSpan(spanName, {
kind: api_1.SpanKind.CLIENT,
attributes: spanCreationAttrs(semantic_convention_1.default.GEN_AI_OPERATION_TYPE_CHAT, requestModel),
}, effectiveCtx);
return api_1.context
.with(api_1.trace.setSpan(effectiveCtx, span), async () => {
return originalMethod.apply(this, args);
})
.then((response) => {
const { stream = false } = args[0];
if (stream) {
return helpers_1.default.createStreamProxy(response, AnthropicWrapper._messageCreateGenerator({
args,
genAIEndpoint,
response,
span,
}));
}
return AnthropicWrapper._messageCreate({ args, genAIEndpoint, response, span });
})
.catch((e) => {
helpers_1.default.handleException(span, e);
base_wrapper_1.default.recordMetrics(span, {
genAIEndpoint,
model: requestModel,
aiSystem: AnthropicWrapper.aiSystem,
serverAddress: AnthropicWrapper.serverAddress,
serverPort: AnthropicWrapper.serverPort,
errorType: e?.constructor?.name || '_OTHER',
});
span.end();
throw e;
});
};
};
}
static async _messageCreate({ args, genAIEndpoint, response, span, }) {
let metricParams;
try {
metricParams = await AnthropicWrapper._messageCreateCommonSetter({
args,
genAIEndpoint,
result: response,
span,
});
return response;
}
catch (e) {
helpers_1.default.handleException(span, e);
throw e;
}
finally {
span.end();
if (metricParams) {
base_wrapper_1.default.recordMetrics(span, metricParams);
}
}
}
static async *_messageCreateGenerator({ args, genAIEndpoint, response, span, }) {
let metricParams;
const timestamps = [];
const startTime = Date.now();
try {
const result = {
id: '',
model: '',
stop_reason: '',
content: [],
usage: {
input_tokens: 0,
output_tokens: 0,
cache_creation_input_tokens: 0,
cache_read_input_tokens: 0,
},
};
let llmResponse = '';
let toolId = '';
let toolName = '';
let toolArguments = '';
for await (const chunk of response) {
timestamps.push(Date.now());
switch (chunk.type) {
case 'message_start':
if (chunk.message) {
result.id = chunk.message.id;
result.model = chunk.message.model;
result.usage.input_tokens = Number(chunk.message.usage?.input_tokens) || 0;
result.usage.output_tokens += Number(chunk.message.usage?.output_tokens) || 0;
result.usage.cache_creation_input_tokens =
Number(chunk.message.usage?.cache_creation_input_tokens) || 0;
result.usage.cache_read_input_tokens =
Number(chunk.message.usage?.cache_read_input_tokens) || 0;
result.stop_reason = chunk.message.stop_reason || '';
}
break;
case 'content_block_start':
if (chunk.content_block?.type === 'tool_use') {
toolId = chunk.content_block.id || '';
toolName = chunk.content_block.name || '';
toolArguments = '';
}
break;
case 'content_block_delta':
if (chunk.delta?.text) {
llmResponse += chunk.delta.text;
}
if (chunk.delta?.partial_json) {
toolArguments += chunk.delta.partial_json;
}
break;
case 'content_block_stop':
break;
case 'message_delta':
result.stop_reason = chunk.delta?.stop_reason || result.stop_reason;
result.usage.output_tokens += Number(chunk.usage?.output_tokens) || 0;
if (chunk.usage?.cache_creation_input_tokens != null) {
result.usage.cache_creation_input_tokens =
Number(chunk.usage.cache_creation_input_tokens) || 0;
}
if (chunk.usage?.cache_read_input_tokens != null) {
result.usage.cache_read_input_tokens =
Number(chunk.usage.cache_read_input_tokens) || 0;
}
break;
case 'message_stop':
break;
}
yield chunk;
}
if (llmResponse) {
result.content.push({ type: 'text', text: llmResponse });
}
if (toolId) {
let parsedInput = {};
try {
parsedInput = JSON.parse(toolArguments);
}
catch { /* keep empty */ }
result.content.push({
type: 'tool_use',
id: toolId,
name: toolName,
input: parsedInput,
});
}
const ttft = timestamps.length > 0 ? (timestamps[0] - startTime) / 1000 : 0;
let tbt = 0;
if (timestamps.length > 1) {
const timeDiffs = timestamps.slice(1).map((t, i) => t - timestamps[i]);
tbt = timeDiffs.reduce((a, b) => a + b, 0) / timeDiffs.length / 1000;
}
metricParams = await AnthropicWrapper._messageCreateCommonSetter({
args,
genAIEndpoint,
result,
span,
ttft,
tbt,
});
return result;
}
catch (e) {
helpers_1.default.handleException(span, e);
throw e;
}
finally {
span.end();
if (metricParams) {
base_wrapper_1.default.recordMetrics(span, metricParams);
}
}
}
static async _messageCreateCommonSetter({ args, genAIEndpoint, result, span, ttft = 0, tbt = 0, }) {
const captureContent = config_1.default.captureMessageContent;
const requestModel = args[0]?.model || 'claude-3-5-sonnet-latest';
const { messages, system, max_tokens = null, seed = null, temperature = 1, top_p, top_k, stop_sequences = null, stream = false, user, } = args[0];
span.setAttribute(semantic_convention_1.default.GEN_AI_REQUEST_TEMPERATURE, temperature);
span.setAttribute(semantic_convention_1.default.GEN_AI_REQUEST_TOP_P, top_p || 1);
if (top_k != null) {
span.setAttribute(semantic_convention_1.default.GEN_AI_REQUEST_TOP_K, top_k);
}
if (max_tokens != null) {
span.setAttribute(semantic_convention_1.default.GEN_AI_REQUEST_MAX_TOKENS, max_tokens);
}
span.setAttribute(semantic_convention_1.default.GEN_AI_REQUEST_IS_STREAM, stream);
if (seed != null) {
span.setAttribute(semantic_convention_1.default.GEN_AI_REQUEST_SEED, Number(seed));
}
if (stop_sequences && Array.isArray(stop_sequences) && stop_sequences.length > 0) {
span.setAttribute(semantic_convention_1.default.GEN_AI_REQUEST_STOP_SEQUENCES, stop_sequences);
}
if (captureContent) {
const systemStr = typeof system === 'string' ? system : undefined;
span.setAttribute(semantic_convention_1.default.GEN_AI_INPUT_MESSAGES, helpers_1.default.buildInputMessages(messages || [], systemStr));
if (system) {
const sysAttr = typeof system === 'string' ? system : JSON.stringify(system);
span.setAttribute(semantic_convention_1.default.GEN_AI_SYSTEM_INSTRUCTIONS, sysAttr);
}
}
span.setAttribute(semantic_convention_1.default.GEN_AI_RESPONSE_ID, result.id);
const responseModel = result.model || requestModel;
const pricingInfo = config_1.default.pricingInfo || {};
const cost = helpers_1.default.getChatModelCost(requestModel, pricingInfo, result.usage.input_tokens, result.usage.output_tokens);
AnthropicWrapper.setBaseSpanAttributes(span, {
genAIEndpoint,
model: requestModel,
user,
cost,
aiSystem: AnthropicWrapper.aiSystem,
serverAddress: AnthropicWrapper.serverAddress,
serverPort: AnthropicWrapper.serverPort,
});
span.setAttribute(semantic_convention_1.default.GEN_AI_RESPONSE_MODEL, responseModel);
const inputTokens = result.usage.input_tokens;
const outputTokens = result.usage.output_tokens;
span.setAttribute(semantic_convention_1.default.GEN_AI_USAGE_INPUT_TOKENS, inputTokens);
span.setAttribute(semantic_convention_1.default.GEN_AI_USAGE_OUTPUT_TOKENS, outputTokens);
if (result.usage.cache_creation_input_tokens) {
span.setAttribute(semantic_convention_1.default.GEN_AI_USAGE_CACHE_CREATION_INPUT_TOKENS, result.usage.cache_creation_input_tokens);
}
if (result.usage.cache_read_input_tokens) {
span.setAttribute(semantic_convention_1.default.GEN_AI_USAGE_CACHE_READ_INPUT_TOKENS, result.usage.cache_read_input_tokens);
}
if (ttft > 0) {
span.setAttribute(semantic_convention_1.default.GEN_AI_SERVER_TTFT, ttft);
}
if (tbt > 0) {
span.setAttribute(semantic_convention_1.default.GEN_AI_SERVER_TBT, tbt);
}
const finishReason = mapFinishReason(result.stop_reason);
span.setAttribute(semantic_convention_1.default.GEN_AI_RESPONSE_FINISH_REASON, [finishReason]);
const toolUseBlocks = (result.content || []).filter((b) => b.type === 'tool_use');
const outputType = toolUseBlocks.length > 0
? semantic_convention_1.default.GEN_AI_OUTPUT_TYPE_JSON
: semantic_convention_1.default.GEN_AI_OUTPUT_TYPE_TEXT;
span.setAttribute(semantic_convention_1.default.GEN_AI_OUTPUT_TYPE, outputType);
if (toolUseBlocks.length > 0) {
const toolNames = toolUseBlocks.map((b) => b.name || '').filter(Boolean);
const toolIds = toolUseBlocks.map((b) => b.id || '').filter(Boolean);
const toolArgs = toolUseBlocks.map((b) => JSON.stringify(b.input || {}));
if (toolNames.length > 0)
span.setAttribute(semantic_convention_1.default.GEN_AI_TOOL_NAME, toolNames.join(', '));
if (toolIds.length > 0)
span.setAttribute(semantic_convention_1.default.GEN_AI_TOOL_CALL_ID, toolIds.join(', '));
if (toolArgs.length > 0)
span.setAttribute(semantic_convention_1.default.GEN_AI_TOOL_ARGS, toolArgs.join(', '));
}
let inputMessagesJson;
let outputMessagesJson;
if (captureContent) {
const textContent = (result.content || [])
.filter((b) => b.type === 'text')
.map((b) => b.text || '')
.join('');
const toolCallsForOutput = toolUseBlocks.length > 0
? toolUseBlocks.map((b) => ({
id: b.id || '',
name: b.name || '',
arguments: b.input || {},
}))
: undefined;
outputMessagesJson = helpers_1.default.buildOutputMessages(textContent, finishReason, toolCallsForOutput);
span.setAttribute(semantic_convention_1.default.GEN_AI_OUTPUT_MESSAGES, outputMessagesJson);
const systemStr = typeof system === 'string' ? system : undefined;
inputMessagesJson = helpers_1.default.buildInputMessages(messages || [], systemStr);
}
if (!config_1.default.disableEvents) {
const eventAttrs = {
[semantic_convention_1.default.GEN_AI_OPERATION]: semantic_convention_1.default.GEN_AI_OPERATION_TYPE_CHAT,
[semantic_convention_1.default.GEN_AI_REQUEST_MODEL]: requestModel,
[semantic_convention_1.default.GEN_AI_RESPONSE_MODEL]: responseModel,
[semantic_convention_1.default.SERVER_ADDRESS]: AnthropicWrapper.serverAddress,
[semantic_convention_1.default.SERVER_PORT]: AnthropicWrapper.serverPort,
[semantic_convention_1.default.GEN_AI_RESPONSE_ID]: result.id,
[semantic_convention_1.default.GEN_AI_RESPONSE_FINISH_REASON]: [finishReason],
[semantic_convention_1.default.GEN_AI_OUTPUT_TYPE]: outputType,
[semantic_convention_1.default.GEN_AI_USAGE_INPUT_TOKENS]: inputTokens,
[semantic_convention_1.default.GEN_AI_USAGE_OUTPUT_TOKENS]: outputTokens,
};
if (captureContent) {
if (inputMessagesJson)
eventAttrs[semantic_convention_1.default.GEN_AI_INPUT_MESSAGES] = inputMessagesJson;
if (outputMessagesJson)
eventAttrs[semantic_convention_1.default.GEN_AI_OUTPUT_MESSAGES] = outputMessagesJson;
}
helpers_1.default.emitInferenceEvent(span, eventAttrs);
}
return {
genAIEndpoint,
model: requestModel,
user,
cost,
aiSystem: AnthropicWrapper.aiSystem,
};
}
}
AnthropicWrapper.aiSystem = semantic_convention_1.default.GEN_AI_SYSTEM_ANTHROPIC;
AnthropicWrapper.serverAddress = 'api.anthropic.com';
AnthropicWrapper.serverPort = 443;
exports.default = AnthropicWrapper;
//# sourceMappingURL=wrapper.js.map