feat: Add ByteStream and ChatMessage to Haystack init

[julian-risch]

Related Issues

I noticed when reviewing this PR that current imports of Document and ChatMessage or ByteStream differ.

from haystack import Answer, Document, ExtractedAnswer, GeneratedAnswer
from haystack.dataclasses.byte_stream import ByteStream
from haystack.dataclasses.chat_message import ChatMessage

As all of these are user facing and not just for internal use, as a user I would expect all of to work with

from haystack import Answer, Document, ExtractedAnswer, GeneratedAnswer, ByteStream, ChatMessage

We shouldn't treat them differently.

Proposed Changes:

• Added ByteStream and ChatMessage imports to Haystack's init.py

How did you test it?

Notes for the reviewer

init.py gets longer with this change but I don't see a reason why we would treat some dataclasses different than others. What do you think?

Checklist

• I have read the contributors guidelines and the code of conduct
• I have updated the related issue with new insights and changes
• I added unit tests and updated the docstrings
• I've used one of the conventional commit types for my PR title: fix:, feat:, build:, chore:, ci:, docs:, style:, refactor:, perf:, test:.
• I documented my code
• I ran pre-commit hooks and fixed any issue

[silvanocerza]

I wouldn't do this to be fair.

I already discussed a bit with @masci regarding this and we both agree that we should stop exporting everything at the root.

We should remove and not add from this __init__.py, or we risk the same situation we have for 1.x that importing haystack takes way too much time.

[julian-risch]

@silvanocerza Okay, I see. Yes the speed is the only downside I could think of. Have you considered removing ExtractiveAnswer from the init? Or any of the other dataclasses currently in there?

[silvanocerza]

I would actually remove most of the imports from there. I would probably leave just Pipeline. 🤔
But even nothing is good in my opinion.