Improve OTel GenAI semantic convention compliance for LLM instrumentations

[jimilp7]

Summary

This PR improves OTel GenAI Semantic Convention compliance across LLM instrumentations.

What's New

1. OpenAI & Anthropic Instrumentation Improvements

Enhanced existing instrumentations to better comply with OTel GenAI Semantic Conventions:

Attribute	Before	After
SpanKind	INTERNAL	CLIENT
SpanName	"Chat Completion with {model!r}"	"chat gpt-4o-mini"
gen_ai.system	"openai" (deprecated)	Removed
gen_ai.provider.name	Missing	"openai" / "anthropic"
server.address	Missing	"api.openai.com"
gen_ai.request.*	Missing	max_tokens, temperature, top_p, etc.
gen_ai.tool.definitions	Missing	Tool schemas when provided

2. LangChain/LangGraph Instrumentation (New)

Adds logfire.instrument_langchain() to capture tool definitions from LangChain agents.

Why: LangSmith's OTEL integration does NOT include tool definitions - they're sent via their proprietary API. This instrumentation patches LangChain's callback system to capture them.

Usage:

import logfire
from langgraph.prebuilt import create_react_agent

logfire.configure()
logfire.instrument_langchain()

agent = create_react_agent(model, tools)
agent.invoke({"messages": [...]})

Captures:

• gen_ai.tool.definitions - Tool schemas (name, description, parameters)
• gen_ai.input.messages / gen_ai.output.messages - OTel format
• gen_ai.system_instructions - Extracted from system messages
• Token usage and model information

Related

• Addresses Migrate OpenAI and Anthropic instrumentations to new OpenTelemetry semantic conventions #1586 (OpenTelemetry Semantic Conventions Migration)
• Builds on Update LLM SDK instrumentations for better otel semconv compliance #1580 and Update response api to gen_ai.input.messages and gen_ai.output.messages #1590