chore: use text instead of content for ChatMessage in Cohere and Anthropic

[anakin87]

Related Issues

• part of Update all components with ChatMessage to use text instead of content #1236

Proposed Changes:

• use text instead of content for ChatMessage in Cohere and Anthropic
• I did the changes in a single PR because for Anthropic we are only changing tests, so we don't need a release

How did you test it?

CI

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:.

[Amnah199]

Not that it impacts the functionality but some comments in anthropic can be updated to replace content with text. In a longer run, this will prevent any inconsistencies in code e.g.

1. small change for Anthropic here
2. and here
3. for AnthropicVertex here

[anakin87]

@Amnah199

Let me explain better...

Currently text just mirrors content (code, deepset-ai/haystack#8588). If you access content directly, a deprecation warning is shown.

But printing or converting to message to a dictionary still shows content. (So I don't think we should update the code you linked)

from haystack.dataclasses import ChatMessage
msg=ChatMessage.from_user("I'm Stefano")
print(msg)
>>> ChatMessage(content="I'm Stefano", role=<ChatRole.USER: 'user'>, name=None, meta={})

print(msg.to_dict())
>>> {'content': "I'm Stefano", 'role': 'user', 'name': None, 'meta': {}}

print(msg.text)
>>> "I'm Stefano"

I understand this can be a bit confusing, but it is done to smooth the transition to the new ChatMessage in 2.9.0 (see deepset-ai/haystack#8583 (comment) for details and motivation)

[Amnah199]

Alright. Thanks for explaining.

[Amnah199]

LG!