[Nithin (nithin42)]

Description

Fixes #3498

OpenAIChatTarget, OpenAICompletionTarget, and LiteLLMChatTarget document presence_penalty and frequency_penalty as taking values between -2.0 and 2.0, but previously neither class validated these parameters in init. Out-of-range values (< -2.0 or > 2.0) were stored in target identifiers and would fail downstream during API requests.

This change:

1. Adds validate_presence_penalty and validate_frequency_penalty helpers in pyrit.prompt_target.common.utils (raising PyritException if the value is not None and outside [-2.0, 2.0]).
2. Calls both validation helpers during init in OpenAIChatTarget, OpenAICompletionTarget, and LiteLLMChatTarget.

Tests and Documentation

• Added unit tests for validate_presence_penalty and validate_frequency_penalty in tests/unit/prompt_target/test_target_utils.py covering valid bounds (-2.0, 0.0, 2.0, None) and out-of-bounds inputs.
• Added parameter rejection unit tests in:
  • tests/unit/prompt_target/target/test_openai_chat_target.py
  • tests/unit/prompt_target/target/test_azure_openai_completion_target.py
  • tests/unit/prompt_target/target/test_litellm_chat_target.py
• Ran targeted pytest suite (258 passed).
• Verified code formatting with ruff check and ruff format --check.