fix wildcard file encoding

[jjmenet]

This pull request modifies order of assumed encoding when opening wildcard files.

Current logic is opening with ISO-8859-1 then, when error happens, falls back to UTF-8.
In this logic, when there are unicode characters in the wildcard file, it breaks unicode character to Latin-1, breaking unicode characters. When there are lora file with unicode character(for example, 漢字.safetensors), then try to load it, (i. e., lora:漢字:1) braking file names thus fails to load lora.

This change fixes this issue.

Below is summary by copilot.

---

This pull request updates the file encoding handling logic for reading wildcard .txt and .yaml files in modules/impact/wildcards.py. The main improvement is standardizing the default file encoding to UTF-8 for initial reads and switching to ISO-8859-1 only if a Unicode-related error occurs. This should improve compatibility and reduce issues with file reading across different environments.

File encoding handling improvements:

• Changed the default encoding to UTF-8 when opening .txt and .yaml wildcard files, and now only fall back to ISO-8859-1 if a yaml.reader.ReaderError or UnicodeDecodeError is encountered. (load_txt_wildcard, load_yaml_wildcard, read_wildcard_dict) [1] [2] [3] [4]
• Updated exception handling to consistently catch both yaml.reader.ReaderError and UnicodeDecodeError when attempting to read files. [1] [2]