feat: support OpenAI 'developer' role in ChatMessage (#12604)

[Taranum01]

Summary

Fixes #12604.

OpenAI exposes a 'developer' chat role that was silently downgraded to ChatRole.SYSTEM by ChatMessage.from_openai_dict_format. _validate_openai_message and to_openai_dict_format already accepted 'developer', so the value was preserved on the wire but lost its identity on the client. This PR adds first-class ChatRole.DEVELOPER support so the role survives a round-trip.

Changes

• haystack/dataclasses/chat_message.py
  • Added ChatRole.DEVELOPER = "developer".
  • Added ChatMessage.from_developer(content, meta, name) classmethod, mirroring from_system.
  • Split from_openai_dict_format: "developer" -> from_developer; "system" -> from_system. "user" / "assistant" / "tool" unchanged.

Tests

• test/dataclasses/test_chat_message.py
  • test_chat_role_developer_value
  • test_from_developer_basic
  • test_from_developer_with_meta_and_name
  • test_from_openai_dict_format_developer_role
  • test_to_openai_dict_format_developer_role
  • test_round_trip_developer_role_preserved
  • test_round_trip_system_role_unchanged

[vercel]

Someone is attempting to deploy a commit to the deepset Team on Vercel.

A member of the Team first needs to authorize it.

[github-actions]

Hi @Taranum01, thanks for your interest in contributing to Haystack! 🙏

⚠️ You currently have 3 open pull requests in this repository (#12993, #12992 and this one). Our review capacity is limited, so please hold off opening more PRs until we've had a chance to review your first 2 open PRs. This helps us give each contribution the attention it deserves. Thank you!

This is an automated message to help us keep the review queue healthy.

[CLAassistant]

All committers have signed the CLA.

[sjrl]

Thanks for the PR! However, we don't want a new ChatRole here. Adding the role would mean updating every place that special-cases ChatRole.SYSTEM (ChatPromptBuilder, the Jinja message tag, compaction, Agent, and the integrations that extract system messages). On the OpenAI side the two roles behave the same: in a quick test against Chat Completions (gpt-4o-mini, gpt-4.1-nano, gpt-5-nano, o4-mini, o1) both were accepted with identical prompt token counts.