feat: add data_model support to .stream() and .stream_async() methods

[cpsievert]

Summary

• Adds data_model parameter to stream() and stream_async() methods
• When provided, constrains the response format to match a Pydantic model
• Streamed chunks are JSON text that can be parsed after consuming the stream

Usage

from chatlas import ChatOpenAI
from pydantic import BaseModel

class Person(BaseModel):
    name: str
    age: int

chat = ChatOpenAI()
chunks = list(chat.stream("John is 25 years old", data_model=Person))
person = Person.model_validate_json("".join(chunks))

Test plan

• Added test_stream_with_data_model() test
• Added test_stream_async_with_data_model() test
• All 450 tests pass
• Type checking passes

🤖 Generated with Claude Code