访问跟踪函数中的当前运行（span）

在某些情况下，您可能希望在跟踪函数中访问当前运行（span）。这对于从当前运行中提取 UUID、标签或其他信息非常有用。

您可以通过分别调用 Python 或 TypeScript SDK 中的 get_current_run_tree/getCurrentRunTree 函数来访问当前运行。

有关 RunTree 对象上可用属性的完整列表，请参阅此参考。

Python

from langsmith import traceable
from langsmith.run_helpers import get_current_run_tree
from openai import Client

openai = Client()

@traceable
def format_prompt(subject):
  run = get_current_run_tree()
  print(f"format_prompt Run Id: {run.id}")
  print(f"format_prompt Trace Id: {run.trace_id}")
  print(f"format_prompt Parent Run Id: {run.parent_run.id}")
  return [
      {
          "role": "system",
          "content": "You are a helpful assistant.",
      },
      {
          "role": "user",
          "content": f"What's a good name for a store that sells {subject}?"
      }
  ]

@traceable(run_type="llm")
def invoke_llm(messages):
  run = get_current_run_tree()
  print(f"invoke_llm Run Id: {run.id}")
  print(f"invoke_llm Trace Id: {run.trace_id}")
  print(f"invoke_llm Parent Run Id: {run.parent_run.id}")
  return openai.chat.completions.create(
      messages=messages, model="gpt-4o-mini", temperature=0
  )

@traceable
def parse_output(response):
  run = get_current_run_tree()
  print(f"parse_output Run Id: {run.id}")
  print(f"parse_output Trace Id: {run.trace_id}")
  print(f"parse_output Parent Run Id: {run.parent_run.id}")
  return response.choices[0].message.content

@traceable
def run_pipeline():
  run = get_current_run_tree()
  print(f"run_pipeline Run Id: {run.id}")
  print(f"run_pipeline Trace Id: {run.trace_id}")
  messages = format_prompt("colorful socks")
  response = invoke_llm(messages)
  return parse_output(response)

run_pipeline()

TypeScript

import { traceable, getCurrentRunTree } from "langsmith/traceable";
import OpenAI from "openai";

const openai = new OpenAI();

const formatPrompt = traceable(
(subject: string) => {
  const run = getCurrentRunTree();
  console.log("formatPrompt Run ID", run.id)
  console.log("formatPrompt Trace ID", run.trace_id)
  console.log("formatPrompt Parent Run ID", run.parent_run.id)
  return [
    {
      role: "system" as const,
      content: "You are a helpful assistant.",
    },
    {
      role: "user" as const,
      content: `What's a good name for a store that sells ${subject}?`,
    },
  ];
},
{ name: "formatPrompt" }
);

const invokeLLM = traceable(
  async (messages: { role: string; content: string }[]) => {
      const run = getCurrentRunTree();
      console.log("invokeLLM Run ID", run.id)
      console.log("invokeLLM Trace ID", run.trace_id)
      console.log("invokeLLM Parent Run ID", run.parent_run.id)
      return openai.chat.completions.create({
          model: "gpt-4o-mini",
          messages: messages,
          temperature: 0,
      });
  },
  { run_type: "llm", name: "invokeLLM" }
);

const parseOutput = traceable(
  (response: any) => {
      const run = getCurrentRunTree();
      console.log("parseOutput Run ID", run.id)
      console.log("parseOutput Trace ID", run.trace_id)
      console.log("parseOutput Parent Run ID", run.parent_run.id)
      return response.choices[0].message.content;
  },
  { name: "parseOutput" }
);

const runPipeline = traceable(
  async () => {
      const run = getCurrentRunTree();
      console.log("runPipline Run ID", run.id)
      console.log("runPipline Trace ID", run.trace_id)
      console.log("runPipline Parent Run ID", run.parent_run?.id)
      const messages = await formatPrompt("colorful socks");
      const response = await invokeLLM(messages);
      return parseOutput(response);
  },
  { name: "runPipeline" }
);

await runPipeline();