FIX: Fix CodeChameleonConverter decryption functions in binary_tree and length modes

[RohithPariki]

Description

Fixes #3164.

In CodeChameleonConverter, the generated decryption functions for binary_tree and length modes contained bugs that caused runtime failures when executed:

1. binary_tree mode (_decrypt_binary_tree):
   string_to_tree uses json.loads(problem_description), deserializing the full nested structure into Python dictionaries. The recursive helper then called json.loads(node_dict['left']) and json.loads(node_dict['right']). Passing a Python dict to json.loads() resulted in TypeError: the JSON object must be str, bytes or bytearray, not dict.

   • Fix: Changed recursion to pass node_dict['left'] and node_dict['right'] directly to helper() without calling json.loads.

2. length mode (_decrypt_length):
   The _encrypt_length method serializes the list of single-entry dictionaries to a JSON string (json.dumps(...)), which is inserted into the prompt template. However, decryption(encrypted_problem) did not import json or parse encrypted_problem. Iterating directly over the JSON string caused AttributeError: 'str' object has no attribute 'items'.

   • Fix: Added import json and parsed encrypted_problem via json.loads if passed as a string.

Tests and Documentation

• Updated existing docstring assertions in tests/unit/converter/test_code_chameleon_converter.py for binary_tree and length modes.
• Added test_codechameleon_converter_decryption_execution parametrized over all built-in encryption types (reverse, odd_even, binary_tree, length) to verify that the generated decryption() function executes cleanly and reconstructs the original prompt.
• Verified all unit tests pass with pytest tests/unit/converter/test_code_chameleon_converter.py (10/10 passed).
• Linted and formatted with ruff check and ruff format.

[Roman Lutz (romanlutz)]

#3072 was technically here first but author didn't accept the CLA so happy to merge this.

[Roman Lutz (romanlutz)]

I'll make a couple of improvements and then merge. Thank you!